A semiconductor temperature control system control method based on myxomycete algorithm

Abstract

The application discloses a semiconductor temperature control system control method based on a slime mold algorithm. The method comprises the following steps: identifying a differential equation of a semiconductor temperature control system by using a least square method to obtain a parameterized transfer function; performing excitation test, inputting a signal into the semiconductor temperature control system, and outputting a temperature change value; inputting an error value into a PID controller to control the semiconductor temperature control system; optimizing and iterating control parameters by using the slime mold algorithm, annealing and mutating the optimization result of the slime mold algorithm, and replacing a global optimal solution with a new solution generated by annealing and mutation until the fitness value of the fitness function of the PID controller response is the global optimal fitness, and then optimizing the PID control parameters of the PID controller. The application improves the response speed, stability and control precision of the semiconductor temperature control system, reduces overshoot and shortens the regulation time of the control system, and has good engineering application value.

Description

TECHNICAL FIELD

[0001] The present application relates to a temperature control system control method, in particular to a semiconductor temperature control system control method based on myxomycete algorithm. BACKGROUND

[0002] With the development of science and technology, refrigeration technology is ubiquitous in our life, widely used in cryopreservation, temperature control and other fields. Traditional refrigeration system often relies on compressor and other equipment for refrigeration, which is bulky and inconvenient to install, and has great limitations in local temperature control field. With the rapid development of semiconductor technology, the refrigeration efficiency of semiconductor refrigerator is continuously improved, and semiconductor refrigeration is more and more applied in industrial production scenes. Semiconductor refrigeration, also known as thermoelectric refrigeration, is a material with thermoelectric energy conversion characteristics, which is affected by Peltier effect and will appear heat absorption or heat release at the joint when direct current is passed through. Compared with traditional compression refrigeration and absorption refrigeration, semiconductor refrigeration has many advantages such as simple structure, small size, no mechanical parts, no pollution and noise pollution. Therefore, it is widely used in semiconductor laser, battery thermal management, sensor temperature control and medical field which have high temperature control requirements but great limitations.

[0003] The semiconductor temperature control system usually adopts PID control algorithm. The requirement of PID control algorithm for control parameters is relatively harsh. It is difficult for engineers to control the coordination among adjustment time, response speed and overshoot in PID parameter modulation. However, the semiconductor refrigeration system usually has certain time lag and nonlinearity, which makes the refrigeration system particularly sensitive to controller parameters, otherwise the system time will greatly increase, resulting in overshoot and decline of system stability. Industry usually adopts Z-N method and trial and error method to set the PID controller, but it is difficult to guarantee the trade-off between overshoot and response speed in semiconductor refrigeration temperature control system, and the operation is relatively cumbersome. Finally, the obtained parameters are difficult to obtain satisfactory dynamic characteristics. With the development of the times, more and more scholars introduce intelligent algorithms into controller optimization, such as particle swarm optimization (PSO) and genetic algorithm (GA). However, these heuristic algorithms often fall into local optimum when searching some complex problems, resulting in certain overshoot and long system stable time in the control process of semiconductor temperature control system, which is difficult to meet the demand of temperature control precision in some scenes. SUMMARY

[0004] In order to solve the problems in the background art, the present application provides a semiconductor temperature control system control method based on myxomycete algorithm, which overcomes the shortcomings of the prior art and solves the problem of limited search ability of conventional optimization algorithm. In the parameter optimization process, the optimal control parameter value can be searched with greater probability, thereby significantly improving the dynamic characteristics of the semiconductor refrigeration temperature control system.

[0005] The technical scheme adopted by the present application is:

[0006] The semiconductor temperature control system control method comprises the following steps:

[0007] Step 1: According to the heat transfer characteristics of the semiconductor refrigerator, a semiconductor voltage-temperature system mathematical model is established, that is, a semiconductor temperature control system differential equation is derived according to heat transfer and energy conservation law; the least square method is used to identify the semiconductor temperature control system differential equation of the semiconductor temperature control system to obtain the parameterized transfer function of the semiconductor temperature control system; the semiconductor temperature control system is excited and tested, a preset pseudo binary sequence PRBS signal is input into the semiconductor temperature control system, and the semiconductor temperature control system outputs the current temperature change value of the semiconductor temperature control system through the parameterized transfer function.

[0008] Step 2: The error value between the current temperature change value of the semiconductor temperature control system and the preset temperature value is input into the PID controller, and the output of the PID controller is input into the parameterized transfer function to control the semiconductor temperature control system, thereby forming a closed-loop control system.

[0009] Step 3: In the process of controlling the semiconductor temperature control system by the PID controller, the PID control parameters of the PID controller are optimized and iterated by the slime mold algorithm SSMA, the slime mold algorithm optimization result is obtained, the annealing variation of the slime mold algorithm optimization result is carried out, and the new solution generated by the annealing variation replaces the global optimal solution in the annealing variation process. At this time, the fitness value of the fitness function of the PID controller response is the global optimal fitness, and the iterated PID control parameters under the global optimal fitness are obtained. The PID control parameters of the PID controller are optimized, and the semiconductor temperature control system is controlled by the PID controller with optimized PID control parameters. The PID control parameters include proportional factor Kp, integral factor Ki and differential factor Kd.

[0010] In step 3, the control parameters of the PID controller are optimized and iterated by the slime mold algorithm SSMA to obtain the slime mold algorithm optimization result, which is specifically as follows:

[0011]

[0012] Wherein, And Respectively represent the slime mold algorithm optimization result of the first iteration and the second iteration; Respectively represent the slime mold algorithm optimization result of the first iteration and the second iteration; Respectively represent the slime mold algorithm optimization result of the first iteration and the second iteration; Represents the global optimal position of the slime mold algorithm SSMA in the first iteration; Represents the global optimal position of the slime mold algorithm SSMA in the first iteration; Represents the slime mold parameter, , Represents the slime mold parameter limit value; represents the feedback factor, which decreases linearly from 1 to 0; represents the smell concentration of the slime mold foraging area; and respectively represent two random slime mold positions in the next iteration; represents a random number between 0 and 1; represents the slime mold control parameter.

[0013] the smell concentration of the slime mold foraging area Specifically as follows:

[0014]

[0015]

[0016] wherein SmellIndex(i) represents the ith fitness sequence; N is the population size; bF represents the optimal fitness value obtained in the current iteration process; wF represents the worst fitness value obtained in the current iteration process.

[0017] the slime mold parameter limit value Specifically as follows:

[0018]

[0019] wherein, represents the maximum number of optimization iterations of the slime mold algorithm.

[0020] In step 3, the optimization result of the slime mold algorithm is annealed and mutated, and in the annealing and mutation process, the probability of accepting the ith new solution generated by annealing and mutation is calculated according to the Metropolis criterion , specifically as follows:

[0021]

[0022] wherein, represents the fitness increment; represents the annealing constant; represents the annealing temperature.

[0023] The determination of the initial annealing temperature is specifically as follows:

[0024]

[0025] wherein, T t represents the annealing temperature of the tth iteration; T t-1 represents the annealing temperature of the (t-1)th iteration.

[0026] the fitness increment Specifically as follows:

[0027]

[0028] in, denoted as the fitness value of the i-th individual in the solution space generated by the roulette wheel strategy during the annealing mutation process; DF represents the global optimal fitness of the population in the slime mold algorithm SSMA.

[0029] The probability of accepting the i-th new solution generated by annealing mutation Greater than random number If the annealing mutation occurs, the new solution generated by the annealing mutation is accepted and replaced with the global optimal solution in the annealing mutation process.

[0030] The slime mold control parameters Specifically as follows:

[0031]

[0032] in, Let represent the fitness value of the i-th individual in the solution space after t iterations.

[0033] The beneficial effects of this invention are:

[0034] This invention employs a slime mold algorithm for PID control, overcoming the shortcomings of existing algorithm optimizations such as long adjustment time, large overshoot, and unstable control process in semiconductor temperature control systems. It achieves rapid parameter tuning of the semiconductor temperature control system, reduces stabilization time and overshoot, and enables rapid tuning to meet the controller's parameters. This improves the response speed, stability, and control accuracy of the semiconductor temperature control system, reduces overshoot, and shortens the control system's settling time, demonstrating significant engineering application value. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the semiconductor temperature control system based on the slime mold algorithm of the present invention;

[0036] Figure 2 This is a flowchart of the target search process using the slime mold algorithm in this invention;

[0037] Figure 3 This is a comparison chart of the fitness convergence curves of this invention;

[0038] Figure 4 This is a comparison chart showing the effect of the optimized PID controller in the semiconductor temperature control system according to the present invention. Detailed Implementation

[0039] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0040] The semiconductor temperature control system control method of the present invention includes the following steps:

[0041] Step 1: Based on the heat transfer characteristics of the semiconductor cooler, establish a mathematical model of the semiconductor voltage-temperature system, that is, derive the differential equation of the semiconductor temperature control system based on heat transfer and the law of conservation of energy; use the least squares method to identify the differential equation of the semiconductor temperature control system and obtain the parameterized transfer function of the semiconductor temperature control system; perform an excitation test on the semiconductor temperature control system by inputting a preset pseudo-binary sequence PRBS signal into the semiconductor temperature control system, and the semiconductor temperature control system outputs the current temperature change value of the semiconductor temperature control system through the parameterized transfer function.

[0042] Step 2: Input the error value between the current temperature change value and the preset temperature value of the semiconductor temperature control system into the PID controller. The output of the PID controller is then input into the parameterized transfer function to control the semiconductor temperature control system, thereby forming a closed-loop control system.

[0043] Step 3: During the control of the semiconductor temperature control system by the PID controller, the PID control parameters of the PID controller are optimized iteratively using the Slime Mutagenesis Algorithm (SSMA) to obtain the optimization results. The optimization results are then subjected to annealing and mutation until a new solution generated by the annealing and mutation replaces the globally optimal solution obtained during the annealing and mutation process. At this point, the fitness function of the PID controller response has the fitness value of the globally optimal fitness. The iteratively obtained PID control parameters under the globally optimal fitness are then used to optimize the PID control parameters of the PID controller. The optimized PID controller then controls the semiconductor temperature control system using these optimized PID control parameters. The PID control parameters include the proportional factor Kp, the integral factor Ki, and the derivative factor Kd.

[0044] In step 3, the control parameters of the PID controller are optimized iteratively using the slime mold algorithm SSMA to obtain the optimization results, as follows:

[0045]

[0046] in, and They represent The next iteration and The optimization results of the slime mold algorithm in the next iteration; express The global optimal position of the slime mold algorithm SSMA in the next iteration; Indicates slime mold parameters, , Indicates the limits for slime mold parameters; This represents the feedback factor, which decreases linearly from 1 to 0. SmellIndex(i) represents the smell concentration of the slime mold foraging area; and respectively represent two random slime mold positions at the next iteration; represents a random number between 0 and 1; represents the slime mold control parameter.

[0047] SmellIndex(i) represents the smell concentration of the slime mold foraging area Specifically as follows:

[0048]

[0049]

[0050] wherein SmellIndex(i) represents the ith fitness sequence; N is the population number; bF represents the optimal fitness value obtained in the current iteration process; wF represents the worst fitness value obtained in the current iteration process.

[0051] slime mold parameter limit value Specifically as follows:

[0052]

[0053] wherein, represents the maximum number of optimization iterations of the slime mold algorithm.

[0054] In step 3, the annealing mutation of the slime mold algorithm optimization result is carried out, and in the annealing mutation process, the probability of accepting the ith new solution generated by the annealing mutation is calculated according to the Metropolis criterion , specifically as follows:

[0055]

[0056] wherein, represents the fitness increment; represents the annealing constant; represents the annealing temperature.

[0057] The determination of the initial annealing temperature is specifically as follows:

[0058]

[0059] wherein, T t represents the annealing temperature of the tth iteration; T t-1 represents the annealing temperature of the (t-1)th iteration.

[0060] fitness increment Specifically as follows:

[0061]

[0062] in, denoted as the fitness value of the i-th individual in the solution space generated by the roulette wheel strategy during the annealing mutation process; DF represents the global optimal fitness of the population in the slime mold algorithm SSMA.

[0063] The probability of accepting the i-th new solution generated by annealing mutation Greater than random number If the annealing mutation occurs, the new solution generated by the annealing mutation is accepted and replaced with the global optimal solution in the annealing mutation process.

[0064] slime mold control parameters Specifically as follows:

[0065]

[0066] in, Let represent the fitness value of the i-th individual in the solution space after t iterations.

[0067] like Figure 2 As shown, the PID controller parameters are optimized based on the slime mold algorithm. The population is initialized, and the slime mold position fitness is evaluated. Slime mold individuals are randomly generated according to boundary conditions, and the position fitness of the randomly generated slime mold individuals is calculated using the objective function. The initial annealing temperature T is set, the flavor concentration W is updated, and parameters p, a, and va are updated. The slime mold population position is updated, and finally, the optimal position and optimal fitness of the slime mold are updated. The fitness of the new slime mold position is evaluated. The slime mold position is updated, and the updated optimal fitness value DF and the global optimal position of the slime mold are determined. Annealing mutation is performed on the optimal position of the slime mold. According to the Mitropolis criterion, the probability of accepting the new solution is calculated, and it is determined whether the annealing process generates a new solution that replaces the global optimum. The annealing temperature and optimal position are updated. The slime mold population is iteratively updated according to the maximum number of iterations, the annealing temperature is updated, it is determined whether the number of iterations has been reached, and the optimal control parameter position is output.

[0068] The fitness function of the system response is selected to obtain satisfactory smoothness and dynamic characteristics, and the objective function is to obtain satisfactory transient response and dynamic characteristics. The absolute value of error and time integral performance are used as the minimum objective functions for fitness calculation. Simultaneously, to avoid overshoot due to excessive control strength, the overshoot is also included as one of the optimal indices of the fitness function. Specifically as follows:

[0069]

[0070]

[0071]

[0072] in, error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration error value between the preset temperature value and the current temperature value of the semiconductor temperature control system in the nth iteration

[0073] Specific embodiments of the present application are as follows:

[0074] 1) Set the transfer function of the semiconductor refrigeration temperature control system as:

[0075]

[0076] 2) The PID controller parameters are optimized by using the method of the present application, as shown in the following formula: Figure 1 Specifically, the following steps are included:

[0077] S1: Initialize the slime mold algorithm, set the maximum number of iterations MaxZ to 100, the number of slime molds N to 50, the search dimension dim to 3, the upper boundary of the slime mold position up to [0, 0, 0], the lower boundary lp to [-30, -1, -1], the initial temperature of annealing, and the annealing constant to 0.99.

[0078] S2: Update the population by the slime mold algorithm, generate the position of the slime mold individual randomly as the parameter of the PID controller, run the closed-loop control system to obtain the response output result, evaluate the output result by the fitness function J, calculate the fitness value of each slime mold position, and obtain the global optimal position and the global optimal fitness DF.

[0079] S3: The fitness evaluation of the population individual is realized according to the fitness function J, and the weights are w1=0.99999, w2=0.01, w3=2, and w4=50, respectively.

[0080] S4: Anneal and mutate the results obtained by the slime mold algorithm, and judge whether the new position can replace the global optimal position obtained by the slime mold algorithm by the Metropolis criterion.

[0081] S5: Determine if the iteration count condition is met. If not, return to S2. If it is met, output the global optimal position.

[0082] Table 1. Comparison of the results of optimizing semiconductor temperature control system parameters using four different methods.

[0083] Parameter Algorithm Kp Ki Kd <![CDATA[t f ]]> δ DF GA -16.224 -0.4353 -0.1424 79 2.54% 229.542 PSO -15.047 -0.8546 -0.0124 75 7.61% 231.304 SMA -20.881 -0.5081 -0.8113 70 3.52% 229.642 SSMA -16.443 -0.125 -0.0395 20 0.12% 229.2962

[0084] Search results such as Figure 3 , Figure 4 As shown in Table 1, the SSMA algorithm used in this invention achieves the best optimization of the control parameters of the semiconductor temperature control system. Its global optimal fitness (DF) is lower than that of the conventional genetic algorithm (GA), particle swarm optimization (PSO), and the common slime mold algorithm (SMA). The stable adjustment time (t) is also lower. f Both the overshoot δ and the overshoot δ are significantly better than the other three methods, and have better control effect.

[0085] The present invention discloses a semiconductor temperature control system control method based on slime mold algorithm. Based on slime mold algorithm, the method optimizes parameters according to target fitness value, combines slime mold search principle and improves the search ability of slime mold during optimization through annealing mutation, and applies the algorithm to semiconductor temperature control system control. The results obtained have the characteristics of smooth transition, fast response speed, low overshoot and short stable adjustment time during the adjustment process.

[0086] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. The present invention can also be applied to the control of other complex control systems. For those skilled in the art, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A control method for a semiconductor temperature control system based on a slime mold algorithm, characterized in that: The method includes the following steps: Step 1: Identify the differential equation of the semiconductor temperature control system using the least squares method to obtain the parameterized transfer function of the semiconductor temperature control system; conduct an excitation test on the semiconductor temperature control system by inputting a preset pseudo-binary sequence PRBS signal into the semiconductor temperature control system, and the semiconductor temperature control system outputs the current temperature change value of the semiconductor temperature control system through the parameterized transfer function; Step 2: Input the error value between the current temperature change value and the preset temperature value of the semiconductor temperature control system into the PID controller. The output of the PID controller is then input into the parameterized transfer function to control the semiconductor temperature control system. Step 3: During the process of the PID controller controlling the semiconductor temperature control system, the PID control parameters of the PID controller are optimized and iterated using the Slime Mutagenesis Algorithm (SSMA) to obtain the optimization results. The optimization results of the Slime Mutagenesis Algorithm are then subjected to annealing and mutation until the new solution generated by the annealing and mutation replaces the global optimal solution in the annealing and mutation process. At this time, the fitness value of the fitness function of the PID controller response is the global optimal fitness. Then, the iterated PID control parameters under the global optimal fitness are obtained, and the PID control parameters of the PID controller are optimized. The semiconductor temperature control system is controlled by the optimized PID controller through the PID control parameters.

2. The semiconductor temperature control system control method based on slime mold algorithm according to claim 1, characterized in that: In step 3, the control parameters of the PID controller are optimized iteratively using the slime mold algorithm SSMA to obtain the optimization results, as follows: in, and They represent The next iteration and The optimization results of the slime mold algorithm in the next iteration; express The global optimal position of the slime mold algorithm SSMA in the next iteration; Indicates slime mold parameters, , Indicates the limits for slime mold parameters; This represents the feedback factor, which decreases linearly from 1 to 0. Indicates flavor intensity; and They represent Two random slime mold locations at the next iteration; Represents a random number between 0 and 1; This indicates the control parameters for slime mold.

3. The semiconductor temperature control system control method based on slime mold algorithm according to claim 2, characterized in that: The aforementioned slime mold parameter limits Specifically as follows: in, This represents the maximum number of optimization iterations for the slime mold algorithm.

4. The semiconductor temperature control system control method based on slime mold algorithm according to claim 3, characterized in that: In step 3, the optimization result of the slime mold algorithm is subjected to annealing mutation. During the annealing mutation process, the probability of accepting the i-th new solution generated by the annealing mutation is calculated according to the Mitropolis criterion. The details are as follows: in, Indicates the fitness increment; Indicates the annealing constant; Indicates the annealing temperature.

5. The semiconductor temperature control system control method based on slime mold algorithm according to claim 4, characterized in that: The fitness increment Specifically as follows: in, denoted as the fitness value of the i-th individual in the solution space generated by the roulette wheel strategy during the annealing mutation process; DF represents the global optimal fitness of the population in the slime mold algorithm SSMA. The probability of accepting the i-th new solution generated by annealing mutation Greater than random number If the annealing mutation occurs, the new solution generated by the annealing mutation is accepted and replaced with the global optimal solution in the annealing mutation process.

6. The semiconductor temperature control system control method based on slime mold algorithm according to claim 5, characterized in that: The slime mold control parameters Specifically as follows: in, Let represent the fitness value of the i-th individual in the solution space after t iterations.

Images