JP6652398B2 - Parameter selection method, parameter selection program and parameter selection device - Google Patents

Description

  The present invention relates to a parameter selection method, a parameter selection program, and a parameter selection device.

  In order to control the state, it is necessary to set a set value governing the state to a suitable value. PID control is known as one of the control methods. PID control is a control method that determines a value such as an applied voltage according to a deviation (P) between an output value and a target value, an integral (I) of the deviation, and a derivative (D) of the deviation.

  In PID control, when parameters are determined for controlling the state of a newly designed product, it is generally difficult to select the parameters. Therefore, an apparatus using PID control sometimes incorporates an automatic PID parameter selection program called auto-tuning. For example, Patent Literature 1 discloses a technique related to auto tuning.

JP-A-58-68106

  Control using the parameters selected by auto-tuning may not always achieve the target state. Auto-tuning is only a theoretical optimum parameter, and may be different from what is actually sought. As a method of selecting the optimum parameters, a method of conducting an experiment with all possible combinations of the values of the respective parameters can be considered, but this method is very time-consuming and inefficient for practical use.

  In order to solve the above problems, the first invention is an optimal value of a total value obtained by integrating a plurality of evaluation values derived from a plurality of factors from a combination of a plurality of parameter values assigned to each of a plurality of factors. A parameter selection method for calculating an optimum value of a parameter that determines the first optimum value, which is a parameter optimum value of a plurality of factors, using a known adjustment rule, A second optimum value candidate calculating step of calculating a plurality of second optimum value candidates for each factor based on the obtained first optimum value, and using the obtained plurality of second optimum value candidates An experiment plan formulation step of formulating an experiment plan using an orthogonal array table based on an experiment design method, and a system in which a plurality of evaluation values obtained by experiments performed based on the established experiment plan are used as variables. A total value calculating step of calculating a total value indicated by the evaluation function; and a plurality of calculated total values, and a total value is optimized from a secondary optimum value candidate which is each parameter value used for calculating the total value. A parameter optimum value selecting step of selecting a parameter optimum value that is each parameter value to be set, wherein the control evaluation function used in the total value calculating step is an experiment performed based on an experimental plan. We propose a parameter selection method, which is a function that weights the degree of improvement of each evaluation value from the predetermined evaluation value standard value for each evaluation value obtained for each experiment using the AHP method for each evaluation value improvement degree. . Further, in the parameter selection method, an experiment plan formulation step and an overall value calculation step are performed by formulating an experiment plan using an orthogonal array table using the obtained plurality of secondary optimal value candidates based on the experiment design method. The present invention proposes a parameter selection method that replaces a total value calculation step of calculating a total value represented by a control evaluation function with a plurality of evaluation values obtained by an experiment performed based on an experiment plan.

  Further, the second invention calculates an optimum value of a parameter that determines an optimum value of a total value obtained by integrating a plurality of evaluation values derived from a plurality of factors from a combination of a plurality of parameter values assigned to each of the plurality of factors. A parameter selection program for calculating a primary optimum value, which is a parameter optimum value of a plurality of factors, using a known adjustment law, an adjustment rule using primary optimum value calculating step, and a calculated primary optimum value A second optimal value candidate calculating step of calculating a plurality of second optimal value candidates for each factor based on the factor, and an experiment plan using an orthogonal array table by using the obtained plurality of second optimal value candidates. Indicated by an experiment plan formulation step based on the experiment design method and a control evaluation function with multiple evaluation values obtained from experiments performed based on the established experiment plan as variables A total value calculating step of calculating a total value to be calculated, a plurality of calculated total values, and each parameter for optimizing the total value from a secondary optimum value candidate which is each parameter value used for calculating the total value. The present invention proposes a parameter selection program in which a parameter optimum value selection step of selecting a parameter optimum value, which is a value, is described in a computer so as to be executable.

  Further, the third invention calculates an optimal value of a parameter that determines an optimal value of a total value obtained by integrating a plurality of evaluation values derived from a plurality of factors from a combination of a plurality of parameter values assigned to each of the plurality of factors. A parameter selecting device for calculating a primary optimal value, which is a parameter optimal value of a plurality of factors, using a known adjustment rule, an adjustment rule using primary optimal value calculating unit, and a determined primary optimal value A second optimal value candidate calculating unit that calculates a plurality of second optimal value candidates for each factor based on the factor, and an experiment plan using an orthogonal array table using the plurality of second optimal value candidates obtained. An experiment plan formulation unit that formulates an experiment based on the experiment design method, and a total that calculates a total value indicated by a control evaluation function that uses a plurality of evaluation values obtained by experiments performed based on the established experiment plan as variables. A calculation unit, and a plurality of calculated total values, and a parameter optimum value that is a parameter value for optimizing the total value is selected from a secondary optimum value candidate that is a parameter value used for calculating the total value. A parameter optimum value selecting unit for performing the control evaluation function used in the total value calculating unit, wherein a control evaluation function for each experiment obtained by an experiment performed based on an experiment plan is determined in advance. We propose a parameter selection device that is a function that weights the degree of improvement from the determined standard value of each evaluation value for each degree of improvement of each evaluation value using the AHP method. Further, in the parameter selection device, the experiment plan formulating unit and the total value calculating unit formulate an experiment plan using the orthogonal arrangement table using the obtained plurality of secondary optimal value candidates based on the experiment design method. We propose a parameter selection device that replaces a total value calculation unit that calculates a total value represented by a control evaluation function using a plurality of evaluation values obtained by an experiment performed based on an experiment plan as variables.

  According to the present invention, it is possible to efficiently calculate a parameter optimum value, and it is possible to efficiently enhance controllability in various situations.

Example of functional block diagram of parameter selection device Example of the second optimal value candidate calculation step An example of the orthogonal array table used in the experiment design step (L9 orthogonal array table) An example of an experiment plan in the experiment step Example of hardware configuration of parameter selection device Example of processing flow of parameter selection device (parameter selection method) Flow chart showing an example of processing in a total value calculation step using a control evaluation function in which weights of each evaluation value are calculated by the AHP method

Hereinafter, embodiments of the present invention will be described with reference to the drawings. It should be noted that the present invention is not limited to the description in the present specification at all, and can be implemented in various modes without departing from the gist thereof.
<Structure>

  FIG. 1 shows an example of a functional block in the parameter selection device. For example, the parameter selection device 0101 includes an adjustment rule utilization primary optimum value calculation unit 0102, a secondary optimum value candidate calculation unit 0103, an experiment plan formulation unit 0104, an overall value calculation unit 0105, a parameter optimum value selection unit Unit 0106.

  Each functional block described above can be realized as hardware, software, or both hardware and software. Further, a software product used for causing a computer to execute software and a recording medium on which the product is recorded are naturally included in the technical scope of the present invention.

Further, each of the functional blocks described above may be configured to function by the same hardware, or may be configured as separate hardware. There are various types of data for making these associations, such as a configuration obtained from an external storage medium, a configuration receiving input via an operation input device, and a configuration obtained from another device on a network via a communication device. Conceivable.
<Configuration of each part>

  The “adjustment rule using primary optimum value calculation unit” 0102 has a function of obtaining a primary optimum value that is a parameter optimum value of a plurality of factors using a known adjustment rule. The “plurality of factors” include, for example, three factors such as a deviation (P) between the output value and the target value, an integral (I) of the deviation, and a derivative (D) of the deviation in the PID control. Although there is no particular limitation on the “known adjustment rule”, it is preferable to use a Ziegler-Nichols adjustment method, an Edgar adjustment method, a CHR method, or the like.

  The “secondary optimum value candidate calculation unit” 0103 calculates a plurality of second optimum value candidates for each factor based on the first optimum value obtained by the adjustment rule use first optimum value calculation unit 0102. It has a function to do. In order to calculate a plurality of secondary optimum value candidates for each factor, for example, the calculation may be performed within an amplitude of + 25% or −25% around the primary optimum value of each factor. Incidentally, any number of levels may be used, but three is more preferable.

  FIG. 2 shows an example of the second optimum value candidate calculation unit 0103. For example, the values of P, I, and D obtained by the auto tuning are set as the second level, the value obtained by decreasing the respective values by 25% is set as the first level, and the value obtained by increasing the value by 25% is set as the third level. Three secondary optimum value candidates are calculated for the factors.

  The “experiment plan formulation unit” 0104 formulates an experiment plan using the orthogonal array table based on the experiment design method, using the plurality of secondary optimal value candidates calculated by the secondary optimal value candidate calculation unit 0103. Has functions. The experiment design method is a statistical method used for efficiently performing an experiment.

  FIG. 3 shows an example of an orthogonal array table used in the experiment plan formulation unit 0104. The orthogonal array shown in FIG. 3 is called an L9 orthogonal array. One line indicates one experiment, and an experiment plan is formulated so that the experiment is performed using the second-order optimal value candidates indicated in the line. The experiments are preferably performed in the order shown at the far right of the figure. As a result of the experiment, a plurality of evaluation values are obtained for each combination of the second optimum value candidates of the plurality of factors. As the evaluation value, for example, an overshoot, a settling time, a rise time, and the like can be considered, but any value may be used and any number may be obtained. Normally, when there are three values as the optimal value candidates for the three parameters, 27 experiments are required if all combinations are tried. However, in the experiment design method using the orthogonal array, it is possible to find the optimum value of the parameter in nine experiments, which is very efficient.

  It should be noted that not only the PID parameter but also other factors such as a reduction amount of the target value filter and a filter time constant can be increased. In that case, an experiment plan is formulated by enlarging the orthogonal array table. For example, by increasing the size from the L9 orthogonal array table to the L27 orthogonal array table and further to the L81 orthogonal array table, it becomes possible to calculate the optimum values of more parameters.

  The experiment plan formulation unit 0104 may formulate an experiment plan using a Taguchi method instead of all or part of the experiment design method. The “Taguchi method” is also called quality engineering, and is known as a technique for efficient technology development similar to the design of experiments. In the Taguchi method, in addition to the use of the orthogonal array table as in the design of experiment, the concept of an SN ratio (signal-to-noise ratio) and a loss function are further introduced.

  The “total value calculation unit” 0105 is a function of calculating a total value represented by a control evaluation function having a plurality of evaluation values obtained by an experiment performed based on the experiment plan formulated by the experiment plan formulation unit 0104 as a variable. Having. In each experiment, as a result of substituting a plurality of obtained evaluation values into a control evaluation function and calculating, a total value is calculated for each combination of second-order optimal value candidates of a plurality of factors.

FIG. 4 shows an outline of an experimental result in the total value calculation unit 0105. The “control evaluation function” is represented, for example, by Expression 1. The evaluation value may be less or more than three. The denominator of each term may be, for example, a plurality of evaluation values obtained by using the primary optimum value obtained by the adjustment rule-based primary optimum value calculation unit 0102 (Y _i ). The numerator of each term indicates a plurality of evaluation values for each combination of the second optimum value candidates of the plurality of factors obtained by the total value calculation unit 0105 (Y _{i, n} ). This control evaluation function is a unit-free control evaluation function that performs unit-free by using a ratio for each of a plurality of evaluation values obtained by experiments. This makes it possible to derive a total value by linearly combining a plurality of evaluation values. As a result, it is possible to incorporate a plurality of evaluation values as parameter selection elements.
<Equation 1>

At the time of linear combination, a weighting coefficient is added to each evaluation value. Each coefficient satisfies the condition of Equation 2. For example, when the user attaches importance to the evaluation value 1, the coefficient α of the evaluation value 1 is set to be larger than the coefficients β and γ of the evaluation values 2 and 3. With this evaluation function, it is possible to select parameters more in accordance with the user's request.
<Equation 2>

  The “parameter optimum value selecting unit” 0106 includes a plurality of total values calculated from the result of the experiment of the total value calculating unit 0105 and a second optimum value candidate which is each parameter value used for calculating the total value. Has a function of selecting a parameter optimum value which is a parameter value for optimizing the total value from the parameters.

  The parameter optimum value selecting unit 0106 may include a parameter optimum value selecting sub-means 0107. The “parameter optimal value selection sub-means” 0107 is a parameter optimal value that is a parameter value for minimizing or maximizing the population average of the total value based on an expression for estimating the population average of the total value from the result of the total value calculation unit 0105. It has a function to select a value.

Equation 3 shows an example of an equation for estimating the population mean of the total value. A _i , B _i , and C _i are second-order optimum value candidates of a plurality of factors, and μ (superscript bar) _{Ai, Bi, Ci} are second-order optimum value candidates shown in A _i , B _i, and C _i Represents the population mean under the combination. In addition, x (superscript bar) _Ai , x (superscript bar) _Bi , x (superscript bar) _Ci is an average value of a plurality of factors for each second-order optimal value candidate, and x (superscript bar) is the overall value. Represents the average of the values. The “average value for each secondary optimum value candidate” is the average value of the total value obtained from the experiment performed by the total value calculation unit 0105 using each secondary optimum value candidate. Calculate the population mean of the combination of the secondary optimal value candidates of all the factors, and use the combination of the secondary optimal value candidates of each of the factors to minimize or maximize the calculated population average as the parameter optimal value. Can be selected.
<Equation 3>

  The formula for estimating the population mean of the total value in the parameter optimum value selection unit 0106 may be based on a data structure formula based on analysis of variance. Analysis of variance is a statistical technique for analyzing the effect of factors on data by decomposing data variability into data variances and data variance factors. By performing an analysis of variance, it is possible to verify whether or not each of the multiple factors has a superior effect on the overall value. Can be taken into account.

  What the above-described parameter optimum value selection sub-means 0107 does is to select a secondary optimum value candidate that gives the maximum or minimum value of the average value for each of the secondary optimum values for a plurality of factors. Is equivalent to selecting That is, it is not necessary to formulate an equation for estimating the population mean.

  Further, when there are three levels of the second optimum value candidates in the second optimum value candidate calculation section 0103, the parameter optimum value selection section 0106 can be performed as follows. That is, for each of the plurality of factors, the average value of the three secondary optimal value candidates is calculated. A quadratic curve passing through the three points is uniquely determined, and a value taking the maximum value or the minimum value of the quadratic curve may be selected as the parameter optimum value of each of the plurality of factors.

  By the way, in the above-mentioned control evaluation function, a weighting coefficient (α, β, γ) is added to each evaluation value such as an overshoot amount and a settling time. By arbitrarily setting these coefficients within a range that satisfies the condition of <Equation 2>, it becomes possible to select parameters more in accordance with the user's request.

  However, a user (operator) who operates a control target has images of a desired overshoot amount, settling time, and the like, but these images are not immediately linked to coefficients to be weighted to respective evaluation values. It is not easy to set an appropriate coefficient for realizing the above control.

  Therefore, a process of calculating a weighting coefficient based on the user's intuitive preference and selecting an optimal parameter value based on the calculated coefficient will be described. Specifically, the control evaluation function used in the above-described total value calculation step is an improvement of the evaluation value for each experiment obtained from the experiment performed based on the experiment plan from each predetermined evaluation value standard value. It is a function that weights the degree for each improvement degree of each evaluation value using the AHP method.

  The AHP (Analytic Hierarchy Process) method was proposed in 1971 by Professor Thomas L. Thurty of the University of Pittsburgh as a decision-making method for various evaluation criteria. This method is one of the problem-solving decision-making methods that mixes subjective judgment and system approach well in problem analysis. One of the features of this method is that a pair comparison is performed for a plurality of evaluation items. In the one-to-one comparison, since the evaluation items are compared one-to-one, it is expected that the evaluator has an honest feeling.

  The AHP method in the present invention means that each evaluation item (each factor) is evaluated based on a paired comparison table of evaluation items (a plurality of factors), and based on the result, each evaluation item (each factor) is weighted. Means to determine In the evaluation based on the comparison table, two evaluation items are bidirectionally compared for all combinations, and the importance is expressed in n stages (n is a natural number), and k (1 ≦ k ≦ n: k is 1 / k (k is a natural number) is given to a relatively inferior person (k is a natural number), and an average is taken for each evaluation item to obtain an evaluation result 1. The mean may be a geometric mean (geometric mean) or any other mean. Then, the ratio of the evaluation result 1 of all evaluation items of each evaluation item to the integrated value of the evaluation results 1 (or a value equivalent thereto) is determined as the final weight.

  Specifically, a pairwise comparison is performed for each item using the “overshoot amount”, “settling time”, “rise time”, and the like, which are listed as evaluation items. The number of evaluation items may be larger, and “overshoot duration”, “amplitude attenuation ratio”, and the like may be added.

In the paired comparison, each of the evaluation items is expressed in a relationship between a row (vertical item column) and a column (horizontal item column) as shown in Table 1, and the pair is compared with a row and a column. In addition, the result of the comparison is evaluated by a numerical value that is set as the next reference. That is,
3: Row items are slightly more important than column items. 5: Row items are more important than column items. 1/3: Column items are slightly more important than row items. 1/5: Column items are more important than row items. The reference of the comparison result may further use “7” and “1/7”, “9” and “1/9”, and the like.
<Table 1>

  In Table 1, in the comparison between the overshoot amount (row) and the settling time (column), if the settling time is slightly more important than the overshoot amount, “枠” is entered in this frame. In the comparison between the overshoot amount (row) and the rise time (column), if the overshoot amount is slightly more important than the rise time, “3” is included in this frame. The table is completed by comparing each item in this way. Note that “1” is entered in the comparison between the same items.

  Then, the average of each value for the row is calculated from the value of the comparison result. In this table, the geometric mean is calculated. Here, since the geometrical average of three values is used, if an overshoot is taken as an example, “1.00” which is the cube root of “1” × “1/3” × “3” is the average. Thus, the average of each item is calculated, and the importance is calculated based on the calculated average. That is, the ratio of the average value of each item when the sum of the average values of the items is “1” is the importance. Note that arithmetic mean (arithmetic mean) or the like may be used instead of geometric mean.

  The degree of importance thus obtained is used as a weight for the control evaluation function. That is, the calculated importance (0.33, 0.28, 0.39) of each item is used as α, β, γ in Expressions 1 and 2.

Here, the control evaluation function weights the degree of improvement with respect to the evaluation value standard value in each evaluation item in obtaining the total value. In addition, as the evaluation value standard value, for example, an evaluation value obtained by an experiment performed using the primary optimum value obtained by the adjustment rule utilizing primary optimum value calculation unit can be used. Equation 4 is the control evaluation function.
<Equation 4>

In this formula, Y ₁ , Y ₂ , and Y ₃ are evaluation value standard values. Then, a plurality of evaluation values obtained by an experiment performed based on the experiment plan formulated by the experiment plan formulation unit are substituted for Y _{1, n} , Y _{2, n} , and Y _{3, n} . Then, an overall value is calculated by the control evaluation function, and the parameter optimum value is selected by the parameter optimum value selecting unit as described above. As described above, the degree of improvement of the evaluation value obtained by the experiment with respect to the evaluation value reference value is reflected in the total value after being weighted.

By using this control evaluation function, the user's sensuous preference is reflected in the weighting of each evaluation value, and a parameter optimum value more suited to the user's preference can be selected.
<Hardware configuration>

  FIG. 5 shows an example of a hardware configuration of the parameter selection device. The parameter selection device 0501 includes, for example, “CPU” 0502, “main storage device” 0503, “program storage device” 0504, “secondary storage device” 0505, “external device I / F” 0506, and “user I / F” 0507. , “Output I / F” 0508, “bus” 0509, and the like.

  In the “CPU” 0502, processes such as an adjustment rule utilizing primary optimum value calculating step and a secondary optimum value candidate calculating step are executed. Instructions for the calculation, parameters necessary for the calculation, and the like are given from another device.

  The “main storage device” 0503 provides a work area necessary for the CPU 0502 to execute a program. In addition, it holds an orthogonal array table, evaluation values, and the like obtained during execution of the program and provides them to the CPU 0502 as necessary.

  The “program storage device” 0504 stores a program required for parameter selection such as an adjustment rule utilization primary optimum value calculation step and a secondary optimum value candidate calculation step, and provides the program to the CPU 0502 and the like.

  The “secondary storage device” 0505 is a storage device in which data can be dynamically rewritten during execution of a program, and stored data is not erased even when the power of the signal processing device is turned off.

  “External device I / F” 0506 indicates a device to be controlled. Upon receiving an instruction from the CPU 0502, processing such as an experiment in the parameter selection method is performed.

  “User I / F” 0507 indicates, for example, a mouse or a keyboard. The user can input a request regarding parameter selection to the main storage device 0503 or the like through the user I / F 0507.

  “Output I / F” 0508 indicates, for example, a display. The user can check the execution status of the program and the calculated parameter optimum value through the output I / F 0508.

Note that the devices indicated by 0502 to 0508 may be connected by a “bus” 0509.
<Process flow>

  FIG. 6 shows an example of the processing flow of the parameter selection device. First, in S0601, a primary optimal value that is a parameter optimal value of a plurality of factors is obtained by a known adjustment rule (a primary optimal value search step using an adjustment rule). Next, in S0602, a plurality of secondary optimal value candidates are calculated for each factor based on the obtained primary optimal values (second optimal value candidate calculating step). Then, in S0603, an experiment plan using the orthogonal arrangement table is formulated based on the experiment design method using the plurality of obtained secondary optimal value candidates (experiment plan formulation step). Thereafter, in S0604, a total value represented by a control evaluation function having a plurality of evaluation values obtained by an experiment performed based on the established experiment plan as a variable is calculated (a total value calculation step). In step S0605, a parameter optimum value, which is a parameter value for optimizing the total value, is selected from a plurality of calculated total values and a secondary optimum value candidate, which is a parameter value used for calculating the total value. (Parameter optimum value selection step).

  Note that the processing flow from S0601 to S0605 can also be implemented as a parameter selection method.

  FIG. 7 is a flowchart showing an example of the process in the total value calculation step in the case of using the control evaluation function in which the weight of each evaluation value is calculated by the AHP method.

As shown in the drawing, first, a weighting coefficient of each evaluation value by the AHP method is calculated based on the input reception of the result of the pair comparison by the user (S0701). Then, a plurality of evaluation values obtained by performing the experiment formulated in the experiment plan formulation step are used as variables, and a total value is calculated by a control evaluation function to which the calculated weighting coefficient is added (S0702).
<Effect>

  With this configuration, it is possible to efficiently select the optimum value of the parameter.

Claims (27)

A parameter selection method for calculating an optimal value of a parameter that determines an optimal value of a total value obtained by integrating a plurality of evaluation values derived from a plurality of factors from a combination of a plurality of parameter values assigned to each of a plurality of factors. ,
An adjustment rule using primary optimal value calculating step of obtaining a primary optimal value that is a parameter optimal value of a plurality of factors by a known adjustment law,
Secondary optimal value candidate calculating step of calculating a plurality of secondary optimal value candidates for each factor based on the determined primary optimal value,
An experiment plan formulation step of formulating an experiment plan using an orthogonal array table based on an experiment design method using a plurality of obtained secondary optimal value candidates,
A total value calculation step of calculating a total value indicated by a control evaluation function having a plurality of evaluation values obtained by an experiment performed based on the established experiment plan as a variable,
A plurality of calculated total values and a parameter optimum value that selects a parameter optimum value that is a parameter value that optimizes the total value from secondary optimum value candidates that are the parameter values used for calculating the total value. A selection step,
A control evaluation function used in the overall value calculation step, wherein the control evaluation function used in the total value calculation step is a predetermined evaluation evaluation value for each experiment obtained by an experiment performed based on an experiment plan. Parameter selection method that is a function that weights the degree of improvement from the standard value using the AHP method for each degree of improvement of each evaluation value. The parameter optimum value selecting step includes selecting a parameter optimum value that is a parameter value that maximizes or minimizes the population mean of the total value based on an equation for estimating the population mean of the total value from the results of the experiment. Including sub-steps,
The parameter selection method according to claim 1. The formula for estimating the population mean of the total value in the parameter optimum value selection sub-step is based on a data structure formula based on analysis of variance,
The parameter selection method according to claim 2. The control evaluation function used in the overall value calculation step is a unitless control evaluation function that performs unitless calculation by using a ratio for each of the plurality of evaluation values.
The parameter selection method according to claim 1. The second optimum value candidate calculation step calculates each factor within an amplitude within + 25% and -25% around the first optimum value.
The parameter selection method according to claim 1. Use Taguchi method instead of experimental design ,
The parameter selection method according to claim 1. The factors include a PID parameter,
The parameter selection method according to claim 1. The known adjustment rule is one of the Ziegler-Nichols adjustment method, the Edgar adjustment method, and the CHR method.
A parameter selection method according to any one of claims 1 to 7. The plurality of evaluation values include overshoot, settling time, and rise time,
9. The parameter selection method according to claim 1. A parameter selection program for calculating an optimal value of a parameter that determines an optimal value of a total value obtained by integrating a plurality of evaluation values derived from a plurality of factors from a combination of a plurality of parameter values assigned to each of a plurality of factors. ,
An adjustment rule using primary optimal value calculating step of obtaining a primary optimal value that is a parameter optimal value of a plurality of factors by a known adjustment law,
Secondary optimal value candidate calculating step of calculating a plurality of secondary optimal value candidates for each factor based on the determined primary optimal value,
An experiment plan formulation step of formulating an experiment plan using an orthogonal array table based on an experiment design method using a plurality of obtained secondary optimal value candidates,
A total value calculation step of calculating a total value indicated by a control evaluation function having a plurality of evaluation values obtained by an experiment performed based on the established experiment plan as a variable,
A plurality of calculated total values and a parameter optimum value for selecting a parameter optimum value which is a parameter value for optimizing the total value from secondary optimum value candidates which are each parameter value used for calculating the total value. A selection step;
Is a parameter selection program that is written in a computer so that it can be read and executed. The parameter optimum value selection step includes selecting a parameter optimum value that is a parameter value that maximizes or minimizes the population mean of the total value based on an equation for estimating the population mean of the total value from the result of the experiment. Including sub-steps,
A parameter selection program according to claim 10. The formula for estimating the population mean of the total value in the parameter optimum value selection sub-step is based on a data structure formula based on analysis of variance,
A parameter selection program according to claim 11. The control evaluation function used in the overall value calculation step is a unitless control evaluation function that performs unitless calculation by using a ratio for each of the plurality of evaluation values.
A parameter selection program according to any one of claims 10 to 12. The second optimum value candidate calculation step calculates each factor within an amplitude within + 25% and -25% around the first optimum value.
A parameter selection program according to any one of claims 10 to 13. Use Taguchi method instead of experimental design ,
A parameter selection program according to any one of claims 10 to 14. The factors include a PID parameter,
A parameter selection program according to any one of claims 10 to 15. The known adjustment rule is one of the Ziegler-Nichols adjustment method, the Edgar adjustment method, and the CHR method.
A parameter selection program according to any one of claims 10 to 16. The plurality of evaluation values include overshoot, settling time, and rise time,
A parameter selection program according to any one of claims 10 to 17. A parameter selection device for calculating an optimal value of a parameter that determines an optimal value of a total value obtained by integrating a plurality of evaluation values derived from a plurality of factors from a combination of a plurality of parameter values assigned to each of a plurality of factors. ,
An adjustment rule-based primary optimum value calculation unit that obtains a primary optimum value that is a parameter optimum value of a plurality of factors using a known adjustment rule,
A second optimum value candidate calculation unit that calculates a plurality of second optimum value candidates for each factor based on the obtained first optimum value,
An experiment plan formulating unit that formulates an experiment plan using an orthogonal array table based on an experiment design method using a plurality of obtained secondary optimal value candidates,
A total value calculation unit that calculates a total value represented by a control evaluation function having a plurality of evaluation values obtained by an experiment performed based on the established experiment plan as a variable,
A plurality of calculated total values and a parameter optimum value that selects a parameter optimum value that is a parameter value that optimizes the total value from secondary optimum value candidates that are the parameter values used for calculating the total value. The selection section,
A control evaluation function used in the total value calculation unit, wherein a predetermined evaluation value standard value for an evaluation value for each experiment obtained by an experiment performed based on an experiment plan. A parameter selection device that is a function that weights the degree of improvement from each evaluation value for each degree of improvement using the AHP method. The parameter optimum value selecting unit selects a parameter optimum value that is a parameter value that maximizes or minimizes the population mean of the total value based on an equation for estimating the population mean of the total value from the result of the experiment. Including sub-means,
The parameter selection device according to claim 19. The formula for estimating the population mean of the total value in the parameter optimum value selection sub-means is based on a data structure formula based on analysis of variance,
The parameter selection device according to claim 20. The control evaluation function used in the overall value calculation unit is a unitless control evaluation function that performs unitless calculation by using a ratio for each of the plurality of evaluation values.
The parameter selection device according to any one of claims 19 to 21. The second optimum value candidate calculation unit calculates each factor within an amplitude within + 25% and within −25% around the first optimum value.
The parameter selection device according to any one of claims 19 to 22. Use Taguchi method instead of experimental design ,
The parameter selection device according to any one of claims 19 to 23. The factors include a PID parameter,
The parameter selection device according to any one of claims 19 to 24. The known adjustment rule is one of the Ziegler-Nichols adjustment method, the Edgar adjustment method, and the CHR method.
The parameter selection device according to any one of claims 19 to 25. The plurality of evaluation values include overshoot, settling time, and rise time,
The parameter selection device according to any one of claims 19 to 26.

Images