JPH10283002A - Optimum load distributor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an optimum load distributor with which a suitable initial value can be automatically set and a value better than the last optimum load distribution value can be provided. SOLUTION: Concerning a requested load, the optimum load distribution of respective elements is calculated from object model characteristic data 3 and initial value data 4 by an optimum load distribution computing element 2. When the optimum load distribution value as the output result of this optimum load distribution computing element 2 does not satisfy the load change request, a characteristic parameter is changed concerning an evaluation function value provided as a result of optimizing calculation by an initial value setting computing element 7, it is judged whether the initial value provided by the characteristic parameter at that time satisfies limitation conditions or not and when the limitation conditions are satisfied, while defining such a relevant initial value as the suitable initial value, initial value data are updated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optimum load distribution device applied to load distribution in a thermal power plant, a cooling plant and the like.

[0002]

2. Description of the Related Art Conventionally, in equipment such as a thermal power plant and a cooling plant, various optimum control methods using mathematical programming or the like have been proposed for setting load distribution of each element for a certain load request. . FIG. 4 shows an example of a conventional optimal control device using the optimal load distribution arithmetic unit.

[0003] In FIG. 4, when a load distribution change request 1 is input to an optimal load distribution computing unit 2, the optimal load distribution computing unit 2 uses the target model characteristic data 3 and the initial value data 4 to determine an optimal load distribution. Calculate and display the calculation result on the monitor screen 5
Output to Here, if the output result read from the monitor screen 5 is satisfactory, the operator changes the load distribution. Otherwise, the initial value is appropriately changed, the information is input to the initial value data 4, and the output result is output to the monitor screen 5 again through the optimum load distribution calculator 2. The operator repeats the same until the output result is satisfactory.

[0004]

As described above, conventionally, when the result of the optimization calculation is an unsatisfactory value, the operator himself / herself appropriately changes the initial value and causes the optimization calculation to be performed again. I was However, in this method,
Since the newly obtained optimal load distribution value does not always obtain a better value than the previous output result, it is necessary to repeat the same process many times until the output result is satisfactory, which causes a problem that burdens the operator. there were.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and has as its object to automatically set a proper initial value and obtain a value better than the previous optimum load distribution value. It is to provide an optimum load distribution device capable of performing the above.

[0006]

According to the present invention, there is provided an optimum load distribution device which calculates an optimum load distribution of each element from target model characteristic data and initial value data for a requested load. When the optimum load distribution value, which is the output result of the optimum load distribution computing unit, does not satisfy the load change request, the characteristic parameter is changed with respect to the evaluation function value obtained as a result of the optimization calculation. An initial value setting computing unit that determines whether or not the initial value obtained by the characteristic parameter satisfies the constraint condition, and updates the initial value data with the initial value as a valid initial value when the constraint condition is satisfied. It is provided.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an optimum load distribution device according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration as one embodiment of the present invention.
In FIG. 1, 1 is a load distribution change request, 2 is an optimal load distribution calculator, 3 is target model characteristic data, 4 is initial value data, 5 is a monitor screen, 6 is an output result judgment, and 7 is an initial value setting calculator. , 8 indicate a load distribution change. Also, 101 is a load distribution change request signal, 102 is an optimal load distribution value, 103
Is an initial value change request signal, 104 is initial value change data, 1
05 indicates each load set value.

In such a configuration, when the load distribution change request 1 is input to the optimal load distribution computing unit 2, the optimal load distribution computing unit 2 uses the target model characteristic data 3 and the initial value data 4 to optimize the load. The distribution is calculated, and the calculation result is output to the monitor screen 5.

Here, FIG. 1 differs from the conventional example shown in FIG. As shown in FIG. 2, this initial value setting computing unit 7
Optimization calculation result data 11, characteristic parameter value setting calculator 12, step width calculator 13, initial value search calculator 1
4. Search direction conversion operation unit 15, constraint condition evaluation operation unit 16
Having.

In FIG. 2, optimization calculation result data 11
Is obtained from the optimal load distribution value 102. The characteristic parameter value setting calculator 12 calculates a characteristic parameter value using the optimization calculation result data 11 and the search direction data obtained by the search direction conversion calculator 15. Step width calculator 13
Is the target model characteristic data 3 and the search direction conversion calculator 15
The step width value is calculated using the search direction data obtained in (1). The initial value search computing unit 14 searches for an initial value based on the characteristic parameter value from the characteristic parameter value computing unit 12 and the step width value from the step width computing unit 13 to calculate a characteristic parameter value. The search direction conversion calculator 15 calculates search direction data.

The constraint condition evaluation computing unit 16 determines whether or not the constraint condition is satisfied by using the characteristic parameter value from the initial value search computing unit 14. The value data 104 is calculated.

In such a configuration, the initial value setting computing unit 7 responds to the request from the initial value change request signal 103 by
The characteristic parameter value calculation unit 12 calculates the characteristic parameter values, the step width calculation unit 13 calculates the step width values, and these values are used as initial value search calculation units 14.
To search for an initial value, and calculate the constraint condition evaluation arithmetic unit 1
It is checked whether all the constraint conditions are satisfied by 6 and the initial value data 104 is calculated.

FIG. 3 is a flowchart showing a processing procedure for minimizing the evaluation function as an example of the initial value setting computing unit 7. When minimizing the evaluation function of the optimization problem, the initial value setting computing unit 7 applies a characteristic parameter Xi (i = 1, 1) to the evaluation function value D obtained as a result of the optimization calculation.
, N) (N: the number of control parameters), one characteristic parameter Xi is set to a step width hi =
(Upper limit value of Xi−lower limit value of Xi) / t, and a suitable initial value is searched. Here, t is an integer arbitrarily given by the operator.

First, the evaluation function value D obtained as a result of the optimization calculation and each characteristic parameter value ai (i = 1, 1)
.., N) (N: the number of control parameters) is input (step S1). Here, ai is a value of each characteristic parameter Xi obtained as a result of the optimization calculation.

Next, Xi is set to a step width h within the upper and lower limit regions.
i, and the evaluation function value F (X) at that time is calculated (steps S2 to S5). Here, the evaluation function value F (X)
Is smaller than D (No in step S6), each characteristic parameter value at that time is input to the constraint condition Gj (X) (j = 1,..., M) (M: the number of constraint conditions) (step S6). S7) It is determined whether all the constraint conditions Gj (X) are satisfied (step S8).

As a result, if all the constraint conditions Gj (X) are satisfied (Yes in step S8), the search is terminated at that time, and each characteristic parameter value at that time is set as initial value data (step S8). S9). Based on the initial value data, the optimum load distribution value is obtained again by using the optimum load distribution calculator 2.

On the other hand, when all the constraints Gj (X) are not satisfied (No in step S8), the evaluation function value F (X) is smaller than D and all the constraints Gj (X) are satisfied. X2,..., XN are repeated in order (steps S10 to S13).

It should be noted that even when the value of XN is reached, the evaluation function value F
If (X) is smaller than D and all the constraint conditions Gj (X) are not satisfied (N in step S13)
o) It is assumed that a valid initial value has not been found, and when searching for a valid initial value again, the value of t is increased and the process is repeated from the beginning (step S14).

As described above, when the optimum load distribution value D calculated so as to satisfy the load request cannot satisfy the load change request, each characteristic parameter Xi is added to the evaluation function value D.
Among (i = 1,..., N) (N: the number of control parameters), one characteristic parameter Xi is changed in the upper and lower limit regions, and the evaluation function value F (X) at that time is smaller than D,
And when all the constraint conditions Gj (X) are satisfied, the time is updated as an appropriate initial point, and the initial value data is updated.
Perform the optimization calculation again. Until a reasonable initial value is found,
Similar processing is performed by sequentially changing the characteristic parameters. As a result, the updating of the initial value data can be automated, and the optimum load distribution value obtained by performing the optimization calculation based on the initial value has a better result than the previous value.

[0020]

As described above, according to the first aspect of the present invention, when the optimum load distribution value calculated to satisfy the load request cannot satisfy the load change request, the optimum function When the characteristic parameter is changed and the evaluation function value at that time is smaller than the optimum load distribution value and satisfies the constraint, the initial value data is updated with that time as a valid initial point. As a result, the updating of the initial value data can be automated, and the optimum load distribution value obtained by performing the optimization calculation based on the initial value has a better result than the previous value.

That is, the following effects can be obtained by applying the optimum load distribution device of the present invention. (1) Conventionally, the change setting of the initial value left to the judgment of the operator can be automated, and the burden on the operator can be reduced. (2) Better results can be obtained compared to the previous optimal load distribution value.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an optimum load distribution device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of an initial value setting computing unit shown in FIG. 1;

FIG. 3 is a flowchart for explaining the processing operation of the initial value setting arithmetic unit shown in FIG. 1;

FIG. 4 is a block diagram showing a configuration of a conventional optimal control device.

[Explanation of symbols]

 1 Load distribution change request 2 Optimal load distribution calculator 3 Target model characteristic data 4 Initial value data 5 Monitor screen 6 Output result judgment 7 Initial value setting calculator 8 Load distribution change 9 Initial value change by operator's judgment 101 Load distribution change request Signal 102 Optimal load distribution value 103 Initial value change request signal 104 Initial value change data 105 Each load set value 11 Optimization calculation result data 12 Characteristic parameter value setting calculator 13 Step width calculator 14 Initial value search calculator 15 Search direction conversion Arithmetic unit 16 Constraint condition evaluation arithmetic unit

Claims (1)

[Claims] 1. An optimum load distribution calculator for calculating an optimum load distribution of each element from target model characteristic data and initial value data for a requested load, and an optimum load distribution calculator output result. When the load distribution value does not satisfy the load change request, the characteristic parameter is changed for the evaluation function value obtained as a result of the optimization calculation, and the initial value obtained by the characteristic parameter at that time satisfies the constraint condition. An optimal value distribution device, comprising: an initial value setting computing unit that determines whether or not the initial value data is a valid initial value when the constraint condition is satisfied.