FI111932B - Method of adjusting the speed of the lift and the lift system - Google Patents

Description

111932 METHOD FOR ADJUSTING THE LIFT SPEED AND THE LIFT SYSTEM

The invention relates to the elevator speed control according to the preamble of claim 1 and to the elevator system mentioned in the preamble of claim 7.

5 A common problem with elevator speed control is the various malfunctions in the speed control feedback loop. Disturbances affect the quality of the response from the elevator motor output and thereby impair the lift's ride comfort.

10

In the elevator application, different motors can be selected as the power source. For example, a DC motor, a short circuit motor, an AC motor or a synchronous motor. The speed control of these motors is implemented in various ways. In a DC-15 current motor, speed control is achieved by adjusting the voltage and current to the motor so that excess power is lost to the resistor, for example. The problem is a large power loss. The conventional speed control of AC motors is implemented as a slip control. This is equivalent to incremental losses and is uneconomical. However, AC motors are * * * significantly better in efficiency than DC motors * 1, but their traditional speed control results in * * ·! · * Horsepower loss. • · 35 *: The economic benefits of AC motors are outweighed by conventional leftover control. The drive operation • · • t l "25 allows the loss of power to be kept quite low.

• * · • »·

The drive inverter drive refers to controlling the AC motor with variable frequency and • • »voltage. The mains voltage is, for example, 50 Hz, 30 sine, substantially sinusoidal alternating current, which · · · is rectified in the inverter. Thereafter, a new sinusoidal alternating current is made in the frequency converter, the frequency and amplitude of which change according to the frequency converter control. The new AC power is supplied to the AC motor.

2 111932

One commonly used method for measuring the speed of an elevator is to use an analog tachometer that provides a voltage proportional to the speed of rotation of the drive wheel. The voltage signal is filtered and scaled by analog components, after which the signal is applied to an analog / digital converter. The converter provides a digital speed signal that is used for speed control.

The problem with velocity measurement with an analog tachometer is that interference with the velocity signal from the voltage occurs due to rotation of the tachometer, electrical disturbances in the engine room and resolution of the analog-to-digital conversion (from voltage to velocity-to-information). Attempts have been made to eliminate the problem by filtering the voltage signal with analog components before 15 AD conversions. Filtration causes a delay, which increases as new disturbances occur.

At low speeds, the analog tachometer provides a very low voltage so that the accuracy requirements for the analog components are in the range of per million. The AD converter easily has tens of millivolts of offset, which is closer to a percentage of the rated speed of the elevator: 4.5V.

• · · * · ': As a speed sensor you can use a resolver that transmits; * ·.! Blue and cosine signals proportional to 25 positions. Of these sig- nals, the * resolver / digital converter (RD converter) gives a pulse as the angle changes. The rate itself is measured by a method of calculating the number of pulses coming from the RD converter over a known period of time (the so-called MT method). From here

»I

* ·. 30 further calculates the speed. When using a resolver, the digital speed signal has a disturbance at the »» »frequency of the resolver and its multiple times. Low speeds cause problems because of the extremely small number of pulses available. The measured velocity value 35 changes easily and typically oscillates between some values. Because the resolver is a position sensor and 3111932 speeds are measured, the conversion becomes its own noise in the digital signal. An attempt is made to filter out interference, but the filtering causes the same problems as when using the tachometric signal.

5

Sometimes an encoder is used for speed feedback, which gives the absolute position of the drive wheel. The MT method provides velocity at the site. It is possible to get digital signals out of the encoders, thus reducing electrical interference.

10 However, low speeds have resolution problems. Otherwise, the use of an encoder leads to problems similar to the use of a resolver, even though the position information is obtained directly in digital form.

The object of the invention is to eliminate the drawbacks of the prior art and in particular to provide a new, more trouble-free elevator speed control. The basic idea of the invention is to generate a feedback signal to be used in the control using the measured speed information of the elevator and the 20 instructions controlling the elevator operation. Detailed Description of the Invention. are disclosed in the claims.

The invention can significantly reduce the noise that is • visible at the measured speed and thus also at the motor output (torque) • ·: *. Specifically, the so-called random faults • * • '·· 25 or better their adverse effects on control can be: eliminated from the feedback signal. This will make the adjustment more precise and stable. The invention is very well suited for use with a variety of discrete speed controls for elevators. Similarly, the invention can be used with various types of elevator drives.

* t '... · The invention achieves a significant improvement in driving quality at relatively low cost, as well as increased robustness of control. Robustness reduces the need for fine-tuning during installation and maintenance, which results in a reduction in work and costs. In addition, the elevator control is able to adapt to changes in elevator conditions, such as how the temperature change affects the response due to, for example, a change in friction between the elevator guides and guides. Likewise, longer-term changes that occur with the aging of the elevator system can be compensated for. This adaptability is achieved by using an adjustable model of the elevator, which can be updated. There are several ways to perform the upgrade. It is preferable to obtain an update schedule that maintains the model sufficiently up-to-date but does not overload the available computing capacity.

The elevator speed control system can simply be said to consist of four components: the control system, the motor drive and motor, the elevator and its mechanics, and the speed measurement. The term "elevator" and its mechanics refers here to a set of equipment mainly for elevator car, mechanical control of elevator car movement along a track in the elevator shaft, means for conveying engine turnover to elevator car movement such as lifting ropes and any counterweight and other equipment. The motor drive transmits fire from the control system. · · · J.. the elevator control commands to the elevator. The control system gets • • * faster. feedback from the elevator via speed measurement.

25 * ·

The boundary between the speed measurement and the control system can be drawn to * * · 'when the speed from the speed sensor is in digital format. It is advantageous to perform the speed feedback transforms and scaling on the same circuit board as the speed control, thus shortening the signal transmission distance between these functions. In speed measurement, filtering is usually required to eliminate interference. The filter causes a delay in the speed feedback, which, before, * *>, results in a vibration of the speed control loop.

I

35 5 111932

Control of the speed and position of the lift is done in the control system, as well as control of any acceleration and acceleration changes. In the control system, the speed feedback from the elevator monitors the realization of the speed reference and, if necessary, corrects the servo and motor control. Motor drive includes drive and motor. These transform the guidance from the steering system into a torque that is transmitted to the drive wheel via the motor shaft and, if necessary, the gear. In the drive, digital instruction 10 is converted to analog motor phase supply voltages. The conversion causes interference that is visible on the drive wheel. The invention provides a substantial reduction in these interferences.

The invention will now be described by way of non-limiting embodiment of the invention with reference to the accompanying drawings, in which Figure 1 schematically illustrates one known elevator speed control loop, Figure 2 schematically illustrates another known elevator speed control loop: the elevator speed control loop according to the invention, Fig. 4 schematically shows another mu-i ', ,, lane speed control loop of the invention, Fig. 5 shows the model identification, Fig. 6 shows the result of the elevator speed measurement and the like. Figure 7 shows the measurement curves of the torque reference and Figure 8 shows the measurement curves of the speed signal.

• * * »i ·

* I

*, “Figure 1 shows schematically a prior art elevator speed control loop. In Fig. 1, the speed control: 1/35 working loop is divided into four blocks, which are speed control 101, drive 104 including drive 102 and motor 103, elevator 106 as process part 105, and fourth speed feedback output v0i0 producing rate control 107. controller 110 generates an instruction for elevator operation, here voltage / frequency reference U0 / reference; his-5 sin to drive the speed reference vref. Conventionally, for elevator operation, one of the following instructions is formed: voltage and frequency reference, current and frequency reference, torque reference, acceleration reference. Output data or information typically includes position, velocity 10, or acceleration reference in this formation.

In speed measurement 107, the speed of the elevator 106 is measured, for example, directly from the drive wheel of the elevator 106 by an analog tachometer 108, which provides a voltage proportional to the speed of rotation of the drive wheel. If required, the voltage signal from the tachometer is filtered and scaled by analog components, then the signal is applied to the A / D converter 109. The A / D converter 109 provides a rate control 101 as a feedback rate digital speed signal v0i0.

Figure 2 schematically shows another prior art elevator speed control loop. In Fig. 2, the velocity control loop is divided into four blocks which are velocity! . control 121 with controller 130, servo 122 and motor 123 * · · *. *. including elevator drive 124, elevator 126 as process part 125, and speed measurement 127 providing a fourth rate actual value v0j0.

• t t

In the speed measurement 127, the speed of the elevator 126 is measured, for example, from the traction sheave of the elevator 126 by a resolver 128 which provides sine and cosine signals proportional to 30 positions. From these signals, the resolver / digital converter 129 outputs the angle variable (ΐι! With the support pulse.) The speed feedback speed value v0i0 is generated by the M / T method, which calculates the number of pulses from the RD converter in a known time period. Here, the speed of the 35 can be further calculated, and the resulting digital velocity signal 7111932 shows interference, for example, with the rotation speed of the resolver and its multiple interruptions, and low speeds cause problems when the number of pulses is remarkably small. Because the resolver is a position sensor and the speed is measured, this becomes its own noise in the digital signal.

Figure 3 is a diagrammatic representation of the hiss speed control loop of the invention. The speed control loop of Fig. 3 is divided into five blocks, which are speed control 1, torque controlled elevator drive 4 with servo 2 and motor 3, elevator 6 as process part 5, speed measurement 7 for generating actual speed v0i0, and fifth feedback signal, wherein the singular processing performing member 11 is preferably an estimator. The feedback signal vest is conventionally directed to the speed control 1 controller 12.

20 In the speed measurement 7, the speed of the elevator 6 is measured by the resolver 8. The resolver provides position-proportional sine and * *: cosine signals. Of these signals, the resolver / digital: converter 8 gives a pulse as the angle changes. The rate feedback value v0i0 of the rate feedback is generated by the M / T method, where * · 25 is the number of pulses from the RD converter over a known period of time. From this you can still calculate the speed. The resulting digital velocity signal v0i0 has various interferences, noise and possibly measurement errors. The velocity signal vo1o is applied to the signal processing block into the estimator 30. The input of the estimator is input torque reference T0hJe from speed control 1 to the drive and the actual speed signal from the speed measurement v0io · Using the elevator • modeling of the dynamics of the elevator and taking into account the different system types. .. · 35 quantization noise and other interference generated by speed estimator 7.

8 111932

Fig. 4 illustrates an alternative control loop option in which an acceleration reference aref is provided to the signal-processing block 10 in place of the instruction Ref. In this case, the parameter-5 must be selected differently, but otherwise the adjustment functions substantially in the same way as the adjustment according to Fig. 3. From the point of view of the invention, the use of acceleration reference instead of the actual instruction to the engine can be considered as a special case facilitating computation.

10

In practice, instead of the estimator of Figures 3 and 4, a state observer could be used to compute the state estimate, which only operates according to a different algorithm, seeking exactly the same effect, i.e. smoothing the rate feedback. An observer could be used instead of Est-15 markets. In the observer, the dynamic model of the elevator cannot be utilized in the same way as in the estimator. Otherwise, the observer's operating principle is close to that of the estimator. Both an estator and an observer require the existence of an elevator model. The elevator model parameters can advantageously be estimated by a Kalman filter which is used as a feedback signal processing element 11 and in which the filter state is extended to include the necessary parameters. Elevator · * · '; the control system has a limited computing capacity, whereby estimation and identification of the elevator model parameters «· j '. _ cannot calculate the computation capacity without the usual adjustment at the same interval. In this case, the identification algorithm can be * * · divided into parts that are calculated one at a time. This way, the identification algorithm does not update the estimator parameters every time » . ··, 30 in between.

• t »>» I | • * * '· * Figure 5 shows a model identification as part of the generation of the feedback signal · · * * ··· field signal vest. Motorized; * .. with the help of u and the velocity information v, estimate: 35 in section 20, a velocity estimate vest which is used as a feedback signal to the controller 22. The velocity estimate 9 111932

Vest and with the help of u the identification part identifies the parameters β of the model describing the elevator, which in turn is input into the estimation part as the basic information for the estimation. Thus, the estimation affects the model and the model estimates. In order to save the computation capacity-5, it is preferable to update the model identification and perform a recalculation of the parameters β less frequently than to update the feedback signal.

Using the estimator, a speed estimate Vest »is generated based on the instruction given by the speed controller Tohjei or a substantially proportional instruction thereof, and the result of the speed measurement v0i0, which is used as a feedback signal in the speed control 4.

15 The computation algorithm that computes the velocity estimate can be solved by the following estimator state model: 9k + 1 - CeXk + l + DeUek + 1, 20 where Ae is the system matrix of the estimator state equation, • *; ·, Be is the gain matrix of the estimator state equation inputs, ,; Ce is the measurement matrix of the estimator equation, • · · • ·

De is the gain matrix of the inputs 25 of the estimator equation, i · t xk is the state vector of the estimator at time k, • · '; yk is the velocity estimate at time k,: Uekko estimator control vector at time k, '. , * U «k ['^' help 'state 0' where T0jjek is the torque reference at k and Voiok is the velocity measurement at k.

10 111932

An elevator model is required when using the estimator. The properties of the model are identified by process tests.

The rope lift can be provided with a slow-mass, un-5 simplified model, which is also sufficient for most practical lifts.

The model is represented by the equation 10 0) = -— a + --Tm, ^ m ** m where ω 'is the angular acceleration of the drive wheel ω is the angular speed of the drive wheel,

Jm is the total inertia of the elevator system

Bm is the total friction of the elevator system and 15 Tm is the steering torque.

Since the feedback is speed information, the angular velocity ω is changed to velocity v. This is not a problem since it is only a scalar (drive wheel radius) multiplier ... 20 nos. This system can write continuous-time !. * state equation and measurement equation • · · »<· · t« t * · '. *. x = Ax + Bu + v »*« • i t ,; * y = Cx + Du + w, »« »« t * I <> 25 where,, A is the system of the general continuous state equation i> »

* I

*.,! matrix, <* * · B is the 'continuous matrix state control' coefficient matrix, '30 C is the general continuous time equation control matrix, coefficient matrix,: D is the general continuous time equation control 11 111932 u is the control moment Tm, x is the drive wheel speed v, v is the stochastic disturbance (noise) in the process and w is the stochastic disturbance (noise) in the measurement.

5

Discretize the system to give ** + i = Akxk + Bkuk + vk yk = ca + A «* + w *.

where

Ak is the system matrix of the general discrete state equation, 10 Bk is the control matrix of the general discrete state equation,

Ck is the measurement matrix of the general discrete measurement equation,

Dk is the coefficient matrix of the control of the general discrete measurement equation 15,

Uk is the steering moment Tm at time k,

Xk is the drive wheel speed v at time k and

Vk, Wk are stochastic disturbances (noises) in process and measurement.

20: Note! Here, footnote k of Ak's Bk's Ck and Dk denotes dis-: · .. matrixes of the dis- crete state model. Matrices are constant. Sen: · '; instead of x *, u *; and y as a footnote, k denotes the time k.

* 25

To identify the model, you need information about the lift speed Voio and the torque reference T0Help · These are available directly in the control loop. The identification must take into account the specific torque requirement due to friction and the effect of the load in question. 30 ink. The effect of the load can be taken into account by using a basket! the horizontal information available from the scale. The measurement 31 and the simulated speed 30 of the elevator are shown in Fig. 6. The measurement has been made in the service condition of the elevator. The simulation is done with a second-order space model and using the elevator's service travel speed while driving up.

35 12 111932 In practice, the Kalman filter is a good solution for implementing the estimator. The speed of the elevator drive wheel, or other speed proportional to the speed of the elevator car, is monitored by the Kalman filter. The Kalman filter equations can be written to mu-5

Xk ~ (^ Ic ~ ^ k ^^ k) Xk + (- ^ / £ - ALDk) uk P, = (C, -C.LC ^ + QLy, + (Dk-C.LD ^ u ,, where

Ajc is the system matrix of the general discrete state equation,

Bk is the coefficient matrix of the general discrete state equation control 10, C 1 c is the measurement matrix of the general discrete measurement equation,

Dk is the control coefficient matrix of the general discrete measurement equation, 15 uk is the control torque Tm at time k, xk is the Kalma filter state vector at k, y is the estimator output (velocity estimate vest) / yk is the measurement (velocity v0i0) L is the Kalman filter gain matrix. 20 The gain matrix L can be computed by a digital linear * least squares method. Measurement noise and! . process noise covariates are obtained by looking at the constant velocity range »I» and using the equation <♦ · ΐ \ ί Q-ql!: ·. £ <r2, * · »25 where Q is the covariance of process noise (moment T0i), R is the covariance of measurement noise (velocity measurement v0io), and

• * t I

g, r are the corresponding standard deviations.

Next, a discrete Kalman filter is generated by a suitable 30 numerical solution method. Assuming process noise and measurement noise to be white and when their covariates; : are well known, the numerical solution can easily be exemplified by any commercially available numerical software. When calculating the required data from the measurement data and simulating, it is found that the estimated and measured speed signal overlap almost overlap.

In practice, the estimator is somewhat system dependent. When the invention was tested in an elevator with a frequency converter-driven permanent magnet synchronous motor and the elevator was driven at a service run speed of 0.3 m / s using an estimator, a clear effect of the control according to the invention on the quality of run 10 was observed. In the frequency range above 10 Hz, both the measured ripple and the ripple of the torque reference decreased significantly, to about 15 dB. Fig. 7 shows the measurement curves of the elevator down-run and torque reference without estimator operation 32 and with estimator adjustment 33 and in Fig. 8 the speed-15 signal from torque reference without estimator operation 34 and estimator adjustment 35. In Figures 7 and 8, the estimator was based on first order model. This kind of driving result is downright dazzling. A significant reduction in torque vibrations directly affects driving comfort.

20

The realization of the estimator itself can easily be done programmatically. In practice, a relatively short program cycle is sufficient. It should be noted that although torque * ** input is considered above, the model does not distinguish torque from acceleration, for example, ti · * · '25. Most importantly, the input quantity, or control • ”· is proportional to the torque. Thus, the realization of the estimator culminates in the identification of the model. There are at least three possibilities for identification> i v ': Pre-calculation and tabulation of parameters, identification setup and real-time · 30-time identification algorithm. Thus, the practical model by which the invention is practiced can be formed in many ways.

In principle, the model does not need to be accurate according to the Stochastic Weather Theory. This may allow constraints - “* Tux is a set of ready parameter sets that can be computed 111932 14 in advance, that is, for parameter pre-computation and tabulation. For example, suitable values according to the weight of the elevator are selected from the parameter table. Although this option may be sufficient in many practical cases, it should be borne in mind that such a model describes the system rather poorly and that, as the model improves, the adjustment is generally improved.

A better way than the previous one, which requires relatively little processor power, is to connect, for example, data-10 gathering to the elevator shaft shaft for identification. A few seconds would be sufficient for the control system to measure the torque or voltage reference and the actual speed. After the shaft caliper, the necessary coefficients for the estimator would be calculated. After that, the elevator can drive normally while the estimator is connected.

15

A more advanced approach is to add to the model and filter calculation algorithm, which periodically updates the model parameters and calculates new coefficients for the estimator. In this way, the design and adjustment are adapted to the changing shaft and load conditions. However, such an arrangement uses more computing capacity, which requires equipment. Repeat some more processor power than the usual • ·!, 'control. Especially in fast lifts with a speed of * · «. 5 m / s or higher, this may require - · · * # 25 to select a more powerful processor. In such repetitive adjustment; · Parameter estimation of mechanical model parameters can be carried out, for example, by the least squares method. Note that model parameter estimation may occur at a significantly slower magnification than the actual speed control. For example, speed control every 4 ms and identification of the coefficients '·· 1' every 100 ms. I saw an update to the elevator-; The interval between consecutive updates of a model is larger than the interval between successive feedback signals: · [.

35 "111932 lb

Selecting some of the variables in the model may be different and may be more economical. For example, using the acceleration reference as an input instead of the actual torque corrects the model for initial torque and friction.

5

It will be apparent to those skilled in the art that the embodiments of the invention are not limited to the exemplified above, but may vary within the scope of the following claims.

10 1 · · r · »· j ·

Claims (9)

A method of speed control of an elevator, in which method the speed control is feedbacked and feedbacked with the feedback signal (Vest) formed by means of the matte speed (V0i0) and the reference (T0hje) controlling the elevator operation (4) or any other substantially The proportional reference, characterized in that a model describing the lift is determined and the model is regularly updated. Method according to claim 1, characterized in that the reference (T0hje) which controls the elevator operation (4) is one of the following: voltage and frequency reference, current and frequency reference, torque reference, acceleration reference. Method according to one or more of the preceding claims, characterized in that a model describing the lift is determined and with the model the feedback signal (Vesi) is generated by the reference (T0hje) and the measured speed (V0i0). Method according to claim 3, characterized in that the model describing the elevator is determined as an invariable part of the speed control of the elevator. 20 Method according to claim 3, characterized in that the model is updated according to a set time schedule. The method according to claim 3, characterized in that the speed control has a control range and the update of the model has a control range, and the control range '. 25 for updating the model is greater than the speed control range. "7. Elevator system comprising a motor drive and a motor, an elevator with associated mechanics, * speed measurement equipment and a control system forming the reference (T0hje) for the motor operation, which control system comprises a digital control device where at least the speed signal generated by the speed measurement equipment ( Voio) is available, and the reference (T0hje) is formed in the control device by using as the feedback signal v, the signal (Vest) which, using the model's parameters (β), is formed from the reference and the speed reference, characterized in that the model describing the operation of the elevator! ' is updated regularly 35 ice 111932 Method according to claim 7, characterized in that the parameters (β) of the model can be updated. Method according to claim 8, characterized in that the time interval between consecutive updates of the model parameters (β) is greater than the time interval between the generation of consecutive feedback signals. I I · · t> · »