A method for determining performance parameters for engine regeneration cooling simulation

Abstract

The application provides a performance parameter determination method for engine regeneration cooling simulation, and relates to the technical field of rocket engine simulation. The application obtains a training data set, the training data set comprises a simulation data set and an experimental data set, a Bayesian neural network model is iteratively trained according to the simulation data set and the experimental data set, and the posterior distribution of the Bayesian neural network model is obtained. Then, based on the posterior distribution, the performance parameters, the confidence scores and the error ranges of each performance parameter can be quickly and accurately obtained according to the design parameters of the engine model, the simulation calculation time of the engine model is effectively reduced, and the simulation efficiency is improved. Moreover, the method provided by the application determines the posterior distribution based on the simulation data set obtained through simulation calculation and the experimental data set obtained through actual experiment, can effectively realize the alignment of the simulation data and the experimental data, and improves the accuracy of the prediction result.

Description

Technical Field

[0001] This invention relates to the field of rocket engine simulation technology, and in particular to a method for determining performance parameters for engine regenerative cooling simulation. Background Technology

[0002] In recent years, with the rapid development of closed-cycle liquid rocket technology, the requirements for accurate calculation of temperature distribution in the thrust chamber of liquid rocket engines, optimization of cooling effects, and assessment of structural strength have been continuously increasing.

[0003] In related technologies, numerical calculations are typically performed using Fluent algorithm software. However, this simulation method reveals many problems in the initial design phase. For example, simulating even a simplified engine model with a geometry of only 1 / 100th scale often requires a mesh distribution of tens of millions, resulting in computation times that can easily reach tens of hours. This high computational load and long computation time severely limits the efficiency of the initial design phase. To circumvent these problems, engineers often rely on personal experience to develop preliminary design schemes and then adjust the details to reduce the amount of simulation computation. However, this experience-based approach makes it difficult to balance accuracy and efficiency, especially in complex design tasks where the trade-off between efficiency and accuracy is even more challenging, resulting in low computational efficiency.

[0004] To meet the rapid iteration requirements of the initial design phase, researchers have recently attempted to introduce machine learning techniques into the liquid rocket engine design process, for example, using neural network models to build rapid prediction models. These models can significantly improve design evaluation efficiency in the early stages. However, despite these breakthroughs, significant drawbacks remain. Current neural network models typically rely on simulation data for training, which inevitably carries model errors and deviates considerably from actual physical experimental data. This deviation manifests as insufficient accuracy in model predictions, making it difficult to accurately reflect the actual conditions under complex operating circumstances. Furthermore, traditional neural network models are mostly "black box models," unable to explain the internal logic of prediction results and lacking explicit quantification of uncertainty, which introduces potential risks in design decisions. Simultaneously, traditional neural network models are usually trained with fixed parameters, making it difficult to flexibly adjust them when faced with new experimental data, thus limiting their applicability in design iterations.

[0005] Beyond the design of liquid rocket engine thrust chambers, similar problems exist in other engineering design fields. For example, in the aerodynamic design of hypersonic vehicles, traditional CFD simulations also face the challenges of high computational costs and long processing times in order to assess the heat flux distribution and aerodynamic performance of the vehicle surface. Existing machine learning-based accelerated prediction methods, when applied to the airspace domain, are limited by insufficient model accuracy, difficulties in integrating experimental data, and the lack of uncertainty assessment. Furthermore, experimental data, a key element for improving model reliability, is often inefficiently utilized in current technologies. Although some studies have attempted to combine simulation and experimental data, a systematic integration mechanism has not yet been established, further restricting the improvement of model performance.

[0006] In summary, the relevant technologies still have limitations in many aspects, including the lack of efficient and rapid prediction capabilities, insufficient integration and utilization of experimental and simulation data, lack of reliability and uncertainty in prediction results, and the problem of flexibility in parameter adjustment. Summary of the Invention

[0007] This invention provides a method for determining performance parameters in engine regenerative cooling simulation. Based on a pre-trained neural network model, it can quickly and accurately obtain performance parameters, along with the confidence score and error range for each parameter, according to the engine model's design parameters. This allows technicians to rapidly assess engine risks using the confidence score and error range of each performance parameter, effectively reducing simulation computation time and improving simulation efficiency. Furthermore, the method assimilates or aligns simulation and experimental data. The established Bayesian neural network model can effectively address the uncertainty of prediction results by simultaneously referencing the knowledge carried by both simulation and experimental data. The Bayesian neural network model provided by this invention not only quantifies the confidence level of prediction results but also reveals possible error ranges, helping decision-makers better understand and assess the model's reliability and risks. In subsequent research, the Bayesian neural network model provided by this invention can be embedded in simulators to replace some computational work, significantly reducing simulation computation time while improving the flexibility and accuracy of predictions.

[0008] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0009] Firstly, a method for determining performance parameters in engine regenerative cooling simulation is provided. The method includes: acquiring a training dataset, which comprises a simulation dataset and an experimental dataset. The simulation dataset includes design parameters of multiple engine models and performance parameters calculated for each engine model based on simulation. The experimental dataset includes design parameters of multiple engine models and performance parameters calculated for each engine model based on simulation experiments. The design parameters include one or more of cooling channel geometric parameters, material parameters, and flow rate. The performance parameters include one or more of pressure drop, wall temperature, pressure, and velocity. Iteratively training a Bayesian neural network model based on the simulation dataset yields a prior distribution of the Bayesian neural network model. Updating the prior distribution based on the experimental dataset yields a posterior distribution of the Bayesian neural network model. Acquiring data to be tested, which includes design parameters of a target engine model. Based on the posterior distribution of the Bayesian neural network model, determining the predicted value and confidence interval of each performance parameter of the target engine model according to the data to be tested.

[0010] In one possible implementation of the first aspect, before iteratively training the Bayesian neural network model based on the simulation dataset to obtain the prior distribution of the Bayesian neural network model, the method further includes: performing preprocessing operations on the training dataset, the preprocessing operations including outlier removal operations, missing value imputation operations and normalization operations, where outliers are values ​​that are not located within a preset threshold range.

[0011] In one possible implementation of the first aspect, the prior distribution is updated based on the experimental dataset to obtain the posterior distribution of the Bayesian neural network model, including: constructing a variational distribution and an evidence lower bound function; iteratively updating the variational distribution based on the experimental dataset according to the evidence lower bound function; and determining the variational distribution as the posterior distribution of the Bayesian neural network model when the value of the evidence lower bound function is greater than a preset threshold.

[0012] In one possible implementation of the first aspect, the variational distribution is iteratively updated based on the experimental dataset according to the evidence lower bound function, including: determining the entropy corresponding to each latent variable of the variational distribution according to the experimental dataset; determining the weight coefficient corresponding to each latent variable according to the entropy corresponding to each latent variable, wherein the weight coefficient corresponding to each latent variable is negatively correlated with the entropy; and iteratively updating the variational distribution based on the evidence lower bound function, the weight coefficient corresponding to each latent variable, and the prior distribution.

[0013] In one possible implementation of the first aspect, the entropy H(θ) corresponding to the i-th latent variable i The formula for determining ) is:

[0014]

[0015] Where, σ i Let be the standard deviation of the i-th latent variable.

[0016] The weight coefficient λ of the i-th latent variable i The formula for determining it is:

[0017] λ i =1 / (1+H(θ) i )).

[0018] In one possible implementation of the first aspect, the lower bound function L of the evidence is:

[0019]

[0020] Among them, E q(θ) [logp(D | [θ)] represents the log-likelihood expectation term; D is the experimental dataset, θ is a latent variable; q φ (θ i ) represents the variational distribution; KL(q) φ (θ i )||p(θ i p(θ) represents the KL divergence value of the i-th latent variable. i ) represents the prior distribution; K represents the number of latent variables; θ i Let λ be the i-th latent variable; i Let be the weight coefficient of the i-th latent variable.

[0021] The beneficial effects of this invention are as follows: The method provided by this invention first trains the prior distribution of a Bayesian neural network model based on the simulation dataset obtained from simulation calculations. Then, it updates the prior distribution based on the experimental dataset obtained from actual experiments to obtain the posterior distribution. This allows for the determination of the predicted value and confidence interval of each performance parameter of the target engine model based on the posterior distribution of the Bayesian neural network model. This enables technicians to quickly assess the risks of the engine using the confidence score and error range of each performance parameter, effectively reducing the simulation calculation time of the engine model and improving simulation efficiency. Furthermore, the method provided by this invention can assimilate or align simulation data with experimental data, and the established Bayesian neural network model can effectively provide information on the uncertainty of the prediction results. The Bayesian neural network model provided by this invention can not only quantify the confidence of the prediction results but also reveal the possible error range, helping decision-makers better understand and assess the reliability and risk of the model. In subsequent research, this model can be embedded in a simulator to replace some computational work, thereby significantly reducing simulation calculation time while improving the flexibility and accuracy of predictions. On the other hand, this invention provides a variational inference method for determining the posterior distribution. By assigning weight coefficients to different latent variables during the updating of the posterior distribution of the Bayesian neural network model, and since these weight coefficients are determined based on the entropy of each latent variable, the accuracy of latent variable calculation can be effectively improved, thereby enhancing the accuracy and reliability of the model. The method provided by this invention further enhances the adaptability to system complexity and changes by modeling the uncertainty of latent variables, improving the model's performance in practical applications. Therefore, the posterior distribution of the Bayesian neural network model provided by this invention can improve the prediction accuracy and robustness of performance parameters.

[0022] Secondly, the present invention provides a performance parameter determination system for engine regenerative cooling simulation. The system includes: a data acquisition module for acquiring a training dataset, which includes a simulation dataset and an experimental dataset. The simulation dataset includes design parameters of multiple engine models and performance parameters calculated by each engine model based on simulation. The experimental dataset includes design parameters of multiple engine models and performance parameters calculated by each engine model based on simulation experiments. The design parameters include one or more of cooling channel geometric parameters, material parameters, and flow rates. The performance parameters include one or more of pressure drop, wall temperature, pressure, and velocity. A priori determination module is used to iteratively train a Bayesian neural network model based on the simulation dataset to obtain the prior distribution of the Bayesian neural network model. A posterior determination module is used to update the prior distribution based on the experimental dataset to obtain the posterior distribution of the Bayesian neural network model. The data acquisition module is also used to acquire data to be tested, which includes design parameters of a target engine model. A parameter determination module is used to determine the predicted value and confidence interval of each performance parameter of the target engine model based on the posterior distribution of the Bayesian neural network model and the data to be tested.

[0023] In one possible implementation of the second aspect, the posterior determination module is specifically used for: constructing a variational distribution and an evidence lower bound function; iteratively updating the variational distribution based on the evidence lower bound function and the experimental dataset; and determining the variational distribution as the posterior distribution of the Bayesian neural network model when the value of the evidence lower bound function is greater than a preset threshold.

[0024] In one possible implementation of the second aspect, the posterior determination module is specifically used to: determine the entropy corresponding to each latent variable of the variational distribution based on the experimental dataset; determine the weight coefficient corresponding to each latent variable based on the entropy corresponding to each latent variable, wherein the weight coefficient corresponding to each latent variable is negatively correlated with the entropy; and iteratively update the variational distribution based on the evidence lower bound function, according to the weight coefficient corresponding to each latent variable and the prior distribution.

[0025] The entropy H(θ) corresponding to the i-th latent variable i The formula for determining ) is:

[0026]

[0027] Where, σ i Let be the standard deviation of the i-th latent variable;

[0028] The weight coefficient λ of the i-th latent variable i The formula for determining it is:

[0029] λ i =1 / (1+H(θ) i )).

[0030] The lower bound function L of the evidence is:

[0031]

[0032] Among them, E q(θ) [logp(D|θ)] represents the log-likelihood expectation term; D is the experimental dataset; θ is a latent variable; q φ (θ i ) represents the variational distribution; KL(q) φ (θ i )||p(θ i p(θ) represents the KL divergence value of the i-th latent variable. i ) represents the prior distribution; K represents the number of latent variables; θ i Let λ be the i-th latent variable; i Let be the weight coefficient of the i-th latent variable.

[0033] Thirdly, an electronic device is provided, the electronic device including a memory and one or more processors; the memory is coupled to the processors; wherein the memory stores computer program code, the computer program code including computer instructions, which, when executed by the processor, cause the electronic device to perform the method as described in any implementation of the first aspect.

[0034] Fourthly, a computer-readable storage medium is provided, including computer instructions that, when executed on an electronic device, cause the electronic device to perform a method as described in any implementation of the first aspect.

[0035] Fifthly, a computer program product is provided that, when run on a computer, causes the computer to perform the method in any implementation of the first aspect.

[0036] Understandably, the beneficial effects achieved by the system of the second aspect, the electronic device of the third aspect, the computer-readable storage medium of the fourth aspect, and the computer program product of the fifth aspect provided above can be referred to with reference to the beneficial effects of the first aspect and any of its possible design embodiments, which will not be repeated here. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention;

[0038] Figure 2 A flowchart illustrating a method for determining performance parameters in engine regenerative cooling simulation, provided by an embodiment of the present invention;

[0039] Figure 3A flowchart illustrating another method for determining performance parameters in engine regenerative cooling simulation provided by an embodiment of the present invention;

[0040] Figure 4 A flowchart illustrating another method for determining performance parameters in engine regenerative cooling simulation provided by an embodiment of the present invention;

[0041] Figure 5 This is a schematic diagram illustrating the change in the weight coefficients of latent variables during the training process, as shown in an embodiment of the present invention.

[0042] Figure 6 This is a schematic diagram of the structure of a determination system provided in an embodiment of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be described below with reference to the accompanying drawings. In the description of the present invention, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. The "or" in the present invention is merely a description of the relationship between the related objects, indicating that three relationships can exist. For example, A or B can represent: A alone, A and B simultaneously, and B alone. A and B can be singular or plural. Furthermore, in the description of the present invention, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items.

[0044] Furthermore, to facilitate a clear description of the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that the terms "first" and "second" are not necessarily different.

[0045] In this embodiment of the invention, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" or "for example" in this embodiment of the invention should not be construed as superior or more advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.

[0046] In recent years, with the rapid development of closed-cycle liquid rocket technology, the requirements for accurate calculation of temperature distribution in the thrust chamber of liquid rocket engines, optimization of cooling effects, and assessment of structural strength have been continuously increasing.

[0047] In related technologies, numerical calculations are typically performed using Fluent algorithm software. However, this simulation method reveals many problems in the initial design phase. For example, simulating even a simplified engine model with a geometry of only 1 / 100th scale often requires a mesh distribution of tens of millions, resulting in computation times that can easily reach tens of hours. This high computational load and long computation time severely limits the efficiency of the initial design phase. To circumvent these problems, engineers often rely on personal experience to develop preliminary design schemes and then adjust the details to reduce the amount of simulation computation. However, this experience-based approach makes it difficult to balance accuracy and efficiency, especially in complex design tasks where the trade-off between efficiency and accuracy is even more challenging, resulting in low computational efficiency.

[0048] To meet the rapid iteration requirements of the initial design phase, researchers have recently attempted to introduce machine learning techniques into the liquid rocket engine design process, for example, using neural network models to build rapid prediction models. These models can significantly improve design evaluation efficiency in the early stages. However, despite these breakthroughs, significant drawbacks remain. Current neural network models typically rely on simulation data for training, which inevitably carries model errors and deviates considerably from actual physical experimental data. This deviation manifests as insufficient accuracy in model predictions, making it difficult to accurately reflect the actual conditions under complex operating circumstances. Furthermore, traditional neural network models are mostly "black box models," unable to explain the internal logic of prediction results and lacking explicit quantification of uncertainty, which introduces potential risks in design decisions. Simultaneously, traditional neural network models are usually trained with fixed parameters, making it difficult to flexibly adjust them when faced with new experimental data, thus limiting their applicability in design iterations.

[0049] Beyond the design of liquid rocket engine thrust chambers, similar problems exist in other engineering design fields. For example, in the aerodynamic design of hypersonic vehicles, traditional CFD simulations also face the challenges of high computational costs and long processing times in order to assess the heat flux distribution and aerodynamic performance of the vehicle surface. Existing machine learning-based accelerated prediction methods, when applied to the airspace domain, are limited by insufficient model accuracy, difficulties in integrating experimental data, and the lack of uncertainty assessment. Furthermore, experimental data, a key element for improving model reliability, is often inefficiently utilized in current technologies. Although some studies have attempted to combine simulation and experimental data, a systematic integration mechanism has not yet been established, further restricting the improvement of model performance.

[0050] In summary, the problems in related technologies are as follows: First, there is a lack of effective assimilation between simulation and experimental data: existing neural network models typically rely solely on a large amount of simulation data for training. However, simulation data often contains model errors and deviates significantly from real physical experimental data, leading to insufficient reliability of model predictions. Second, existing neural network models cannot dynamically update model parameters: existing neural network-based models are often "black box models" with fixed parameters, making it difficult to dynamically adjust them based on new experimental data after training, thus limiting the model's ability to continuously optimize during practical design. Third, existing neural network models lack sufficient understanding of prediction uncertainty: traditional neural networks and support vector machines cannot effectively quantify the uncertainty of prediction results, which can easily lead to risks in design decisions, especially when the prediction results deviate significantly from the actual data. Finally, neural network models in related technologies do not make sufficient effective use of experimental data. Although experimental data is of great value in improving model reliability, existing methods lack a systematic approach to combining simulation and experimental data, making it difficult to achieve effective fusion of the two.

[0051] In summary, the relevant technologies still have limitations in many aspects, including the lack of efficient and rapid prediction capabilities, insufficient integration and utilization of experimental and simulation data, lack of reliability and uncertainty in prediction results, and the problem of flexibility in parameter adjustment.

[0052] In view of this, embodiments of the present invention provide a method for determining performance parameters in engine regenerative cooling simulation. The method includes: acquiring a training dataset, which includes a simulation dataset and an experimental dataset. The simulation dataset includes design parameters of multiple engine models and performance parameters calculated by each engine model based on simulation. The experimental dataset includes design parameters of multiple engine models and performance parameters calculated by each engine model based on simulation experiments. The design parameters include one or more of cooling channel geometric parameters, material parameters, and flow rate. The performance parameters include one or more of pressure drop, wall temperature, pressure, and velocity. Iteratively training a Bayesian neural network model based on the simulation dataset to obtain the prior distribution of the Bayesian neural network model. Updating the prior distribution based on the experimental dataset to obtain the posterior distribution of the Bayesian neural network model. Acquiring data to be detected, which includes design parameters of a target engine model. Based on the posterior distribution of the Bayesian neural network model, determining the predicted value and confidence interval of each performance parameter of the target engine model based on the data to be detected.

[0053] The method provided by this invention first trains a prior distribution of a Bayesian neural network model based on a simulation dataset obtained from simulation calculations. Then, it updates the prior distribution based on an experimental dataset obtained from actual experiments to obtain a posterior distribution. This allows for the determination of the predicted value and confidence interval of each performance parameter of the target engine model based on the posterior distribution of the Bayesian neural network model. This enables technicians to quickly assess the risks associated with the engine using the confidence score and error range of each performance parameter, effectively reducing simulation calculation time and improving simulation efficiency. Furthermore, the method provided by this invention can assimilate or align simulation data with experimental data, and the established Bayesian neural network model can effectively provide information on the uncertainty of the prediction results. The Bayesian neural network model provided by this invention not only quantifies the confidence level of the prediction results but also reveals the possible error range, helping decision-makers better understand and assess the reliability and risks of the model. In subsequent research, this model can be embedded in simulators to replace some computational work, thereby significantly reducing simulation calculation time while improving the flexibility and accuracy of predictions. On the other hand, this invention provides a variational inference method for determining the posterior distribution. By assigning weight coefficients to different latent variables during the updating of the posterior distribution of the Bayesian neural network model, and since these weight coefficients are determined based on the entropy of each latent variable, the accuracy of latent variable calculation can be effectively improved, thereby enhancing the accuracy and reliability of the model. The method provided by this invention further enhances the adaptability to system complexity and changes by modeling the uncertainty of latent variables, improving the model's performance in practical applications. Therefore, the posterior distribution of the Bayesian neural network model provided by this invention can improve the prediction accuracy and robustness of performance parameters.

[0054] In some embodiments, the performance parameter determination method for engine regenerative cooling simulation provided by the present invention can be executed by a performance parameter determination system 100 for engine regenerative cooling simulation (hereinafter referred to as determination system 100).

[0055] As an example, the determination system 100 can be any electronic device 200 with data processing capabilities, such as a general-purpose computer, personal computer, laptop computer, switch, or tablet computer. The specific implementation of the determination system 100 is not limited here.

[0056] Figure 1 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention is shown. The electronic device 200 includes a processor 210, a memory 220, and a communication interface 230.

[0057] Processor 210 may include one or more processing cores. Processor 210 connects to various parts within electronic device 200 using various interfaces and lines, and performs various functions and processes data of electronic device 200 by running or executing instructions, programs, code sets, or instruction sets stored in memory 220, and by calling data stored in memory 220. Optionally, processor 210 may be implemented using at least one of the following hardware forms: Central Processing Unit (CPU), Graphics Processing Unit (GPU), Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA).

[0058] The memory 220 may include random access memory (RAI) or read-only memory (ROI). Optionally, the memory 220 may include non-transitory computer-readable storage ledger. The memory 220 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 220 may include a program storage area. The program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as video acquisition, feature extraction, and process detection), and instructions for implementing the various method embodiments described above.

[0059] The communication interface 230 is used to communicate with other devices, equipment, or communication networks, such as data storage devices, image processing devices, or Ethernet, wireless access networks (RAN), wireless local area networks (WLAN), etc.

[0060] In terms of physical implementation, the aforementioned devices (such as processor 210, memory 220, and communication interface 230) can each be devices within the same device (such as a laptop computer). Alternatively, at least two of these devices can be located within the same device, i.e., as different devices within the same device, similar to the deployment of devices or components in a distributed system.

[0061] It is understood that the structure illustrated in this embodiment does not constitute a specific limitation on the electronic device 200. In other embodiments of the present invention, the electronic device 200 may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0062] The following description, in conjunction with the accompanying drawings, illustrates a method for determining performance parameters in engine regenerative cooling simulation provided by an embodiment of the present invention.

[0063] Figure 2 This is a flowchart illustrating a method for determining performance parameters in engine regenerative cooling simulation, provided as an embodiment of the present invention. Optionally, this method can be... Figure 1 The illustrated electronic device 200 performs this operation. The method may include the following steps:

[0064] S1. Obtain the training dataset, which includes the simulation dataset and the experimental dataset.

[0065] Specifically, the simulation data includes design parameters of multiple engine models and performance parameters of each engine model calculated based on simulation. The experimental dataset includes design parameters of multiple engine models and performance parameters of each engine model obtained from simulation experiments. The design parameters include one or more of the following: cooling channel geometry parameters, material parameters, and flow rate. The performance parameters include one or more of the following: pressure drop, wall temperature, pressure, and velocity.

[0066] S2. Iteratively train the Bayesian neural network model based on the simulation dataset to obtain the prior distribution of the Bayesian neural network model.

[0067] Specifically, a Bayesian Neural Network (BNN) is a model that combines Bayesian probabilistic methods with neural networks, aiming to provide a measure of uncertainty for the weights and outputs of the neural network. Unlike traditional deterministic neural networks, BNNs introduce probability distributions to describe the uncertainty of model parameters, thus providing not only the result but also an assessment of the confidence level of the result during prediction. In other words, a Bayesian neural network model is a method that introduces Bayesian inference into neural networks; it not only learns the mapping relationships in the data but also models the uncertainty of the network parameters. Traditional neural networks obtain a fixed set of connection parameters (weights and biases) during training and then make deterministic predictions on new data. Bayesian neural network models, however, treat these parameters as random variables, assigning a prior distribution to each parameter and updating it to the corresponding posterior distribution through observation of data. Therefore, the output of a BNN contains not only the predicted value but also its uncertainty information (such as confidence intervals or variance). The core principle of the Bayesian neural network model is: using Bayes' theorem, combining prior knowledge and observed data to obtain the posterior distribution of the parameters, and then passing this uncertainty to the prediction result to form a prediction distribution. This enables the model to make more robust judgments when faced with unknown or noisy data.

[0068] The prior distributions are the connection parameters between different layers in the Bayesian neural network model, where the connection parameters include the weight distribution and the bias distribution.

[0069] In some embodiments, the Bayesian neural network model includes multiple input layers, multiple hidden layers, and multiple output layers. The prior distribution includes the connection parameters between any two connected layers in the Bayesian neural network model. For example, the Bayesian neural network model includes input layer 1, input layer 2, hidden layer 1, and output layer 1. The prior distribution includes the weight distribution and bias distribution between input layer 1 and hidden layer 1, the weight distribution and bias distribution between input layer 2 and hidden layer 1, and the weight distribution and bias distribution between hidden layer 1 and output layer 1.

[0070] It should be understood that a Bayesian neural network model may include more or fewer input layers, hidden layers, and output layers than in the examples above, and the embodiments of the present invention do not impose any particular limitation on this.

[0071] In one possible implementation, prior to S2 above, the method provided by the embodiments of the present invention further includes:

[0072] The training dataset is preprocessed, including outlier removal, missing value imputation, and normalization. Outliers are values ​​that are not within a preset threshold range.

[0073] Specifically, the system first filters out outliers in the training dataset, then fills in missing values, and finally normalizes the training dataset to a preset range to complete the preprocessing of the training dataset.

[0074] S3. Update the prior distribution based on the experimental dataset to obtain the posterior distribution of the Bayesian neural network model.

[0075] In some embodiments, see Figure 3 The aforementioned S3 specifically includes:

[0076] S31. Construct variational distribution and evidence lower bound function.

[0077] The lower bound function L of the evidence is:

[0078]

[0079] Among them, E q(θ) [logp(D|θ)] represents the log-likelihood expectation term; D is the experimental dataset; θ is a latent variable; q φ (θ i ) represents the variational distribution; KL(q) φ (θ i )||p(θ i p(θ) represents the KL divergence value of the i-th latent variable. i ) represents the prior distribution; K represents the number of latent variables; θ i Let λ be the i-th latent variable; i Let be the weight coefficient of the i-th latent variable.

[0080] Specifically, the lower bound function of evidence measures the closeness between the variational distribution and the true posterior distribution, the log-likelihood logp(D|θ) measures the degree of match between the model's predicted values ​​and the true values ​​carried in the experimental dataset, and the expected value E... q(θ) This is used to enable the distribution to cover the parameter range of interpretable data.

[0081] S32. Based on the lower bound function of evidence, the variational distribution is iteratively updated according to the experimental dataset.

[0082] In one possible implementation, see Figure 4 The aforementioned S32 specifically includes:

[0083] S321. Determine the entropy corresponding to each latent variable of the variational distribution based on the experimental dataset;

[0084] Specifically, the prior distribution includes the connection parameters between any two connected layers in the Bayesian neural network model. These connection parameters include the weight distribution and the bias distribution. The weight distribution includes the weight mean and standard deviation, and the bias distribution includes the bias mean and standard deviation. In other words, the latent variables are the distribution parameters of the weights and biases of each layer (such as the mean and standard deviation in a Gaussian distribution). Entropy is used to characterize the determinism of the latent variables.

[0085] In some embodiments, the entropy H(θ) corresponding to the i-th latent variable i The formula for determining ) is:

[0086]

[0087] Where, σ i Let be the standard deviation of the i-th latent variable.

[0088] S322. Determine the weight coefficient of each latent variable based on the entropy corresponding to each latent variable, wherein the weight coefficient of each latent variable is negatively correlated with the entropy;

[0089] Specifically, the weight coefficient λ of the i-th latent variable i The formula for determining it is:

[0090] λ i =1 / (1+H(θ) i )).

[0091] S323. Based on the evidence lower bound function, the variational distribution is iteratively updated according to the weight coefficients and prior distributions corresponding to each latent variable.

[0092] Specifically, the variational distribution is defined by all latent variables. Iteratively updating the variational distribution can be understood as adjusting its shape (such as moving the mean or scaling the variance).

[0093] S33. When the value of the lower bound function of evidence is greater than a preset threshold, the variational distribution is determined as the posterior distribution of the Bayesian neural network model.

[0094] Specifically, when the value of the evidence lower bound function is greater than a preset threshold, the error between the variational distribution and the true posterior distribution is less than a preset error interval. Therefore, the variational distribution is determined as the posterior distribution of the Bayesian neural network model.

[0095] S4. Obtain the data to be tested, which includes the design parameters of the target engine model.

[0096] S5. Based on the posterior distribution of the Bayesian neural network model, determine the predicted value and confidence interval of each performance parameter of the target engine model according to the data to be detected.

[0097] For example, when the performance parameters of the target engine model include wall temperature and pressure drop, the predicted value and confidence interval of the wall temperature are 314±4K and 0.2±0.01MPa.

[0098] As described in S1-S5 above, the method provided by this invention first trains the prior distribution of a Bayesian neural network model based on the simulation dataset obtained from simulation calculations. Then, it updates the prior distribution based on the experimental dataset obtained from actual experiments to obtain the posterior distribution. This allows the method to determine the predicted value and confidence interval of each performance parameter of the target engine model based on the posterior distribution of the Bayesian neural network model. This enables technicians to quickly assess the risks of the engine using the confidence score and error range of each performance parameter, effectively reducing the simulation calculation time of the engine model and improving simulation efficiency. Furthermore, the method provided by this invention can assimilate or align simulation data with experimental data, and the established Bayesian neural network model can effectively provide information on the uncertainty of the prediction results. The Bayesian neural network model provided by this invention can not only quantify the confidence of the prediction results but also reveal the possible error range, helping decision-makers better understand and assess the reliability and risk of the model. In subsequent research, this model can be embedded in a simulator to replace some computational work, thereby significantly reducing simulation calculation time while improving the flexibility and accuracy of predictions.

[0099] In one possible implementation, the method provided by the embodiments of the present invention further includes:

[0100] The preset interface displays the curve of the change of the weight coefficient of each latent variable during the iterative training process.

[0101] For example, see Figure 5 , Figure 5 This is a schematic diagram illustrating the changes in the weight coefficients of latent variables during training, as shown in an embodiment of the present invention. It includes two latent variables, parameter w1-loc and parameter w2-loc. The curve that gradually increases with the number of iterations and remains relatively stable represents the weight coefficient of parameter w1-loc in each iteration across multiple iterations, while the curve that gradually decreases with the number of iterations represents the weight coefficient of parameter w2-loc in each iteration across multiple iterations.

[0102] The method provided in this invention, by displaying the change curves of the weight coefficients of different latent variables during the training process, enables technicians to quickly and accurately determine the weight coefficients of different latent variables, and then, based on the weight coefficients...

[0103] To facilitate understanding of this solution, the specific implementation process of this embodiment is explained below through an example. First, a Bayesian neural network model is defined, with its weights and biases defined as random variables. The initial prior distribution is a standard Gaussian distribution. Then, the Bayesian neural network model is iteratively trained using a simulation dataset to obtain the trained prior distribution. Alternatively, this can be understood as optimizing and updating the initial prior distribution based on the simulation dataset to obtain an updated prior distribution, which incorporates the knowledge carried by the simulation dataset. Next, an experimental dataset is obtained, and the updated prior distribution is iteratively trained using this dataset to obtain the posterior distribution. During this iterative training process, the objective function is a lower bound function of evidence, and each latent variable in the lower bound function is dynamically assigned a different weight coefficient based on entropy. Finally, based on the posterior distribution, the Bayesian neural network model can determine the performance parameters and confidence intervals corresponding to each piece of data to be detected.

[0104] In other words, the method provided in this invention completes the initial update of the prior distribution using a simulation dataset, and then further updates the prior distribution based on a more realistic experimental dataset to obtain the posterior distribution. Thus, the performance parameters of the target engine model determined by the posterior distribution closely approximate the knowledge carried by both the simulation and experimental datasets, effectively improving accuracy and robustness, and meeting the requirements for accurate calculation of the temperature distribution in the thrust chamber of liquid rocket engines, optimization of cooling effects, and assessment of structural strength. In other words, this invention provides a variational inference method for determining the posterior distribution. By assigning weight coefficients to different latent variables during the update of the posterior distribution of the Bayesian neural network model, and since the weight coefficients are determined based on the entropy of each latent variable, the accuracy of solving the latent variables can be effectively improved, thereby enhancing the accuracy and reliability of the model. The method provided in this invention further enhances the adaptability to system complexity and changes by modeling the uncertainty of latent variables, improving the model's performance in practical applications. Therefore, the posterior distribution of the Bayesian neural network model provided by this invention can improve the prediction accuracy and robustness of performance parameters.

[0105] The foregoing primarily describes the solutions of the embodiments of the present invention from a methodological perspective. It is understood that, to achieve the aforementioned functions, the system 100 includes at least one of the hardware structures and software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments disclosed herein, the embodiments of the present invention can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the embodiments of the present invention.

[0106] In this embodiment of the invention, the system 100 can be divided into functional units according to the above method example. For example, the system 100 can be divided into functional units corresponding to various functions, or two or more functions can be integrated into one processing unit. The integrated unit can be implemented in hardware or as a software functional unit. It should be noted that the unit division in this embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0107] For example, Figure 6 A schematic diagram of the hardware structure of a determination system provided by an embodiment of the present invention is shown. The determination system 100 includes: a data acquisition module 110, used to acquire a training dataset, which includes a simulation dataset and an experimental dataset. The simulation dataset includes design parameters of multiple engine models and performance parameters of each engine model calculated based on simulation. The experimental dataset includes design parameters of multiple engine models and performance parameters of each engine model obtained based on simulation experiments. The design parameters include one or more of cooling channel geometric parameters, material parameters, and flow rate. The performance parameters include one or more of pressure drop, wall temperature, pressure, and velocity. A prior determination module 120 is used to iteratively train a Bayesian neural network model based on the simulation dataset to obtain the prior distribution of the Bayesian neural network model. A posterior determination module 130 is used to update the prior distribution based on the experimental dataset to obtain the posterior distribution of the Bayesian neural network model. The data acquisition module 110 is also used to acquire data to be detected, which includes design parameters of a target engine model. A parameter determination module 140 is used to determine the predicted value and confidence interval of each performance parameter of the target engine model based on the posterior distribution of the Bayesian neural network model and the data to be detected.

[0108] Optionally, the posterior determination module 130 is specifically used for: constructing a variational distribution and an evidence lower bound function; iteratively updating the variational distribution based on the evidence lower bound function and the experimental dataset; and determining the variational distribution as the posterior distribution of the Bayesian neural network model when the value of the evidence lower bound function is greater than a preset threshold.

[0109] Optionally, the posterior determination module 130 is specifically used for: determining the entropy corresponding to each latent variable of the variational distribution based on the experimental dataset; determining the weight coefficient corresponding to each latent variable based on the entropy corresponding to each latent variable, wherein the weight coefficient corresponding to each latent variable is negatively correlated with the entropy; and iteratively updating the variational distribution based on the evidence lower bound function, according to the weight coefficient corresponding to each latent variable and the prior distribution.

[0110] The entropy H(θ) corresponding to the i-th latent variable i The formula for determining ) is:

[0111]

[0112] Where, σ i Let be the standard deviation of the i-th latent variable;

[0113] The weight coefficient λ of the i-th latent variable i The formula for determining it is:

[0114] λ i =1 / (1+H(θ) i ));

[0115] The lower bound function L of the evidence is:

[0116]

[0117] Among them, E q(θ) [logp(D|θ)] represents the log-likelihood expectation term; D is the experimental dataset; θ is a latent variable; q φ (θ i ) represents the variational distribution; KL(q) φ (θ i )||p(θ i p(θ) represents the KL divergence value of the i-th latent variable. i ) represents the prior distribution; K represents the number of latent variables; θ i Let λ be the i-th latent variable; i Let be the weight coefficient of the i-th latent variable.

[0118] It should be understood that a detailed description of the above-mentioned optional methods can be found in the foregoing method embodiments, and will not be repeated here. Furthermore, explanations of any of the determination systems 100 provided above, as well as descriptions of their beneficial effects, can be found in the corresponding method embodiments described above, and will not be repeated here.

[0119] This invention also provides a computer-readable storage medium storing at least one computer instruction, which is loaded and executed by a processor to implement the methods of the various embodiments described above. Explanations of the relevant content and descriptions of the beneficial effects of any of the computer-readable storage media provided above can be found in the corresponding embodiments described above, and will not be repeated here.

[0120] This invention also provides a chip. This chip integrates a control circuit for implementing the functions of the aforementioned determining system 100 and one or more ports. Optionally, the functions supported by this chip are as described above and will not be repeated here.

[0121] Those skilled in the art will understand that the program for implementing all or part of the steps of the above embodiments, which can be executed by a program instructing related hardware, can be stored in a computer-readable storage medium. The storage medium mentioned above can be a read-only memory, a random access memory, etc. The processing unit or processor mentioned above can be a central processing unit, a general-purpose processor, an application-specific integrated circuit (ASIC), a microprocessor (DSP), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof.

[0122] This invention also provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform any of the methods described in the above embodiments. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this invention is generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions may be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., SSD), etc.

[0123] It should be noted that the devices for storing computer instructions or computer programs provided in the embodiments of the present invention, such as, but not limited to, the aforementioned memory, computer-readable storage medium, and communication chip, are all non-transitory. Those skilled in the art should recognize that the functions described in the embodiments of the present invention in one or more of the above examples can be implemented using hardware, software, firmware, or any combination thereof. When implemented using software, these functions can be stored in a computer-readable storage medium or transmitted as one or more instructions or code on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein communication media include any medium that facilitates the transmission of computer programs from one place to another. Storage media can be any available medium accessible to general-purpose or special-purpose computers.

[0124] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for determining performance parameters for engine regeneration cooling simulation, characterized in that, The method comprises: obtaining a training data set, the training data set comprising a simulation data set and an experimental data set, the simulation data comprising design parameters of a plurality of engine models and performance parameters of each engine model obtained according to simulation calculation, the experimental data set comprising design parameters of a plurality of engine models and performance parameters of each engine model obtained according to simulation experiment, the design parameters comprising one or more of cooling channel geometric parameters, material parameters and flow velocity, and the performance parameters comprising one or more of pressure drop, wall temperature, pressure and velocity; iteratively training a Bayesian neural network model according to the simulation data set to obtain a prior distribution of the Bayesian neural network model; updating the prior distribution according to the experimental data set to obtain a posterior distribution of the Bayesian neural network model; obtaining to-be-detected data, the to-be-detected data comprising design parameters of a target engine model; based on the posterior distribution of the Bayesian neural network model, determining a predicted value and a confidence interval of each performance parameter of the target engine model according to the to-be-detected data.

2. The method of claim 1, wherein, Before the step of iteratively training a Bayesian neural network model according to the simulation data set to obtain a prior distribution of the Bayesian neural network model, the method further comprises: performing a preprocessing operation on the training data set, the preprocessing operation comprising an outlier deletion operation, a missing value filling operation and a normalization processing operation, the outliers being values not located in a preset threshold interval.

3. The method of claim 2, wherein, The step of updating the prior distribution according to the experimental data set to obtain a posterior distribution of the Bayesian neural network model comprises: constructing a variational distribution and an evidence lower bound function; iteratively updating the variational distribution according to the experimental data set based on the evidence lower bound function; in a case where a value of the evidence lower bound function is greater than a preset threshold, determining the variational distribution as the posterior distribution of the Bayesian neural network model.

4. The method of claim 3, wherein, The step of iteratively updating the variational distribution according to the experimental data set based on the evidence lower bound function comprises: determining an entropy corresponding to each latent variable of the variational distribution according to the experimental data set; determining a weight coefficient corresponding to each latent variable according to the entropy corresponding to each latent variable, wherein the weight coefficient corresponding to each latent variable is negatively correlated with the entropy; iteratively updating the variational distribution according to the weight coefficient corresponding to each latent variable and the prior distribution based on the evidence lower bound function.

5. The method of claim 4, wherein, The determination formula of the entropy H(θ i ) corresponding to the i-th hidden variable is as follows: where σ i is the standard deviation of the ith latent variable; The weight coefficient λ of the i-th hidden variable i The determination formula is: λ i = 1 / (1 + H(θi)).

6. The method of claim 5, wherein, The evidence lower bound function L is: where E q(θ) [logp(D | θ)] is the log-likelihood expectation term; D is the experimental data set, and θ is the latent variable. φ (θ i ) is the variational distribution; KL(q φ (θ i ) || p(θ i )) is the KL divergence value of the i-th latent variable, and p(θ i ) is the prior distribution; K is the number of latent variables; θ i is the i-th latent variable; and λ i is the weight coefficient of the i-th latent variable.

7. A system for determining performance parameters of an engine regeneration cooling simulation, comprising: The system comprises: a data acquisition module configured to obtain a training data set, the training data set comprising a simulation data set and an experimental data set, the simulation data comprising design parameters of a plurality of engine models and performance parameters of each engine model obtained according to simulation calculation, the experimental data set comprising design parameters of a plurality of engine models and performance parameters of each engine model obtained according to simulation experiment, the design parameters comprising one or more of cooling channel geometric parameters, material parameters and flow velocity, and the performance parameters comprising one or more of pressure drop, wall temperature, pressure and velocity; The prior determination module is configured to perform iterative training on the Bayesian neural network model according to the simulation data set, to obtain a prior distribution of the Bayesian neural network model. The posterior determination module is configured to update the prior distribution according to the experimental data set, to obtain a posterior distribution of the Bayesian neural network model. The data acquisition module is further configured to acquire to-be-detected data, the to-be-detected data including design parameters of a target engine model. The parameter determination module is configured to determine, based on the posterior distribution of the Bayesian neural network model, a predicted value and a confidence interval of each performance parameter of the target engine model according to the to-be-detected data.

8. The system of claim 7, wherein, The posterior determination module is specifically configured to: construct a variational distribution and an evidence lower bound function; perform iterative updating on the variational distribution according to the experimental data set based on the evidence lower bound function; determine the variational distribution as the posterior distribution of the Bayesian neural network model in a case where a value of the evidence lower bound function is greater than a preset threshold.

9. The system of claim 8, wherein, The posterior determination module is specifically configured to: determine an entropy corresponding to each latent variable of the variational distribution according to the experimental data set; determine a weight coefficient corresponding to each latent variable according to the entropy corresponding to each latent variable, wherein the weight coefficient corresponding to each latent variable is negatively correlated with the entropy; perform iterative updating on the variational distribution according to the weight coefficient corresponding to each latent variable and the prior distribution based on the evidence lower bound function. The determination formula of the entropy H(θ i ) corresponding to the i-th hidden variable is as follows: where σ i is the standard deviation of the ith latent variable; The weight coefficient λ of the i-th hidden variable i The determination formula is: λ i = 1 / (1 + H(θ i )) ; The evidence lower bound function L is: where E q(θ) [logp(D|θ)] is the log-likelihood expected term; D is the experimental data set, and θ is the latent variable; q φ (θ i ) is the variational distribution; KL(q φ (θ i )||p(θ i )) is the KL divergence value of the i-th latent variable, and p(θ i ) is the prior distribution; K is the number of latent variables; θ i is the i-th latent variable; and λ i is the weight coefficient of the i-th latent variable.

10. An electronic device, comprising: include: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement the performance parameter determination method for engine regeneration cooling simulation according to any one of claims 1-6.

Images