CN111478312A - Comprehensive energy cluster coordination control method for improving power grid stability - Google Patents

Abstract

The invention discloses a comprehensive energy cluster coordination control method for improving the stability of a power grid, which comprises the following steps: the dynamic game based disturbance rejection state monitoring system comprises a disturbance rejection state observation method for improving system stability, a measurement-correlation-prediction MCP modeling and evaluation method, a source network load integrated energy cluster coordination fast response disturbance rejection control method and an optimization simulation platform, and a multi-benefit subject optimization operation and transaction risk hedging strategy based on a dynamic game. In the aspect of complex system theoretical analysis, a general disturbance rejection state observer with engineering use value and an MCP modeling evaluation method are provided, the general disturbance rejection state observer can be used for a multi-time scale cold/heat/electricity integrated energy coupling system in a county, urban and rural overall planning area, a multi-energy power generation coordination optimization control method based on cloud interaction and an energy transaction simulation technology guided by cluster benefit optimization can be provided for wide-area multi-energy power generation clusters such as photovoltaic/wind power/thermal power generation/gas turbines, and a matched edge computing sensor and an intelligent controller are developed.

Description

Comprehensive energy cluster coordination control method for improving power grid stability

Technical Field

The invention belongs to the technical field of energy system management, electricity utilization efficiency and comprehensive energy optimization and control, and particularly relates to a comprehensive energy cluster coordination control method for improving the stability of a power grid.

Background

With the continuous advancing of generator sets in the power grid towards the directions of high capacity and high parameters, the higher the automation degree is, the higher the requirement on power grid frequency modulation control in the wide area energy internet is provided. The intelligent power grid system is gradually mature in all aspects, and aiming at the fact that a large-scale intelligent power grid system needs to be subjected to practical theoretical research and demonstration verification when participating in power frequency modulation, on one hand, modeling, simulation and analysis based on a traditional frequency modulation power supply and a grid frame are needed, and on the other hand, the intelligent power grid system needs to focus on the following aspects: firstly, the influence of a novel distributed intermittent power supply on the safe and stable operation of a power grid is discussed on the background of large-scale grid connection of renewable energy sources such as wind and light, and the advantages, the economy and the feasibility of the novel distributed intermittent power supply for assisting the new energy sources to participate in power frequency modulation are discussed; secondly, the difference of the intelligent power grid system and the traditional frequency modulation unit in the frequency modulation capabilities such as the adjustment precision and the adjustment rate is researched from the technical comparison angle of the frequency modulation power supply, and the frequency modulation efficiency of the intelligent power grid system and the traditional frequency modulation unit is compared based on the economic and environmental benefits; thirdly, the mechanism of the output of the energy storage power supply is researched from the angle cut-in of the energy storage power supply model, a corresponding equivalent model is established by considering the basic characteristics and various limiting factors of the energy storage power supply, or the energy storage power supply is described by using a constant frequency regulation effect coefficient and a first-order inertia link, and the influence of the energy storage power supply participating in frequency modulation on the suppression of frequency fluctuation and the exchange power deviation of a contact line is researched through small-load disturbance analysis based on a frequency modulation simulation model established by a region equivalent method; fourthly, from the economic aspect, the characteristics, the limits and various benefits brought by the participation of different types of micro sources in frequency modulation are combined, and the economic evaluation is carried out on the participation of the smart grid system in the grid frequency modulation; and fifthly, controlling the traditional frequency modulation power supply, the energy storage power supply and the distributed power supply to participate in frequency modulation based on the regional power grid, the energy storage power supply and the distributed power supply described by the equivalent model or the state space equation, or controlling the traditional frequency modulation power supply, the energy storage power supply and the distributed power supply by traditional hysteresis control, and focusing on optimizing a controller to improve the control performance, or completing coordination control among the three by adopting the advanced predictive control.

Meanwhile, the county, urban and rural overall planning area is a link of urban and rural integration, and presents the characteristics of complete industrial structure, close urban and rural regional connection and deep diversification of the energy system participating main body, wherein the county, urban and rural overall comprehensive energy market becomes an important unitary of energy business direction.

The county, city and county overall planning region is organically composed of administrative regions such as a county new region, an ancient city region and a county, compared with cities and industrial energy, the county new region, an old city region and a county overall planning region are unbalanced in public resource allocation and weak in infrastructure such as cooling, heating and power, and the comprehensive energy market has the characteristics of large-range wide-area distribution dispersion, strong local-region comprehensive energy supply requirements and large load periodic distribution difference. The old urban area has complete energy but difficult infrastructure transformation, the internal energy systems in the new area are respectively planned, the regional load demand periodic difference is obvious, the county is seriously lack of an energy service foundation, the comprehensive energy utilization efficiency is seriously influenced, and the county comprehensive energy market development is restricted. The intelligent-green operation mode of the comprehensive energy system is urgently needed to develop commercial service modes such as comprehensive energy cluster regulation, energy transaction, trusteeship operation and maintenance, promote urban infrastructure to extend to rural areas, promote urban social service business to cover rural areas, and guarantee urban and rural overall planning high-reliability stable collaborative development.

Taking county, urban and rural overall comprehensive energy cluster regulation and control and service as starting points, mainly realizing a comprehensive energy system cluster coordination control technology, an energy simulation transaction technology, a comprehensive energy cluster cloud-side-end information interaction technology, a comprehensive energy body cluster intelligent decision operation and maintenance technology, constructing a county, urban and rural overall comprehensive energy cluster whole life cycle operation regulation and control demonstration project based on a cloud-net-side-end architecture, creating county, urban and rural overall comprehensive energy system clustered operation typical samples through technical research and project demonstration in a mode of 'landing one batch, talking one batch and conspirating one batch', meeting multi-investment main body and multi-user requirements of county, new county, old town-rural area and rural overall energy project, realizing low-cost and large-scale comprehensive energy system operation, the method creates a low-cost, universal and open type comprehensive energy commercial operation service mode with the capability of falling to the ground.

Disclosure of Invention

The invention provides a comprehensive energy cluster coordination control method for improving the stability of a power grid, which is suitable for complex control systems including but not limited to comprehensive energy coordination control systems of power systems and aims to provide a modeling simulation, a control technology, a trading strategy and an implementation device for improving the stability of the systems, extracting disturbance rejection analysis, measurement-correlation-prediction, multi-energy power generation coordination optimization and comprehensive energy trading market risk hedging from data.

In order to achieve the purpose, the technical scheme adopted by the invention is as follows:

a comprehensive energy cluster coordination control method for improving power grid stability is characterized by comprising the following steps: the method comprises an anti-interference state observation method for improving the system stability, a measurement-correlation-prediction MCP modeling and evaluation method, an integrated energy cluster coordination fast response anti-interference control method, an integrated energy primary frequency modulation method, a digital source/network/charge multi-energy power generation model, an integrated energy coordination control optimization simulation control platform and an integrated energy electric power market trading risk hedging method based on an energy-energy space-time game theory.

As a further improvement of the invention, the disturbance rejection state observation method for improving the system stability utilizes an ESO observer for on-line data modeling, and the observer, the modeling model and the inverse modeling model all adopt a discrete mode to operate; the disturbance rejection state observation method has universality, is suitable for a source side system, is also suitable for a network side system, is also suitable for a cold/heat/electric coupled multi-energy power generation load side system, and is also suitable for modeling and disturbance rejection analysis of various systems.

As a further improvement of the invention, the measurement-correlation-prediction MCP method is to establish a correlation model between the wide-area comprehensive energy short-term source network charge energy flow data of the measurement position and the synchronous source network charge energy flow data of the reference position to predict the long-term energy flow data of the measurement position. Wherein, the measurement position refers to a target station position for modeling and evaluation; the reference position refers to a data source station with long-term measured data and simulation data enhanced excitation fusion. When using the MCP method, the resource characteristic values at the measurement location and the reference location should have a strong correlation.

(1) Technical requirements of MCP method

When using the MCP method, the following requirements should be observed:

the coordinates of the reference position and the measurement position, the data source and the effectiveness of the source network load energy flow characteristic value are recorded, and data cleaning is realized based on an anti-interference state observation method for improving the system stability. For long-term data, the technical requirements on boundary conditions in a source network load energy flow data model are met, namely

a) And field data of the wide-area comprehensive energy source network load power generation unit. The method comprises the steps of carrying out real-time mass data on the energy flow of the wide area source network. An industrial data full-process automatic checking device is developed, data cleaning and conversion are carried out, the reliability of source data is improved, a time constant and characteristic parameters of a unit are identified, a real-time transfer function model is determined, the problem of online data utilization of a traditional simulation model is solved, and the data level reaches ten thousand points per second;

b) the power grid frequency control simulation data comprises power grid L FC frequency response system models, frequency control response characteristics and dynamic characteristic data of primary frequency modulation and secondary frequency modulation, and the data level reaches one second to thousands of points;

c) synthesizing energy simulation data; and (3) synthesizing simulation data (including gas turbine, wind power and photovoltaic) of the energy unit. The micro-grid simulation experiment system is from a micro-grid simulation laboratory actual physical device and a simulation model. The dynamic operation data of the new energy frequency characteristics of a gas turbine, wind power, photovoltaic and the like are contained, and the data level reaches thousands of points in one second.

d) Reporting data after a wide area comprehensive energy source network load energy flow model in a wider range;

e) the above four data are combined.

If the boundary condition data is derived from the observed data, erroneous and invalid data should be rejected and any known system errors should be eliminated. The observation data time span is preferably more than 2 years, and the data qualification rate reaches more than 70% of the observation total amount. If different units of observation data are used, the source of the data, the observation conditions and the processing method should be specified.

As further improvement of the invention, the method for controlling the disturbance rejection of the large-amplitude variable-working-condition network source coordination and rapid response coordination comprises a modeling method based on online data verification preprocessing and a power grid frequency rapid response coordination optimization control method for improving the source network load coordination response characteristic.

As a further improvement of the invention, the modeling method based on online data verification preprocessing comprises the following processes of providing a universal industrial system online data verification preprocessing method based on an excitation-enhanced simulation genetic optimization method (Z L201510883631.3) and a universal excitation-enhanced simulation data verification method (Z L201510883385.5) in the issued invention patents, improving the signal precision through normalization and filtering noise reduction preprocessing calibration, designing a signal with continuous excitation characteristics and dynamic characteristics fully excited by combining the test data of the existing simulation system, determining the order and time delay parameters of an actual object by depending on an excitation-enhanced dynamic simulation platform, extracting a simulation model aiming at the production test flow, and realizing the combination of online data and a simulation modeling technology;

a variable working condition modeling method is provided, based on a data verification preprocessing technology, on the basis of a variable working condition simulation experiment in a range of 50% -100% of rated load of a thermal power unit simulation system, a large amount of experiment data are obtained, a PSO particle swarm algorithm is adopted to identify system parameters, time delay parameters and system orders, a deviation correction model is built according to model output and actual output deviation, a model correction value is output, a dynamic model of the unit is obtained by adding the model correction value and a basic model obtained by PSO identification, the parameter identification and model precision reach 95%, and a transfer function model for coordinately controlling the variable working condition and the static working condition of a supercritical unit is built.

As a further improvement of the invention, the multi-energy power generation coordination optimization control method based on the comprehensive energy cluster cloud interaction technology comprises the following steps:

step 1, developing cloud-edge-end information interaction standardized modeling, establishing a cloud platform model, a cloud-edge interaction model, an edge-end interaction model and an end equipment model, and realizing the basis of cloud distributed computing;

step 2, developing a miniaturized, modularized and passive multi-energy intelligent information sensor, realizing an open and universal access technology suitable for cross-region, large-range and scene energy supply and utilization nodes, realizing massive multi-dimensional data collection of coupling energy such as cold/heat/electricity and the like, and quickly sensing the real-time running state of a comprehensive energy body;

step 3, giving play to the advantages of 5G fast networking and high efficiency, providing a multi-stream convergence and multi-mode heterogeneous cooperative communication protocol conversion technology based on 5G, namely a county comprehensive energy system information exchange use case based on UM L, developing a comprehensive energy cluster system communication protocol, supporting REST and MQTT information exchange modes, realizing cloud-edge-end communication protocol standardization, providing a service scene cloud customization and unified management technology aiming at the characteristic of difference of application scene requirements of different comprehensive energy bodies, and realizing seamless cloud switching and dynamic interaction among comprehensive energy cluster scenes;

step 4, realizing a multi-energy power generation unit coordination control algorithm library; the method comprises the steps of realizing a cluster control method facing energy requirements and operation and maintenance requirements based on a distributed observer, providing a distributed optimization control algorithm with super-linear convergence speed, deploying the algorithm to an intelligent controller to form local closed-loop control, providing a cluster efficient distributed feedback optimization control technology for decomposition and coordination, and forming cluster closed-loop control by a cloud platform and a plurality of intelligent controllers;

step 5, based on the coordination control algorithm library in the step 4, realizing L ADRC linear active disturbance rejection control strategy of multi-energy power generation by taking an energy balance signal as a feedforward signal, and replacing a traditional PID control module with L ADRC, wherein the control strategy is strictly realized according to DCS/P L C control logic paging, and the control strategy runs in a flow mode on an NI Compact RIO controller in a 200MS scanning period, further develops a plug-and-play autonomous high-performance intelligent controller, receives a cloud platform control instruction, and adjusts internal equipment of the comprehensive energy body to realize energy and operation and maintenance demand response;

step 6, realizing active defect identification and risk alarm facing to cold/heat/electricity and other coupling energy sources and complex network mass operation data based on a cloud platform, and carrying out decarbonization economic operation evaluation;

step 7, carrying out research aiming at the characteristics of large service span, wide range and the like of the wide-area comprehensive energy body, and realizing the on-line prediction of the energy supply demand of each comprehensive energy body; and establishing a monthly/quarterly multi-scene global operation and maintenance optimization decision mechanism of the comprehensive energy body, and providing specialized high-level operation and maintenance service for users in the region.

As a further improvement of the invention, the method comprises the following steps:

step 1, a multi-energy power generation micro-grid participates in modeling of a thermal power generation/gas turbine coordinated control frequency modulation control architecture and analyzes the influence of the modeling;

step 2, researching and verifying a coordinated prediction control algorithm of the multi-energy power generation micro-grid participating in thermal power generation/gas turbine;

step 3, developing a cooperative control device suitable for a multi-energy power generation micro-grid to participate in frequency modulation of a thermal power generation/gas turbine power system;

and 4, building a regional interconnection system management and control platform containing the renewable energy intelligent micro-grid and participating in thermal power generation/gas turbine.

As a further improvement of the invention, a digital source/grid/load model and a comprehensive energy coordination control optimization simulation control platform use StarSim or RT L ab power system simulation software to build a source/grid/load power system model comprising a thermal power generation boiler coordination control system, a steam turbine coordination control system, an energy storage system or a pumped storage and load system;

aiming at a high-capacity unit coordination optimization control technology and a primary frequency modulation technology route under the condition of extra-high voltage power grid operation, aiming at the on-line hardware closed loop simulation technology development of source grid load coordination control performance, a hybrid logic dynamic programming method or an intelligent method is used for simulating a digital source/grid/load model and a comprehensive energy source energy coordination optimization control technology according to the overall system architecture and parameters of a thermal power generation boiler system, a steam engine system, a gas turbine control system, an energy storage system and a micro-grid load system;

the method applies a mixed integer linear programming method, a genetic algorithm and an improved complex process global optimization evolutionary algorithm to the energy coordination optimization management of the source/network/load.

As a further improvement of the invention, the process of establishing the energy coordination optimization management model of the source/network/load system comprises the following steps:

on the basis of ensuring the power supply of the local load, the minimum operation cost of a source/network/load is taken as a target, wherein the minimum operation cost comprises the cost of purchasing electricity from a power grid, the income obtained by selling the electricity to the power grid and the maintenance cost and depreciation loss of a storage battery;

in the formula, eSell (t) is the real-time electricity purchasing price of a power grid, eBuy (t) is the real-time electricity selling price of the power grid, ebat (t) is the operation management cost of a storage battery, PgBuy (t) is the electric power rate absorbed by the large power grid at the t moment, the sign is negative, PgSell (t) is the electric power generated by the large power grid at the t moment, the sign is positive, Pbat (t) is the active power of the storage battery at the t moment, △ t is the system operation time interval, and the value is 1 hour;

the objective function comprises the cost for purchasing electricity from the power grid and the income obtained by selling electricity to the power grid, and how to use Pg (t) to represent the main grid to source/grid/load output power when the value is positive, and to represent the input power when the value is negative, which is specifically represented by the following formula;

wherein, when Pg (t) is positive, the electricity purchasing cost is expressed as eSell (t) and Pg (t); when Pg (t) is negative, the electricity selling cost is expressed as-eBuy (t) Pg (t).

Adopting a predictive control framework based on rolling time domain optimization, solving the optimization problem of minimum consumption cost from an external power grid when a model operates in a rolling time domain [ t t + tp ] at the moment t, and enabling a target function in the rolling time domain to be minimum by calculating an optimal control sequence in [ t t + tp ]; on the basis of the formula, a rolling optimization range is added, rolling is performed by taking one hour as a period, the step length of the rolling range is assumed to be tp, and the consumption cost in the time from t to t + tp is calculated and used as a new objective function, which is expressed as:

under the new objective function, the rolling time domain not only considers the current step, but also puts the system operation state in the future period into the calculation range.

As further improvement of the invention, the comprehensive energy power market trading risk hedging method based on the energy-energy space-time game theory and the comprehensive energy cluster multi-interest subject optimization operation and trading strategy comprise the following processes:

(1) firstly, classifying users in the power market into three categories: large power users and medium-sized users. Collecting historical data related to an integrated energy trading market from the establishment of varieties by using an existing platform, and summarizing and subsuming seasonal rules of arbitrage combination according to different time scales by using the SVM-ARIMA characteristic sequence prediction method and the trading data characteristic sequence pattern recognition technology in claim 5;

(2) an open universal energy transaction mechanism based on a block chain is provided, point-to-point energy transaction settlement of green energy credit is realized, user and service resources are integrated, an intelligent contract with simplified flow and safe transaction is deployed, and transaction matching and contract management are realized;

(3) combining the seasonal regularity of commodities of the comprehensive energy phase power market, analyzing and counting technical indexes such as volume of transaction, position quantity, time scale, MACD deviation and the like and relevant game characteristics based on the volume energy space-time characteristics of the wide-area comprehensive energy flow and the service flow, exploring a commodity futures hedging implementation method, carrying out backtracking test on a trading system, evaluating the advantages and disadvantages of different strategies, and proposing a trading strategy for improving the trading rate to more than 50%;

(4) based on the average lines and golden section points of different periodic scales, researching and calculating a win-loss control technology of the transaction, realizing entrance and stand management with optimal cost performance and quantification of profit and loss prevention, carrying out backtracking test on a transaction system, and evaluating the advantages and disadvantages of different strategies;

(5) calculating correlation degrees of hedge arbitrage varieties of statistical arbitrage, developing regression analysis, calculating a win-loss threshold value, performing statistical induction on each hedge arbitrage combination from different time scales, performing backtracking test on a trading system, and evaluating the performance of different strategies by using a mathematical statistical method, adopting ideas such as a normal distribution mathematics original idea and a least square method and the like;

(6) a tool of a CTP interface is used for developing a commodity futures arbitrage hedging monitoring software system, and a method for realizing 90% arbitrage hedging variety combined coverage is researched;

(7) a commodity futures arbitrage hedging method related to a comprehensive energy power market based on seasonal regularity, and a rolling multi-strategy combination hedging transaction risk management system for improving the profit rate of the repeated profit;

(8) a standardized energy storage device is designed, and a corresponding control device is designed for the purpose of delivering electric energy with standard quality.

(9) Based on a non-cooperative dynamic game theory, providing comprehensive energy body optimization operation, realizing comprehensive benefit maximization, constructing an energy simulation transaction control system which takes green low-carbon value as a guide active management and global optimal strategy, and realizing the full life cycle management of an energy transaction combined bill;

(10) establishing a multi-energy-station energy network simulation model facing to energy supply and energy supply difference, realizing a multi-energy-station cluster regulation and control simulation method considering interactive transaction, and realizing county-area comprehensive energy system cluster regulation and control simulation evaluation software module;

(11) based on the content of (10), a comprehensive energy whole industrial chain transaction data sharing technology based on a cloud platform is established, and comprehensive energy transaction data analysis service is provided for governments at all levels; the method is oriented to new energy increasing customers, and provides energy operation and maintenance service through modes of investment, total package construction, financing lease and the like; the system is oriented to the storage energy utilization customers, provides services such as energy collection monitoring, energy utilization scheme optimization, energy-saving effect sharing and the like, and develops commercial profit model innovation practice of comprehensive energy service.

Compared with the prior art, the invention has the following beneficial effects:

the invention provides an anti-interference state observer for improving the stability of a system, which has the advantages of providing an on-line/off-line general anti-interference state observer with engineering use value for improving the stability of the system, and being suitable for data analysis and modeling, internal information extraction and disturbance information extraction of a complex system; in the aspect of electric power system analysis, a measurement-correlation-prediction MCP modeling and evaluation method is provided, the method can be used for modeling and analyzing a multi-time scale cold/heat/electricity integrated energy coupling system in a county, city and county overall planning area, and can also be used for analyzing a wide-area multi-energy power generation cluster system such as a photovoltaic system, a wind power system, a thermal power system, a gas turbine and the like, a multi-energy power generation frequency coordination optimization control method and a simulation platform based on an integrated energy cluster cloud interaction technology are provided, and an integrated energy cluster operation control system based on a cloud-network-side-end architecture is developed and comprises a multi-energy intelligent information sensor and an integrated energy intelligent controller; and further developing an energy trading simulation technology with comprehensive energy cluster benefit optimization as guidance, providing a comprehensive energy power market trading risk hedging method based on an energy-energy space-time game theory, providing commercial value-added services by a comprehensive energy cluster multi-benefit subject optimization operation and trading strategy based on a non-cooperative dynamic game, and supporting the integrated visual, calculated and controllable comprehensive energy cluster cold/heat/electricity interconnection and mutual aid demonstration engineering construction in a region.

Drawings

FIG. 1 is an ESO-based on-line modeling schematic;

FIG. 2 is an ESO-based off-line modeling schematic;

FIG. 3 is a logic diagram of an L ADCR linear active disturbance rejection coordination control strategy;

FIG. 4 is a schematic diagram of an external control device and a simulation system;

FIG. 5 is a diagram of a multi-energy complementary user-level microgrid topology;

FIG. 6 is a diagram of a coordination control scheme for a micro-grid to participate in power frequency modulation;

FIG. 7 is a schematic diagram of an MPC control module;

FIG. 8 is a schematic diagram of a source/grid/load system energy coordination optimization management model;

FIG. 9 is a flow chart of energy coordination optimization;

FIG. 10 is a graph of a primary frequency modulation optimization design based on stored energy;

FIG. 11 is a diagram of an energy management system high level application software architecture;

FIG. 12 is a schematic diagram of an energy storage power station measurement and control system;

FIG. 13 is a schematic diagram of a miniaturized, modular, passive, multi-energy intelligent information sensor;

FIG. 14 is a cloud-web-edge-end system architecture diagram;

FIG. 15 is a diagram of ARIMA and SVM combined prediction principle;

fig. 16 is a technical scheme of the present invention.

Detailed Description

The technical solutions in the embodiments of the present application will be described below clearly and completely with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the application, its application, or uses. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

A comprehensive energy cluster coordination control method for improving power grid stability comprises the following steps: the method comprises an anti-interference state observation method for improving system stability, a measurement-correlation-prediction MCP modeling and evaluation method, a source network load cluster coordination fast response anti-interference control method, a source/network/load comprehensive energy coordination control optimization simulation control platform, a comprehensive energy power market trading risk hedging method based on an energy-energy space-time game theory, and a comprehensive energy cluster multi-benefit subject optimization operation and trading strategy based on a non-cooperative dynamic game. Fig. 13 is a schematic diagram of a miniaturized, modular, passive, multi-energy intelligent information sensor, fig. 14 is a cloud-network-edge-end system architecture diagram, and fig. 16 is a technical route diagram of the present invention.

As shown in fig. 1-2, the disturbance rejection state observation method for improving the system stability proposes a method for modeling online data by using an ESO observer. In an actual control system, the disturbance of the system can cause great influence on system modeling, a system model without a large amount of prior information can be obtained, an accurate inverse model of the system is further obtained, and a new means is provided for further overcoming the disturbance and improving the system performance. The observer, the modeling model and the inverse modeling model all operate in a discretization mode, and can be applied to hardware fast calculation in a large scale, so that the system performance and the precision are improved. The method has universality, is suitable for a source side system and a network side system, and can be further popularized and popularized to various system modeling and anti-interference analysis.

The large-amplitude variable-working-condition network source coordination quick response coordination disturbance rejection control method comprises a modeling method based on online data verification pretreatment and a large-scale unit AGC coordination optimization control method based on the coordination response characteristic of a hoisting machine network.

The process of the enhanced excitation modeling method based on online data verification preprocessing is as follows: providing an online data enhancement excitation check preprocessing method for a general industrial system, and calibrating and improving the signal precision through normalization and filtering noise reduction preprocessing; designing a signal with continuous excitation characteristics and full excitation of dynamic characteristics by combining test data of the conventional simulation system, determining order and time delay parameters of an actual object by relying on an enhanced excitation dynamic simulation platform, extracting a simulation model aiming at a production test flow, realizing the combination of online data and a simulation modeling technology, and solving the problem of mismatching of multi-time scale characteristics of a multi-energy power generation unit;

providing a comprehensive energy cluster power supply and energy utilization prediction and aggregation perception model based on multivariate data analysis, scientifically and reliably describing correlation relations between response time, response rate and response precision of a target station and a reference station on probability density distribution or time sequence in a form of a mathematical model based on the measurement-correlation-prediction MCP modeling and evaluation method of claim 3 and an MCP correlation model; fitting a prediction model of measurement-correlation-prediction MCP by using an SVM-ARIMA time characteristic sequence model, and displaying the scheduling available capability of the in-network multi-energy power generation unit to a scheduling center in a visual form in real time, wherein the indexes are a power grid frequency response regulation rate K1, a power grid frequency response regulation precision K2 and a power grid frequency response time K3; then, determining a source network load energy flow coordination feedforward quick response control method based on time sequence prediction;

the ARIMA and the SVM have advantages in processing the linear model and the nonlinear model respectively, the advantages of the ARIMA and the SVM are complementary, the ARIMA and the SVM are combined for price prediction, and a better result is received;

assuming that the time series Y can be considered as a combination of the linear autocorrelation portion L and the nonlinear residual N, i.e., Y L + N, the following steps are to be taken herein to construct a combined prediction model:

(1) modeling a linear part by utilizing an ARIMA model, setting a prediction result as L, setting residual errors of a sequence assembly as N and N, and including a nonlinear relation of a sequence Y;

(2) reconstructing the N sequence obtained in the previous step to obtain an SVM sample set, and predicting a residual error by using an SVM to obtain a prediction result N;

(3) l obtained by linear prediction is combined with N obtained by a non-linear set to obtain a prediction result Y which is L and + N, wherein a combined prediction principle, such as an ARIMA and SVM combined prediction principle diagram, is shown in FIG. 15;

according to the fast response requirement of the power grid frequency, determining the main response characteristic of the source grid load energy flow as response time, and improving the response capability of the power grid frequency scheduling instruction in real time; finally, a power grid frequency response control technology online hardware rapid deployment method is realized, an external-hanging controller is developed, and the problems that a big data technology cannot be practically applied and popularized in a multi-energy power generation comprehensive energy system and intelligent scheduling is realized are solved.

As shown in fig. 3, a coordinated control method of fast response in the furnace-to-machine (CCBF) mode in the Automatic Generation Control (AGC) mode centered on the power grid is proposed. The method overcomes the inherent pure hysteresis and large inertia characteristic of a boiler part of a unit (especially the serious influence of the hysteresis of main steam pressure on the boiler and AGC response rate), realizes the coordinated quick response control of the unit under the variable working condition of 50-100 percent in a large range, has strong variable load characteristic, improves the assessment index of a power plant, solves the problem that a coordinated object of the unit cannot quickly respond to the energy requirement of a power grid, and improves the stability of the power grid.

(1) And realizing a coordination control algorithm library.

(2) Based on the coordination control algorithm library in step 1, a network source coordination L ADRC linear active disturbance rejection control strategy taking an energy balance signal as a feedforward signal is realized, L ADRC replaces a traditional PID control module, the control strategy is realized strictly according to DCS control logic paging, the NI Compact RIO controller runs in a flow mode in a 100MS or 200MS scanning period, the output boundary of the network source coordination L ADRC linear active disturbance rejection control strategy is the output of a boiler main control signal and a steam turbine main control signal, the input boundary is a scheduling AGC energy demand signal, and a main steam pressure set value and a power set value.

As shown in fig. 4, the mainframe group system has the characteristics of variable working conditions and pure hysteresis and large inertia, and exhibits a characteristic of variation along with different operating conditions, which causes great difficulty in realizing coordinated control and fast response in the combustion process, and further has no problem in realizing the network-computer coordinated AGC control.

The invention provides a variable working condition fast response anti-interference control strategy of feedforward control, active anti-interference and offline modeling correction signals. The method is divided into three parts:

(1) the method comprises the steps of adopting an ESO observer to respectively sample and observe the output of a controlled object, adopting a positive model (such as a neural network or a linear filter) as an online modeling method to realize online modeling on the controlled object, and correcting the model by utilizing the output error of the ESO observer and the model.

(2) Copying an online modeling model, selecting a white noise signal close to the actual object frequency as an input signal of the offline modeling by adopting an offline modeling mode, correcting an inverse model by utilizing input and output errors to obtain the inverse of the object model, copying the inverse of the object model into a control system as a correction signal, comparing the correction signal with the output of an active disturbance rejection controller, and correcting an output signal of the active disturbance rejection controller.

(3) Based on the load and main steam pressure characteristic model of the supercritical unit and based on the corresponding relation of the load, main steam pressure error and real-time feedback of the opening of the regulating valve, the model is used as a feedforward instruction, so that oscillation in the optimizing process is avoided, the control precision is improved, the influence of pure lag and inertia is corrected, and the response speed is accelerated.

(4) And an RJ45 interface is used as a communication mode to realize the network communication between the RIO controller and the PXI platform, and the communication protocol adopts a general industrial protocol MODBUS-TCP or OPC protocol.

(5) A PXI platform is used as a carrier, and control logics of a combustion control system, a feedwater control system and a main steam temperature control system are realized based on L ABVIEW.

The method can greatly improve the response speed and the accuracy of the unit to the load instruction, and meet the control requirements of main operating parameters such as main steam pressure and the like.

The CCS coordination control logic list is as follows:

the comprehensive energy primary frequency modulation method is developed from five steps of model establishment- > mechanism analysis- > control method- > real-time simulation- > system implementation in order to solve the problem of frequency modulation coordination control of a thermal power generation power system participated by a renewable energy smart micro-grid, and the detailed contents of the adopted research scheme, method and route are as follows:

(1) as shown in FIG. 5, the intelligent micro-grid participates in modeling of a coordinated control frequency modulation control architecture of the thermal power generator grid and the influence of the modeling is analyzed

Firstly, a regional equivalent method in a grid frequency modulation simulation model is systematically analyzed by combining the frequency characteristic and the load characteristic of a power system of a traditional coal-fired thermal generator set Coordinated Control System (CCS), and a regional equivalent model for conventional power system frequency modulation is constructed. The heterogeneous and dispersive characteristics of a control communication network architecture and a coordination control system are researched, a looped network communication architecture model of the thermal power generation local control communication equipment containing renewable energy sources is constructed, and basic theoretical verification of functions of data monitoring, economic dispatching planning, power flow analysis, frequency and voltage stabilization, self power generation planning, real-time monitoring and the like is completed.

The supply and demand balance in the primary and secondary frequency modulation control loops of the interconnected power system consists of three parts, namely a generator, a load and the exchange power of a tie line. The key for constructing the frequency response model required by the primary and secondary frequency modulation control of the generator set is the modeling of the speed regulating system and the prime motor. The thermal power generating unit is provided with a primary frequency modulation control loop and a secondary frequency modulation control loop on the structures of a speed regulating system and a prime motor. In addition to this, consideration is also required for the load of the system and the links between control areas. The unit, the load, the tie line, the primary frequency modulation control ring and the secondary frequency modulation control ring are respectively modeled, and the frequency modulation dynamic model of the power system can be obtained by combining according to the structural relationship in the actual power system.

Secondly, considering external characteristic function models of distributed energy and loads and three-phase steady-state models of a power distribution system including a thermal power generation model, a power transmission line model, a distribution transformer model, a parallel capacitor model and a load model, and analyzing the influence of related models on the system stability of the self-optimizing power distribution network aiming at the characteristics of high permeability of distributed energy such as photovoltaic power, wind power and the like and demand side response and the like.

When the electric power system is in operation, the control of the frequency and the exchange power deviation of the tie line is mainly completed by secondary frequency modulation. Under the conditions of the intelligent microgrid with renewable energy and frequency modulation, each frequency modulation power supply is reasonably coordinated to control and adjust the output power of each generator and the intelligent microgrid with renewable energy so that the system frequency meets the requirement of the power grid. The strategy of using the intelligent microgrid containing renewable energy sources to participate in primary frequency modulation is as follows: and a frequency conversion regulation effect coefficient strategy.

The method for determining the fixed frequency regulation effect coefficient comprises the following steps: primary frequency modulation in a power grid is a regulation mode for preventing the frequency of the system from deviating from a standard by utilizing the inherent load frequency characteristic of the system and the action of a speed regulator of a generator set. Therefore, the intelligent microgrid with renewable energy can participate in primary frequency modulation of the power grid in a mode of simulating a generator set to adjust self output according to the frequency deviation and the frequency adjustment effect coefficient. Namely, the relationship between the current increment delta ib of the intelligent microgrid with renewable energy and the primary frequency modulation control signal delta f is shown as the following formula:

wherein, K_bAnd R_bRespectively a frequency regulation effect coefficient (or unit regulation current) and a static droop coefficient (or regulation) of the intelligent microgrid containing renewable energy sourcesDifference coefficient).

Frequency conversion adjustment effect coefficient strategy: if the intelligent microgrid with renewable energy sources has a large amount of energy surplus, the frequency of the power grid is reduced, and on the premise of not exceeding the rated power of the intelligent microgrid with renewable energy sources, the frequency regulation effect coefficient of the microgrid is selected to be a value larger than that determined by a traditional method; if the grid frequency is higher than 50Hz, the micro-grid should be selected to discharge small current with a certain reasonable frequency regulation effect coefficient from the viewpoint of protecting the micro-grid. This is beneficial to the energy management of the micro-grid and can influence the frequency modulation of the grid as little as possible.

(2) Research and verification of coordination prediction control algorithm of thermal power generator network participated by intelligent micro-grid

The system is a layered distribution type framework structure and comprises an upper layer of regional coordination controller and a lower layer of local frequency modulation controller. The method is introduced into a regional coordination controller, and based on a dynamic model of an interconnected power system and a system containing a renewable energy intelligent micro-grid, an optimization model taking output error and control increment weighting as a target function and thermal power generating unit, micro-grid and tie line regulation capacity as constraint conditions is constructed by establishing a relation between an intelligent micro-grid containing renewable energy and frequency modulation control input quantity and frequency and tie line power prediction quantity. And the local frequency modulation controller of each area receives the instruction of the upper layer controller, and can correct the frequency and the tie line power deviation of the system.

The task of the coordinating controller is mainly the following three major parts, as shown in fig. 6.

(1) Monitoring state

The regional coordination controller receives system real-time state information monitored by each region, and the system real-time state information comprises inter-region tie line exchange power deviation, output power change of generators in each region, output power change of intelligent micro-grids containing renewable energy sources in each region, discharge depth of energy storage power supplies of batteries, load power change of each region, system frequency deviation of each region and the like.

(2) Predicting the future

The regional coordination controller predicts the future dynamic track on the basis of the system dynamic model, establishes the relation between the control quantity and the prediction quantity and lays a foundation for the optimal strategy.

(3) Optimal control decision

And (3) on the basis of the result of the step (2), solving by the regional coordination controller according to the set objective function and the constraint of each frequency modulation device. After comparing the predicted values with the actual values, errors can be derived to further correct the prediction model.

The corresponding state equation of the system predictive control is basically formed as the formula:

y(t)＝Cx(t)+D_vv(t)+D_dd(t)

wherein x (t) is a state variable of the system; u (t) is a control input amount; v (t) is the measurable disturbance variable; d (t) is the amount of unmeasurable disturbance; y (t) is the output of the system; A. bu, Bv, Bd, C, D and Dd are coefficient matrixes corresponding to the system quantities. The operation of the MPC control module is shown in FIG. 7.

(3) Development of frequency modulation cooperative control device suitable for renewable energy-containing intelligent micro-grid participating thermal power generation power system

The intelligent micro-grid cooperative optimization control device suitable for thermal power generation and containing renewable energy is developed, and the device can collect instrument data of an energy system; voltage, current, active and reactive closed-loop control functions; controlling reactive power, voltage and optimal power flow of a local converter; centralized and distributed real-time network communication functions are provided; regional distribution network autonomous function, etc. Meanwhile, an embedded operation platform is provided, an advanced control algorithm can be realized, and the optimized value of the power supply quantity of each distributed power generation unit in the self-optimization-trending system is calculated by using the instrument data and the load demand; and feeding the optimized value back to the system so that the system adjusts the operation parameters of each distributed power generation unit according to the optimized value. Design parameters of specific equipment: rated voltage: 220V; rated frequency: 50 Hz; power consumption: less than 45W. The annual availability ratio of the system is more than or equal to 99.9 percent; and (3) measuring the comprehensive error by an analog quantity: less than or equal to 1 percent; mean barrier free time: more than or equal to 8760 h; data sampling scanning period: less than or equal to 10 seconds; the system controls the response time of the operation (from pressing the execution key to the execution of the charger): <10 seconds; the insulation resistance should be not less than 5M omega; working temperature: -25 ℃ to 55 ℃.

(4) Building regional interconnection system management and control platform containing renewable energy intelligent micro-grid and participating in thermal power generation

And (3) building a regional interconnection system management and control platform containing a multi-energy power generation micro-grid participating in thermal power generation/gas turbine power generation, and integrating the original thermal power generation system. Meanwhile, a distributed energy monitoring system (photovoltaic and wind power), a flexible load monitoring system (electric automobile, coal-to-electricity and demand side response) and a self-optimization distributed operation coordination control function are added to the platform, so that the economic and technical requirements of thermal power generation and renewable energy multi-source coordination optimization, optimal power flow calculation, comprehensive reactive power optimization and voltage control are met, the energy utilization rate of a power grid is improved, the operation cost is reduced and the like. Managing and controlling platform picture calling time: <3 seconds; and (3) refreshing time of real-time data of the picture: 5-30 seconds; real-time data query response time: <3 seconds; historical data query response time: <10 seconds; sampling interval of distribution power data historical curve: 1 min-30 min, adjustable; sampling interval of charging data historical curve: 1 s-30 min, adjustable; the storage time of the historical trend curve, the daily newspaper, the monthly newspaper and the annual newspaper is more than or equal to 1 year.

As shown in fig. 8-13, a digital source/grid/load model and a comprehensive energy coordination control optimization simulation control platform use StarSim power system simulation software to build a source/grid/load power system model including a thermal power generation boiler coordination control system, a steam turbine coordination control system, an energy storage system or a pumped storage and load system;

aiming at a high-capacity unit coordination optimization control technology and a primary frequency modulation technology route under the condition of ultra-high voltage power grid operation, aiming at the online hardware closed loop simulation technology development of the unit coordination control performance, a hybrid logic dynamic programming method or an intelligent method is used for simulating the digital source/network/load model and the comprehensive energy source energy coordination optimization control technology according to the overall system architecture and parameters of a thermal power generation boiler system, a steam turbine system, an energy storage system and a load system.

As further improvement of the invention, a mixed integer linear programming method, a genetic algorithm and an improved complex process global optimization evolutionary algorithm are applied to the energy coordination optimization management of the source/network/load.

As a further improvement of the invention, the process of establishing the energy coordination optimization management model of the source/network/load system comprises the following steps:

on the basis of ensuring the power supply of the local load, the minimum operation cost of a source/network/load is taken as a target, wherein the minimum operation cost comprises the cost of purchasing electricity from the power grid, the income obtained by selling the electricity to the power grid and the maintenance cost and depreciation loss of a storage battery;

in the formula, eSell (t) is the real-time electricity purchasing price of a power grid, eBuy (t) is the real-time electricity selling price of the power grid, ebat (t) is the operation management cost of a storage battery, PgBuy (t) is the electric power rate absorbed by the large power grid at the t moment, the sign is negative, PgSell (t) is the electric power generated by the large power grid at the t moment, the sign is positive, Pbat (t) is the active power of the storage battery at the t moment, △ t is the system operation time interval, and the value is 1 hour;

the objective function comprises the cost for purchasing electricity from the power grid and the income obtained by selling electricity to the power grid, and how to use Pg (t) to represent the main grid to source/grid/load output power when the value is positive, and to represent the input power when the value is negative, which is specifically represented by the following formula;

wherein, when Pg (t) is positive, the electricity purchasing cost is expressed as eSell (t) and Pg (t); when Pg (t) is negative, the electricity selling cost is expressed as-eBuy (t) Pg (t).

Adopting a predictive control framework based on rolling time domain optimization, solving the optimization problem of minimum consumption cost from an external power grid when a model operates in a rolling time domain [ t t + tp ] at the moment t, and enabling a target function in the rolling time domain to be minimum by calculating an optimal control sequence in [ t t + tp ]; on the basis of the formula, a rolling optimization range is added, rolling is performed by taking one hour as a period, the step length of the rolling range is assumed to be tp, and the consumption cost in the time from t to t + tp is calculated and used as a new objective function, which is expressed as:

under the new objective function, the rolling time domain not only considers the current step, but also puts the system operation state in the future period into the calculation range.

As further improvement of the invention, the comprehensive energy power market trading risk hedging method based on the energy-energy space-time game theory and the comprehensive energy cluster multi-interest subject optimization operation and trading strategy comprise the following processes:

(1) firstly, classifying users in the power market into three categories: large power users and medium-sized users. Collecting historical data related to an integrated energy trading market from the establishment of varieties by using an existing platform, and summarizing and subsuming seasonal rules of arbitrage combination according to different time scales by using the SVM-ARIMA characteristic sequence prediction method and the trading data characteristic sequence pattern recognition technology in claim 5;

(2) an open universal energy transaction mechanism based on a block chain is provided, point-to-point energy transaction settlement of green energy credit is realized, user and service resources are integrated, an intelligent contract with simplified flow and safe transaction is deployed, and transaction matching and contract management are realized;

(3) combining the seasonal regularity of commodities of the comprehensive energy phase power market, analyzing and counting technical indexes such as volume of transaction, position quantity, time scale, MACD deviation and the like and relevant game characteristics based on the volume energy space-time characteristics of the wide-area comprehensive energy flow and the service flow, exploring a commodity futures hedging implementation method, carrying out backtracking test on a trading system, evaluating the advantages and disadvantages of different strategies, and proposing a trading strategy for improving the trading rate to more than 50%;

(4) based on the average lines and golden section points of different periodic scales, researching and calculating a win-loss control technology of the transaction, realizing entrance and stand management with optimal cost performance and quantification of profit and loss prevention, carrying out backtracking test on a transaction system, and evaluating the advantages and disadvantages of different strategies;

(5) calculating correlation degrees of hedge arbitrage varieties of statistical arbitrage, developing regression analysis, calculating a win-loss threshold value, performing statistical induction on each hedge arbitrage combination from different time scales, performing backtracking test on a trading system, and evaluating the performance of different strategies by using a mathematical statistical method, adopting ideas such as a normal distribution mathematics original idea and a least square method and the like;

(6) a tool of a CTP interface is used for developing a commodity futures arbitrage hedging monitoring software system, and a method for realizing 90% arbitrage hedging variety combined coverage is researched;

(7) a commodity futures arbitrage hedging method related to a comprehensive energy power market based on seasonal regularity, and a rolling multi-strategy combination hedging transaction risk management system for improving the profit rate of the repeated profit;

(8) a standardized energy storage device is designed, and a corresponding control device is designed for the purpose of delivering electric energy with standard quality.

(9) Based on a non-cooperative dynamic game theory, providing comprehensive energy body optimization operation, realizing comprehensive benefit maximization, constructing an energy simulation transaction control system which takes green low-carbon value as a guide active management and global optimal strategy, and realizing the full life cycle management of an energy transaction combined bill;

(10) establishing a multi-energy-station energy network simulation model facing to energy supply and energy supply difference, realizing a multi-energy-station cluster regulation and control simulation method considering interactive transaction, and realizing county-area comprehensive energy system cluster regulation and control simulation evaluation software module;

(11) based on the content of (10), a comprehensive energy whole industrial chain transaction data sharing technology based on a cloud platform is established, and comprehensive energy transaction data analysis service is provided for governments at all levels; the method is oriented to new energy increasing customers, and provides energy operation and maintenance service through modes of investment, total package construction, financing lease and the like; the system is oriented to the storage energy utilization customers, provides services such as energy collection monitoring, energy utilization scheme optimization, energy-saving effect sharing and the like, and develops commercial profit model innovation practice of comprehensive energy service.

Examples are: power marketization hedging analysis of source side and network side

Basic logic: seasonal regular risk hedging, price fluctuation alternately formed in light and busy seasons.

And (5) carrying out hedging analysis on the marketization risk of the power generation side by the power grid. The latest coal-electricity linkage price mechanism is a risk hedging method based on spot goods, and requires that 100% of power grid companies and power generation enterprises purchase electric coal from spot goods markets and cannot provide price signals reflecting long-term supply and demand relations. Once a long-term drop situation occurs, long-term adverse effects are caused to the power grid enterprises and the power generation enterprises.

Therefore, through analysis, the futures price trends of the power coal and the coking coal are analyzed by adopting the futures contract of the power coal and the futures contract of the coking coal, and the futures price discovery function is utilized to reasonably participate in the futures of the power coal and the coking coal according to the seasonal rule and statistical arbitrage regression, so that the risk of coal charging and electric linkage can be avoided.

Because industrial electricity is far higher than civil electricity, electricity utilization enterprises take industrial electricity utilization enterprises as research objects, coal enterprises as special electricity utilization enterprises for additional discussion, and power generation enterprises mainly analyze thermal power generation. In the actual social production system, the relationship is as follows: coal enterprise-thermal power plant-power grid transmission and distribution-industrial power utilization enterprise

According to the production supply and demand relationship, there is an associated combination that can embody electricity prices. The seasonal operation rules of the chemical prices of the coal steel coke can be analyzed by applying a seasonal chart method, the strength relation of the power coal and the prices of the upstream and downstream of the power coal in different months in a year is summarized, and the seasonal conduction among the power coal, the power coal and the upstream and downstream of the power coal and the downstream of the power coal in different months in the year and the.

Claims (10)

1. A comprehensive energy cluster coordination control method for improving power grid stability is characterized by comprising the following steps: the method comprises an anti-interference state observation method for improving system stability, a measurement-correlation-prediction MCP modeling and evaluation method, a source network and load cluster coordination fast response anti-interference control method, a multi-energy power generation comprehensive energy primary frequency modulation method, a digital source/network/load model, a comprehensive energy coordination control optimization simulation platform, a comprehensive energy power market trading risk hedging method based on an energy space-time game theory and a seasonal law, and a comprehensive energy cluster multi-interest subject optimization operation and trading strategy based on a non-cooperative dynamic game.

2. The integrated energy cluster coordination control method for improving the power grid stability according to claim 1, is characterized in that an anti-interference state observation method for improving the system stability is used for online data modeling by utilizing an ESO observer, and the observer, a modeling model and an inverse modeling model are operated in a discretization mode; the disturbance rejection state observation method has universality, is suitable for a source side system, is also suitable for a network side system, is also suitable for a cold/heat/electric coupled multi-energy power generation load side system, and is also suitable for various system modeling and disturbance rejection analysis.

3. The measure-correlation-predict MCP modeling and evaluation method of claim 1, wherein the measure-correlation-predict MCP method predicts long-term energy flow data of a measured location by correlating wide area integrated energy short-term source network charge energy flow data of the measured location with contemporaneous source network charge energy flow data of a reference location, wherein the measured location refers to a modeled and evaluated target site; the reference position refers to a data source station with long-term actual measurement data and simulation data enhanced excitation fusion; when the MCP method is used, the resource characteristic values at the measurement position and the reference position have strong correlation;

(1) technical requirements of MCP method

When using the MCP method, the following requirements should be observed:

the coordinates of the reference position and the measurement position, the data source and the effectiveness of the source network load energy flow characteristic value are recorded, and data cleaning is realized based on an anti-interference state observation method for improving the system stability; for long-term data, the technical requirements on boundary conditions in a source network load energy flow data model are met, namely

1) On-site data of the wide area comprehensive energy source network load power generation unit; the method comprises the steps of containing real-time mass data of the wide area source network charge energy flow; an industrial data full-process automatic checking device is developed, data cleaning and conversion are carried out, the reliability of source data is improved, a time constant and characteristic parameters of a unit are identified, a real-time transfer function model is determined, the problem of online data utilization of a traditional simulation model is solved, and the data level reaches ten thousand points per second;

2) the power grid L FC frequency control simulation data come from a power grid L FC frequency response system model and comprise frequency control response characteristics and dynamic characteristic data of primary frequency modulation and secondary frequency modulation, and the data level reaches one second to thousands of points;

3) synthesizing energy simulation data; the simulation data of the comprehensive energy unit comprise a gas turbine, wind power and photovoltaic; the micro-grid simulation laboratory actual physical device and the simulation model comprise dynamic operation data of the frequency characteristics of a gas turbine, wind power and photovoltaic new energy, the data level is one second to thousands of points,

4) reporting data after a wide area comprehensive energy source network load energy flow model in a wider range;

5) combining the four data;

if the boundary condition data comes from the observed data, eliminating error and invalid data, eliminating any known system error, wherein the time span of the observed data is more than 2 years, the data qualification rate reaches more than 70% of the observed total amount, and if the observed data provided by different units is used, the data source, the observation condition and the processing method are determined;

(2) establishing MCP classification correlation model

Establishing an MCP classification correlation model by adopting a scattered point fitting model; describing a definite function relation of the target station energy flow sequences at the same time with respect to the reference station energy flow sequences, and fitting by using a least square method, a moment estimation method, a maximum likelihood estimation method or other mathematical statistics methods to obtain a correlation function model of the two energy flow sequences;

(3) correlation coefficient

When selecting the reference station, the energy flow sequences of the target station and the reference station are required to have higher correlation coefficients; when the correlation coefficient is used for representing the similarity degree of the inter-station wind characteristics, the commonly used correlation coefficient is a Pearson linear correlation coefficient R and a Spearman rank correlation coefficient RS; because the cold/heat/electric coupling comprehensive energy flow belongs to fluid due to inconsistent time scale and physical distance characteristics, and the energy flow behavior between two stations has time delay according to the characteristics of the cold/heat/electric coupling energy flow, the resolution and delay time of measurement data used in the MCP process can cause certain influence on the correlation coefficient of the energy flow between the stations, and the correlation coefficient of the energy flow between the stations can be improved by introducing time delay;

in order to improve the prediction precision, direction grouping needs to be carried out on energy flow data, and correlation and prediction are respectively carried out;

(4) scattered point fitting model

The basic idea of the scattered point fitting model is that in an energy flow scattered point data graph, a functional relation between a target station energy flow and a reference station energy flow is obtained through a fitting method; according to the practical situation, the functional relation between the energy flows between the stations is in the form of a first-order linear function, a second-order or multi-order function, an exponential type or a power function; this patent adopts the linear regression method, and the linear regression method assumes that the energy flow of target comprehensive energy scene satisfies linear distribution about the energy flow of reference station, satisfies promptly:

y＝βx+b

wherein x is the reference station energy flow, y is the target station energy flow, β and b are the slope and intercept of the linear model respectively;

(5) method of variance ratio

Representing the slope and intercept of a linear regression model by means of the mean and variance ratio of the synchronization data between the stations

In the formula, mu_x,μ_yAnd σ_x,σ_yRespectively the average value and the standard deviation of the short-term energy flows of the reference station and the target station;

in order to achieve a more refined fit to the linear energy flow model, the data may be segmented in groups according to the diaphysics.

4. The integrated energy cluster coordination control method for improving power grid stability as claimed in claim 1, wherein the source grid load cluster coordination fast response disturbance rejection control method comprises a SVM-ARIMA time sequence fitting method based on online data verification preprocessing and on measurement-correlation-prediction MCP model and a multi-energy power generation coordination optimization control method based on an integrated energy cluster cloud interaction technology.

5. The comprehensive energy cluster coordination control method for improving power grid stability according to claim 4, wherein the process of the enhanced excitation modeling method based on online data verification preprocessing is as follows: providing a general industrial system online data enhancement excitation check preprocessing method, and calibrating and improving the signal precision through normalization and filtering noise reduction preprocessing; designing a signal with continuous excitation characteristics and full excitation of dynamic characteristics by combining test data of the conventional simulation system, determining order and time delay parameters of an actual object by relying on an enhanced excitation dynamic simulation platform, extracting a simulation model aiming at a production test flow, realizing the combination of online data and a simulation modeling technology, and solving the problem of mismatching of multi-time scale characteristics of a multi-energy power generation unit;

providing a comprehensive energy cluster power supply and energy utilization prediction and aggregation perception model based on multivariate data analysis, scientifically and reliably describing correlation relations between response time, response rate and response precision of a target station and a reference station on probability density distribution or time sequence in a form of a mathematical model based on the measurement-correlation-prediction MCP modeling and evaluation method of claim 3 and an MCP correlation model; fitting a prediction model of measurement-correlation-prediction MCP by using an SVM-ARIMA time characteristic sequence model, and displaying the scheduling available capability of the in-network multi-energy power generation unit to a scheduling center in a visual form in real time, wherein the indexes are a power grid frequency response regulation rate K1, a power grid frequency response regulation precision K2 and a power grid frequency response time K3; then, determining a source network load energy flow coordination feedforward fast response control method based on time sequence prediction;

the ARIMA and the SVM have advantages in processing the linear model and the nonlinear model respectively, the advantages of the ARIMA and the SVM are complementary, the ARIMA and the SVM are combined for price prediction, and a better result is received;

assuming that the time series Y can be considered as a combination of the linear autocorrelation portion L and the nonlinear residual N, i.e., Y L + N, the following steps are to be taken herein to construct a combined prediction model:

(1) modeling a linear part by utilizing an ARIMA model, setting a prediction result as L, setting residual errors of a sequence assembly as N and N, and including a nonlinear relation of a sequence Y;

(2) reconstructing the N sequence obtained in the previous step to obtain an SVM sample set, and predicting a residual error by using an SVM to obtain a prediction result N;

(3) l obtained by linear prediction is combined with N obtained by a nonlinear set to obtain a prediction result Y, which is L and + N;

according to the fast response requirement of the power grid frequency, determining the main response characteristic of the source grid load energy flow as response time, and improving the response capability of the power grid frequency scheduling instruction in real time; finally, a power grid frequency response control technology online hardware rapid deployment method is realized, an external-hanging controller is developed, and the problems that a big data technology cannot be practically applied and popularized in a multi-energy power generation comprehensive energy system and intelligent scheduling is realized are solved.

6. The comprehensive cluster energy coordination control method for improving power grid stability according to claim 4, wherein the multi-energy power generation coordination optimization control method based on the comprehensive energy cluster cloud interaction technology comprises the following steps:

step 1, developing cloud-edge-end information interaction standardized modeling, establishing a cloud platform model, a cloud-edge interaction model, an edge-end interaction model and an end equipment model, and realizing the basis of cloud distributed computing;

step 2, developing a miniaturized, modularized and passive multi-energy intelligent information sensor, realizing an open and universal access technology suitable for cross-region, large-range and scene energy supply and utilization nodes, realizing massive multi-dimensional data collection of coupling energy sources such as cold/heat/electricity and the like, and quickly sensing the real-time running state of a comprehensive energy body;

step 3, giving play to the advantages of 5G fast networking and high efficiency, providing a multi-stream convergence and multi-mode heterogeneous cooperative communication protocol conversion technology based on 5G, namely a county comprehensive energy system information exchange case based on UM L, developing a comprehensive energy cluster system communication protocol, supporting REST and MQTT information exchange modes, realizing cloud-edge-end communication protocol standardization, providing a service scene cloud end customization and unified management technology aiming at the characteristic of difference of application scene requirements of different comprehensive energy bodies, and realizing cloud end seamless switching and dynamic interaction among comprehensive energy cluster scenes;

step 4, realizing a multi-energy power generation unit coordination control algorithm library; the method comprises the steps of realizing a cluster control method facing energy requirements and operation and maintenance requirements based on a distributed observer, providing a distributed optimization control algorithm with super-linear convergence speed, deploying the algorithm to an intelligent controller to form local closed-loop control, providing a cluster high-efficiency distributed feedback optimization control technology for decomposition and coordination, and forming cluster closed-loop control by a cloud platform and a plurality of intelligent controllers;

step 5, based on the coordination control algorithm library in the step 4, realizing L ADRC linear active disturbance rejection control strategy of multi-energy power generation by taking an energy balance signal as a feedforward signal, and replacing a traditional PID control module with L ADRC, wherein the control strategy is realized strictly according to DCS/P L C control logic paging, and the control strategy runs in a stream mode on an NI Compact RIO controller in a 100/200MS scanning period, so that a plug-and-play autonomous high-performance intelligent controller is further developed, a cloud platform control instruction is received, and internal equipment of the comprehensive energy body is adjusted to realize energy and operation and maintenance demand response;

step 6, realizing active defect identification and risk alarm facing coupled energy sources such as cold/heat/electricity and the like and complex network massive operation data based on a cloud platform, and developing decarbonization economic operation evaluation;

step 7, carrying out research aiming at the characteristics of large service span, wide range and the like of the wide-area comprehensive energy body, and realizing the on-line prediction of the energy supply demand of each comprehensive energy body; and establishing a monthly/quarterly multi-scene global operation and maintenance optimization decision mechanism of the comprehensive energy body, and providing specialized high-level operation and maintenance service for users in the region.

7. The integrated energy cluster coordination control method for improving power grid stability according to claim 1, characterized in that: a comprehensive energy primary frequency modulation method aims at the problem that a multi-energy power generation microgrid participates in frequency modulation coordination control of a thermal power generation power system, and comprises the following steps:

step 1, a multi-energy power generation micro-grid participates in modeling of a thermal power generation/gas turbine coordinated control frequency modulation control architecture and analyzes the influence of the modeling;

step 2, researching and verifying the multi-energy power generation micro-grid participating thermal power generation/gas turbine coordination prediction control algorithm;

step 3, developing a cooperative control device suitable for a multi-energy power generation micro-grid to participate in frequency modulation of a thermal power generation/gas turbine power system;

and 4, building a regional interconnection system control platform containing the renewable energy intelligent micro-grid and participating in thermal power generation/gas turbine.

8. The integrated energy cluster coordination control method for improving the power grid stability as claimed in claim 1, wherein a digitalized source/grid/charge model and an integrated energy coordination control optimization simulation control platform use power system simulation software such as StarSim/RT L ab to build a source/grid/charge power system model comprising a thermal power generation coordination control system, a gas turbine control system, an energy storage system or a pumped storage and microgrid load system;

aiming at a high-capacity unit coordination optimization control technology and a primary frequency modulation technical route under the condition of ultra-high voltage power grid operation, aiming at the on-line hardware closed loop simulation technology development of source grid load coordination control performance, a hybrid logic dynamic programming method or an intelligent method is used for simulating a digital source/grid/load model and a comprehensive energy source energy coordination optimization control technology according to the overall system architecture and parameters of a thermal power generation boiler system, a steam turbine system, a gas turbine control system, an energy storage system and a multi-energy power generation micro-grid load system;

the method applies a mixed integer linear programming method, a genetic algorithm and an improved complex process global optimization evolutionary algorithm to the energy coordination optimization management of the source/network/load.

9. The integrated energy cluster coordination control method for improving power grid stability according to claim 7, characterized in that: the process of establishing the energy coordination optimization management model objective function of the source/network/load system is as follows:

on the basis of ensuring the power supply of the local load, the minimum operation cost of a source/network/load is taken as a target, wherein the minimum operation cost comprises the cost of purchasing electricity from the power grid, the income obtained by selling the electricity to the power grid, and the maintenance cost and depreciation loss of a storage battery;

in the formula, eSell (t) is the real-time electricity purchasing price of a power grid, eBuy (t) is the real-time electricity selling price of the power grid, ebat (t) is the operation management cost of a storage battery, PgBuy (t) is the electric power absorbed by the large power grid at the t moment, the sign is negative, PgSell (t) is the electric power generated by the large power grid at the t moment, the sign is positive, Pbat (t) is the active power of the storage battery at the t moment, △ t is the system operation time interval, and the value is 1 hour;

the objective function comprises the cost for purchasing electricity from the power grid and the income obtained by selling electricity to the power grid, and how to use Pg (t) to represent the main grid to source/grid/load output power when the value is positive, and the input power when the value is negative, which is specifically represented by the following formula;

wherein, when Pg (t) is positive, the electricity purchasing cost is expressed as eSell (t) and Pg (t); when Pg (t) is a negative value, the electricity selling cost is expressed as-eBuy (t) and Pg (t);

adopting a predictive control framework based on rolling time domain optimization, solving an optimization problem with minimum consumption cost from an external power grid when a model operates in a rolling time domain [ tt + tp ] at the moment t, and minimizing a target function in the rolling time domain by calculating an optimal control sequence in the [ tt + tp ]; on the basis of the above formula, a rolling optimization range is added, rolling is performed by taking one hour as a period, the step length of the rolling range is assumed to be tp, and the consumption cost in the time from t to t + tp is calculated and used as a new objective function, which is expressed as:

under the new objective function, the rolling time domain not only considers the current step, but also puts the system operation state in the future period into the calculation range.

10. The integrated energy cluster coordination control method for improving power grid stability according to claim 1, characterized in that: the comprehensive energy electric power market trading risk hedging method based on the energy space-time game theory comprises the following steps:

(1) firstly, classifying users in the power market into three categories: large electricity users and medium-scale users; collecting historical data related to an integrated energy trading market from the establishment of varieties by using an existing platform, and summarizing and inducing seasonal rules of arbitrage combination according to different time scales by using the SVM-ARIMA characteristic sequence prediction method and the trading data characteristic sequence pattern recognition technology in claim 5;

(2) an open universal energy transaction mechanism based on a block chain is provided, point-to-point energy transaction settlement of green energy credit is realized, user and service resources are integrated, an intelligent contract with simplified flow and safe transaction is deployed, and transaction matching and contract management are realized;

(3) combining the seasonal regularity of commodities of the comprehensive energy phase power market, analyzing and counting technical indexes such as volume of transaction, quantity of taken positions, time scale, MACD deviation and the like and relevant game characteristics based on the volume energy space-time characteristics of the wide-area comprehensive energy flow and the service flow, exploring a commodity futures hedging implementation method, carrying out backtracking test on a trading system, evaluating the advantages and disadvantages of different strategies, and proposing a trading strategy for improving the trading rate to more than 50%;

(4) based on the average lines and golden section points of different periodic scales, a win-loss control technology for calculating transactions is researched and calculated, entrance and stand management with optimal cost performance is realized, quantification of profit and loss prevention is realized, a backtracking test is carried out on a transaction system, and the advantages and the disadvantages of different strategies are evaluated;

(5) calculating correlation degrees of hedge arbitrage varieties of statistical arbitrage, developing regression analysis, calculating a win damage threshold value, performing statistical induction on each hedge arbitrage combination from different time scales, performing backtracking test on a trading system, and evaluating the performance of different strategies by using a mathematical statistical method, adopting ideas such as a normal distribution mathematics original method, a least square method and the like;

(6) a tool of a CTP interface is used for developing a commodity futures arbitrage hedging monitoring software system, and a method for realizing 90% arbitrage hedging variety combined coverage is researched;

(7) a commodity futures arbitrage hedging method related to a comprehensive energy power market based on seasonal regularity, and a rolling multi-strategy combination hedging transaction risk management system for improving the profit rate of the repeated profit;

(8) designing a standardized energy storage device, and designing a corresponding control device with the aim of delivering electric energy with standard quality;

(9) based on a non-cooperative dynamic game theory, comprehensive energy body optimization operation is provided, comprehensive benefit maximization is realized, an energy simulation transaction control system which takes green low-carbon value as a guide active management and global optimal strategy is constructed, and full life cycle management of an energy transaction combined bill is realized;

(10) establishing a multi-energy-station energy network simulation model facing to energy supply and energy supply difference, realizing a multi-energy-station cluster regulation and control simulation method considering interactive transaction, and realizing county-area comprehensive energy system cluster regulation and control simulation evaluation software module;

(11) based on the content of (10), establishing a comprehensive energy whole industrial chain transaction data sharing technology based on a cloud platform, and providing comprehensive energy transaction data analysis service for governments at all levels; the energy operation and maintenance service is provided for new energy increase customers through investment, total package construction and financing lease modes; the system is oriented to the storage energy utilization customers, provides energy collection monitoring, energy utilization scheme optimization and energy-saving effect sharing service, and develops commercial profit model innovation practice of comprehensive energy service.

Images