Chip temperature stability control method

Abstract

The invention relates to a chip temperature stability control method. The method comprises the following steps: acquiring a difference value between the temperature of a chip and a preset working temperature; constructing a fractional order PID controller, and respectively constructing a multi-target band constraint fitness function and a single-target fitness function based on the regulation and control time and the driving power; based on the fitness function, parameters of the fractional order PID controller are optimized through a multi-target self-adaptive Hereum algorithm; according to the multi-target self-adaptive Heman algorithm, through population classification, corresponding fitness functions are allocated to different populations so as to carry out position optimization, and global optimization of controller parameters is realized; generating a control parameter based on a difference value between the temperature of the chip and a preset working temperature by using the fractional order PID controller after parameter optimization; and the chip is heated to a preset working temperature based on the control parameters, and chip temperature stability control is completed. According to the method, the problems that an optimization algorithm in the prior art is easy to converge too early and falls into local optimum are solved, and the chip temperature stability control effect is improved.

Description

Technical Field

[0001] The invention belongs to the field of semiconductor technology, and in particular relates to a chip temperature stability control method. Background Art

[0002] The operating frequency of on-board chips such as clocks and inertial navigation will drift due to changes in ambient temperature. In severe cases, it will cause synchronization errors in the sensor network, affecting the normal operation and safety of the on-board system. Studies have shown that temperature control is one of the most effective ways to solve the frequency drift of related chips. Fractional-order PID extends the order of traditional PID from the integer domain to the real domain, so as to obtain better temperature control effect, which is expected to further improve the temperature stability of the chip. However, its parameter tuning is difficult, and a method that can optimize the parameters of fractional-order PID is needed. Common optimization methods include particle swarm optimization, whale optimization algorithm, hippo optimization algorithm, etc. For example, in 2019, Wei Lixin et al. proposed to use particle swarm optimization to improve the parameter selection process of fractional-order PID control of inverted pendulum, which improved the stability and robustness of the control effect. In 2024, Zhang Yanli et al. proposed to use the whale optimization algorithm to optimize the design of fractional-order PID controllers for first-order, second-order and other systems, which confirmed the effectiveness of swarm intelligence optimization algorithm in solving the parameter tuning problem of fractional-order controllers. In 2024, MohammadHussien Amiri and others proposed the hippopotamus optimization algorithm (HO), which simulates the natural behavior of hippos and has the characteristics of strong evolutionary ability, fast search speed and strong optimization ability.

[0003] However, the above methods all have problems such as premature convergence and falling into local optimality, which limits the optimization effect of control parameters. If the existing optimization algorithm is directly used, it will affect the control effect of chip temperature stability. Summary of the invention

[0004] In view of the above analysis, the present invention aims to disclose a chip temperature stability control method, which solves the problems in the prior art that the optimization algorithm is prone to converge prematurely and fall into local optimality, and extends the standard Hippo algorithm to multi-objective and constrained conditions, thereby increasing the applicability and practical value of the algorithm and improving the chip temperature stability control effect. Moreover, the method does not require redesign of the existing chip, thereby improving the versatility of the method and reducing the chip temperature control cost.

[0005] The purpose of the present invention is mainly achieved through the following technical solutions:

[0006] The present invention discloses a chip temperature stability control method, comprising:

[0007] Obtaining a difference between the temperature of the chip and a preset operating temperature; the preset operating temperature is a rated maximum operating temperature of the chip;

[0008] A fractional-order PID controller is constructed, and a multi-objective constrained fitness function and a single-objective fitness function are constructed based on the control time and driving power respectively;

[0009] Based on the fitness function, the parameters of the fractional-order PID controller are optimized by a multi-objective adaptive Hippo algorithm; the improved Hippo algorithm classifies the population and assigns corresponding fitness functions to different populations for position optimization, thereby achieving global optimization of the controller parameters;

[0010] Using the parameter-optimized PID controller, control parameters are generated based on the difference between the temperature of the chip and the preset operating temperature; and based on the control parameters, the chip is heated to the preset operating temperature to complete the chip temperature stability control.

[0011] Furthermore, the optimization of the fractional-order PID controller parameters by using a multi-objective adaptive Hippo algorithm includes:

[0012] S1: constructing a corresponding hippo population based on each controller parameter of the fractional-order PID, and randomly initializing each hippo position in the hippo population, wherein the hippo position is a possible solution of the corresponding controller parameter;

[0013] S2: randomly dividing the hippo population into a male hippo population and an immature hippo population, and selecting a dominant male hippo from the male hippo population based on a multi-objective constrained fitness function, and selecting a temporary leader from the immature hippo population based on a single-objective fitness function;

[0014] S3: updating the position of the male hippo population based on the multi-objective constrained fitness function; and updating the position of the immature hippo population based on the single-objective fitness function;

[0015] S4: exchange the population of the temporary leader with the non-dominant male hippo after the position update; and repeat S2-S3 to iteratively optimize the position of each hippopotamus to obtain the optimal solution of the hippopotamus position corresponding to each controller parameter, which is the optimal PID controller parameter.

[0016] Furthermore, based on the multi-objective constrained fitness function, the position of the male hippo population is updated, including:

[0017] Based on the current positions of each hippopotamus in the male hippopotamus population, the predicted positions of each hippopotamus are generated by sequentially using the exploration phase position update strategy, the defense position update strategy and the escape position update strategy in the hippopotamus algorithm;

[0018] Based on the multi-objective fitness function value corresponding to each predicted position, the optimal solution for the position of each hippopotamus is obtained; wherein, the optimal solution obtained by the previous position update strategy is the initial position of the next update strategy. After each update strategy obtains the optimal solution for the position of each hippopotamus, the position of each hippopotamus is updated based on the optimal solution.

[0019] Furthermore, the updating of the position of the immature hippopotamus population based on the single-objective fitness function includes:

[0020] Generate a predicted exploration phase position of the immature hippo population based on the temporary leader position, the dominant male hippo position and a preset adaptive factor, and update the exploration phase position based on the corresponding single-target fitness value;

[0021] Based on the updated exploration phase position, each hippopotamus position is updated in turn through the defense position update strategy and escape position update strategy of the hippopotamus algorithm and the fitness function.

[0022] Furthermore, the multi-objective constrained fitness function is expressed as:

[0023]

[0024] The single objective fitness function is expressed as:

[0025] ITAE=∫t|e(t)|dt;

[0026] Where, e(t) is the difference between the chip temperature at time t and the preset operating temperature, λ1 and λ2 are the weight coefficients of the control time and the driving power; v c (t) is the control quantity at each moment; P is the power, P max is the maximum power limit; Vcc is the supply voltage; ts is the adjustment time; t0 is the stabilization time required by the chip design; Ess is the steady-state error; E0 is the maximum steady-state error required by the chip design; σ tc The overshoot of the system, T max It is the maximum overshoot that the chip can accept.

[0027] Furthermore, the temporary leader position in the immature hippo population is obtained by the following formula:

[0028]

[0029] The predicted location of the immature hippo population during the exploration phase is generated by the following formula:

[0030]

[0031] Among them, X DFB is the position of the temporary leader, is the optimal hippo position in the immature hippo population, ITAEFBhippo is the optimal fitness value of the immature hippo obtained by the current number of iterations, and ITAEbest is the global optimal fitness value in the immature hippo population before this update; is the exploration stage position of the immature hippopotamus, X i,j is the current position of the i-th hippopotamus in the j-th parameter dimension, X Dhippo is the optimal hippo position in the current iteration, c1 is the gathering factor, ω(t) is the adaptive factor, G i is the average position of multiple randomly selected hippos.

[0032] Furthermore, the temporary leader and the non-dominant male hippopotamus exchange populations, which is expressed as:

[0033]

[0034] in, To exchange the position of the non-dominant male hippopotamus to the immature hippo population, For the non-dominant male hippopotamus to exchange positions before the population, is the position of the temporary leader after exchanging with the male hippo population, imaxFB and imaxM are the number of hippopotamuses in the immature hippo population and the male hippo population respectively; X DFB is the position of the temporary leader; X Dhippo is the global optimal hippo position; r7 and r8 are random disturbance factors with equal probability in the interval [0,1].

[0035] Furthermore, the controller parameters include parameter K p , K i , K d , λ and μ; the fractional-order PID is expressed as:

[0036] u(t)=K p e(t)+K i s -λ e(t)+K d s -μ e(t);

[0037] Wherein, u(t) is the output of the fractional-order PID, e(t) is the difference between the chip temperature at time t and the preset operating temperature, K p is the proportionality coefficient, K i is the integration coefficient, K d is the differential coefficient, S is a complex frequency domain variable, and λ and μ are non-integer orders with respect to S.

[0038] Furthermore, the temperature of the chip is obtained by a temperature measuring resistor, and the chip is heated by a heating resistor;

[0039] The temperature measuring resistor and the heating resistor are both arranged on the surface of the chip in the form of metal films through secondary packaging.

[0040] Furthermore, based on the temperature measuring resistor and three low temperature coefficient resistors, a temperature measuring bridge is constructed to detect the chip temperature;

[0041] Based on the control parameters, the heating resistor is heated by a driving element to achieve heating of the chip.

[0042] The present invention can achieve at least the following beneficial effects:

[0043] 1. The present invention proposes an improved Hippopotamus optimization algorithm that adds population division, population communication and target decomposition to solve the practical engineering problems faced in the fractional-order PID parameter tuning of the on-board chip temperature control system under multi-objective and constrained conditions; the present invention first divides the immature hippopotamus and male hippopotamus of the Hippopotamus optimization algorithm into two sub-populations, so as to realize population division and target decomposition for optimization targets with different degrees of constraints, thereby releasing the exploration characteristics of immature hippopotamus; a temporary leader is set for the immature hippopotamus population and an adaptive gathering factor is introduced to pull the immature hippopotamus population to move; and population communication is added, and the temporary leader is added to the dominant male hippopotamus population, continuously providing a wide range of search capabilities for the male hippopotamus population, realizing population communication, improving the algorithm convergence speed and avoiding falling into local optimality. The improved optimization algorithm of the present invention is used to tune the fractional-order PID parameters of the stable temperature control of the on-board chip, thereby improving the temperature control system effect of the chip under environments such as sudden temperature changes and rapid changes. In addition, the present invention extends the standard Hippopotamus algorithm to multi-objective and constrained conditions, thereby increasing the applicability and practical value of the algorithm;

[0044] 2. The present invention improves the temperature stability of the chip by arranging temperature measuring resistors and heating resistors on the chip surface and keeping the chip always operating at the rated maximum operating temperature, thereby improving the temperature control effect without changing the chip structure and reducing the control cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings are only used for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. In the entire drawings, the same reference symbols represent the same components;

[0046] Figure 1 is a flow chart of a chip temperature stability control method in an embodiment of the present invention;

[0047] Figure 2A schematic diagram of the positions of the temperature measuring resistor and the heating resistor in an embodiment of the present invention;

[0048] Figure 3 A schematic diagram of a multi-objective adaptive Hippo algorithm flow in an embodiment of the present invention;

[0049] Figure 4 Schematic diagram of a chip temperature control system in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used to illustrate the principles of the present invention together with the embodiments of the present invention.

[0051] The embodiment of the present invention discloses a chip temperature stability control method, such as Figure 1 As shown, the following steps are included:

[0052] Step S1: obtaining a difference between the temperature of the chip and a preset operating temperature; the preset operating temperature is a rated maximum operating temperature of the chip;

[0053] Specifically, this embodiment takes the vehicle-mounted clock chip as an example. In order to keep the clock chip at a constant temperature at all times, in the absence of effective heat dissipation and cooling means, the temperature of the clock chip needs to be kept at a high level at all times to avoid chip temperature fluctuations. For automotive-grade chips, their operating temperature range needs to meet -40°C to 125°C. Therefore, for automotive-grade chips, the operating temperature of the clock chip can be kept in an environment with a rated maximum operating temperature of 125°C at all times to avoid the problem of unstable chip temperature due to the inability to effectively cool down due to the increase in ambient temperature. In order to achieve the above functions, this embodiment obtains the temperature of the chip through a temperature measuring resistor and heats the chip through a heating resistor; the temperature measuring resistor and the heating resistor are both arranged on the surface of the chip in the form of a metal film through secondary packaging. Based on the temperature measuring resistor and three low temperature coefficient resistors, a temperature measuring bridge is constructed to detect the chip temperature; and the heating resistor is heated by a driving element to heat the chip; that is, in this embodiment, a temperature measuring resistor RT and a heating resistor RH are set on the chip surface, and the temperature measuring resistor RT is used to measure the temperature of the chip, provide an input signal for the chip temperature control circuit, and drive the heating resistor RH through the output signal of the chip temperature control circuit to provide a constant temperature for the chip, thereby improving its temperature stability. Figure 2As shown, in this embodiment, the clock chip is packaged as a top module on the bottom module, and the bottom module is used to provide support. The temperature measuring resistor RT and the heating resistor RH are attached to the upper surface of the bottom module in the form of a metal film, contacting the core part of the clock chip. The sensing function of the temperature measuring resistor RT is realized by virtue of the sensitivity of the metal to temperature, and the heating resistor RH is heated by applying a DC voltage to the heating resistor RH, and the clock chip is heated by the electrothermal effect. Since the physical scale inside the clock chip is in the micron order and has a good heat-insulating shell, the temperature inside the chip can be approximately regarded as uniform.

[0054] Step S2: construct a fractional-order PID controller, and construct a multi-objective constrained fitness function and a single-objective fitness function based on the control time and the driving power respectively;

[0055] Specifically, in control system design, PID controllers are widely used due to their simple implementation, good control effect and strong robustness, and fractional-order PID expands the order of traditional PID from the integer domain to the real domain, thereby achieving better temperature control effect; in order to further improve the temperature stability of the chip, this embodiment uses a fractional-order PID controller to perform temperature control based on the difference between the temperature of the chip obtained by detecting the temperature measuring resistor and the preset operating temperature.

[0056] Among them, the fractional-order PID controller parameters include parameter K p , K i , K d , λ and μ; the fractional-order PID is expressed as:

[0057] u(t)=K p e(t)+K i s -λ e(t)+K d s -μ e(t);

[0058] Wherein, u(t) is the output of the fractional-order PID, e(t) is the difference between the chip temperature at time t and the preset operating temperature, K p is the proportionality coefficient, K i is the integration coefficient, K d is the differential coefficient, s is the complex variable in the Laplace transform, and λ and μ are non-integer orders with respect to s.

[0059] It can be seen that integer-order PID is a special case of fractional-order PID where λ and μ are 1, and fractional-order PID can achieve more sophisticated control performance settings. In a control system, steady-state error and adjustment time are the main parameters of system performance. ITAE (time absolute error score) is a commonly used indicator in control system design. This standard can be used to measure the dynamic and steady-state performance of the system. The smaller its value, the better the system performance; therefore, ITAE is often used in existing control systems to measure the performance of the control system. This embodiment uses ITAE as a single-objective fitness function for iterative optimization of the Hippo algorithm; the mathematical expression of ITAE is expressed as:

[0060] ITAE=∫t|e(t)|dt;

[0061] Wherein, t is the system operation time, e(t) is the error at each moment, and in this embodiment, it is the difference between the chip temperature at moment t and the preset operating temperature.

[0062] If the ITAE of the system is too high, it means that the control effect is poor. Therefore, this embodiment uses ITAE as a single-objective fitness function, optimizes all parameters in the system through iterative calculation, and achieves the minimum ITAE to achieve the best control effect.

[0063] However, for the chip temperature control application of this embodiment, in addition to the control effect, power consumption cost is also an issue that needs to be considered in chip optimization. If only ITAE is used as the objective function, the optimization algorithm often requires huge heating and driving power in order to make the chip reach stability at the fastest speed, which is contrary to the power consumption cost design requirements of the chip. Therefore, the purpose of this embodiment is to make the best compromise between the control effect and the power consumption cost, so a multi-objective constrained fitness function is proposed to iteratively optimize the Hippo algorithm at the same time;

[0064] Among them, the multi-objective constrained fitness function is expressed as:

[0065]

[0066] Where e(t) is the difference between the chip temperature at time t and the preset operating temperature, λ1 and λ2 are the weight coefficients of the control time and driving power, which are used to balance the control effect and power consumption cost; v c (t) is the control quantity at each moment; P is the power, P max is the maximum power limit; Vcc is the supply voltage; ts is the adjustment time; t0 is the stabilization time required by the chip design; Ess is the steady-state error, E0 is the maximum steady-state error required by the chip design; σ tc The overshoot of the system, T max It is the maximum overshoot that the chip can accept.

[0067] Step S3: Based on the fitness function, the parameters of the fractional-order PID controller are optimized by a multi-objective adaptive Hippo algorithm; the improved Hippo algorithm classifies the populations and assigns corresponding fitness functions to different populations for position optimization, thereby achieving global optimization of the controller parameters;

[0068] Specifically, through the above analysis, it can be seen that the fractional-order PID parameter tuning in this article is difficult, and it is necessary to complete the controller parameter tuning under multi-objective and multi-constraint conditions. In order to solve the problem that the parameters of fractional-order PID are difficult to tune in chip temperature control systems. This embodiment proposes to tune the parameters of fractional-order PID through an improved Hippo optimization algorithm. The Hippo optimization algorithm is proposed by simulating the natural behavior of hippopotamus. It is a meta-heuristic algorithm. The algorithm has been proven to have the advantages of fast search and high precision. Existing studies have shown that its performance is better than traditional particle swarm, gray wolf and other algorithms. The standard Hippo algorithm optimizes and updates the control parameters through three stages: exploration phase position update, defense position update and escape position update.

[0069] Specifically, the multi-objective adaptive Hippo algorithm optimizes the parameters of the fractional-order PID controller by the following steps:

[0070] S31: constructing a corresponding hippo population based on each controller parameter of the fractional-order PID, and randomly initializing each hippo position in the hippo population, wherein the hippo position is a possible solution of the corresponding controller parameter;

[0071] S32: randomly dividing the hippo population into a male hippo population and an immature hippo population, and selecting a dominant male hippo from the male hippo population based on a multi-objective constrained fitness function, and selecting a temporary leader from the immature hippo population based on a single-objective fitness function;

[0072] S33: updating the position of the male hippo population based on the multi-objective constrained fitness function; and updating the position of the immature hippo population based on the single-objective fitness function;

[0073] S34: exchanging populations of the temporary leader with the non-dominant male hippopotamus after the position update;

[0074] S35: Repeat S33-S34, iteratively optimize the position of each hippopotamus, and obtain the optimal solution of the male hippopotamus position corresponding to the minimized multi-objective constrained fitness function value, which is the optimal fractional-order PID controller parameter.

[0075] More specifically, in the Hippo algorithm, the position of the Hippo represents the possible solutions for each PID parameter, and its initialization process is expressed as:

[0076] X i,j =Xi,j min +r1×(X i,j max -X i,j min );

[0077] Where, X i,j is the possible position of the i-th hippopotamus in the j-th control parameter dimension, j represents the number of variables, i.e., the j-th control parameter; r1 is a random number between [0,1]; X i,j max , X i,j min are the upper and lower bounds of the hippopotamus position respectively.

[0078] Furthermore, the location of the male hippo population is updated by the following method:

[0079] Based on the current positions of each hippopotamus in the male hippopotamus population, the predicted positions of each hippopotamus are generated by sequentially using the exploration phase position update strategy, the defense position update strategy and the escape position update strategy in the hippopotamus algorithm;

[0080] Based on the multi-objective fitness function value corresponding to each predicted position, the optimal solution for the position of each hippopotamus is obtained; wherein, the optimal solution obtained by the previous position update strategy is the initial position of the next update strategy. After each update strategy obtains the optimal solution for the position of each hippopotamus, the position of each hippopotamus is updated based on the optimal solution.

[0081] The location of the immature hippo population is updated by the following method:

[0082] Generate a predicted exploration phase position of the immature hippo population based on the temporary leader position, the dominant male hippo position and a preset adaptive factor, and update the exploration phase position based on the corresponding single-target fitness value;

[0083] Based on the updated exploration phase position, each hippopotamus position is updated in turn through the defense position update strategy and escape position update strategy of the hippopotamus algorithm and the single-objective fitness function.

[0084] More specifically, in the iterative process of the standard hippopotamus algorithm, the hippopotamus position is first updated through the exploration phase position update strategy of the first segment; in the exploration phase update strategy, each male hippopotamus X Mhippo The dominant male hippo in the iteration, that is, the current extreme point X Dhippo Update its position to show the gathering characteristics. The update formula is as follows:

[0085]

[0086] In the formula, is the position of the male hippopotamus; r2 is a random number between [0,1]; X Dhippo is the optimal hippo position before this iteration, that is, the position of the dominant male hippo; I1 is a random number in the interval [1, 2].

[0087] In addition, for immature Hippopotamus X FBhippo For example, due to their innate curiosity, they often show a tendency to move away from the group, and their position update in the exploration phase is expressed as:

[0088]

[0089] in, t is the current iteration number, τ is the maximum iteration number; r3 is a random number between [0,2]; I2 is a random number in the interval [1,2]; G i is the average position of multiple randomly selected hippos.

[0090] However, in this embodiment, in order to optimize the chip temperature fractional-order PID controller parameters, it is necessary to meet the two objectives of system control performance and minimum power consumption cost, and there are actual constraints. The existence of multiple objectives and constraints will lead to the discontinuity of the feasible solution domain, so that the algorithm cannot cross the boundary of the feasible domain and fall into the local optimum. The standard Hippo algorithm is proposed for single-objective function problems, and there is no relevant research on the multi-objective Hippo algorithm. For this reason, Figure 3 As shown, this embodiment proposes to improve the standard Hippo algorithm by adding population division, population communication and target decomposition. The male hippopotamus and immature hippopotamus in the standard Hippopotamus algorithm are divided into two populations to achieve target decomposition by optimizing the objective function respectively, and the male hippopotamus population is used as the leading one to ensure the convergence of the results; the immature hippopotamus are used as the sub-population, and the objective function is simplified to a single-objective ITAE without considering the constraints, so as to further utilize the exploration characteristics of the immature hippopotamus to explore more areas, and the population division is achieved through the parallel operation of the two populations.

[0091] Specifically, before updating the position of the immature hippo population in the exploration phase, this embodiment first sets a temporary leader X for the immature hippo population. DFB , causing immature hippos to gradually move closer to temporary leaders and move away from male populations to explore more areas.

[0092] Specifically, this embodiment obtains the temporary leader position in the immature hippo population through the following formula:

[0093]

[0094] The exploration phase position of the immature hippo population is generated by the following formula:

[0095]

[0096] Among them, X DFB is the position of the temporary leader, is the optimal hippo position in the immature hippo population, ITAEFBhippo is the optimal fitness value of the immature hippo obtained by the current number of iterations, and ITAEbest is the global optimal fitness value in the immature hippo population before this update; Predicted position for the exploration phase of immature hippos, X i,j is the current position of the i-th hippopotamus in the j-th parameter dimension, X Dhippo is the optimal hippo position in the current iteration, c1 is the gathering factor, ω(t) is the adaptive factor, G i is the average position of multiple randomly selected hippos.

[0097] After obtaining the current exploration phase position of each hippo, the multi-objective constrained fitness function is used to determine whether the position of the male hippo obtained this time is better than the optimal fitness value before the update of this stage, and the single-objective fitness function is used to determine whether the position of the immature hippo obtained this time is better than the optimal fitness value before the update of this stage. If it is better, the hippo position obtained in this iteration is retained. If the result is worse, the update of this stage is abandoned, as shown in the following formula:

[0098]

[0099] in, is the optimal position of the male hippopotamus after each iteration, is the optimal position of the immature hippo after each iteration, g represents the fitness function, gMhippo is the fitness value of the male hippo position obtained in the current stage, gbest is the fitness value of the optimal hippo position before this iterative update, ITAEFBhippo is the optimal fitness value of the immature hippo obtained in the current number of iterations, and ITAEbest is the global optimal fitness value in the immature hippo population before this update.

[0100] Furthermore, for both immature and male hippo populations, this embodiment uses the second-stage defense update strategy of the standard hippo algorithm to update the hippo position; the defense update strategy means that when hippos become targets of predators due to deviation from the group, hippos rotate themselves toward predators and use their powerful jaws and vocalizations to block and repel attackers. At this stage, hippos may behave in a way that approaches predators to induce them to retreat, and the predator's position in the search space is It is expressed as:

[0101]

[0102] Where r4 is a random number between [0,1].

[0103] Hippopotamus position during the defense update strategy phase It is expressed as:

[0104]

[0105] Where RL is a random vector with Lévy distribution; ζ, c, d, g, and r4 are random numbers between [2,4], [1,1.5], [2,3], and [-1,1] respectively; is the fitness g or ITAE of the predator position in the population, g i For each hippo in its population, the corresponding fitness gMhippo or ITEAFBhippo, during which the hippopotamus To limit the range of activity, if it is less than g i , it means that the predator is very close to the hippopotamus, and the hippopotamus quickly turns to the predator and moves towards it, causing it to retreat. On the contrary, the hippopotamus's defensive action is small; D is the distance from the hippopotamus to the predator, defined as follows:

[0106]

[0107] Among them, for male hippo populations, a is M, and for immature hippo populations, a is FB;

[0108] After obtaining the hippopotamus position of the defense update strategy, the multi-objective constrained fitness function is still used to determine whether the male hippopotamus position obtained this time is better than the optimal fitness value before the update of this stage, and the single-objective fitness function is used to determine whether the immature hippopotamus position obtained this time is better than the optimal fitness value before the update of this stage. If it is better, the male hippopotamus and immature hippopotamus positions of this iteration are retained respectively. If the result is worse, the update of this stage is abandoned, as shown in the following formula:

[0109]

[0110] Among them, gMhippor is the fitness value of the male hippo position obtained at the current iteration number in the defense phase, gbest is the fitness value of the optimal hippo position before this iterative update, ITAEFBhippor is the optimal fitness value of the immature hippo obtained at the current iteration number in the defense phase, and ITAEbest is the global optimal fitness value in the immature hippo population before this update.

[0111] Furthermore, the third stage is the escape position update strategy stage. Another option for the hippopotamus is that the hippopotamus will try to leave the danger and try to run to a nearby cluster to avoid harm from the predator. This strategy enables the hippopotamus to find a safe location near its current location, thereby enhancing the utilization ability in local search. The calculation method is:

[0112]

[0113] Where X local min , X local max are the upper and lower bounds of the current safe community position, r5 is a random number between [0,1]; r6 is a random number with normal distribution.

[0114] After obtaining the hippopotamus position in the escape position update strategy stage, the multi-objective constrained fitness function is still used to determine whether the male hippopotamus position obtained this time is better than the optimal fitness value before the update in this stage, and the single-objective fitness function is used to determine whether the immature hippopotamus position obtained this time is better than the optimal fitness value before the update in this stage. If it is better, the male hippopotamus and immature hippopotamus positions of this iteration are retained respectively. If the result is worse, the update of this stage is abandoned, as shown in the following formula:

[0115]

[0116] Among them, gMhippoε is the fitness value of the male hippo position obtained at the current iteration number of the escape stage, gbest is the fitness value of the optimal hippo position before this iterative update, ITAEFBhippoε is the optimal fitness value of the immature hippo obtained at the current iteration number of the escape stage, and ITAEbest is the global optimal fitness value in the immature hippo population before this update.

[0117] In order to prevent the overall parameter optimization process from falling into the local optimum, the present embodiment uses immature hippos to provide new individuals for male hippos, thereby generating information interaction and realizing population communication; that is, after the escape update strategy stage, a population communication link is added, the temporary leader of the immature hippopotamus is exchanged with one of the male hippopotamus population, the temporary leader is added to the dominant male hippopotamus population, and a large-scale search capability is continuously provided for the male hippopotamus population, and a non-dominant male hippopotamus is added to the immature hippopotamus population for the next round of iterative update;

[0118] Preferably, the temporary leader and the non-dominant male hippopotamus are exchanged for populations by the following formula:

[0119]

[0120] in, To exchange the position of non-dominant male hippopotamus after entering the immature hippo population, For the non-dominant male hippopotamus to exchange positions before the population, is the position of the temporary leader after exchanging with the male hippo population, imaxFB and imaxM are the number of hippopotamuses in the immature hippo population and the male hippo population respectively; X DFB is the position of the temporary leader; X Dhippo is the global optimal hippo position; r7 and r8 are random disturbance factors with equal probability in the interval [0,1].

[0121] After multiple iterations, the optimal solution of the male hippo position corresponding to the minimized multi-objective constrained fitness function value is obtained as the optimal fractional-order PID controller parameter. This embodiment uses a single-objective unconstrained fitness function to increase the exploratory nature of the overall algorithm, helping the hippo algorithm to jump out of the feasible domain island caused by constraints and multiple objectives, and continuously transmits the feasible solution of immature hippopotamus to the male hippo population through population communication to provide more feasible domain information, thereby improving the optimization effect of the algorithm.

[0122] Step S4: Based on the difference between the temperature of the chip and the preset working temperature, a control parameter is generated by using the parameter-optimized PID controller; and based on the control parameter, the chip is heated to the preset working temperature to complete the temperature stability control of the chip;

[0123] Specifically, Figure 4 As shown, the parameters of the optimized fractional-order PID are K p , K i , K d , λ and μ, when there is an error between the preset chip operating temperature and the chip temperature fed back by the measuring resistor, the fractional-order PID controller is based on the input temperature error and the optimized controller parameter K p , K i , K d , λ and μ, calculate the control amount and drive the subsequent heating resistors to stabilize the temperature of the controlled object, i.e. the chip, at the preset temperature.

[0124] In summary, the chip temperature stability control method of the present invention improves the temperature stability of the chip by setting a temperature measuring resistor and a heating resistor on the chip surface, and makes the chip always work at the rated maximum operating temperature, and through fractional-order PID control, it improves the temperature control effect without changing the chip structure, and reduces the control cost. And the present invention proposes an improved hippo optimization algorithm that adds population division, population communication and target decomposition to solve the practical engineering problems faced in the fractional-order PID parameter setting of the vehicle-mounted chip temperature control system under multi-objective and constraint conditions; the present invention first divides the immature hippopotamus and mature hippopotamus of the hippo optimization algorithm into two sub-populations, so as to realize population division and target decomposition for optimization targets with different degrees of constraints, respectively, and release the exploration characteristics of immature hippopotamus; a temporary leader is set for the immature hippopotamus population and an adaptive gathering factor is introduced to pull the immature hippopotamus population to move; and population communication is added, and the temporary leader is added to the dominant male hippopotamus population, and a large-scale search capability is continuously provided for the male hippopotamus population, so as to realize population communication, improve the algorithm convergence speed and avoid falling into the local optimum. The improved optimization algorithm of the present invention is used to adjust the fractional-order PID parameters of the vehicle-mounted chip temperature stability control, thereby improving the temperature control effect of the chip under the environment of sudden temperature change and rapid change.

[0125] Those skilled in the art will appreciate that all or part of the process of the method in the above embodiment can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0126] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A chip temperature stability control method, characterized in that: include: Obtaining a difference between the temperature of the chip and a preset operating temperature; the preset operating temperature is a rated maximum operating temperature of the chip; A fractional-order PID controller is constructed, and a multi-objective constrained fitness function and a single-objective fitness function are constructed based on the control time and driving power respectively; Based on the fitness function, the parameters of the fractional-order PID controller are optimized by a multi-objective adaptive Hippo algorithm; the improved Hippo algorithm classifies the population and assigns corresponding fitness functions to different populations for position optimization, thereby achieving global optimization of the controller parameters; Using the fractional-order PID controller after parameter optimization, control parameters are generated based on the difference between the temperature of the chip and the preset working temperature; and based on the control parameters, the chip is heated to the preset working temperature to complete the temperature stability control of the chip.

2. The chip temperature stability control method according to claim 1, characterized in that: The optimization of the fractional-order PID controller parameters by using a multi-objective adaptive Hippo algorithm includes: S1: constructing a corresponding hippo population based on each controller parameter of the fractional-order PID, and randomly initializing each hippo position in the hippo population, wherein the hippo position is a possible solution of the corresponding controller parameter; S2: randomly dividing the hippo population into a male hippo population and an immature hippo population, and selecting a dominant male hippo from the male hippo population based on a multi-objective constrained fitness function, and selecting a temporary leader from the immature hippo population based on a single-objective fitness function; S3: updating the position of the male hippo population based on the multi-objective constrained fitness function; and updating the position of the immature hippo population based on the single-objective fitness function; S4: exchanging populations between the temporary leader after the position update and the non-dominant male hippopotamus; S5: Repeat S3-S4, iteratively optimize the position of each hippopotamus, and obtain the optimal solution of the male hippopotamus position corresponding to the minimized multi-objective constrained fitness function value, which is the optimal fractional-order PID controller parameters.

3. The chip temperature stability control method according to claim 2, characterized in that: Based on the multi-objective constrained fitness function, the position of the male hippo population is updated, including: Based on the current positions of each hippopotamus in the male hippopotamus population, the predicted positions of each hippopotamus are generated by sequentially using the exploration phase position update strategy, the defense position update strategy and the escape position update strategy in the hippopotamus algorithm; Based on the multi-objective fitness function value corresponding to each predicted position, the optimal solution for the position of each hippopotamus is obtained; wherein, the optimal solution obtained by the previous position update strategy is the initial position of the next update strategy. After each update strategy obtains the optimal solution for the position of each hippopotamus, the position of each hippopotamus is updated based on the optimal solution.

4. The chip temperature stability control method according to claim 2, characterized in that: The method of updating the position of the immature hippopotamus population based on the single-objective fitness function includes: Generate a predicted exploration phase position of the immature hippo population based on the temporary leader position, the dominant male hippo position and a preset adaptive factor, and update the exploration phase position based on the corresponding single-target fitness value; Based on the updated exploration phase position, each hippopotamus position is updated in turn through the defense position update strategy and escape position update strategy of the hippopotamus algorithm and the fitness function.

5. The chip temperature stability control method according to claim 2, characterized in that: The multi-objective constrained fitness function is expressed as: The single objective fitness function is expressed as: ITAE=∫t|e(t)|dt; Where, e(t) is the difference between the chip temperature at time t and the preset operating temperature, λ1 and λ2 are the weight coefficients of the control time and the driving power; v c (t) is the control quantity at each moment; P is the power, P max is the maximum power limit; Vcc is the supply voltage; ts is the adjustment time; t0 is the stabilization time required by the chip design; Ess is the steady-state error, E0 is the maximum steady-state error required by the chip design; σ tc The overshoot of the system, T max It is the maximum overshoot that the chip can accept.

6. The chip temperature stability control method according to claim 4, characterized in that: The temporary leader position in the immature hippo population is obtained by the following formula: The predicted location of the immature hippo population during the exploration phase is generated by the following formula: Among them, X DFB is the position of the temporary leader, is the position of the immature hippo in the exploration phase, ITAEFBhippo is the optimal fitness value of the immature hippo obtained in the current iteration, and ITAEbest is the global optimal fitness value of the immature hippo population before this update; X i,j is the current position of the i-th hippopotamus in the j-th parameter dimension, X Dhippo is the dominant hippo position in the current iteration, c1 is the gathering factor, ω(t) is the adaptive factor, G i is the average position of multiple randomly selected hippos.

7. The chip temperature stability control method according to claim 2, characterized in that: The temporary leader and the non-dominant male hippopotamus exchange populations, expressed as: in, To exchange the position of non-dominant male hippopotamus after entering the immature hippo population, For the non-dominant male hippopotamus to exchange positions before the population, is the position of the temporary leader after exchanging with the male hippo population, imaxFB and imaxM are the number of hippopotamuses in the immature hippo population and the male hippo population respectively; X DFB is the position of the temporary leader; Dhippo is the global optimal hippo position; r7 and r8 are random disturbance factors with equal probability in the interval [0,1].

8. The chip temperature stability control method according to claim 2, characterized in that: The controller parameters include parameter K p , K i , K d , λ and μ; the fractional-order PID is expressed as: u(t)=K p e(t)+K i s -λ e(t)+K d s -μ e(t); Wherein, u(t) is the output of the fractional-order PID, e(t) is the difference between the chip temperature at time t and the preset operating temperature, K p is the proportionality coefficient, K i is the integration coefficient, K d is the differential coefficient, s is the complex variable in the Laplace transform, and λ and μ are non-integer orders with respect to s.

9. The chip temperature stability control method according to claim 1, characterized in that: The temperature of the chip is obtained by a temperature measuring resistor, and the chip is heated by a heating resistor; The temperature measuring resistor and the heating resistor are both arranged on the surface of the chip in the form of metal films through secondary packaging.

10. The chip temperature stability control method according to claim 2, characterized in that: Based on the temperature measuring resistor and three low temperature coefficient resistors, a temperature measuring bridge is constructed to detect the chip temperature; Based on the control parameters, the heating resistor is heated by a driving element to achieve heating of the chip.

Images