A Temperature Control Method for Thermoelectric Coolers Based on Improved BP Algorithm

Abstract

The present invention discloses a temperature control method for a thermoelectric cooler based on an improved BP algorithm, belonging to the technical field of computing, inferring or counting. This method first establishes an equivalent trend model between the cold-end temperature and the control current based on the step response, uses a gradient safety factor to act on the weight adjustment process to solve the problems of network weight gradient explosion and disappearance in the BP algorithm, and adopts the starling optimization algorithm to determine the initial network weights of the BP algorithm to solve the problem of initial value dependence in the BP algorithm. Then, the improved BP algorithm is used to adaptively adjust the control parameters of the incremental PID method, and finally the incremental PID method is used to output the control current to control the temperature of the thermoelectric cooler to reach the target state. This method makes full use of the advantage of the BP algorithm's self-learning to approximate any non-linear function, solves the problem that the parameters of the PID controller are difficult to be adaptively adjusted according to the changes in the working environment and the target temperature in temperature control, and realizes fast, robust and accurate control of the cold-end temperature of the thermoelectric cooler.

Description

Technical Field

[0001] The present invention relates to the temperature control of a time-delay controlled object, specifically to the temperature control technology of a thermoelectric cooler and a semiconductor refrigeration chip, and belongs to the technical field of calculation, extrapolation or counting. Background Art

[0002] Thermoelectric refrigeration technology realizes the conversion between electrical energy and thermal energy by utilizing the thermoelectric effect to meet the requirements of temperature change of a thermoelectric module. Due to its advantages such as simple structure, no mechanical vibration, no environmental pollution, reliable operation, and fast response, thermoelectric coolers have extensive and important applications in the fields of infrared detection, aerospace, electronic industry, biochemical research, medical treatment, and civil use. A reasonable cold-end temperature not only helps to protect electronic components and meet high-precision temperature requirements, but also ensures the experimental accuracy and the normal operation of equipment in extreme environments. The temperature reaching time, process overshoot, and temperature reaching error for achieving the cold-end temperature target directly affect the working stability, efficiency, and service life of the thermoelectric cooler. However, when the cold-end temperature of the thermoelectric cooler relies on traditional PID technology to achieve a target with gradient change, there are problems such as a significant overshoot, a long temperature reaching time, and a large steady-state error. These problems are more obvious in a complex application environment, which poses challenges to the temperature control of the thermoelectric cooler.

[0003] In response to the above challenges, traditional PID control technology cannot meet the temperature control target requirements in a complex environment. Using a neural network to online adjust the PID parameters for adaptive control of the cold-end temperature is a solution. The BP algorithm is a commonly used neural network training method, which has the advantages of autonomous learning and approximating any nonlinear function. However, due to excessive control error and the time-delay accumulation of control error, when the BP algorithm is applied to the temperature control of the thermoelectric cooler, there are problems of network weight gradient explosion and gradient disappearance. Gradient explosion and gradient disappearance cause the network weight of the BP algorithm to saturate prematurely and seriously affect the efficiency of the BP algorithm to adjust the PID control parameters. In addition, the initial value sensitivity problem leads to a slow or fast learning convergence rate of the BP algorithm.

[0004] In summary, the present invention aims to propose an improved BP algorithm to meet the temperature target control requirements of a thermoelectric cooler with a wide range and various gradient changes. Summary of the Invention

[0005] The object of the present invention is to provide a method for controlling the cold-end temperature of a thermoelectric cooler and a semiconductor chip due to the deficiencies of the above-mentioned background technology. By improving the BP algorithm and applying the improved BP algorithm to online adaptively adjust the proportional parameter, integral parameter, and differential parameter, the invention aims to achieve the cold-end temperature of the thermoelectric cooler reaching multiple target states in a manner of small overshoot, high speed, and high precision, and solve the technical problems that the traditional PID control technology cannot meet the requirements of time-delay, nonlinear, and unsteady temperature control of the thermoelectric cooler in a complex environment.

[0006] The present invention adopts the following technical solutions to achieve the above object:

[0007] A method for controlling the temperature of a thermoelectric cooler based on an improved BP algorithm, comprising the following steps:

[0008] Step 1, initialize the parameters of the BP algorithm;

[0009] Step 2. Initialize the network weights of the BP algorithm according to the starling optimization algorithm;

[0010] Step 3, sample the actual value of the cold-end temperature at time k, the target temperature at time k, and the temperature error at time k;

[0011] Step 4, construct a BP neural network according to the determined initial value of the network weights, and through the forward propagation of the BP neural network of the actual value of the cold-end temperature at time k, the target temperature at time k, and the temperature error at time k, obtain the proportional parameter at time k, the integral parameter at time k, and the differential parameter at time k;

[0012] Step 5, calculate the control current at time k according to the sampling data of the previous k moments, the proportional parameter at time k, the integral parameter at time k, the differential parameter at time k, and the incremental PID control law, and apply the control current at time k to the thermoelectric cooler;

[0013] Step 6, sample the actual value of the cold-end temperature at time k + τ, and calculate the temperature error at time k + τ;

[0014] Step 7, backpropagate the temperature error at time k + τ in the BP neural network, and introduce a gradient safety factor to update the network weights during the backpropagation process;

[0015] Step 8, when the actual value of the cold-end temperature at time k + τ reaches the target temperature at time k, end the entire method process; otherwise, k = k + 1, and return to Step 3.

[0016] As a further optimized scheme of a method for controlling the temperature of a thermoelectric cooler based on an improved BP algorithm, the parameters of the BP algorithm initialized in Step 1 include but are not limited to the learning rate and the inertia coefficient.

[0017] As a further optimization scheme of the thermoelectric cooler temperature control method based on the improved BP algorithm, the specific method for initializing the network weights of the BP algorithm according to the starling optimization algorithm in step 2 is: construct the objective function cost, cost = t s +rmse, t s is the adjustment time required for the actual cold-end temperature to enter the ±1% error band, and rmse is the root mean square error between the actual cold-end temperature and the target temperature after the adjustment time t s Search for the network weight W I connecting the input layer and the intermediate layer when the objective function obtains the minimum value, and the network weight W O .

[0018] As a further optimization scheme of the thermoelectric cooler temperature control method based on the improved BP algorithm, in step 5, according to the sampling data of the previous k moments, the proportional parameter at the k moment, the integral parameter at the k moment, the differential parameter at the k moment, and the incremental PID control law, the expression for calculating the control current at the k moment is: u(k) = K P (e(k) - e(k - 1)) + K I e(k) + K D (e(k) - 2e(k - 1) + e(k - 2)) + u(k - 1)), where u(k) and u(k - 1) are the control quantities at the k moment and the k - 1 moment, e(k), e(k - 1), and e(k - 2) are the deviation signals at the k moment, the k - 1 moment, and the k - 2 moment, K P is the proportional coefficient, K I is the integral coefficient, K D is the differential coefficient.

[0019] As a further optimization scheme of the thermoelectric cooler temperature control method based on the improved BP algorithm, in step 6, the value of τ is obtained by identifying the delay link parameter in the equivalent trend transfer function between the cold-end temperature and the control current of the thermoelectric cooler.

[0020] As a further optimization scheme of the thermoelectric cooler temperature control method based on the improved BP algorithm, in step 7, the expression for updating the network weights by introducing the gradient safety factor during the backpropagation process is: Among them, W I (k), W I (k - 1) are the network weights connecting the input layer and the intermediate layer at the k moment and the k - 1 moment, ΔW I (k) is the adjustment value of the network weight connecting the input layer and the intermediate layer at the k moment according to the chain rule of error propagation in the BP algorithm, W O (k), W O(k - 1) is the network weight connecting the middle layer and the output layer at the k-th and (k - 1)-th moments, and ΔW O (k) is the adjustment value of the network weight connecting the middle layer and the output layer at the k-th moment according to the chain rule of error propagation in the BP algorithm. s is the gradient safety factor. E max is to control the maximum temperature difference between the target temperature and the actual temperature at the initial moment.

[0021] As a further optimization scheme of a thermoelectric cooler temperature control method based on the improved BP algorithm, the adjustment time t s The root mean square error between the actual cold-end temperature and the target temperature after y(k) refers to the actual value of the cold-end temperature at the k-th moment, Y(k) refers to the target temperature at the k-th moment, and N is the total number of sampling moments.

[0022] As a further optimization scheme of a thermoelectric cooler temperature control method based on the improved BP algorithm, the equivalent trend transfer function between the cold-end temperature and the control current of the thermoelectric cooler is: Among them, G I is the equivalent trend transfer function between the cold-end temperature and the control current of the thermoelectric cooler. are the time-domain functions of the cold-end temperature and the control current of the thermoelectric cooler respectively. s is the Laplace operator, k1, k2 are the coefficients of the time-domain function of the cold-end temperature of the thermoelectric cooler, and p1, p2, p3 are the coefficients of the time-domain function of the control current.

[0023] An electronic device includes a memory and a processor. A computer program is stored on the memory and runs on the processor. When the processor runs the computer program, it executes the steps of the above thermoelectric cooler temperature control method.

[0024] A computer-readable storage medium stores a computer program. When the computer program runs, it executes the steps of the above thermoelectric cooler temperature control method.

[0025] The present invention adopts the above technical solutions and has the following beneficial effects:

[0026] (1) The present invention improves the BP algorithm, introduces a gradient safety factor to update the network weight, and the network weight will not have problems of gradient explosion and gradient disappearance, and can maintain the ability of learning convergence in complex application backgrounds.

[0027] (2) The present invention improves the BP algorithm, and uses the starling optimization algorithm with both local optimization and global search capabilities to iteratively search for the best initial network weight, overcomes the defect that the BP algorithm depends on the initial network weight, and the neural network can adjust to obtain the PID control parameters with the optimal efficiency.

[0028] (3) The improved BP algorithm proposed by the present invention can adapt to the working requirements of various target temperatures and ambient temperatures, self-adaptively adjust the PID control parameters, and provide accurate and robust control effects for the cold-end temperature of the thermoelectric cooler. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a working principle block diagram of the temperature control strategy proposed by the present invention.

[0030] Figure 2 It is a neural network structure diagram of the BP algorithm proposed by the present invention.

[0031] Figure 3 It is an implementation flowchart of the temperature control strategy proposed by the present invention.

[0032] Figure 4 It is a flowchart of the temperature control method for the thermoelectric cooler proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] The present invention will be further described below with reference to the accompanying drawings.

[0034] As Figure 1 shown, the present invention proposes a temperature control strategy for optimizing the PID control parameters through an improved BP algorithm, so as to realize the adaptive control of the temperature of the thermoelectric cooler. To implement the temperature control strategy proposed in this paper, first, a reasonable transfer function relationship of the controlled object needs to be established, then the incremental PID method is used as the basic control method, then the BP algorithm is improved, and finally, the improved BP algorithm is used to online learn and converge to reasonable PID control parameters.

[0035] The neural network structure of the BP algorithm proposed by the present invention is specifically as Figure 2 shown. The BP algorithm is internally composed of an input layer, an intermediate layer, and an output layer. The input layer consists of three neurons, which respectively receive the target temperature value, the actual temperature value, and the temperature error value. The intermediate layer consists of five neurons. The output layer consists of three neurons, which respectively output the PID control parameters: the proportional parameter K P , the integral parameter K I , and the differential parameter K D . Among them, W I is the network weight connecting the input layer and the intermediate layer, and W O is the network weight connecting the intermediate layer and the output layer.

[0036] Figure 3 In order to successfully apply this temperature control strategy to the thermoelectric cooler, its implementation process is as Figure 3 shown. The following will sequentially elaborate on the four parts of establishing a reasonable transfer function relationship of the controlled object, using the incremental PID method as the basic control method, improving the BP algorithm, and using the improved BPPID method to control the temperature of the thermoelectric cooler.

[0037] Part 1: Establishing a reasonable transfer function relationship for the controlled object

[0038] Based on thermoelectric effects such as the Peltier effect, Joule effect, Thomson effect, and Fourier effect, a thermoelectric cooler can achieve the conversion of electrical energy into thermal energy. Due to the time-delay and time-varying characteristics of the cold-end temperature of the thermoelectric cooler, the transfer function relationship between the cold-end temperature and the control current cannot be easily obtained, but its equivalent trend transfer function can be obtained through system identification. Through the small-signal linearization method, a transfer function relationship approximately maintaining 1 zero and 2 poles can be obtained between the cold-end temperature and the control current. The transfer function structure between the cold-end temperature and the control current in the thermoelectric cooler is as follows:

[0039]

[0040] Furthermore, due to the non-negligible thermal conductivity of the thermal medium, heat conduction has the characteristics of hysteresis and inertia. Therefore, the equivalent trend transfer function structure between the cold-end temperature and the control current of the thermoelectric cooler is as follows:

[0041]

[0042] To obtain the six parameters k1, k2, p1, p2, p3, and τ in the structure, the system identification method needs to be used. At the working environmental temperature t0, a step-type current excitation is given to the thermoelectric cooler, and the response data t{t1, t2... t i ...} of the cold-end temperature is collected. After subtracting the environmental temperature, the change values T{T1, T2... T i ...} of the cold-end temperature under the time series can be obtained.

[0043] T{T1, T2... T i ...} = t{t1, t2... t i ...} - t0 (3)

[0044] Using the nonlinear least squares fitting method to perform parameter identification on the change values T{T1, T2... T i ...} of the cold-end temperature - control current data under the time series, the equivalent trend transfer function between the cold-end temperature T and the control current I can be obtained.

[0045] Part 2: Using the incremental PID method as the basic control method

[0046] Since computer control is a sampling control, it calculates the control quantity based on the deviation value at the sampling moment. At this time, the PID control law is expressed as:

[0047] u(k) = K P e(k) + KI ∑e(k)+K D (e(k)-e(k - 1)) (4)

[0048] In the above formula, K P is the proportional coefficient, K I is the integral coefficient, K D is the differential coefficient, k is the sampling time, k = 1, 2, 3, …; e(k - 1) and e(k) are the deviation signals obtained at the (k - 1)-th and k-th moments respectively. Classical PID control is divided into positional PID and incremental PID. Since there is no need for accumulation in the incremental control algorithm, the control increment Δu(k) is only related to the most recent k samplings, and its expression rule is:

[0049] u(k) = u(k - 1)+Δu(k)(5)

[0050] When there are incremental misoperations, the impact is small, and it is relatively easy to obtain a better control effect through weighted processing. Combining equations (4) and (5), the expression rule of incremental PID is:

[0051] u(k) = K P (e(k)-e(k - 1))+K I e(k)+K D (e(k)-2e(k - 1)+e(k - 2))+u(k - 1))(6)

[0052] Part 3: Improvement of the BP algorithm

[0053] The improvement of the BP algorithm in the present invention includes the following two aspects: adjusting the weights of the BP neural network using the gradient safety factor s and iteratively optimizing the initial values of the BP neural network using the starling optimization algorithm. The optimization and improvement process is elaborated in detail below.

[0054] 3.1. Introducing the gradient safety factor s to improve the BP algorithm

[0055] The classical BP algorithm consists of two parts: the forward propagation of the working signal and the backward propagation of the error signal. The BP neural network adjusts the weights W I and W O of each layer through error feedback, so as to achieve the purpose that the actual output of the network gradually approaches the expected output. The formula for adjusting the network weights is:

[0056]

[0057] Among them, W I , W O refers to the connection weights of the BP neural network, ΔW I , ΔW OIt is the self - adjustment value of the network weights of the BP algorithm. Affected by the structure of the BP algorithm itself, the self - adjustment value Δw is positively correlated with the error. When controlling the excessive error and the accumulation of error time delay, the self - adjustment value Δw will have problems of gradient explosion and gradient disappearance, resulting in the inability of the BP algorithm to work properly. Therefore, a gradient safety factor s is proposed to act on the self - weight adjustment process of the BP neural network, and its action formula is:

[0058]

[0059] Among them, the gradient safety factor s can normalize the error to avoid problems of gradient explosion and gradient disappearance in weight adjustment. The value of the gradient safety factor s is:

[0060]

[0061] Among them, E max refers to the maximum value of the temperature difference between the target temperature and the actual temperature at the initial moment of control, and τ is the parameter of the delay link obtained by system identification in the first part.

[0062] 3.2. Improve the BP algorithm again using the starling optimization algorithm

[0063] Since different network initial values lead to different learning efficiencies, the BP algorithm has the problem of being sensitive to the initial network weights. It is necessary to use the starling optimization algorithm to iteratively optimize the network initial values of the BP algorithm to achieve the secondary improvement of the BP neural network algorithm. During the iterative optimization of the connection initial values of the BP neural network by the starling optimization algorithm, it depends on a reasonable objective function cost. In order to balance the response speed and steady - state error of PID control within a fixed observation time, the objective function cost is defined as:

[0064] cost = t s +rmse(10)

[0065] Among them, t s refers to the adjustment time required for the actual temperature at the cold end to enter the ±1% error band, and rmse refers to the root - mean - square error between the actual temperature at the cold end and the target temperature after the adjustment time t s The expression is:

[0066]

[0067] Among them, y(k) refers to the actual value of the cold-end temperature at time k, Y(k) refers to the target temperature at time k, k is the sampling time, and N is the total number of sampling times. During the iterative optimization of the network weights by the Starling optimization algorithm, the network weights corresponding to the minimum value of the objective function cost are the optimal initial network weights. The optimal initial network weights can accelerate the learning efficiency and convergence speed of the BP algorithm, thereby optimizing the PID control parameters and further optimizing the temperature control effect of the thermoelectric cooler.

[0068] Part 4: Controlling the temperature of the thermoelectric cooler using the improved BPPID method

[0069] Introducing the gradient safety factor s and determining the best initial network value for the classical BP algorithm can form an improved BP algorithm. As Figure 4 shown, the temperature control method for the thermoelectric cooler proposed in the present invention includes the following 8 steps:

[0070] Step 1, determine key parameters such as the learning rate and inertia coefficient, and initialize the BP algorithm;

[0071] Step 2, initialize the network weights W I connecting the input layer and the middle layer, and the network weights W O connecting the middle layer and the output layer, that is, search for the initial values of W I and W O when the objective function shown in formulas (10) and (11) obtains the minimum value;

[0072] Step 3, sample the actual value y(k) of the cold-end temperature at time k, the target temperature Y(k) at time k, and the temperature error e(k) at time k, where e(k) = y(k) - Y(k);

[0073] Step 4, construct a BP neural network according to the determined initial network weights, input the actual value of the cold-end temperature at time k, the target temperature at time k, and the temperature error at time k into the BP neural network. The actual value of the cold-end temperature, the target temperature, and the temperature error at time k are propagated forward through the BP neural network to obtain the proportional parameter at time k, the integral parameter at time k, and the differential parameter at time k;

[0074] Step 5, calculate the control current at time k according to the sampling data of the previous k times, the proportional parameter at time k, the integral parameter at time k, the differential parameter at time k, and the incremental PID control law shown in formula (6), and apply the control current at time k to the thermoelectric cooler;

[0075] Step 6, sample the actual value y(k τ ) of the cold-end temperature at time k + τ, and calculate the temperature error e(k τ ), where τ is the delay link parameter obtained by system identification in Part 1;

[0076] Step 7, the temperature error e(k) at the moment of k+τ τ ) propagates backward in the BP neural network, and a gradient safety factor s is introduced during the backward propagation to update the network weights, that is, the network weights are updated according to formula (8);

[0077] Step 8, when the actual cold-end temperature y(k) at the moment of k+τ τ ) reaches the target temperature Y(k) at moment k, the entire method process ends; otherwise, k = k + 1, and return to Step 3. By looping through Steps 3 to 8, the BP neural network is enabled to learn the actual cold-end temperature, target temperature, temperature error change, and predict the proportional parameter, integral parameter, and differential parameter.

[0078] It can be seen that using the improved BP algorithm to adaptively adjust the PID control parameters, quickly learn and converge to reasonable P, I, and D control parameters online, can meet the working requirements of various target temperatures and environmental temperatures, and enable the cold-end temperature of the thermoelectric cooler to accurately and robustly reach the target state.

[0079] In an embodiment of the present invention, an electronic device is further provided, including a memory and a processor. A computer program is stored on the memory and runs on the processor. When the processor runs the computer program, it executes the steps of the above-mentioned thermoelectric cooler temperature control method.

[0080] In an embodiment of the present invention, a computer-readable storage medium is further provided, on which a computer program is stored. When the computer program runs, it executes the steps of the above-mentioned thermoelectric cooler temperature control method.

[0081] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above specific embodiments. The above specific embodiments and the descriptions in the specification are only for further explaining the principles and preparation effects of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope claimed by the present invention. The scope claimed by the present invention is defined by the claims and their equivalents.

Claims

1. A temperature control method for a thermoelectric cooler based on an improved BP algorithm, characterized in that, It includes the following steps: Step 1, initialize the parameters of the BP algorithm; Step 2. Initialize the network weights of the BP algorithm according to the starling optimization algorithm: construct the objective function cost, cost = t s + rmse, where t s is the adjustment time required for the actual cold-end temperature to enter the ±1% error band, and rmse is the root mean square error between the actual cold-end temperature and the target temperature after the adjustment time t s . Search for the network weights W I connecting the input layer and the middle layer and the network weights W O connecting the middle layer and the output layer when the objective function reaches the minimum value; Step 3, sample the actual value of the cold-end temperature at time k, the target temperature at time k, and the temperature error at time k; Step 4, construct a BP neural network according to the determined initial values of the network weights, and perform forward propagation of the actual value of the cold-end temperature at time k, the target temperature at time k, and the temperature error at time k through the BP neural network to obtain the proportional parameter at time k, the integral parameter at time k, and the derivative parameter at time k; Step 5, calculate the control current at time k based on the sampling data of the previous k moments, the proportional parameter at time k, the integral parameter at time k, the derivative parameter at time k, and the incremental PID control law, and apply the control current at time k to the thermoelectric cooler; Step 6, sample the actual value of the cold-end temperature at time k + τ, and calculate the temperature error at time k + τ; Step 7: Backpropagate the temperature error at time k + τ in the BP neural network, and introduce a gradient safety factor during the backpropagation process to update the network weights. The expression for updating the network weights by introducing a gradient safety factor during the backpropagation process is as follows: where, W I (k), W I (k - 1) are the network weights connecting the input layer and the middle layer at time k and time k - 1, ΔW I (k) is the adjustment value of the network weights connecting the input layer and the middle layer at time k according to the chain rule of error propagation in the BP algorithm, W O (k), W O (k - 1) are the network weights connecting the middle layer and the output layer at time k and time k - 1, ΔW O (k) is the adjustment value of the network weights connecting the middle layer and the output layer at time k according to the chain rule of error propagation in the BP algorithm, s is the gradient safety factor, E max is to control the maximum temperature difference between the target temperature and the actual temperature at the initial moment; Step 8, when the actual value of the cold-end temperature at time k + τ reaches the target temperature at time k, end the entire method process; otherwise, k = k + 1, and return to Step 3.

2. The temperature control method of a thermoelectric cooler based on an improved BP algorithm according to claim 1, characterized in that, The parameters of the BP algorithm initialized in Step 1 include the learning rate and the inertia coefficient.

3. The method for controlling the temperature of a thermoelectric cooler based on an improved BP algorithm according to claim 1, characterized in that, In Step 5, the expression for calculating the control current at time k based on the sampling data of the previous k moments, the proportional parameter at time k, the integral parameter at time k, the derivative parameter at time k, and the incremental PID control law is: u(k) = K P (e(k) - e(k - 1)) + K I e(k) + K D (e(k) - 2e(k - 1) + e(k - 2)) + u(k - 1)), where, u(k) and u(k - 1) are the control variables at time k and time k - 1, and e(k), e(k - 1), and e(k - 2) are the deviation signals at time k, time k - 1, and time k - 2. K P is the proportionality coefficient, K I is the integral coefficient, K D is the differential coefficient.

4. The temperature control method of a thermoelectric cooler based on an improved BP algorithm according to claim 3, characterized in that, In Step 6, the value of τ is obtained by identifying the delay link parameter in the equivalent trend transfer function between the cold-end temperature and the control current of the thermoelectric cooler.

5. A temperature control method for a thermoelectric cooler based on an improved BP algorithm according to claim 1, characterized in that, The adjusted time t s The root mean square error between the actual cold end temperature and the target temperature after y(k) refers to the actual cold end temperature at time k, Y(k) refers to the target temperature at time k, and N is the total number of sampling times.

6. The temperature control method of a thermoelectric cooler based on an improved BP algorithm according to claim 4, characterized in that The equivalent trend transfer function between the cold-end temperature of the thermoelectric cooler and the control current is as follows: where G I is the equivalent trend transfer function between the cold-end temperature of the thermoelectric cooler and the control current, are the time-domain functions of the cold-end temperature and the control current of the thermoelectric cooler respectively, s is the Laplace operator, k1 and k2 are the coefficients of the time-domain function of the cold-end temperature of the thermoelectric cooler, and p1, p2, and p3 are the coefficients of the time-domain function of the control current.

7. An electronic device, comprising a memory and a processor, wherein the memory stores a computer program that runs on the processor, and when the processor runs the computer program, it executes the steps of the thermoelectric cooler temperature control method according to Claim 1.

8. A computer-readable storage medium, on which a computer program is stored, and when the computer program runs, it executes the steps of the thermoelectric cooler temperature control method according to Claim 1.

Images