Optimum value search control device, optimum value search control method, and optimum value search control program

Abstract

To provide a controller apparatus which properly operates extreme value control for optimizing a plurality of operation amounts.SOLUTION: An optimum value search controller apparatus according to an embodiment of the present invention includes: a first search control unit 200 having first signal generator units 202, 206 for generating a first signal for use in vibration by a first operation amount u1 and a first normalization processing unit 209 for normalizing inclination of an evaluation function value, and searching an operation point of the first operation amount u1 which makes the evaluation function value optimal based on a response of the evaluation function value obtained by vibration by the first operation amount u1 based on the first signal; and a second search control unit 300 having second signal generator units 302, 306 for generating a second signal used in vibration by a second operation amount u2 and different from the first signal and a second normalization processing unit 309 for normalizing inclination of the evaluation function value, and searching for an operation point of the second operation amount u2 which makes the evaluation function value optimal based on the response of the evaluation function value obtained by vibration by the second operation amount u2 based on the second signal.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] Embodiments of the present invention relate to an optimal value search control device, an optimal value search control method, and an optimal value search control program.

Background Art

[0002] In recent years, as a method of plant control, a technique called extremum control has attracted attention. Extremum control is a model-free real-time optimal control technique that does not use a complex model of a plant, and calculates from online sensor information that can directly measure the value of an evaluation function (cost function, performance index) to be optimized (minimized or maximized), and adaptively searches while changing the manipulated variable so as to maintain the evaluation function value at a (local) optimal value (local minimum value = minimum value or local maximum value = maximum value).

[0003] That is, the value of the evaluation function representing the performance index of the plant to be maximized (maximized) or minimized (minimized) is measured online. The evaluation function is defined, for example, as a unified index obtained by multiplying a plurality of output values of the plant by predetermined coefficients and adding them together. Searching in the direction of reducing the evaluation function value corresponds to automatically searching in the direction of the manipulated variable in which each output is reduced (balanced at a low value). Extremum control generally features a single input-output, and was mainly applicable to plants in which there is one manipulated variable to be optimized.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] On the other hand, for some plants, the operation may be determined by multiple operation amounts. When applying extreme value control to such a case, there is a possibility that control for optimizing the entire plant operation can be realized (the exploration of operation amounts becomes possible). However, since the conventional extreme value control is designed to optimize a single operation amount, there is no guarantee that appropriate control can be obtained even if the configuration is reorganized to optimize multiple operation amounts as it is, and there is a possibility of falling into partial optimization rather than optimizing the entire plant. This is because problems specific to extreme value control for optimizing multiple operation amounts, which could not occur in the conventional extreme value control, occur, and appropriate behavior could not be obtained in the extreme value control for optimizing multiple operation amounts without solving these problems.

[0006] An embodiment of the present invention has been made in view of the above circumstances, and an object thereof is to provide an optimal value search control device, an optimal value search control method, and an optimal value search control program that appropriately operate extreme value control for optimizing multiple operation amounts.

Means for Solving the Problems

[0007] The optimal value search control device according to the embodiment is a device applied to an arbitrary target process that takes a plurality of operation amounts including at least a first operation amount and a second operation amount as inputs and outputs one evaluation function value, and includes a first signal generation unit that takes the evaluation function value measured in real time as an input and generates a first signal that vibrates the first operation amount, and a first normalization processing unit that normalizes the gradient of the evaluation function value, and based on the response of the evaluation function value due to vibrating the first operation amount by the first signal, a first extreme value search unit that searches for an operating point of the first operation amount at which the evaluation function value becomes an optimal value, and a second signal generation unit that takes the evaluation function value as an input and generates a second signal that is a signal for vibrating the second operation amount and is different from the first signal, and a second normalization processing unit that normalizes the gradient of the evaluation function value, and based on the response of the evaluation function value due to vibrating the second operation amount by the second signal, a second extreme value search unit that searches for an operating point of the second operation amount at which the evaluation function value becomes an optimal value. , the first signal and the second signal are signals that oscillate periodically and have different phases from each other. 。

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

[0009] Hereinafter, the optimal value search control device, the optimal value search control method, and the optimal value search control program according to the embodiments will be described with reference to the drawings. FIG. 1 is a block diagram schematically showing a configuration example of the optimal value search control device according to the first embodiment. The optimal value search control device according to the present embodiment optimizes a plurality of manipulated variables u1 and u2 input to the plant 101 by using the value of the evaluation function y measured in real time from the plant 101.

[0010] As an optimization control method, extremum control is a control method for adaptively searching for the optimal value of an evaluation function based on the change in the evaluation function value according to the change in the manipulated variable. The evaluation function value is an index value determined based on the controlled variable of the controlled process, and the relationship between the evaluation quantity and the controlled variable is represented by a predetermined function. This evaluation function may be set based on any evaluation criterion as long as it is based on the controlled variable. Also, the evaluation quantity may be the controlled variable itself. In the controlled process in extremum control, the evaluation function may be an unknown function with respect to the manipulated variable.

[0011] In extreme value control, for example, the manipulated variable is continuously vibrated by a dither signal, and the change (increase or decrease) in the evaluation function value is observed. The dither signal is often given as a sine wave, but is not limited to a sine wave. Based on the change in the evaluation function value with respect to the change in the manipulated variable, the manipulated variable is changed in a direction such that the evaluation function value approaches the optimum value. By repeating such a change in the manipulated variable, the optimum value of the evaluation function is searched for.

[0012] The optimum value search control device includes a first search control unit 200 that optimizes a first manipulated variable u1 and a second search control unit 300 that optimizes a second manipulated variable u2. The first search control unit 200 includes a high-pass filter (HPF: High-Pass Filter) 201, a first signal generation unit, a multiplier 203, a low-pass filter (LPF: Low-Pass Filter) 204, an integrator 205, a multiplier 207, an adder 208, and a normalization processing unit (first normalization processing unit) 209. The first signal generation unit includes a first dither signal generator 206 and a second dither signal generator 202.

[0013] The high-pass filter 201 can remove a bias of a constant value corresponding to the optimum value from the evaluation function y fed back from the plant 101. The first dither signal generator 206 and the second dither signal generator 202 output perturbation signals (dither signals) to be added to the manipulated variables u1 and u2. The first dither signal generator 206 and the second dither signal generator 202 can cause fluctuations in the evaluation function value.

[0014] The signal output from the first dither signal generator 206 is input to the adder 208 after being multiplied by the amplitude a by the multiplier 207. The signal output from the second dither signal generator 202 is input to the multiplier 203. The multiplier 203 outputs a product of the evaluation function y that has passed through the high-pass filter 201 and the dither signal.

[0015] In this embodiment, the first dither signal generator 206 and the second dither signal generator 202 output a dither signal including a first oscillation period and a first pause period. The dither signals output by the first dither signal generator 206 and the second dither signal generator 202 are, for example, signals that alternately repeat a first oscillation period that outputs a sine wave for one wavelength and a first pause period for one wavelength. The method of applying such a signal is called the "cyclic" method, and in the following description, the dither signal in this embodiment is referred to as a cyclic sine wave.

[0016] The waveform of the signal output from the first dither signal generator 206 and the waveform of the signal output from the second dither signal generator 202 can be set to the same waveform if the time delay and waste time in the plant 101 can be ignored. When there is a time delay or waste time in the plant 101, the waveforms of the sine waves output from the first dither signal generator 206 and the second dither signal generator 202 may be different by incorporating predetermined phase compensation and waste time compensation. When the first dither signal generator 206 and the second dither signal generator 202 output the same dither signal, the optimum value search control device may include one dither signal generator and be configured to output the dither signal from one dither signal generator to the multiplier 203 and the multiplier 207.

[0017] The low-pass filter 204 has the role of extracting the low-frequency component from the output of the multiplier 203, and from this, it can be known whether the evaluation function value has increased or decreased. The normalization processing unit 209 normalizes the gradient of the low-frequency component of the evaluation function y extracted by the low-pass filter 204. The normalization processing in the normalization processing unit 209 will be described in detail later.

[0018] The integrator 205 functions as an estimator that estimates the direction of the operation amount to be moved in order to approximate the evaluation function value to the optimum value by integrating the evaluation function y after the normalization processing by the normalization processing unit 209. The adder 208 adds the value output from the integrator 205 and the dither signal output from the multiplier 207, and outputs it to the plant 101 as the first manipulated variable u1.

[0019] The second search control unit 300 includes a high-pass filter 301, a second signal generation unit, a multiplier 303, a low-pass filter 304, an integrator 205, a multiplier 307, an adder 308, and a normalization processing unit (second normalization processing unit) 309. The second signal generation unit includes a third dither signal generator 306 and a fourth dither signal generator 302.

[0020] In the present embodiment, the second search control unit 300 is such that the dither signals (second signals) output from the third dither signal generator 306 and the fourth dither signal generator 302 are different from the dither signals (first signals) output from the first dither signal generator 202 and the second dither signal generator 206. The second signal is, for example, a signal whose timing (or pause timing) of vibrating with respect to the first signal is shifted. The high-pass filter 301 can remove a bias of a constant value corresponding to the optimum value from the evaluation function y fed back from the plant 101.

[0021] The third dither signal generator 306 and the third dither signal generator 302 output a perturbation signal (dither signal) to be added to the manipulated variable u2. The signal output from the third dither signal generator 306 is input to the adder 308 after being multiplied by the amplitude a by the multiplier 307. The signal output from the fourth dither signal generator 302 is input to the multiplier 303. The multiplier 303 outputs a product of the evaluation function y that has passed through the high-pass filter 301 and the dither signal.

[0022] The third dither signal generator 306 and the fourth dither signal generator 302 output a dither signal including a second oscillation period and a second pause period. The dither signals output by the first dither signal generator 206 and the second dither signal generator 202 are, for example, cyclic sine waves that alternately repeat a second oscillation period that outputs a sine wave for one wavelength and a second pause period for one wavelength.

[0023] If the waveform of the signal output from the third dither signal generator 302 and the waveform of the signal output from the fourth dither signal generator 306 can ignore the time delay and waste time in the plant, they can be set to the same waveform. When there is a time delay or waste time in the plant, the waveforms of the signal output from the third dither signal generator 306 and the signal output from the fourth dither signal generator 302 may be different by incorporating predetermined phase compensation and waste time compensation. When the third dither signal generator 306 and the fourth dither signal generator 302 output the same dither signal, the optimum value search control device may include one dither signal generator and be configured to output the dither signal from the one dither signal generator to the multiplier 203 and the multiplier 207.

[0024] FIG. 2 is a diagram schematically showing an example of a dither signal used in the optimum value search control device of the first embodiment. Here, the waveform of the dither signal (first signal) applied to the first manipulated variable u1 and the waveform of the dither signal (second signal) applied to the second manipulated variable u2 are shown. In the present embodiment, the timings of the oscillation period and the pause period of the first signal applied to the first manipulated variable u1 and the oscillation period and the pause period of the second signal applied to the second manipulated variable u2 are shifted.

[0025] That is, the first signal includes a first pause period T1, and the second signal includes a second pause period T2. The waveform of the first signal becomes a sine wave at least in part of the second pause period T2, and is a constant value (=0) that does not oscillate during the first pause period T1. The waveform of the second signal is a sine wave at least in part of the first pause period T1, and is a constant value (=0) that does not oscillate during the second pause period T2.

[0026] In this embodiment, the first pause period T1 of the first signal is a period corresponding to one wavelength of the sine wave (oscillation waveform) of the second signal. The second pause period T2 of the second signal is a period corresponding to one wavelength of the sine wave (oscillation waveform) of the first signal. Therefore, even if sine waves of the same period are set as the oscillation waveforms of the first signal and the second signal, since the application timings are different, it is possible to exclude the change due to the other operation amount at the timing of extracting the change in one operation amount.

[0027] Note that the application time (oscillation period) of the dither signal applied to each of the operation amounts u1 and u2, and the time to pause may be arbitrarily determined. For example, it may be set such that the period during which the dither signal applied to one operation amount oscillates is the pause period of the dither signal applied to the other operation amount.

[0028] Also, in this embodiment, the oscillation waveforms of the first signal and the second signal are sine waves, but the waveforms that can be adopted as the dither signal are not limited to sine waves. For example, any signal can be adopted as long as it is an excitation signal that causes periodic oscillation, such as a cosine wave, a rectangular wave, a triangular wave, or a sawtooth wave. Then, it may be set such that the periods during which the first signal used for the optimization of the first operation amount u1 and the second signal used for the optimization of the second operation amount u2 oscillate are shifted.

[0029] Furthermore, in this embodiment, although the extremum control algorithm for optimizing the two manipulated variables u1 and u2 has been described, similarly, even when optimizing three or more manipulated variables, by shifting the periods during which the dither signals used for the respective optimizations oscillate relative to each other, it is possible to avoid the response caused by moving the two manipulated variables u1 and u2 from interfering with each other. In this case, it is necessary to set the oscillation period and the pause period of the dither signal applied to each of the plurality of manipulated variables according to the number of manipulated variables.

[0030] The low-pass filter 304 has the role of extracting the low-frequency component from the output of the multiplier 303, from which it is possible to tell whether the evaluation function value has increased or decreased. The normalization processing unit 309 normalizes the gradient of the low-frequency component of the evaluation function y extracted by the low-pass filter 304. The normalization processing in the normalization processing unit 309 will be described in detail later.

[0031] The integrator 305 functions as an estimator that estimates the direction of the manipulated variable to be moved in order to bring the evaluation function value closer to the optimum value by integrating the evaluation function y after the normalization processing by the normalization processing unit 309. The adder 308 adds the value output from the integrator 305 and the dither signal output from the multiplier 307, and outputs it to the plant 101 as the second manipulated variable u2.

[0032] Next, the normalization processing performed by the normalization processing units 209 and 309 will be described. The normalization processing units 209 and 309 perform normalization processing for normalizing the gradient of the low-frequency component of the evaluation function y extracted by the low-pass filters 204 and 304. In this embodiment, for example, assuming that the gradient information of the low-frequency component of the evaluation function is G[t], the normalized gradient information Gn[t] of the evaluation function is expressed as Gn[t]=G[t] / |G[t]|. Due to this function, the normalized gradient information Gn[t] always becomes -1 or more and 1 or less (-1≦Gn[t]≦1), and it is possible to prevent the gradient of the evaluation function from taking an excessive value.

[0033] In addition, in the normalization processing units 209 and 309, when the magnitude of the gradient changes according to the situation, if it is possible to avoid the magnitude of the gradient from becoming excessive and prevent the magnitude of the gradient from exceeding a certain level, the same effect can be obtained even with processing other than the above normalization processing. For example, the above normalization processing units 209 and 309 perform normalization processing to make the magnitude of the absolute value of the signal less than 1. However, in addition to this, an arbitrary constant α or the like may be set, and the normalization processing may be configured to perform normalization processing when the absolute value of the signal becomes less than that value (constant α).

[0034] Let the gradient information component of the extremum control algorithm for optimizing the manipulated variable u1 be G1[t] (the component of the gradient of the evaluation function with respect to the manipulated variable u1), the normalized gradient information of the evaluation function be Gn1[t], the gradient information of the extremum control algorithm for optimizing the manipulated variable u2 be G2[t] (the component of the gradient of the evaluation function with respect to the manipulated variable u2), and the normalized gradient information of the evaluation function be Gn2[t]. Then, each normalized signal is expressed as follows.

[0035] Gn1[t]=k1G1[t] / |G1[t]+ε1| Gn2[t]=k2G2[t] / |G2[t]+ε2| Here, ε1 and ε2 are values provided to prevent division by zero. k1 and k2 are arbitrary coefficients that follow the following relational expressions. k1 2 +k2 2 =1

[0036] This is a process for creating a normalized signal of the gradient of the evaluation function y when the evaluation function y is composed of the sum of the squares of the manipulated variable u1 component and the manipulated variable u2 component. As described above, when the gradient of the evaluation function is to be kept within a predetermined constant α, the right side of the above equation is set to α (k1 2 +k2 2 =α).

[0037] Similarly, in the extreme value control of the evaluation function using three or more operation amounts, the same effect can be obtained by obtaining and applying the components of the gradient information of the evaluation function by the number of optimized operation amounts. In the extreme value control algorithm for optimizing the operation amount, the sum of the squares of the coefficients is set to 1 (or a constant α).

[0038] Furthermore, as the above normalization method, it is a normalization process using a so-called soft sign function. In addition to this, for example, a sigmoid function, a Gompertz function, an inverse tangent function (arctan), a Guterman function, a function x / (ε + x p ) (1 / p) etc. can be used to perform the normalization process.

[0039] These functions are functions that change smoothly between -1 and 1 (functions that change between -α and α when suppressing within a constant α), and any function with similar characteristics can be used for the normalization process. For example, the above normalization process may be realized by using a function obtained by adding upper and lower limit limits to an arbitrary monotonically increasing function.

[0040] Next, the effects of the optimum value search control device, the optimum value search control method, and the optimum value search control program of the present embodiment will be described. FIGS. 3 and 4 are examples of the results of simulating the optimization of a plurality of operation amounts by the optimum value search control device of the first embodiment.

[0041] Here, an example will be described in which the minimum value of the evaluation function y is the optimum operating point, the optimum value of the first operation amount u1 is 2, and the optimum value of the operation amount u2 is 5. FIG. 3 shows an example of the optimum value search result for the first operation amount u1 along the time axis. FIG. 4 shows an example of the optimum value search result for the second operation amount u2 along the time axis.

[0042] According to the simulation results, it can be seen that the optimal value search control device of the first embodiment can converge to the optimal value in a predetermined time without the first operation amount u1 and the second operation amount u2 interfering with each other and without causing a sudden change in the operation amount transiently. From the above, by applying the algorithm of the present embodiment, appropriate control of the convergence speed in multivariable extremum control becomes possible. That is, according to the optimal value search control device, the optimal value search control method, and the optimal value search control program of the present embodiment, extremum control for optimizing a plurality of operation amounts can be appropriately operated.

[0043] Hereinafter, the effects of the present embodiment will be further described in comparison with the operation of the optimal value search control device of the comparative example. FIG. 5 is a block diagram schematically showing a configuration example of the optimal value search control device of the first comparative example.

[0044] The evaluation function value of the optimal value search control device of this comparative example is one and is determined by the operating conditions of a plurality of operation amounts u1 and u2. The optimal value search control device of this comparative example is different from the optimal value search control device of the first embodiment described above in that the dither signal applied to the first operation amount u1 and the dither signal applied to the second operation amount u2 have the same waveform and that the normalization processing units 209 and 309 are not provided.

[0045] The optimal value search control device of this comparative example includes a first search control unit 200 that optimizes the operation amount u1 and a second search control unit 300 having the same function as the first search control unit 200, and the first search control unit 200 and the second search control unit 300 are connected in parallel to the plant 101.

[0046] First, an example of searching for an optimal operation amount (an operation amount at which the evaluation function is minimized) when the relationship between an unknown evaluation function and an operation amount is in the shape of a downwardly convex quadratic function in extremum control for optimizing one operation amount will be described.

[0047] In the optimization process of searching for the manipulated variable that minimizes the evaluation function while driving the manipulated variable with a dither signal, the behavior varies depending on the operating point of the manipulated variable. Since the relationship between the manipulated variable and the evaluation function is unknown, in order to determine the direction in which the manipulated variable is to be moved next, it is necessary to add a dither signal to the manipulated variable and know the response of the resulting evaluation function.

[0048] For example, when the manipulated variable is at a certain operating point A, assume that increasing the manipulated variable results in a decrease in the evaluation function value. Therefore, when the manipulated variable is driven with a sine wave at this operating point A, the response of the resulting evaluation function is out of phase with the movement of the manipulated variable. On the other hand, when the manipulated variable is at another operating point B, assume that increasing the manipulated variable results in an increase in the evaluation function value. Therefore, when the manipulated variable is driven with a sine wave at this operating point B, the response of the resulting evaluation function is in phase with the movement of the manipulated variable.

[0049] Information on the response of the evaluation function to the movement of this manipulated variable is obtained from the results of signals obtained by multiplying the value of the evaluation function by the dither signal passing through the low-pass filters 204 and 304 in the multipliers 203 and 303. The signal in this part mainly corresponds to the gradient information of the evaluation function, and together with the information on the direction in which the manipulated variable is to be moved, it also extracts information on the degree (magnitude) of how much the manipulated variable should be moved. In this way, a mechanism is provided to obtain information on the direction and magnitude of the manipulated variable to be moved next from the dither signal added to the manipulated variable and the shape of the evaluation function which is the response waveform.

[0050] On the other hand, in the optimal value search control device of the comparative example, two manipulated variables u1 and u2 are inputs to the plant 101, and one evaluation function value is output from the plant 101. For example, consider the case of aiming for the minimum value of the evaluation function formed by the two manipulated variables u1 and u2 by the optimal value search control device of this comparative example. Also in this example, since the relationship between the first manipulated variable u1, the second manipulated variable u2, and the evaluation function cannot be grasped in advance, the minimum manipulated variables u1 and u2 are searched for based on the response waveforms of the evaluation function accompanying each of the manipulated variables u1 and u2 while varying them.

[0051] For example, when the operating points of the first operation amount u1 and the second operation amount u2 are at arbitrary positions, it is assumed that for the first operation amount u1, the evaluation function has an inverse phase relationship, and for the second operation amount u2, the evaluation function has a same phase relationship. At this time, for example, when the waveforms of the dither signals of the first operation amount u1 and the second operation amount u2 are set to the same waveform, such as a sine wave, if there is no time delay or the like, the response of the evaluation function with respect to the first operation amount u1 and the response of the evaluation function with respect to the second operation amount u2 will be in an inverse response. Therefore, the responses of the evaluation function due to the movements of the plurality of operation amounts u1 and u2 may cancel each other out, and the response waveform may disappear.

[0052] When such a situation occurs, information on the direction and magnitude in which the first operation amount u1 and the second operation amount u2 should be moved next cannot be extracted from the response of the evaluation function. Therefore, depending on the operating point, an event may occur where the optimum value stops. This phenomenon occurs because interference in the responses between the operation amounts u1 and u2 occurs, and the interference can be avoided by changing the waveforms of the dither signals applied to the operation amounts u1 and u2. For example, by setting the waveforms of the dither signals applied to the operation amounts u1 and u2 as in the first embodiment, the occurrence of the above event can be avoided.

[0053] FIG. 6 is a block diagram schematically showing a configuration example of the optimum value search control device of the second comparative example. The evaluation function value of the optimum value search control device of this comparative example is one and is determined by the operating conditions of a plurality of operation amounts u1 and u2. The optimum value search control device of this comparative example is different from the optimum value search control device of the first embodiment described above in that it does not include the normalization processing units 209 and 309.

[0054] The optimum value search control device of this comparative example includes a first search control unit 200 that optimizes the operation amount u1 and a second search control unit 300 that has the same function as the first search control unit 200, and the first search control unit 200 and the second search control unit 300 are connected in parallel to the plant 101.

[0055] In the optimal value search control device of this comparative example, as in the first embodiment, the first signal and the second signal are set such that the vibration period and the pause period are shifted from each other. Therefore, in this embodiment, it is possible to avoid interference between the response of the first manipulated variable u1 and the response of the second manipulated variable u2.

[0056] On the other hand, for example, when aiming for the minimum value of the evaluation function formed by the two manipulated variables u1 and u2 by the optimal value search control device of this comparative example, the manipulated variables u1 and u2 may change rapidly immediately after the start of control.

[0057] Figures 7 and 8 are an example of the result of simulating the optimization of a plurality of manipulated variables by the optimal value search control device of the second comparative example. Here, an example will be described in which the minimum value of the evaluation function y is the optimal operating point, the optimal value of the first manipulated variable u1 is 2, and the optimal value of the manipulated variable u2 is 5. Figure 7 shows an example of the optimal value search result for the first manipulated variable u1 along the time axis. Figure 8 shows an example of the optimal value search result for the second manipulated variable u2 along the time axis.

[0058] According to this simulation result, by the optimal value search control device of the second comparative example, the first manipulated variable u1 and the second manipulated variable u2 were able to finally converge to the optimal values without interfering with each other. However, the first manipulated variable u1 and the second manipulated variable u2 tended to change rapidly immediately after the start of control (for example, A1 and A3 in Figures 7 and 8), and then gradually converged to the optimal values (for example, A2 and A4 in Figures 7 and 8). Assuming that the optimal value search control device of this comparative example is actually applied to a plant, such a large change in the manipulated variable immediately after the start of control is not preferable for operating the plant, and it also takes a long time to converge to the optimal operating conditions.

[0059] Such a tendency is because in multivariable extremum control, the gradient of the evaluation function is affected by the operating point since the extremum control moves the first manipulated variable u1 and the second manipulated variable u2 based on the gradient information of the evaluation function. In extremum control, it is theoretically possible to smooth the sudden change in the manipulated variable immediately after the start of control by changing the magnitude of the integral gain. However, when the magnitude of the integral gain is changed, the behavior after the operating points of the first manipulated variable u1 and the second manipulated variable u2 move to a region where the value (gradient) of the evaluation function is small becomes an even gentler change, resulting in a large amount of time required for convergence to near the optimal manipulated variable. Also, in extremum control, since there is no guideline for changing the integral gain in real time, it has been difficult to appropriately control the convergence speed of extremum control.

[0060] On the other hand, for example, in the optimal value search control device of the first embodiment, the normalization processing units 209 and 309 can normalize the magnitude of the gradient of the evaluation function. By this, according to the optimal value search control device of the first embodiment, it becomes possible to appropriately control the convergence speed without changing the integral gain in real time.

[0061] As described above, according to the optimal value search control device, the optimal value search control method, and the optimal value search control program of the first embodiment, it is possible to appropriately operate the extremum control for optimizing a plurality of manipulated variables.

[0062] Next, the optimal value search control device, the optimal value search control method, and the optimal value search control program of the second embodiment will be described in detail with reference to the drawings. In the following description, the same components as those of the above-described first embodiment will be denoted by the same reference numerals and the description thereof will be omitted.

[0063] FIG. 9 is a block diagram schematically showing a configuration example of the optimal value search control device of the second embodiment. The optimal value search control device of this embodiment includes a first search control unit 200 that optimizes the first operation amount u1 and a second search control unit 300 that optimizes the second operation amount u2, and uses the evaluation function y obtained from the plant 101 to optimize a plurality of operation amounts u1 and u2 input to the plant 101. The optimal value search control device of this embodiment is different from the first embodiment in the first signal applied to the first operation amount u1 and the second signal applied to the second operation amount u2.

[0064] The first dither signal generator 206 and the second dither signal generator 202 output perturbation signals (dither signals) to be added to the operation amounts u1 and u2. The signal output from the first dither signal generator 206 is input to the multiplier 203. The signal output from the second dither signal generator 202 is input to the adder 208 after being multiplied by the amplitude a by the multiplier 207. The multiplier 203 outputs the product of the evaluation function y that has passed through the high-pass filter 201 and the dither signal.

[0065] In this embodiment, the first dither signal generator 206 and the second dither signal generator 202 output sine waves as dither signals. The waveform of the sine wave output from the first dither signal generator 206 and the waveform of the sine wave output from the second dither signal generator 202 can be set to the same waveform if the time delay and waste time in the plant 101 can be ignored. When there are time delay and waste time in the plant 101, the waveforms of the sine wave output from the first dither signal generator 206 and the sine wave output from the second dither signal generator 202 may be different by incorporating predetermined phase compensation and waste time compensation. When the first dither signal generator 206 and the second dither signal generator 202 output the same dither signal, the optimal value search control device may be configured to include one dither signal generator and output the dither signal from the one dither signal generator to the multiplier 203 and the multiplier 207.

[0066] In this embodiment, the second search control unit 300 has the dither signals (second signals) output from the third dither signal generator 306 and the fourth dither signal generator 302 being different signals from the dither signals (first signals) output from the first dither signal generator 202 and the second dither signal generator 206, and the second signal is, for example, a signal whose phase is shifted with respect to the first signal.

[0067] The third dither signal generator 306 and the third dither signal generator 302 output perturbation signals (dither signals) to be added to the manipulated variable u2. The signal output from the third dither signal generator 306 is input to the multiplier 303. The signal output from the fourth dither signal generator 302 is input to the adder 308 after being multiplied by the amplitude a by the multiplier 307. The multiplier 303 outputs a product of the evaluation function y that has passed through the high-pass filter 301 and the dither signal.

[0068] The third dither signal generator 306 and the fourth dither signal generator 302 output cosine waves as dither signals. The waveform of the cosine wave output from the third dither signal generator 302 and the waveform of the cosine wave output from the fourth dither signal generator 306 can be set to the same waveform if the time delay and dead time in the plant can be ignored. When there are time delay and dead time in the plant, the waveforms of the cosine wave output from the third dither signal generator 306 and the cosine wave output from the fourth dither signal generator 302 may be different by incorporating predetermined phase compensation and dead time compensation. When the third dither signal generator 306 and the fourth dither signal generator 302 output the same dither signal, the optimal value search control device may be provided with one dither signal generator and configured to output the dither signal from the one dither signal generator to the multiplier 203 and the multiplier 207.

[0069] FIG. 10 is a diagram schematically showing an example of the dither signal used in the optimal value search control device of the first embodiment. In the optimal value search control device of this embodiment, the phase difference between the dither signal used for optimizing the first manipulated variable u1 and the dither signal used for optimizing the second manipulated variable u2 is 90°. In this embodiment, by shifting the phases of the dither signals for each of the plurality of manipulated variables, interference between the responses of the plurality of manipulated variables u1 and u2 is suppressed. Note that the phase difference between the dither signals for each of the plurality of manipulated variables is not limited to 90°, and any phase difference may be provided as long as non-interference can be effectively achieved.

[0070] Also, in this embodiment, the dither signal used for optimizing the first manipulated variable u1 is a sine wave, and the dither signal used for optimizing the second manipulated variable u2 is a cosine wave. However, the dither signal used for optimizing the second manipulated variable u2 may be set as a sine wave, and in that case, the dither signal used for optimizing the first manipulated variable u1 may be set as a cosine wave.

[0071] Furthermore, the waveforms that can be adopted as the dither signal are not limited to sine waves and cosine waves. For example, any signal that is an excitation signal causing periodic vibration, such as a rectangular wave, a triangular wave, or a sawtooth wave, can be adopted. Then, it may be set so that a phase difference occurs between the dither signal used for optimizing the first manipulated variable u1 and the dither signal used for optimizing the second manipulated variable u2.

[0072] Furthermore, in this embodiment, the extremum control algorithm for optimizing the two manipulated variables u1 and u2 has been described. However, even when optimizing three or more manipulated variables, similarly, by shifting the phases of the dither signals used for each optimization from each other, it is possible to avoid interference between the responses caused by moving each of the two manipulated variables u1 and u2. In this case, it is necessary to set the phase difference of the dither signals applied to the plurality of manipulated variables according to the number of manipulated variables. Next, the effects of the optimal value search control device, the optimal value search control method, and the optimal value search control program of this embodiment will be described.

[0073] Figures 11 and 12 are examples of the results of simulating the optimization of a plurality of manipulated variables by the optimum value search control device according to the second embodiment. Here, an example will be described in which the minimum value of the evaluation function y is the optimum operating point, the optimum value of the first manipulated variable u1 is 2, and the optimum value of the manipulated variable u2 is 5. FIG. 11 shows an example of the result of searching for the optimum value of the first manipulated variable u1 along the time axis. FIG. 12 shows an example of the result of searching for the optimum value of the second manipulated variable u2 along the time axis.

[0074] According to the results of this simulation, it can be seen that the optimum value search control device according to the second embodiment can converge to the optimum value in a predetermined time without the first manipulated variable u1 and the second manipulated variable u2 interfering with each other and without causing a sudden change in the manipulated variable transiently. From the above, by applying the algorithm of the present embodiment, it is possible to appropriately control the convergence rate in multivariable extremum control. That is, according to the optimum value search control device, the optimum value search control method, and the optimum value search control program of the present embodiment, it is possible to appropriately operate the extremum control for optimizing a plurality of manipulated variables.

[0075] Next, the optimum value search control device, the optimum value search control method, and the optimum value search control program according to the third embodiment will be described in detail with reference to the drawings. In the following description, the same components as those in the above-described first embodiment will be denoted by the same reference numerals and the description thereof will be omitted.

[0076] FIG. 13 is a block diagram schematically showing a configuration example of the optimum value search control device according to the third embodiment. The optimum value search control device of the present embodiment includes a first search control unit 200 that optimizes the first manipulated variable u1 and a second search control unit 300 that optimizes the second manipulated variable u2, and uses the evaluation function y obtained from the plant 101 to optimize a plurality of manipulated variables u1 and u2 input to the plant 101. The first signal applied to the first manipulated variable u1 and the second signal applied to the second manipulated variable u2 of the optimum value search control device of the present embodiment are different from those in the above-described first and second embodiments.

[0077] The first dither signal generator 206 and the second dither signal generator 202 output perturbation signals (dither signals) to be added to the operation amounts u1 and u2. The signal output from the first dither signal generator 206 is input to the multiplier 203. The signal output from the second dither signal generator 202 is input to the adder 208 after being multiplied by the amplitude a by the multiplier 207. The multiplier 203 outputs a product of the evaluation function y that has passed through the high-pass filter 201 and the dither signal.

[0078] In this embodiment, the first dither signal generator 206 and the second dither signal generator 202 output sine waves as dither signals. If the waveforms of the sine waves output from the first dither signal generator 206 and the sine waves output from the second dither signal generator 202 can ignore the time delay and dead time in the plant 101, they can be set to the same waveform. When there are time delay and dead time in the plant 101, the waveforms of the sine waves output from the first dither signal generator 206 and the sine waves output from the second dither signal generator 202 may be different by incorporating predetermined phase compensation and dead time compensation. When the first dither signal generator 206 and the second dither signal generator 202 output the same dither signal, the optimal value search control device may be provided with one dither signal generator and configured to output the dither signal from the one dither signal generator to the multiplier 203 and the multiplier 207.

[0079] In this embodiment, the second search control unit 300 has the dither signals (second signals) output from the third dither signal generator 306 and the fourth dither signal generator 302 different from the dither signals (first signals) output from the first dither signal generator 202 and the second dither signal generator 206, and the second signal is, for example, a sine wave having a different period from the first signal.

[0080] The third dither signal generator 306 and the fourth dither signal generator 302 output a perturbation signal (dither signal) to be added to the manipulated variable u2. The signal output from the third dither signal generator 306 is input to the multiplier 303. The signal output from the fourth dither signal generator 302 is input to the adder 308 after being multiplied by the amplitude a in the multiplier 307. The multiplier 303 outputs the product of the evaluation function y that has passed through the high-pass filter 301 and the dither signal.

[0081] The third dither signal generator 306 and the fourth dither signal generator 302 output sine waves as dither signals. If the time delay and dead time in the plant can be ignored, the waveforms of the sine waves output from the third dither signal generator 302 and the sine waves output from the fourth dither signal generator 306 can be set to the same waveform. When there are time delays and dead times in the plant, the waveforms of the sine waves output from the third dither signal generator 306 and the sine waves output from the fourth dither signal generator 302 may be different by incorporating predetermined phase compensation and dead time compensation. When the third dither signal generator 306 and the fourth dither signal generator 302 output the same dither signal, the optimum value search control device may be provided with one dither signal generator and configured to output the dither signal from the one dither signal generator to the multiplier 203 and the multiplier 207.

[0082] FIG. 14 is a diagram schematically showing an example of a dither signal used in the optimum value search control device of the third embodiment. In the optimal value search control device of this embodiment, the period of the dither signal used for optimizing the first operation amount u1 is set shorter than the period of the dither signal used for optimizing the second operation amount u2. In this embodiment, by applying dither signals with different periods to each of the plurality of operation amounts u1 and u2, interference between the responses of the plurality of operation amounts u1 and u2 is suppressed. When the first signal and the second signal are set to have the same waveform, interference occurs between the responses of the plurality of operation amounts u1 and u2 as in the above-described first comparative example. However, for example, by setting the period of one dither signal to be longer than the other, interference of the responses can be avoided. For example, when the period of the sine wave of the first signal is TA and the period of the sine wave of the second signal is TB, by setting TB = 10 × TA, interference between the responses of the plurality of operation amounts u1 and u2 can be avoided.

[0083] Also, in this embodiment, the dither signals applied to the first operation amount u1 and the second operation amount u2 are sine waves, but the present invention is not limited thereto. The waveform that can be adopted as the dither signal is not limited to a sine wave. For example, any signal can be adopted as long as it is an excitation signal that causes periodic vibration, such as a cosine wave, a rectangular wave, a triangular wave, or a sawtooth wave. Then, the periods of the dither signal applied to the first operation amount u1 and the dither signal applied to the second operation amount u2 may be set to be different.

[0084] Furthermore, in this embodiment, the extremum control algorithm for optimizing the two operation amounts u1 and u2 has been described. However, even when optimizing three or more operation amounts, similarly, by shifting the periods of the dither signals used for each optimization from each other, interference of the responses caused by moving each of the two operation amounts u1 and u2 can be avoided. In this case, it is necessary to set the periods of the dither signals applied to the plurality of operation amounts according to the number of operation amounts.

[0085] Next, the effects of the optimal value search control device, the optimal value search control method, and the optimal value search control program of this embodiment will be described. Figures 15 and 16 are an example of the result of simulating the optimization of a plurality of manipulated variables by the optimal value search control device of the third embodiment.

[0086] Here, an example will be described in which the minimum value of the evaluation function y is the optimal operating point, and the optimal value of the first manipulated variable u1 is 2 and the optimal value of the manipulated variable u2 is 5. Figure 15 shows an example of the result of searching for the optimal value of the first manipulated variable u1 along the time axis. Figure 16 shows an example of the result of searching for the optimal value of the second manipulated variable u2 along the time axis.

[0087] According to this simulation result, it can be seen that the optimal value search control device of the third embodiment can converge to the optimal value in a predetermined time without the first manipulated variable u1 and the second manipulated variable u2 interfering with each other and without causing a sudden change in the manipulated variable transiently. From the above, by applying the algorithm of this embodiment, appropriate control of the convergence speed in multivariable extremum control becomes possible. That is, according to the optimal value search control device, the optimal value search control method, and the optimal value search control program of this embodiment, extremum control for optimizing a plurality of manipulated variables can operate appropriately.

[0088] In the optimal value search control device of each of the above-described embodiments, the above functions may be realized by hardware or by software. Further, when the functions of the optimal value search control device are realized by software, the optimal value search control device may include at least one processor and a memory in which a program executed by the processor is recorded. Further, the optimal value search control device may be configured to store setting information such as dither signals and normalization processing functions applied to each of a plurality of manipulated variables in a storage device such as a magnetic hard disk device or a semiconductor storage device, and to acquire the setting information from the storage device. Further, in this case, the optimal value search control device may be configured to include an input unit that accepts change operations of the dither signal and the normalization processing, and a setting update unit that updates the setting information of the dither signal and the normalization processing in response to the input. Further, in this case, the optimal value search control device may be configured to include a display unit that displays the content of the setting information.

[0089] Although some embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalent scope thereof.

Description of Reference Numerals

[0090] 101... plant, 200... first search control unit, 201... high-pass filter, 202... dither signal generator, 203... multiplier, 204... low-pass filter, 205... integrator, 206... dither signal generator, 207... multiplier, 208... adder, 209... normalization processing unit, 300... search control unit, 301... high-pass filter, 302... dither signal generator, 303... multiplier, 304... low-pass filter, 305... integrator, 306... dither signal generator, 307... multiplier, 308... adder, 309... normalization processing unit

Claims

1. An apparatus applicable to any target process, which receives a plurality of operation variables including at least a first operation variable and a second operation variable as input and outputs one evaluation function value, a first signal generating unit that receives the evaluation function value measured in real time as an input and generates a first signal for oscillating the first manipulated variable, and a first normalization processing unit that normalizes a gradient of the evaluation function value, and a first search control unit that searches for an operating point of the first manipulated variable at which the evaluation function value becomes an optimal value, based on a response of the evaluation function value caused by oscillating the first manipulated variable by the first signal; a second signal generating unit configured to receive the evaluation function value as an input and generate a second signal that is different from the first signal and that vibrates the second manipulated variable, and a second normalization processing unit configured to normalize a gradient of the evaluation function value, and a second search control unit configured to search for an operating point of the second manipulated variable at which the evaluation function value is an optimal value based on a response of the evaluation function value resulting from vibrating the second manipulated variable by the second signal, An optimum value search control device, wherein the first signal and the second signal are periodically oscillating signals having mutually different phases.

2. An apparatus applicable to any target process, which inputs a plurality of operation variables including at least a first operation variable and a second operation variable, and outputs one evaluation function value, a first signal generating unit that receives the evaluation function value measured in real time as an input and generates a first signal for oscillating the first manipulated variable, and a first normalization processing unit that normalizes a gradient of the evaluation function value, and a first search control unit that searches for an operating point of the first manipulated variable at which the evaluation function value becomes an optimal value, based on a response of the evaluation function value caused by oscillating the first manipulated variable by the first signal; a second signal generating unit configured to receive the evaluation function value as an input and generate a second signal that is different from the first signal and that vibrates the second manipulated variable, and a second normalization processing unit configured to normalize a gradient of the evaluation function value, and a second search control unit configured to search for an operating point of the second manipulated variable at which the evaluation function value is an optimal value based on a response of the evaluation function value resulting from vibrating the second manipulated variable by the second signal, An optimum value search control device, wherein the first signal and the second signal are periodically oscillating signals having different periods.

3. An apparatus applicable to any target process, which inputs a plurality of operation variables including at least a first operation variable and a second operation variable, and outputs one evaluation function value, a first signal generating unit that receives the evaluation function value measured in real time as an input and generates a first signal for oscillating the first manipulated variable, and a first normalization processing unit that normalizes a gradient of the evaluation function value, and a first search control unit that searches for an operating point of the first manipulated variable at which the evaluation function value becomes an optimal value, based on a response of the evaluation function value caused by oscillating the first manipulated variable by the first signal; a second signal generating unit configured to receive the evaluation function value as an input and generate a second signal that is different from the first signal and that vibrates the second manipulated variable, and a second normalization processing unit configured to normalize a gradient of the evaluation function value, and a second search control unit configured to search for an operating point of the second manipulated variable at which the evaluation function value is an optimal value based on a response of the evaluation function value resulting from vibrating the second manipulated variable by the second signal, The first signal is a signal that oscillates during a first vibration period and has a constant value during a first pause period, The second signal is a signal that oscillates during a second vibration period and has a constant value during a second pause period, An optimum value search control device, wherein a first vibration period is included in the second rest period, and the second vibration period is included in the first rest period.

4. The method is applied to an arbitrary target process that receives a plurality of operation variables including at least a first operation variable and a second operation variable as inputs and outputs one evaluation function value, Acquire the evaluation function value measured in real time; A method for searching for an operating point of the first manipulated variable and the second manipulated variable at which the evaluation function value becomes an optimal value, comprising the steps of: The searching for the operating point of the first manipulated variable includes generating a first signal for oscillating the first manipulated variable, normalizing a gradient of the evaluation function value, and searching for an operating point of the first manipulated variable at which the evaluation function value is an optimal value based on a response of the evaluation function value caused by oscillating the first manipulated variable by the first signal; Searching for an operating point of the second manipulated variable includes generating a second signal that is a signal for oscillating the second manipulated variable and is different from the first signal, normalizing a gradient of the evaluation function value, and searching for an operating point of the second manipulated variable at which the evaluation function value is an optimal value based on a response of the evaluation function value caused by vibrating the second manipulated variable by the second signal; An optimum value search control method, wherein the first signal and the second signal are periodically oscillating signals having different periods.

5. The present invention is applied to any target process that receives a plurality of operation variables including at least a first operation variable and a second operation variable as input and outputs one evaluation function value, Acquire the evaluation function value measured in real time; A method for searching for an operating point of the first manipulated variable and the second manipulated variable at which the evaluation function value becomes an optimal value, comprising the steps of: The searching for the operating point of the first manipulated variable includes generating a first signal for oscillating the first manipulated variable, normalizing a gradient of the evaluation function value, and searching for an operating point of the first manipulated variable at which the evaluation function value is an optimal value based on a response of the evaluation function value caused by oscillating the first manipulated variable by the first signal; Searching for an operating point of the second manipulated variable includes generating a second signal that is a signal for oscillating the second manipulated variable and is different from the first signal, normalizing a gradient of the evaluation function value, and searching for an operating point of the second manipulated variable at which the evaluation function value is an optimal value based on a response of the evaluation function value caused by vibrating the second manipulated variable by the second signal; An optimum value search control method, wherein the first signal and the second signal are periodically oscillating signals having different periods.

6. A method for applying to any target process that inputs a plurality of operation variables including at least a first operation variable and a second operation variable and outputs one evaluation function value, Acquire the evaluation function value measured in real time; A method for searching for an operating point of the first manipulated variable and the second manipulated variable at which the evaluation function value becomes an optimal value, comprising the steps of: The searching for the operating point of the first manipulated variable includes generating a first signal for oscillating the first manipulated variable, normalizing a gradient of the evaluation function value, and searching for an operating point of the first manipulated variable at which the evaluation function value is an optimal value based on a response of the evaluation function value caused by oscillating the first manipulated variable by the first signal; Searching for an operating point of the second manipulated variable includes generating a second signal that is a signal for oscillating the second manipulated variable and is different from the first signal, normalizing a gradient of the evaluation function value, and searching for an operating point of the second manipulated variable at which the evaluation function value is an optimal value based on a response of the evaluation function value caused by vibrating the second manipulated variable by the second signal; The first signal is a signal that oscillates during a first vibration period and has a constant value during a first pause period, The second signal is a signal that oscillates during a second vibration period and has a constant value during a second pause period, An optimum value search control method, wherein a first vibration period is included in the second rest period, and the second vibration period is included in the first rest period.

7. 7. An optimum value search control program for causing a computer to execute the optimum value search control method according to claim 4.

Images