Heat flow real-time inversion method and device based on annealing gradient algorithm

Abstract

The invention discloses a heat flow real-time inversion method and device based on an annealing gradient algorithm, and relates to the technical field of numerical calculation, and the method comprises the steps: carrying out the numerical calculation based on an external surface heat flow initial value, obtaining a temperature field, selecting a temperature value of a measurement point position, constructing an objective function, randomly generating a new solution, comparing the objective function constructed by the new solution and an old solution, and obtaining a heat flow real-time inversion result; if yes, the new solution serves as a candidate solution, if not, the new solution serves as the candidate solution according to the Monte Carlo judgment criterion by means of probability decision, and local iterative optimization is carried out by means of the gradient information optimization search process till the target function value reaches a predefined threshold value or the temperature parameter is reduced to the lower temperature limit; and ending under the condition that an end condition is met to obtain a final thermal flow vector. Through the randomness of the annealing algorithm and the Monte Carlo judgment criterion, local optimum is avoided, the multi-peak problem of nonlinear heat flow boundary condition identification can be solved, and an accurate identification result of the heat flow on the surface of the thermal protection structure is obtained.

Description

Technical Field

[0001] The present application relates to the field of numerical computing technology, and in particular to a method and device for real-time inversion of heat flow based on an annealing gradient algorithm. Background Art

[0002] In the process of thermal flow parameter identification on the surface of an aircraft, the method usually adopted is to install temperature sensors inside the thermal protection layer, and reversely infer the heat flow on its outer wall based on the temperature change curve of specific points inside the thermal protection layer recorded by these sensors. For thermal protection materials that are nearly homogeneous or do not have microstructure design, traditional identification technology can achieve a relatively ideal identification effect. However, with the development of aerospace technology, a new generation of active thermal protection solutions such as sweat cooling systems have been widely studied and applied. Such systems contain a complex microchannel network for the circulation and heat exchange of cooling fluids.

[0003] Due to special design requirements, the temperature detection equipment used for heat flow identification often needs to be placed on the inner wall, which means that the traditional heat flow inversion algorithm based on cold end temperature may not provide accurate identification results because it is easy to fall into the local optimum. At the same time, compared with using hot end data, relying solely on cold end temperature to estimate heat flow parameters will significantly increase the non-well-posedness of the problem, that is, the uniqueness, stability and existence of the solution are difficult to guarantee. Especially in applications that pursue high time resolution, the identification accuracy of traditional methods often cannot meet actual needs. Summary of the invention

[0004] The purpose of this application is to provide a real-time inversion method and device for heat flow based on an annealing gradient algorithm in order to overcome the existing technical defects. The conjugate gradient algorithm can effectively overcome the ill-posed problem in the inverse heat transfer problem. The randomness of the annealing algorithm and the Monte Carlo judgment criterion avoid falling into the local optimum, and effectively solve the multi-peak problem of nonlinear heat flow boundary condition identification, thereby obtaining accurate identification results of the heat flow on the surface of the thermal protection structure.

[0005] The purpose of this application is achieved through the following technical solutions: In a first aspect, the present application proposes a real-time heat flow inversion method based on an annealing gradient algorithm, the method comprising: Step S1, establishing a fluid-solid convection heat transfer equation based on the thermal non-equilibrium assumption and giving boundary conditions and initial values ​​of the outer surface heat flux, and performing numerical calculation based on the initial values ​​of the outer surface heat flux to obtain a temperature field; Step S2, selecting the temperature value of the measuring point based on the temperature field and constructing the objective function to randomly generate a new solution, and recalculating the objective function corresponding to the new solution; Step S3, comparing the objective function constructed by the new solution and the old solution, if the objective function constructed by the new solution is smaller than the objective function constructed by the old solution, the new solution is used as a candidate solution, and the heat flow to be identified is updated as a new solution for the next iteration step, if the objective function constructed by the new solution is not smaller than the objective function constructed by the old solution, the probability is used to determine whether the new solution is used as a candidate solution according to the Monte Carlo judgment criterion; Step S4, perform local search near the neighborhood of the candidate solution, and use the gradient information to optimize the search process to perform local iterative optimization, and repeat the local search process; Step S5, after completing a local search at an annealing temperature each time, update the iteration counter, check whether the current iteration number has reached the preset maximum iteration number, if not, continue the search process at the current annealing temperature, if reached, lower the annealing temperature as the new annealing temperature, execute steps S2 to S5 at the new annealing temperature, and end when the termination condition is met to obtain the final heat flow vector.

[0006] In a possible implementation, in step S2, the constructed objective function for: ,in represents the heat flux to be identified, is the temperature field obtained by numerical calculation, is the temperature value at the measuring point, is the number of discrete points in space, is the number of discrete time points, is the number of inner loop iterations, is the number of outer loop iterations, are discrete points in space, are discrete points in time.

[0007] In one possible implementation, in step S3, the probability for: , The objective function constructed for the old solution is: The objective function constructed for the new solution is: is the Boltzmann constant, is the annealing temperature, is the difference between the objective functions corresponding to the new solution and the old solution.

[0008] In a possible implementation, in step S4, the process of local iterative optimization is: ,in represents the next heat flux to be identified, is the gradient decreasing direction, The expression is: ,in is the conjugate coefficient, is the previous gradient reduction direction, is the step length, The expression is: , , , the conjugate gradient iteration direction is: .

[0009] In a second aspect, the present application also provides a real-time heat flow inversion device based on an annealing gradient algorithm, the device comprising: The first calculation module is used to establish the fluid-solid two-phase convection heat transfer equation according to the thermal non-equilibrium assumption and give the boundary conditions and the initial value of the outer surface heat flux, and perform numerical calculation based on the initial value of the outer surface heat flux to obtain the temperature field; The second calculation module is used to select the temperature value of the measuring point position based on the temperature field and construct the objective function to randomly generate a new solution, and then calculate the objective function corresponding to the new solution again; A comparison module is used to compare the objective function constructed by the new solution and the old solution. If the objective function constructed by the new solution is smaller than the objective function constructed by the old solution, the new solution is used as a candidate solution, and the heat flow to be identified is updated as the new solution for the next iteration. If the objective function constructed by the new solution is not smaller than the objective function constructed by the old solution, the probability is used to determine whether the new solution is used as a candidate solution according to the Monte Carlo judgment criterion; Iteration module, used to perform local search near the neighborhood of candidate solutions, and use gradient information to optimize the search process for local iterative optimization and repeat the local search process; The generation module is used to update the iteration counter after completing a local search at an annealing temperature each time, check whether the current number of iterations has reached the preset maximum number of iterations, if not, continue the search process at the current annealing temperature, if reached, lower the annealing temperature as the new annealing temperature, execute the second calculation module to the generation module at the new annealing temperature, and end when the termination condition is met to obtain the final heat flow vector.

[0010] In one possible implementation, the objective function constructed is for: ,in represents the heat flux to be identified, is the temperature field obtained by numerical calculation, is the temperature value at the measuring point, is the number of discrete points in space, is the number of discrete time points, is the number of inner loop iterations, is the number of outer loop iterations, are discrete points in space, are discrete points in time.

[0011] In one possible implementation, the probability for: , The objective function constructed for the old solution is: The objective function constructed for the new solution is: is the Boltzmann constant, is the annealing temperature, is the difference between the objective functions corresponding to the new solution and the old solution.

[0012] In a possible implementation, the process of local iterative optimization is: ,in represents the next heat flux to be identified, is the gradient decreasing direction, The expression is: ,in is the conjugate coefficient, is the previous gradient reduction direction, is the step length, The expression is: , , , the conjugate gradient iteration direction is: .

[0013] The above main scheme of the present application and its further options can be freely combined to form multiple schemes, all of which are schemes that can be adopted and claimed for protection in the present application; and in the present application, (non-conflicting options) options and other options can also be freely combined. After understanding the scheme of the present application, those skilled in the art can understand that there are many combinations based on the existing technology and common knowledge, all of which are technical schemes to be protected by the present application, and they are not exhaustively listed here.

[0014] The present application discloses a real-time heat flow inversion method and device based on an annealing gradient algorithm, which relates to the field of numerical calculation technology. First, the temperature field is obtained by numerical calculation based on the initial value of the external surface heat flow, the temperature value of the measuring point position is selected and the objective function is constructed to randomly generate a new solution, and the new solution is compared with the objective function constructed by the old solution. If it is less than, the new solution is used as a candidate solution. If it is not less than, the new solution is used as a candidate solution according to the probability according to the Monte Carlo judgment criterion. The gradient information optimization search process is used to perform local iterative optimization until the objective function value reaches a predefined threshold or the temperature parameter is reduced to the lower limit of the temperature. When the termination condition is met, the final heat flow vector is obtained. The randomness of the annealing algorithm and the Monte Carlo judgment criterion avoid falling into the local optimum, solve the multi-peak problem of nonlinear heat flow boundary condition identification, and obtain accurate identification results of the surface heat flow of the thermal protection structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 A flow chart of a real-time heat flow inversion method based on an annealing gradient algorithm proposed in an embodiment of the present application is shown.

[0017] Figure 2 The diagram shows the layout of temperature sensors for real-time inversion of spatiotemporal alternating heat flux on the surface of the sweating cooling structure. DETAILED DESCRIPTION

[0018] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0019] Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of this application.

[0020] In the prior art, compared with using the hot end temperature, using only the cold end temperature will lead to increased instability in heat flow identification. The traditional iterative method has low identification accuracy at the same time resolution and is difficult to meet the time accuracy requirements for heat flow identification on the aircraft surface.

[0021] In view of this, it is necessary to make innovative improvements to the identification method of the heat flux on the outer wall according to the characteristics of the sweat cooling thermal protection system. Specifically, it is necessary to develop a set of technical solutions that can effectively invert the law of the change of the external heat flux over time and space under given inner wall temperature conditions, so as to overcome the limitations of existing technologies and improve the accuracy and reliability of heat flux parameter identification.

[0022] Therefore, the embodiment of the present application proposes a real-time inversion method and device for heat flow based on the annealing gradient algorithm, which is used to solve the problem of identifying the heat flow on the surface of the heat protection structure. Specifically, through the relationship between the temperature data of the internal temperature measurement points and the heat flow to be identified, an optimization problem is constructed, which effectively avoids the problem of ill-posedness encountered when directly solving the heat flow of the sweating cooling thermal protection system. In the algorithm implementation, a hybrid strategy combining the annealing algorithm and the conjugate gradient method is adopted. Among them, the conjugate gradient method can effectively deal with the ill-posedness in the inverse problem of heat transfer, and the randomness of the annealing algorithm and the Monte Carlo judgment criterion help to avoid the local optimal trap that occurs during the optimization process, thereby ensuring that the global optimal solution can be found from multiple possible solutions, and finally realizing the accurate identification of the heat flow on the surface of the thermal protection structure. Next, the real-time inversion method and device for heat flow proposed in the embodiment of the present application are described in detail.

[0023] Please refer to Figure 1 , Figure 1 A schematic flow chart of a heat flow real-time inversion method based on an annealing gradient algorithm proposed in an embodiment of the present application is shown, and the method comprises the following steps: Step S1, based on the thermal non-equilibrium assumption, establish the fluid-solid two-phase relative heat transfer equation and give the boundary conditions and the initial value of the outer surface heat flux, and perform numerical calculation based on the initial value of the outer surface heat flux to obtain the temperature field.

[0024] Figure 2 The diagram shows the layout of the temperature sensor for real-time inversion of the space-time alternating heat flow on the surface of the sweating cooling structure, which shows the internal layout of the sweating cooling structure. Multiple temperature measuring points are set in the solid domain of the inner wall. These temperature measuring points are distributed at different thickness positions. The position of each temperature measuring point is clearly marked to monitor the temperature change of the inner wall surface, so as to invert the heat flow of the outer wall surface. The heat flow of the outer wall surface acts on the solid and fluid domains according to the thermal conductivity ratio to cool the inner wall temperature of the sweating cooling thermal protection system. Invert the heat flow on its outer surface For example, considering the heat conduction and convection effects in the sweating cooling process, the thermal non-equilibrium hypothesis is used to describe the heat transfer between the fluid and the solid, which involves the temperature difference between the fluid phase and the solid phase. The fluid-solid convection heat transfer equation is based on the principle of heat transfer, and a mathematical model describing the convection heat transfer between the fluid and the solid in the sweating cooling area is established.

[0025] The boundary condition can be set as a heat flow boundary condition, that is, considering the influence of the convection heat transfer coefficient and the ambient temperature, assuming an initial outer surface heat flux distribution as the initial value of the outer surface heat flux , perform numerical calculations, use numerical methods (such as finite difference method, finite element method, etc.) to solve the fluid-solid two-phase convection heat transfer equation established above, input the assumed initial value of the external surface heat flux, perform numerical simulation, and calculate the temperature field distribution of the entire thermal protection system under the heat flux condition , the internal temperature field distribution of the thermal protection system under the given initial value of the external surface heat flux can be obtained, providing basic data for optimization and inversion, and the heat flux distribution of the external surface can be inverted through the optimization algorithm to ensure the accuracy and reliability of the results.

[0026] Step S2: based on the temperature field, select the temperature value of the measuring point and construct the objective function to randomly generate a new solution, and calculate the objective function corresponding to the new solution again.

[0027] In the temperature field obtained by numerical calculation, find the temperature value corresponding to the actual temperature measurement point position. Assuming that the position of the temperature measurement point is known, then construct an objective function to measure the difference between the temperature obtained by numerical calculation and the temperature measured at the actual temperature measurement point. In step S2, the constructed objective function for: ,in represents the heat flux to be identified, is the temperature field obtained by numerical calculation, is the temperature value at the measuring point, is the number of discrete points in space, is the number of discrete time points, is the number of inner loop iterations, is the number of outer loop iterations, are discrete points in space, are discrete points in time.

[0028] Then, a new solution is randomly generated to generate a set of new external surface heat flux distribution as candidate solutions, which can be achieved by a random number generator or other randomization methods. The new external surface heat flux distribution (new solution) is used to recalculate the numerical value to obtain a new temperature field, substitute the new temperature field into the objective function, calculate the new objective function value, and compare the new and old objective function values.

[0029] Step S3, compare the objective function constructed by the new solution and the objective function constructed by the old solution. If the objective function constructed by the new solution is smaller than the objective function constructed by the old solution, the new solution is used as a candidate solution, and the heat flow to be identified is updated as the new solution for the next iteration step. If the objective function constructed by the new solution is not smaller than the objective function constructed by the old solution, the Monte Carlo judgment criterion is used to use probability to decide whether to use the new solution as a candidate solution.

[0030] Calculate the objective function value of the new solution and the old solution, and compare the two. If ( , that is, the new solution is better than the old solution, directly accept the new solution as the current optimal solution, and update the objective function value. If , that is, the new solution has not become better, and it is necessary to decide whether to accept the new solution based on the Monte Carlo judgment criterion. The probability adopted by the Monte Carlo judgment criterion is for: , The objective function constructed for the old solution is: The objective function constructed for the new solution is: is the Boltzmann constant, is the annealing temperature, is the difference between the objective functions corresponding to the new solution and the old solution.

[0031] Repeat the above steps, continuously generate new solutions and compare and accept the objective function values ​​until the termination condition is met: the predetermined maximum number of iterations is reached, and the change in the objective function value is less than a preset minimum threshold, indicating that the optimization process has converged. In addition, the termination condition can also include the annealing temperature dropping to a certain set value, indicating that the optimization process is close to the end. Through this iterative process, the external surface heat flux distribution is gradually optimized, and finally a more accurate external surface heat flux distribution is obtained, realizing the accurate identification of the heat flux on the surface of the thermal protection structure.

[0032] Step S4: perform a local search near the neighborhood of the candidate solution, and use the gradient information to optimize the search process for local iterative optimization, and repeat the local search process.

[0033] After obtaining the candidate solution and the corresponding objective function value, a local search is performed to generate a series of new solutions in the neighborhood of the candidate solution. The solution can be generated by adding a small perturbation. Then, the search process is optimized using gradient information, and the gradient of the objective function is calculated to determine the search direction, so that the objective function value decreases faster and the candidate solution is updated at the same time. Then, local iterative optimization is performed to generate new candidate solutions, and new objective function values ​​are calculated for comparison until the termination conditions are met (the objective function value reaches the predefined threshold, the annealing temperature parameter is reduced to its lower limit, and the predetermined maximum number of iterations is reached).

[0034] In step S4, the process of local iterative optimization is: ,in represents the next heat flux to be identified, is the gradient decreasing direction, The expression is: ,in is the conjugate coefficient, is the previous gradient reduction direction, is the step length, The expression is: , , , conjugate gradient iteration direction: .

[0035] Step S5, after completing a local search at an annealing temperature each time, update the iteration counter, check whether the current iteration number has reached the preset maximum iteration number, if not, continue the search process at the current annealing temperature, if reached, lower the annealing temperature as the new annealing temperature, execute steps S2 to S5 at the new annealing temperature, and end when the termination condition is met to obtain the final heat flow vector.

[0036] After completing a local search at an annealing temperature each time, increase the iteration counter n (n = n + 1), then check whether the current iteration number n has reached the preset maximum iteration number N, if the iteration number is greater than or equal to the preset maximum iteration number, reduce the annealing temperature, take the new annealing temperature as the temperature value, re-execute steps S2 to S5, generate a new candidate solution, perform numerical calculations, obtain the temperature field corresponding to the new solution, calculate the objective function value corresponding to the new solution for comparison, update the current solution according to the acceptance criterion, perform local search near the neighborhood of the candidate solution, use gradient information to optimize the search process, perform local iterative optimization, repeat the above process until the termination condition is met to obtain the final heat flow vector, if the termination condition is not met, slowly reduce the annealing temperature, and continue to execute step S2.

[0037] In summary, the real-time heat flow inversion method proposed in the embodiment of the present application utilizes the basic theories of heat transfer and convex optimization to transform the heat flow identification problem on the surface of the heat protection structure into an optimization problem based on the temperature of the internal temperature measurement points and the heat flow structure to be identified. An annealing-gradient algorithm (SA-CG) is formed by combining the annealing algorithm and the conjugate gradient algorithm to solve the optimization problem. In the establishment of the mathematical model of the thermal protection layer, the thermal non-equilibrium hypothesis is used to describe the temperature of the fluid phase and the solid phase of the sweating cooling system. At the outer wall, the heat flow acts on the solid and fluid domains according to the ratio of thermal conductivity. Temperature measurement points are set in the solid domain of the inner wall, and measurement points are set at positions of different thicknesses of the three layers of the inner wall. The heat flow on the outer wall is inverted through the temperature data of these measurement points.

[0038] Compared with the prior art, the embodiments of the present application have the following beneficial effects: First, the outer surface heat flux can be obtained by only measuring the inner surface cold end temperature, without measuring the outer surface hot end temperature of the thermal protection layer, thus simplifying the measurement process.

[0039] Second, the conjugate gradient algorithm can effectively overcome the ill-posedness problem in the inverse heat transfer problem and improve the accuracy of heat flow identification.

[0040] Third, through the randomness of the annealing algorithm and the Monte Carlo judgment criterion, it has the ability to jump out of the local optimal trap, avoid the inversion process from falling into the local optimal, and effectively solve the multi-peak problem of nonlinear heat flow boundary condition identification.

[0041] Fourth, the higher temporal resolution of the heat flow identification results is conducive to the rapid assessment of the aerodynamic thermal environment and supports the design and engineering application of aircraft thermal protection systems.

[0042] A possible implementation of a heat flow real-time inversion device based on an annealing gradient algorithm is given below, which is used to execute each execution step and corresponding technical effect of the heat flow real-time inversion method shown in the above embodiment and possible implementation. The device includes: The first calculation module is used to establish the fluid-solid two-phase convection heat transfer equation according to the thermal non-equilibrium assumption and give the boundary conditions and the initial value of the outer surface heat flux, and perform numerical calculation based on the initial value of the outer surface heat flux to obtain the temperature field; The second calculation module is used to select the temperature value of the measuring point position based on the temperature field and construct the objective function to randomly generate a new solution, and then calculate the objective function corresponding to the new solution again; A comparison module is used to compare the objective function constructed by the new solution and the old solution. If the objective function constructed by the new solution is smaller than the objective function of the old solution, the new solution is used as a candidate solution, and the heat flow to be identified is updated as the new solution for the next iteration step. If the objective function constructed by the new solution is not smaller than the constructed objective function, the probability is used to determine whether the new solution is used as a candidate solution according to the Monte Carlo judgment criterion; Iteration module, used to perform local search near the neighborhood of candidate solutions, and use gradient information to optimize the search process for local iterative optimization and repeat the local search process; The generation module is used to update the iteration counter after each local search at an annealing temperature is completed, and check whether the current number of iterations has reached the preset maximum number of iterations. If not, the search process at the current annealing temperature is continued. If reached, the annealing temperature is lowered and the search process is repeated at the new annealing temperature. The search ends when the termination condition is met to obtain the final heat flow vector.

[0043] In one possible implementation, the objective function constructed is for: ,in represents the heat flux to be identified, is the temperature field obtained by numerical calculation, is the temperature value at the measuring point, is the number of discrete points in space, is the number of discrete time points, is the number of inner loop iterations, is the number of outer loop iterations, are discrete points in space, are discrete points in time.

[0044] In one possible implementation, the probability for: , The objective function constructed for the old solution is: The objective function constructed for the new solution is: is the Boltzmann constant, is the annealing temperature, is the difference between the objective functions corresponding to the new solution and the old solution.

[0045] In a possible implementation, the process of local iterative optimization is: ,in represents the next heat flux to be identified, is the gradient decreasing direction, The expression is: ,in is the conjugate coefficient, is the previous gradient reduction direction, is the step length, The expression is: , , , conjugate gradient iteration direction: .

[0046] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A real-time heat flow inversion method based on annealing gradient algorithm, characterized in that: The method comprises: Step S1, establishing a fluid-solid convection heat transfer equation based on the thermal non-equilibrium assumption and giving boundary conditions and initial values ​​of the outer surface heat flux, and performing numerical calculation based on the initial values ​​of the outer surface heat flux to obtain a temperature field; Step S2, selecting the temperature value of the measuring point based on the temperature field and constructing the objective function to randomly generate a new solution, and recalculating the objective function corresponding to the new solution; Step S3, comparing the objective function constructed by the new solution and the old solution, if the objective function constructed by the new solution is smaller than the objective function constructed by the old solution, the new solution is used as a candidate solution, and the heat flow to be identified is updated as a new solution for the next iteration step, if the objective function constructed by the new solution is not smaller than the objective function constructed by the old solution, the probability is used to determine whether the new solution is used as a candidate solution according to the Monte Carlo judgment criterion; Step S4, perform local search near the neighborhood of the candidate solution, and use the gradient information to optimize the search process to perform local iterative optimization, and repeat the local search process; Step S5, after completing a local search at an annealing temperature each time, update the iteration counter, check whether the current iteration number has reached the preset maximum iteration number, if not, continue the search process at the current annealing temperature, if reached, lower the annealing temperature as the new annealing temperature, execute steps S2 to S5 at the new annealing temperature, and end when the termination condition is met to obtain the final heat flow vector.

2. The real-time heat flow inversion method according to claim 1, characterized in that: In step S2, the objective function is constructed for: ,in represents the heat flux to be identified, is the temperature field obtained by numerical calculation, is the temperature value at the measuring point, is the number of discrete points in space, is the number of discrete time points, is the number of inner loop iterations, is the number of outer loop iterations, are discrete points in space, are discrete points in time.

3. The real-time heat flow inversion method according to claim 2, characterized in that: In step S3, the probability for: , The objective function constructed for the old solution is: The objective function constructed for the new solution is: is the Boltzmann constant, is the annealing temperature, is the difference between the objective functions corresponding to the new solution and the old solution.

4. The real-time heat flow inversion method according to claim 3, characterized in that: In step S4, the process of local iterative optimization is: ,in represents the next heat flux to be identified, is the gradient decreasing direction, The expression is: ,in is the conjugate coefficient, is the previous gradient reduction direction, is the step length, The expression is: , , , the conjugate gradient iteration direction is: 。 5. A real-time heat flow inversion device based on annealing gradient algorithm, characterized in that: The device comprises: The first calculation module is used to establish the fluid-solid two-phase convection heat transfer equation according to the thermal non-equilibrium assumption and give the boundary conditions and the initial value of the outer surface heat flux, and perform numerical calculation based on the initial value of the outer surface heat flux to obtain the temperature field; The second calculation module is used to select the temperature value of the measuring point position based on the temperature field and construct the objective function to randomly generate a new solution, and then calculate the objective function corresponding to the new solution again; A comparison module is used to compare the objective function constructed by the new solution and the old solution. If the objective function constructed by the new solution is smaller than the objective function constructed by the old solution, the new solution is used as a candidate solution, and the heat flow to be identified is updated as the new solution for the next iteration. If the objective function constructed by the new solution is not smaller than the objective function constructed by the old solution, the probability is used to determine whether the new solution is used as a candidate solution according to the Monte Carlo judgment criterion; Iteration module, used to perform local search near the neighborhood of candidate solutions, and use gradient information to optimize the search process for local iterative optimization and repeat the local search process; The generation module is used to update the iteration counter after completing a local search at an annealing temperature each time, check whether the current number of iterations has reached the preset maximum number of iterations, if not, continue the search process at the current annealing temperature, if reached, lower the annealing temperature as the new annealing temperature, execute the second calculation module to the generation module at the new annealing temperature, and end when the termination condition is met to obtain the final heat flow vector.

6. The real-time heat flow inversion device according to claim 5, characterized in that: Constructed objective function for: ,in represents the heat flux to be identified, is the temperature field obtained by numerical calculation, is the temperature value at the measuring point, is the number of discrete points in space, is the number of discrete time points, is the number of inner loop iterations, is the number of outer loop iterations, are discrete points in space, are discrete points in time.

7. The real-time heat flow inversion device according to claim 6, characterized in that: Probability for: , The objective function constructed for the old solution is: The objective function constructed for the new solution is: is the Boltzmann constant, is the annealing temperature, is the difference between the objective functions corresponding to the new solution and the old solution.

8. The heat flow real-time inversion device according to claim 7, characterized in that: The process of local iterative optimization is: ,in represents the next heat flux to be identified, is the gradient decreasing direction, The expression is: ,in is the conjugate coefficient, is the previous gradient reduction direction, is the step length, The expression is: , , , the conjugate gradient iteration direction is: .

Images