JP2023179851A - Device state evaluating system and device state evaluating method - Google Patents

Abstract

To provide a device state evaluating technique that can evaluate a presence or absence of an abnormality in a device from actual measurement data of an appliance, even if the device includes the appliance which has not experienced the abnormality in the past.SOLUTION: A device state evaluating system 1 comprises: an analysis unit 10 that performs a calculation by reflecting a parameter in an equation formulated based on a physical model; a physical quantity measuring unit 12 that acquires a physical quantity actually measured by each appliance 21 as measurement data; a setting unit 16 that sets reference information for determining a state of the device 20; and a state determining unit 15 that determines the state of the device 20 based on reference information when the measurement data is reflected in the parameter of the physical model and the measurement data is analyzed.SELECTED DRAWING: Figure 1

Description

Embodiments of the present invention relate to device condition evaluation techniques.

BACKGROUND ART Conventionally, there is a technique for determining the presence or absence of abnormality in a motor or estimating the cause from measurement data detected by a sensor. For example, there is a technique that determines the presence or absence of an abnormality by comparing measured data with a preset threshold value.

Patent No. 6968323 Patent No. 6316510

When setting appropriate thresholds, utilize data related to past abnormalities. For example, when determining the presence or absence of abnormality symptoms or the cause of an abnormality, abnormality data obtained when the abnormality occurs is required. However, since it is difficult to obtain abnormality data from devices that rarely experience abnormalities, it is difficult to determine abnormalities from measurement data. For example, important equipment such as large plants that may be shut down for a long period of time, or equipment that is important for safety such as nuclear power plants, is periodically overhauled and repaired or replaced before an abnormality occurs. be exposed. Therefore, there are extremely few cases of abnormalities occurring. Furthermore, various abnormal modes may occur in one device, and data is required for all such abnormal modes.

For electric motors, technology may be used to monitor vibration and determine signs of abnormality from changes in level and frequency characteristics. Judgments about the presence or absence of abnormal signs are made by inspectors based on their experience, and the validity of the judgment results is poor. Further, changes in vibration depend on the frequency characteristics of the equipment or the installation state of the sensor. Therefore, with diagnosis based only on predetermined data, it is difficult to determine the presence or absence of an abnormality or to estimate the cause.

The embodiments of the present invention have been made in consideration of such circumstances, and are capable of evaluating whether or not there is an abnormality in the device based on actual measurement data of the device, even if the device includes a device that has not experienced abnormalities in the past. The purpose is to provide equipment condition evaluation technology that enables

A device state evaluation system according to an embodiment of the present invention reproduces the operation of a device composed of a plurality of devices including at least a rotating electric machine, and a physical model that reproduces the state of the device that occurs in connection with the operation. A model construction unit that constructs a model, an analysis unit that performs calculations by reflecting parameters in equations formulated based on the physical model, and physical quantity measurement that acquires physical quantities actually measured by each of the devices as actual measurement data. a setting unit configured to set reference information for determining the state of the device; and a setting unit configured to set reference information for determining the state of the device; a state determination unit that determines the state of the device based on the state of the device.

Embodiments of the present invention provide a device state evaluation technique that can evaluate the presence or absence of an abnormality in a device based on actual measurement data of the device, even if the device includes a device that has not experienced any abnormality in the past.

FIG. 2 is a block diagram showing the system configuration of the device status evaluation system. FIG. 2 is a configuration diagram showing a device status evaluation system connected to the device. FIG. 2 is an explanatory diagram showing a mode of converting a physical quantity acquired from a sensor into data for analysis. FIG. 4 is an explanatory diagram showing a first example of poor coupling installation. Explanatory diagram showing a second example of poor coupling installation. An explanatory diagram showing a mode of learning based on parameters, calculation results, and abnormal labels. An explanatory diagram showing an example of setting an abnormality label. A graph showing the relationship between fatigue and time. A graph showing the relationship between stress amplitude and number of rupture cycles. 5 is a flowchart showing a device state evaluation method. FIG. 7 is an explanatory diagram showing a modification example of setting a threshold value.

Hereinafter, embodiments of an apparatus condition evaluation system and an apparatus condition evaluation method will be described in detail with reference to the drawings.

Reference numeral 1 in FIG. 1 is an apparatus status evaluation system of this embodiment. This device condition evaluation system 1 supports maintenance activities for the device 20 by optimizing the determination of the presence or absence of an abnormality in a predetermined device 20 (FIG. 2), the estimation of the cause of the abnormality, and the like. In particular, the device condition evaluation system 1 is used to evaluate the condition of the device 20 based on actual measurement data for devices 20 that have not experienced abnormalities in the past. Note that the state of the device 20 includes at least one of the presence or absence of an abnormality, the presence or absence of an abnormality sign, an abnormality event, an abnormality cause, an abnormality degree, and an abnormality type.

The device status evaluation system 1 of this embodiment has hardware resources such as a CPU, ROM, RAM, and HDD, and when the CPU executes various programs, information processing by software is realized using the hardware resources. It consists of computers. Furthermore, the device state evaluation method of this embodiment is realized by causing a computer to execute various programs.

First, the system configuration of the device condition evaluation system 1 will be explained with reference to the block diagram shown in FIG.

The device condition evaluation system 1 includes a plurality of sensors 2 and an analysis computer 3. The analysis computer 3 includes an input section 4, an output section 5, a communication section 6, a storage section 7, and a main control section 8.

Predetermined information is input to the input unit 4 in response to an operation by a user using the analysis computer 3. This input unit 4 includes an input device such as a mouse or a keyboard. That is, predetermined information is input to the input unit 4 in response to operations on these input devices.

The output unit 5 outputs predetermined information. The device condition evaluation system 1 of this embodiment includes a device that displays images, such as a display that outputs analysis results. That is, the output unit 5 controls the image displayed on the display. Note that the display may be separate from the computer main body, or may be integrated with the computer main body.

Note that the device status evaluation system 1 of this embodiment may control images displayed on a display included in another computer connected via a network. In that case, the output unit 5 included in another computer may control the output of the analysis results of this embodiment.

Note that in this embodiment, a display is exemplified as a device that displays an image, but other embodiments may be used. For example, images may be displayed using a head-mounted display or a projector. Furthermore, a printer that prints information on a paper medium may be used in place of the display. That is, the objects controlled by the output unit 5 may include a head-mounted display, a projector, or a printer.

The communication unit 6 communicates with other computers via communication lines such as the Internet. Note that in this embodiment, the device status evaluation system 1 and another computer are connected to each other via the Internet, but other embodiments may be used. For example, the device status evaluation system 1 and another computer may be connected to each other via a LAN (Local Area Network), a WAN (Wide Area Network), or a mobile communication network.

The storage unit 7 stores various types of information necessary for evaluating the state of the device 20. Note that this storage unit 7 includes a predetermined database. These are collections of information stored in memory, HDD, or cloud and organized so that they can be searched or stored.

The main control unit 8 includes a model construction unit 9, an analysis unit 10, a learning unit 11, a physical quantity measurement unit 12, an actual measurement data processing unit 13, a parameter identification unit 14, a state determination unit 15, a setting unit 16, and a remaining life evaluation unit 17. and an inspection estimation section 18. These are realized by the CPU executing programs stored in the memory or HDD.

Note that each component of the analysis computer 3 does not necessarily need to be provided in one computer. For example, one analysis computer 3 may be realized by a plurality of computers connected to each other via a network. For example, the learning section 11 may be provided in a computer different from the analysis computer 3.

As shown in FIG. 2, an example of the device 20 to be diagnosed in this embodiment is one installed in a nuclear power plant or the like. The device 20 is composed of a plurality of devices 21. For example, the device 20 includes a control panel 21A, an electric motor 21B, a rotating body 21C, a load-side equipment group 21D, and the like.

Here, the control panel 21A corresponds to a component (control unit) related to controlling the behavior of the electric motor 21B. For example, the control panel 21A controls the rotational speed or torque of the electric motor 21B.

Further, the rotating body 21C is a component that transmits the energy output from the electric motor 21B to the load-side equipment group 21D.

The load-side equipment group 21D corresponds to a component that consumes the energy transmitted from the rotating body 21C. Note that the load-side equipment group 21D includes, for example, a pump, a fan, a gear, and the like.

Here, the electric motor 21B is illustrated as the rotating electric machine of this embodiment. In other words, the device 20 is composed of a plurality of devices 21 including at least a rotating electric machine. Note that the rotating electric machine may be a generator. This embodiment can be applied whether the rotating electric machine is the electric motor 21B or a generator.

For example, assume that there is a predetermined frequency parameter regarding a rotating electric machine. This frequency may be either the frequency of the electric power output from the rotating electric machine or the frequency of the electric power supplied to the rotating electric machine. When the rotating electric machine is the electric motor 21B, this is the frequency of the alternating current supplied to the electric motor 21B from the outside. Furthermore, if the rotating electric machine is a generator, the frequency is the frequency of alternating current output from the generator to the outside.

Note that each sensor 2 of the device condition evaluation system 1 is installed in the device 20. The sensor 2 detects at least one of vibration, current, voltage, magnetic flux, temperature, pressure, rotational torque, and load as a physical quantity. In this way, the physical quantity can be acquired by the sensor 2 for any type of device 21.

For example, the sensor 2 is provided in the control panel 21A, the electric motor 21B, the rotating body 21C, and the load-side equipment group 21D. Predetermined information acquired by these sensors 2 is input to the analysis computer 3. Note that the sensor 2 may not only be installed in the device 20 in advance, but also be installed each time an inspection is performed. Furthermore, the installation location of the already installed sensor 2 may be changed.

These sensors 2 detect physical quantities that change depending on the state of the device 20. Specific examples of the physical quantities include the current and voltage in the control panel 21A and the electric motor 21B, the temperature of the electric motor 21B, the vibration of the electric motor 21B and the rotating body 21C, the magnetic flux, load, and torque of the electric motor 21B.

In this embodiment, the model construction unit 9 (FIG. 1) constructs a physical model that reproduces the operation of the device 20 and reproduces the state of the device 21 that occurs in relation to the operation. Note that the physical model is a virtual model that can simulate physical quantities related to the device 20. By using the physical model, it is possible to not only simulate the device 21 that is operating normally, but also the device 21 that is operating abnormally.

The physical model reproduces the behavior of the device 20 through a combination of parameters. For example, the physical model includes at least one of a mechanical model, a control model, an electrical model, a magnetic model, a thermal model, a fluid model, a physical property model, and a structural model.

Here, the dynamic model represents the behavior of a dynamic system regarding at least one of rotational motion and translational motion of each device 21, such as the electric motor 21B, the rotating body 21C, and the load-side device group 21D. The control model is for a control panel 21A serving as a control unit constituting at least one device 21 to control the behavior of the entire device 20. The electrical model represents the behavior of the electrical system of device 20. The magnetic model represents the behavior of the magnetic system of device 20. The thermal model represents the thermal behavior of device 20. The fluid model represents the behavior of at least one of the flow rate and pressure of a fluid flowing through piping connected to a pump constituting at least one device 21. The physical property model represents the material properties of each device 21. The structural model has information on at least one of the dimensions and shape of each device 21. In this way, the behavior of the device 20 can be precisely reproduced according to various models.

In other words, the physical model is one in which the structure or function of the device 20 is modeled and formulated. A group of equations corresponding to these has been formulated.

Further, the physical model includes an abnormality model that indicates an abnormality in each device 21 that constitutes the device 20. For example, as shown in FIGS. 4 and 5, consider a coupling as the rotating body 21C. The effect of eccentricity e _c or deviation angle d _c from the center of rotation caused by poor installation of the coupling is formulated as a model. Furthermore, parameters representing the degree of abnormality, such as the magnitude of the eccentricity e _c or the deviation angle d _c , are defined.

Further, as a part of the model, a mathematical model in which inputs and outputs are associated may be used. In this way, the functions of devices for which there is no design information can be incorporated into the model and analyzed. Note that the data necessary for constructing the mathematical model may be actual machine data or may be obtained in advance through element tests. Alternatively, results of analysis conducted separately may be used.

In addition, abnormalities in the rotating body 21C include misalignment of the bearing, breakage of the rotor bar or short-circuit ring of the motor 21B, protrusion of the rotor bar, protrusion of the coil in the stator or slot, bending or cracking of the rotating body 21C, and parts. These include missing parts, residual unbalance, stator winding short circuit or burnout, power supply voltage imbalance, inverter irregular waves, and rolling bearing (inner ring, outer ring, rolling elements) flaws. Parameters related to each abnormality are defined.

The analysis unit 10 (FIG. 1) performs calculations by reflecting parameters on equations (group of equations) formulated based on a physical model. This calculation includes structural analysis using the finite element method (FEM) or the like.

Here, the physical quantity measurement unit 12 (FIG. 1) acquires physical quantities actually measured by each device 21 as actual measurement data. The physical quantity measurement unit 12 acquires the physical quantity detected by the sensor 2 (FIG. 1) as time series data. For example, as shown in FIG. 3, the physical quantity measurement unit 12 acquires time-series data including physical quantities indicating current, magnetic flux, acceleration, angular velocity, torque, displacement, pressure, etc. from each device 21 as actual measurement data.

In this embodiment, the physical quantity measurement unit 12 uses the sensor 2 to acquire the physical quantity in real time while the device 20 is operating, but other embodiments may be used. For example, the sensor 2 may be attached to the device 21 to obtain and record physical quantities during periodic inspection, and the physical quantity measurement unit 12 may obtain the records at a later date.

In other words, the physical quantity measuring unit 12 converts at least one of the physical quantities detected by each sensor 2 and the physical quantities included in the test data or inspection data obtained when the equipment 21 was tested or inspected in the past into actually measured data. Get as. In this way, it is possible to determine whether there is an abnormality in the device 20 based on at least one of the current physical quantity and the past physical quantity.

Note that the test data or inspection data also includes information indicating whether or not the device 21 has failed. That is, the physical quantity of this embodiment includes information indicating whether or not the device 21 has failed in the past. In addition, the physical quantity includes information indicating the date and time, location, and state of failures in the past.

Note that it is preferable that the data acquired by the physical quantity measurement unit 12 be of a plurality of types of physical quantities. For example, since vibration is affected by the characteristics of the support structure of the device 21 or the installation state of the sensor 2, it is difficult to determine whether the change in physical quantity is due to an abnormality based only on vibration data. In such cases, comprehensive judgment can be made from physical quantities such as electric current in addition to vibration.

Furthermore, the actual measurement data processing unit 13 (FIG. 1) converts the actual measurement data acquired by the physical quantity measurement unit 12 into analysis data in a format compatible with the calculations of the analysis unit 10. Here, the calculations executed by the analysis section 10 include calculations for outputting the analysis data in the same format as the analysis data output from the measured data processing section 13.

For example, as shown in FIG. 3, when the measured data processing section 13 outputs vibration spectrum data, the calculation results of the corresponding analysis section 10 are also Fourier-transformed to output vibration spectrum data. (Fast Fourier Transform: FFT) calculation is performed. Here, time series data as actually measured data is converted into data for analysis. In this way, the measured data can be converted into a format most suitable for calculation, and the analysis section 10 can perform the calculation at high speed.

The measured data processing section 13 converts, for example, measured data such as vibration, current, voltage, and magnetic flux acquired by the physical quantity measuring section 12 into spectral data by Fourier transformation. Furthermore, the actual measurement data processing unit 13 performs filter processing as necessary. Furthermore, the actual measurement data processing unit 13 performs a process of converting time series data such as temperature and pressure into maximum values, minimum values, average values, effective values, overall values, etc. as necessary. Through these processes, it is possible to extract the feature amount of the measured data and remove unnecessary noise. Further, it is possible to process errors and variations caused by measurement. For example, since there are measurement errors or statistical errors of measuring instruments, processing is performed for these errors.

The parameter identification unit 14 (FIG. 1) identifies actual measurement parameters, which are parameters when the calculation result of the analysis unit 10 reproduces actual measurement data, based on the teacher data (training data) learned by the learning unit 11. For example, the parameter identification unit 14 identifies actual measurement parameters, which are parameters when the calculation result of the analysis unit 10 reproduces actual measurement data, based on the learned learning model. In this embodiment, analysis data converted from actual measurement data is used. That is, parameters are identified based on the teacher data learned by the learning section 11 so that the calculation result of the analysis section 10 reproduces the data for analysis. In other words, the analysis unit 10 performs an analysis using a physical model so that the analysis data derived as a result of the reproduction matches the measured data.

For example, when the operation of the device 20 in an abnormal state is reproduced using a physical model, parameters at the time of the abnormality can be obtained. In this way, the parameters at the time of an abnormality can be used to determine whether or not there is an actual abnormality in the device 20.

Further, the parameter identification unit 14 infers the optimal combination of parameters for the physical model of the device 20 to reproduce the analysis data output from the measured data processing unit 13, and outputs these as measured parameters. At this time, estimation is performed taking into account errors and variations.

Furthermore, the analysis unit 10 executes calculations by reflecting the parameters identified by the parameter identification unit 14. For example, the analysis unit 10 obtains calculation results regarding time changes in the load acting on each device 21 constituting the device 20 or the position of the rotating body 21C. It also evaluates numerical errors caused by calculations.

In the device condition evaluation system 1 of this embodiment, the parameter identification unit 14 uses, for example, a learned model generated by a neural network to determine actual measurement parameters for reproducing the analysis data output from the actual measurement data processing unit 13. demand.

For example, in the case of the electric motor 21B, the parameters of this embodiment include those related to the dimensions and shape of the electric motor 21B, the number of coil turns, the number of slots, and the like. Parameters also include properties of components such as electrical resistance. Further, the parameters include the operating conditions of the device 20 or the surrounding environment. These parameters may be obtained based on the specifications of each device 21 that constitutes the device 20, may be obtained through an actual test, or may be obtained through simulation.

The learning unit 11 (FIG. 1) learns each parameter corresponding to the calculation result of the analysis unit 10 as training data. For example, the learning unit 11 performs machine learning by inputting parameters corresponding to the calculation results of the analysis unit 10 into a learning model as teacher data. The learning unit 11 inputs each parameter and the calculation results corresponding to these parameters, which are obtained when the analysis unit 10 performs calculations corresponding to a combination of a plurality of parameters in advance, to the learning model as teacher data. In this way, the accuracy of learning can be improved.

For example, a parameter when the device 20 is determined to be abnormal based on the calculation result of the analysis unit 10 is associated with an abnormality label. The learning unit 11 inputs this parameter to the learning model as teacher data. Specifically, the analysis unit 10 may perform a calculation corresponding to a combination of a plurality of parameters in advance, and the result may be determined to be abnormal in the device 20. In this case, the learning unit 11 performs machine learning by associating each parameter and the abnormality label corresponding to the calculation result corresponding to these parameters with the teacher data. In this way, the abnormality label is set based on the result of determining that the device 20 is abnormal, so that the abnormality of the device 20 can be appropriately evaluated.

As shown in FIG. 6, the learning unit 11 generates a learning model by machine learning using the results calculated in advance under various parameter conditions in the analysis unit 10 as teacher data (training data). For example, there are learning models generated by machine learning using neural networks. Here, the learning unit 11 creates a learning model using the parameters used in the calculation by the analysis unit 10 and the calculation results as training data.

The device status evaluation system 1 of this embodiment includes a computer equipped with an algorithm that performs machine learning. In other words, it includes computers equipped with artificial intelligence (AI) that performs machine learning. For example, this device condition evaluation system 1 may be configured with one computer equipped with a neural network, or may be configured with multiple computers equipped with neural networks.

Note that the device state evaluation system 1 may include a deep learning section that extracts a specific pattern from a plurality of patterns based on deep learning.

Analysis technology based on artificial intelligence learning can be used for the computer-based analysis of this embodiment. For example, a learning model generated by machine learning using a neural network, a learning model generated by other machine learning, a deep learning algorithm, a mathematical algorithm such as regression analysis, etc. can be used. Further, forms of machine learning include forms such as clustering and deep learning.

Here, the neural network is a mathematical model that expresses the characteristics of brain function through computer simulation. For example, we will show a model in which artificial neurons (nodes) that form a network through synaptic connections change the strength of synaptic connections through learning and acquire problem-solving ability. Furthermore, neural networks acquire problem-solving ability through deep learning.

For example, a neural network is provided with an intermediate layer having multiple layers. Each layer of this intermediate layer is composed of a plurality of units. Furthermore, by allowing a multilayer neural network to learn in advance using training data, it is possible to automatically extract features in patterns of changes in the state of a circuit or system. Note that for a multilayer neural network, any number of intermediate layers, any number of units, any learning rate, any number of learning times, and any activation function can be set on the user interface.

Note that deep reinforcement learning may be used in the neural network, in which a reward function is set for various information items to be learned, and an information item with the highest value is extracted based on the reward function.

For example, CNN (Convolution Neural Network), which has a proven track record in image recognition, is used. In this CNN, the middle layer is composed of a convolution layer and a pooling layer. The convolutional layer obtains a feature map by filtering nearby nodes in the previous layer. The pooling layer further reduces the feature map output from the convolution layer to create a new feature map. At this time, by obtaining the maximum value of pixels included in the region of interest in the feature map, it is possible to absorb some deviation in the position of the feature amount.

The convolution layer extracts local features of an image, and the pooling layer performs a process of combining local features. In these processes, the image is reduced while maintaining the characteristics of the input image. In other words, the convolutional neural network (CNN) can significantly compress (abstract) the amount of information contained in an image. Then, using the abstracted image stored in the neural network, the input image can be recognized and the image can be classified.

Note that machine learning includes various methods such as autoencoder, LSTM (Long Short-Term Memory), SDF (Signed Distance Function), GAN (Generative Adversarial Network), and RNN (Recurrent Neural Network). These methods may be applied to the machine learning of this embodiment.

The setting unit 16 sets reference information for determining the state of the device 20. The reference information of this embodiment is set based on the calculation result of the analysis unit 10, and includes at least one of the following: presence or absence of an abnormality in the device 20, presence or absence of an abnormality sign, abnormality event, abnormality cause, abnormality degree, and abnormality type. Contains an anomaly label indicating one.

For example, the setting unit 16 sets an abnormality label for determining whether or not there is an abnormality in the device 20. By performing machine learning using this abnormality label, the learning model is an analysis result corresponding to a combination of multiple parameters, and whether or not the calculation result of the analysis unit 10 indicates an abnormal state of the device 20. Be able to judge.

The setting unit 16 sets an abnormality label based on the combination of parameters when the device 20 is determined to be abnormal and the calculation result of the analysis unit 10. In this way, abnormalities in the device 20 can be appropriately evaluated.

As a criterion for judgment, the results of calculations performed on predetermined parameters in the analysis unit 10 are based on requirements such as functional requirements, material properties, and laws and regulations (legal restrictions) for the device 20 composed of a plurality of devices 21. The question is whether or not it satisfies the requirements.

Here, the functional requirements include, for example, rotation speed, torque, flow rate, pressure, drive speed, response, temperature, voltage, and current. Furthermore, examples of material properties include fatigue life, thermal cycle life, plastic deformation, and load leading to breakage. Examples of laws and regulations include legal restrictions on sound volume, chemical substance emissions, temperature, and the like.

The setting unit 16 of the present embodiment sets an abnormality label (reference information) based on whether at least one of the required functions of the device 20, material properties, and laws and regulations is satisfied. In this way, in determining whether or not there is an abnormality in the device 20 that includes the equipment 21 that has not experienced any abnormality in the past, it becomes possible to set a rational evaluation standard instead of a standard based on the experience of the inspector.

Note that the abnormality label may be set not only for the calculation result of the analysis unit 10 but also for the parameters of the physical model. In this way, it is possible to determine whether the device 20 is abnormal based on the parameters identified by the parameter identifying section 14.

There are variations, errors, etc. in sensing, simulation, parameter identification, and material properties. Therefore, the abnormality label is set in consideration of the statistical distribution. Note that the abnormality label may be set with a margin given to the required conditions. In this way, the symptoms of an abnormality can be evaluated immediately before it occurs, and the device 21 can be repaired or replaced.

The remaining life evaluation unit 17 (FIG. 1) calculates the remaining life of each device 21 based on the calculation results of the analysis unit 10 when the parameters whose actual measurement data has been identified by the parameter identification unit 14 are reflected in the physical model. do. In this way, it is possible to evaluate the remaining life for a combination of multiple parameters.

The remaining life evaluation unit 17 also calculates the load, stress, voltage, current, and temperature associated with each device 21 obtained by the analysis unit 10 when the measured parameters identified by the parameter identification unit 14 are reflected in the physical model. The remaining life is evaluated based on the result of determining at least one of the following. In this way, the remaining life of the device 21 can be evaluated in consideration of the load or stress applied to the device 21.

For example, the remaining life evaluation unit 17 calculates the time changes in the load and stress of each device 21 constituting the device 20 by calculation, and determines the remaining life from the life evaluation formula and material properties. In addition, in this embodiment, it is also possible to judge the remaining life not only from load and stress but also from the deterioration of insulating materials due to time changes in voltage or current, and the deterioration of materials due to thermal cycles (time changes in temperature). It has become.

As a method for determining the abnormality label, for example, as shown in FIG. 7, it is assumed that vibrations due to centrifugal force occur due to misalignment of the bearing of the device 21 or unbalance of the rotating body 21C. Here, there is a basic rating life L10 when a radial load P is applied to the rolling bearing that supports the rotating body 21C. This radial load P is added to a combination of multiple parameters based on a physical model. Then, the radial load P is calculated by the analysis section 10. For example, the basic rating life L10 of a rolling bearing is calculated using the following remaining life evaluation formula.

L10=(C/P) ^p

Here, C is the basic dynamic load rating, which represents the dynamic load capacity of the rolling bearing. For example, it refers to a constant load that provides a basic rating life L10 of 1 million revolutions. Further, p is an index that changes depending on the type of rolling bearing. For example, in the case of a ball bearing, p=3. In the case of roller bearings, p=10/3.

Based on the calculation result of this radial load P, an abnormality label is set for the combination of parameters and the calculation result when the life of the rolling bearing no longer satisfies the required operating period. At this time, the abnormality event and information regarding its cause may be associated with each other in the abnormality label. In this case, the abnormal event can be set as an abnormality in the target bearing. Furthermore, the cause of the abnormality can be set from a combination of parameters. In this way, when an abnormality occurs, the location and cause of the abnormality can be quickly estimated.

As shown in FIG. 7, for example, the cause of the abnormality can be selected based on the magnitude relationship between parameters. Alternatively, the correlation between the cause of the abnormality and the parameters may be calculated by the analysis unit 10 performing a sensitivity analysis or the like in advance. In this way, it is possible to improve the accuracy of estimating the cause when an abnormality occurs.

In addition, when the rotating body 21C is connected to the electric motor 21B, the physical quantity acquired by the sensor 2 may be a physical quantity other than vibrations such as the winding current and leakage magnetic flux of the electric motor 21B, or It may be a combination of two or more. In this way, an abnormality label can be set for a device 20 in which a sensor 2 for detecting vibration cannot be installed, or for an event that cannot be determined based on vibration. Further, physical quantities or parameters suitable for abnormality determination may be selected in advance by sensitivity analysis or the like.

Further, as a method for evaluating the remaining life of any device 20, the remaining life evaluation unit 17 uses, for example, the minor rule (modified minor rule) determined from the material properties with respect to the fluctuating stress determined by the analysis unit 10. , the remaining life may be evaluated. At this time, the stress frequency distribution of the actual stress can be determined using the rainflow method or the like.

The remaining life evaluation method using Minor's rule will be explained with reference to the graphs of FIGS. 8 and 9. Note that the solid line graph G1 in FIG. 9 is a fatigue limit curve (material properties). The broken line graph G2 in FIG. 9 is a graph obtained by extending the fatigue limit curve using the modified Miner's law.

Here, in the SN curve of the target material, the number of repetitions at which rupture occurs due to repeated stress σi of a constant stress amplitude is assumed to be Ni. For example, assume that k repeated stresses σi (i=1 to k) with different amplitudes are each repeated ni (i=1 to k) times on a predetermined object. At this time, the fatigue damage D accumulated on this object is given by the following formula. Here, the time when D reaches 1 is defined as the life span.

The stress frequency distribution can be calculated by the analysis unit 10 based on the physical model. Here, if it is determined that the time until the fatigue damage D reaches 1 does not satisfy the required operation period, an abnormality label may be set.

Further, when the load result calculated by the analysis unit 10 is subject to material property conditions, that is, when fracture, plastic deformation, etc. occur, an abnormality label may be set for the physical quantity or parameter at that time.

Furthermore, in this embodiment, the remaining life is determined not only from load and stress but also from deterioration of the insulating material due to time changes in voltage or current, deterioration of the material due to thermal cycles, and the like. Then, an abnormality label may be set based on this determination.

Further, the abnormality label may be set based on the functional requirements of the device 20 (FIG. 2). For example, assume that the electric motor 21B is required to have a predetermined rotational speed in its design. Here, when the calculation result of the analysis unit 10 does not realize the required rotation speed, an abnormality label may be set for the physical quantity or parameter at that time.

Further, an abnormality label may be set based on requirements due to legal restrictions. For example, when an acoustic analysis is performed on the electric motor 21B and the noise level exceeds the level stipulated by law, an abnormality label may be set for the physical quantity or parameter at that time.

The above example illustrates a method of setting an abnormality label when there is no data regarding an abnormality of the device 20. On the other hand, if there is data regarding an abnormality in the device 20, the abnormality label may be set using this data. That is, when there is data that was determined to be abnormal in the past, and when this data is reproduced as a calculation result of the analysis unit 10, an abnormality label may be set for the physical quantity or parameter at that time.

Moreover, although the above-mentioned example illustrates a method of setting an abnormality label regarding the mechanical behavior of the device 20, abnormality labels may also be set for other systems. For example, physical quantities to be set include temperature, current, voltage, magnetic flux (density), sound, flow rate, pressure, and the like. Abnormality labels may be set for parameters related to these.

The state determining unit 15 (FIG. 1) of this embodiment determines the state of the device 20 based on reference information when the measured data is reflected in the parameters of the physical model and the measured data is analyzed. For example, when the measured parameter identified by the parameter identification unit 14 is analyzed, the state determination unit 15 determines the state of the device 20 based on a learning model trained using teacher data associated with an abnormality label. . In this way, the parameters of the device 20 having an abnormality can be learned from the learning model. Then, the device 20 can be appropriately evaluated using this trained learning model.

Further, the setting unit 16 sets an abnormality label based on a comparison between the remaining life calculated by the remaining life evaluation unit 17 and a pre-planned operation period. In this way, it is possible to evaluate whether or not an abnormality occurs in the device 20 during the pre-planned operation period. For example, the presence or absence of an abnormality can be evaluated by comparing the remaining life of each device 21 with the planned operating time.

Further, the state determining unit 15 determines whether or not there is an abnormality in the device 20 by comparing the measured parameters identified by the parameter identifying unit 14 with an arbitrarily set permissible range for parameters. In this way, it is possible to easily determine whether or not there is an abnormality in the device 20 based on the arbitrarily set permissible range for the parameter.

Further, the state determining unit 15 determines whether or not there is an abnormality in the device 20 by comparing the measured data acquired by the physical quantity measuring unit 12 with an arbitrarily set data tolerance range. In this way, it is possible to easily determine whether or not there is an abnormality in the device 20 based on the arbitrarily set data tolerance range.

For example, the state determining unit 15 determines whether there is an abnormality by comparing the actual measurement data acquired by the physical quantity measuring unit 12 and the data tolerance range. Further, the state determining unit 15 determines whether there is an abnormality by comparing the analysis data output from the data processing unit with the allowable range.

Parameter tolerance ranges and data tolerance ranges used for determination by the state determination unit 15 are set in advance. These allowable ranges may be arbitrarily set by the user in advance. Further, all of these tolerance ranges may be the same for each device 20 or may be different for each device 20.

Further, these tolerance ranges may be set based on actual measurement data, specifications regarding the behavior of the device 20, or actual test results, or may be set based on actual test results. It may be set based on simulation results, or may be set based on a combination of these results.

When it is determined that the device 20 is abnormal, the state determining unit 15 (FIG. 1) identifies the degree and cause of the abnormality. In this way, the degree and cause of the abnormality can be identified based on the measured data. Further, the degree of abnormality can be evaluated from the combination of measured parameters and the magnitude of the value.

The inspection estimation unit 18 (FIG. 1) estimates the progress of the abnormality from the combination and magnitude of the measured parameters when the condition determination unit 15 determines that the device 20 is normal, and also estimates the progress of the abnormality from the combination and size of the measured parameters The next inspection location and inspection timing are estimated from the remaining life of the device 21 calculated in . In this way, the inspection points and inspection timing can be automatically presented to the user.

Note that the main control section 8 outputs the evaluation results of the device state evaluation system 1 by controlling the output section 5 . For example, the main control section 8 outputs the results derived by the state determination section 15 and the inspection estimation section 18.

Next, the device condition evaluation method (processing) executed by the device condition evaluation system 1 of this embodiment will be explained using the flowchart of FIG. Note that the above-mentioned drawings will be referred to as appropriate. The following steps are at least part of the processing included in the device condition evaluation method, and other steps may be included in the device condition evaluation method.

First, in step S1, the model construction unit 9 creates a physical model that reproduces the operation of the device 20 composed of a plurality of devices 21 including at least a rotating electrical machine, and reproduces the state of the device 21 that occurs in connection with the operation. To construct.

In the next step S2, the analysis unit 10 performs calculations by reflecting parameters in equations formulated based on the physical model.

In the next step S3, the setting unit 16 sets reference information for determining the state of the device 20. Here, the setting unit 16 sets an abnormality label as reference information and performs a process of associating the abnormality label with each teacher data.

In the next step S4, the learning unit 11 inputs the abnormality label set by the setting unit 16, the calculation result of the analysis unit 10, and each corresponding parameter to the learning model as teacher data. This is where machine learning takes place.

In the next step S5, the physical quantity measurement unit 12 acquires physical quantities actually measured by each device 21 as actual measurement data.

In the next step S6, the actual measurement data processing section 13 converts the actual measurement data acquired by the physical quantity measurement section 12 into analysis data in a format compatible with the calculations of the analysis section 10. This analysis data is reflected in the parameters of the physical model, and the analysis unit 10 analyzes the measured data.

In the next step S7, the parameter identification unit 14 identifies actual measurement parameters, which are parameters when the calculation result of the analysis unit 10 reproduces the analysis data (actual measurement data), based on the teacher data learned by the learning unit 11. do.

In the next step S8, the state determining unit 15 determines the state of the device 20 based on the learning model trained using the teacher data associated with the abnormality label as reference information. For example, the state determining unit 15 determines whether or not there is an abnormality in the device 20 at the present time based on each of the measured parameters identified by the parameter identifying unit 14. Here, if there is no abnormality (NO in step S8), the process advances to step S9. On the other hand, if there is an abnormality (YES in step S8), the process advances to step S11.

In step S9, which proceeds in the case of NO in step S8, the remaining life evaluation section 17 calculates based on the calculation result of the analysis section 10 when the analysis data (actual measurement data) reflects the parameters identified by the parameter identification section 14. , the remaining life of each device 21 is calculated.

In the next step S10, the inspection estimation unit 18 estimates the progress of the abnormality based on the combination and magnitude of the measured parameters, and determines the next inspection location and inspection timing based on the remaining life of the equipment 21 calculated by the remaining life evaluation unit 17. Estimate. Note that the main control section 8 outputs the evaluation results of the device state evaluation system 1 by controlling the output section 5 . Then, the device state evaluation method ends.

In step S11, which is proceeded in the case of YES in step S8, the state determining unit 15 identifies the degree and cause of the abnormality when determining that the device 20 is abnormal. Note that the main control section 8 outputs the evaluation results of the device state evaluation system 1 by controlling the output section 5 . Then, the device state evaluation method ends.

Next, a modification will be described using FIG. 11. Incidentally, with appropriate reference to the above-mentioned drawings, the same reference numerals are given to the same constituent parts as those shown in the above-described embodiment, and redundant explanation will be omitted. The configuration applied in this modification may be applied to the above-described embodiment, or may be appropriately combined with the configuration of the above-described embodiment.

The reference information of the modified example includes a predetermined threshold value. This threshold value is set based on a data tolerance range that indicates the tolerance range of actually measured data. Note that the data tolerance range itself may be treated as reference information. Furthermore, in the modified example, the state of the device 20 is evaluated without using the learning model.

The setting unit 16 (FIG. 1) of the modified example sets a threshold value of an allowable range for the measured data as a method for determining whether or not there is an abnormality in the device 20. Further, the state determining unit 15 (FIG. 1) of the modified example determines the state of the device 20 based on the threshold value when the measured data is analyzed. In this way, the evaluation standard becomes clear by setting the threshold value, and the state of the device 20 can be appropriately evaluated. Furthermore, even when parameters cannot be identified, it is possible to determine the presence or absence of an abnormality by assuming all abnormal events. For example, when there is little actual measurement data that can be obtained from the sensor 2 and the parameter identification accuracy is poor, it is possible to determine whether there is an abnormality based on the actual measurement data.

Further, when the reference information is a data tolerance range, the state determining unit 15 determines the state of the device 20 by comparing the measured data with the data tolerance range. In this way, the evaluation standard becomes clear by setting the data tolerance range, and the state of the device 20 can be appropriately evaluated.

Note that the threshold value or data tolerance range set by the setting unit 16 may be arbitrarily set by the user in advance. Further, all of the threshold values or data tolerance ranges may be the same for all devices 20 or may be different for each device 20.

For example, as shown in FIG. 11, the analysis unit 10 (FIG. 1) calculates the bearing life using a plurality of parameters in advance. If the result is shorter than the required operating period, a threshold value indicating the permissible range of vibration is set for at least one vibration. At this time, the threshold for vibration may be imposed, for example, on the amplitude of vibration, or on the spectrum for each frequency of FFT.

The setting unit 16 calculates a threshold value indicating an allowable range of vibration based on various requirements. Then, among the threshold values calculated for the target physical quantity or parameter, the threshold value with the minimum allowable range is set as the threshold value used for determining the sign of abnormality. In this way, even if parameters cannot be identified, it is possible to determine the presence or absence of an abnormality by assuming all abnormal events.

Although FIG. 11 is an example in which vibration is used as the measured data, other measured data may be used, or these measured data may be combined. Furthermore, the actual measurement data may be, for example, current, voltage, magnetic flux, temperature, etc.

In addition, although the flowchart of the above-mentioned embodiment illustrates a form in which each step is executed in series, the sequential relationship of each step is not necessarily fixed, and the sequential relationship of some steps may be swapped. Also good. Also, some steps may be executed in parallel with other steps.

The device status evaluation system 1 of the above-described embodiment includes a control device in which a processor such as an FPGA (Field Programmable Gate Array), a GPU (Graphics Processing Unit), a CPU (Central Processing Unit), and a dedicated chip is highly integrated; Storage devices such as ROM (Read Only Memory) and RAM (Random Access Memory), external storage devices such as HDD (Hard Disk Drive) and SSD (Solid State Drive), display devices such as displays, mice and keyboards, etc. input device and a communication interface. This device status evaluation system 1 can be realized with a hardware configuration using a normal computer.

Note that the program executed by this device condition evaluation system 1 is provided by being pre-installed in a ROM or the like. Additionally or alternatively, this program may be installed as a file in installable or executable format on a computer-readable non-temporary computer such as a CD-ROM, CD-R, memory card, DVD, or floppy disk (FD). It is stored and provided on a standard storage medium.

Further, the program executed by this device condition evaluation system 1 may be stored in a computer connected to a network such as the Internet, and may be provided by being downloaded via the network. Further, the device status evaluation system 1 can also be configured by combining separate modules that independently perform the functions of the constituent elements by interconnecting them via a network or a dedicated line.

According to the embodiment described above, by providing the state determining unit 15 that determines the state of the device 20 based on reference information when the measured data is reflected in the parameters of the physical model and the measured data is analyzed. Even if the device 20 includes a device 21 that has not experienced any abnormality in the past, it is possible to evaluate the presence or absence of an abnormality in the device 20 from the actual measurement data of the device 21.

Although several embodiments of the invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, changes, and combinations can be made without departing from the gist of the invention. These embodiments or modifications thereof are within the scope and gist of the invention, as well as within the scope of the invention described in the claims and its equivalents.

DESCRIPTION OF SYMBOLS 1... Equipment condition evaluation system, 2... Sensor, 3... Computer for analysis, 4... Input section, 5... Output section, 6... Communication section, 7... Storage section, 8... Main control section, 9... Model construction section, 10 ...Analysis section, 11...Learning section, 12...Physical quantity measurement section, 13...Actual measurement data processing section, 14...Parameter identification section, 15...Status determination section, 16...Setting section, 17...Remaining life evaluation section, 18...Inspection estimation Part, 20... Device, 21... Equipment, 21A... Control panel, 21B... Electric motor, 21C... Rotating body, 21D... Load side equipment group.

Claims (14)

a model construction unit that constructs a physical model that reproduces the operation of a device composed of a plurality of devices including at least a rotating electric machine, and reproduces the state of the device that occurs in connection with the operation;
an analysis unit that performs calculations by reflecting parameters in equations formulated based on the physical model;
a physical quantity measurement unit that acquires physical quantities actually measured by each of the devices as actual measurement data;
a setting unit that sets reference information for determining the state of the device;
a state determining unit that determines the state of the device based on the reference information when the measured data is reflected in the parameters of the physical model and the measured data is analyzed;
Equipped with
Equipment condition evaluation system. a learning unit that performs machine learning by inputting the parameters corresponding to the calculation results of the analysis unit into a learning model as teaching data;
a parameter identification unit that identifies a measured parameter that is the parameter when the calculation result reproduces the measured data based on the learned learning model;
Equipped with
The reference information is set based on the calculation result, and includes an abnormality label indicating at least one of the following: the presence or absence of an abnormality in the device, the presence or absence of an abnormality sign, an abnormality event, a cause of the abnormality, a degree of abnormality, and a type of abnormality. including,
The state determination unit determines the state of the device based on the trained model using the teacher data associated with the abnormality label when the measured parameter identified by the parameter identification unit is analyzed. determine,
The device condition evaluation system according to claim 1. The learning unit calculates the learning model by using each of the parameters and the calculation results corresponding to these parameters as the teacher data when calculations corresponding to the combinations of the plurality of parameters are performed in advance in the analysis unit. enter,
The device condition evaluation system according to claim 2. The learning unit associates the parameter when the device is determined to be abnormal based on the calculation result with the abnormality label, and inputs this parameter to the learning model as the teacher data.
The apparatus condition evaluation system according to claim 2 or 3. The setting unit sets the abnormality label based on the combination of the parameters and the calculation result when the device is determined to be abnormal.
The apparatus condition evaluation system according to claim 2 or 3. The reference information includes a predetermined threshold value,
The state determining unit determines the state of the device based on the threshold when the measured data is analyzed.
The device condition evaluation system according to claim 1. The reference information includes a data tolerance range indicating a tolerance range of the actual measurement data,
The state determination unit determines the state of the device by comparing the measured data and the data tolerance range.
The device condition evaluation system according to claim 1. comprising a remaining life evaluation unit that calculates the remaining life of each of the devices based on the calculation results when the parameters whose actual measurement data has been identified by the parameter identification unit are reflected in the physical model;
The device condition evaluation system according to claim 2. The setting unit sets the abnormality label based on a comparison between the remaining life calculated by the remaining life evaluation unit and a pre-planned operation period.
The device condition evaluation system according to claim 8. The remaining life evaluation section calculates the load, stress, voltage, current, and temperature associated with each of the devices obtained by the analysis section when the measured parameters identified by the parameter identification section are reflected in the physical model. Evaluating the remaining life from the results of determining at least one
The device condition evaluation system according to claim 8 or 9. The setting unit sets the reference information based on whether at least one of the required functions of the device, material properties, and laws and regulations is satisfied.
The device condition evaluation system according to claim 1 or 2. comprising an actual measurement data processing unit that converts the actual measurement data acquired by the physical quantity measurement unit into analysis data in a format compatible with the calculations of the analysis unit;
The device condition evaluation system according to claim 1 or 2. The physical model is
a mechanical model representing the behavior of a mechanical system regarding at least one of rotational motion and translational motion of each of the devices;
a control model for a control unit forming at least one of the devices to control the behavior of the entire device;
an electrical model representing the behavior of the electrical system of the device;
a magnetic model representing the behavior of the magnetic system of the device;
a thermal model representing thermal behavior of the device;
a fluid model representing the behavior of at least one of the flow rate and pressure of a fluid flowing through piping connected to a pump constituting at least one of the devices;
a physical property model representing the material properties of each of the devices;
a structural model having information on at least one of dimensions and shape of each of the devices;
including at least one of
reproducing the behavior of the device by a combination of the parameters;
The device condition evaluation system according to claim 1 or 2. a step in which the model construction unit constructs a physical model that reproduces the operation of a device composed of a plurality of devices including at least a rotating electric machine, and reproduces the state of the device that occurs in connection with the operation;
an analysis unit performing calculations by reflecting parameters in equations formulated based on the physical model;
a step in which the physical quantity measurement unit acquires physical quantities actually measured by each of the devices as actual measurement data;
a step in which the setting unit sets reference information for determining the state of the device;
When the measured data is reflected in the parameters of the physical model and the measured data is analyzed, a state determining unit determines the state of the device based on the reference information;
including,
Equipment condition evaluation method.

Images