An engine system weight prediction method, system, device and storage medium

Abstract

The present application belongs to the field of engine system weight prediction, and particularly relates to an engine system weight prediction method, system, device and storage medium, which establishes a mapping relationship between engine design parameters and system weights corresponding to the design parameters to obtain an original data set; then is input into a generative adversarial network for training, and is optimized through evaluation values generated by a generator and discrimination values generated by a discriminator to obtain an optimized generative adversarial network; generates a training data set, which is input into a neural network model to obtain a loss value, and through minimizing the loss value, an optimized neural network model is obtained and latent features are extracted; the latent features are input into a prediction model to minimize an objective function, and an optimal prediction model is obtained; and actual engine design parameters are input into the optimized prediction model to obtain an actual system weight. Under the condition of scarce engine design data, the present application can significantly improve the prediction accuracy of engine weight prediction by enhancing training samples and extracting latent features.

Description

Technical Field

[0001] This invention belongs to the field of engine system weight prediction, specifically relating to an engine system weight prediction method, system, device, and storage medium. Background Technology

[0002] The essence of flight lies in overcoming gravity, making system-level weight one of the most critical performance parameters in aircraft design. Due to the strong coupling between design variables, errors in early-stage weight prediction can propagate to misjudgments in structural dimensions, propulsion system requirements, and payload capacity. A well-known example is the F-35B program, whose empty weight ultimately exceeded predictions by approximately 10% due to an underestimation of the weight of its new equipment. This directly led to reduced payload capacity and limitations on the load factor. As a crucial component of the aircraft system, the weight of the engine system plays a central role in influencing center of gravity, lift, range, and fuel efficiency. Therefore, accurate and efficient prediction of engine system weight during the conceptual design phase is essential for ensuring the effectiveness of design and development efforts.

[0003] With the rapid development of artificial intelligence technology, the application of machine learning methods in aero-engine performance prediction and analysis has attracted widespread attention. These methods demonstrate significant advantages in terms of modeling simplicity and computational efficiency. Unlike traditional regression or physical models, machine learning techniques do not require explicit construction of physical laws or feature engineering; instead, they learn statistical regularities directly from data, enabling efficient and potentially more accurate predictions in high-dimensional, complex design spaces. Furthermore, many machine learning models have low computational overhead during the inference phase, making them ideal for applications with high timeliness requirements, such as design iteration and real-time performance prediction.

[0004] For the aforementioned technologies, machine learning requires a large number of publicly available high-quality datasets. However, due to confidentiality reasons, there are very few publicly available datasets for engine system weight, resulting in poor model training performance and low prediction accuracy. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method, system, device and storage medium for predicting the weight of an engine system. By imposing constraints on the generative adversarial network, the generated data sets are made more realistic. Then, the prediction model is trained and optimized using the generated data sets to obtain an optimized prediction model, which makes the prediction results more accurate.

[0006] Obtain the engine design parameters and the corresponding system weight;

[0007] A mapping relationship is established between the engine design parameters and the corresponding system weights to obtain the original data set.

[0008] The original data set is input into a generative adversarial network (GAN) to obtain a generated data set. The GAN includes a generator and a discriminator. Constraints are set on the generator to enable it to evaluate the generated data set and obtain an evaluation value. The discriminator discriminates the generated data set and obtains a discrimination value. Based on the discrimination value and the evaluation value, the GAN is optimized to obtain an optimized GAN.

[0009] Based on the optimized generative adversarial network and the original data set, an optimized generated data set is obtained, and the optimized generated data set and the original data set are combined into a training data set.

[0010] The training data set is input into the neural network model, hidden layer features are extracted, the hidden layer features are input into the activation function to obtain the output value, the loss value is calculated based on the output value and the true value, and the model parameters of the neural network model are optimized by minimizing the loss value to obtain the optimized neural network model. The true value is the system weight in the training data set.

[0011] Based on the optimized neural network model and the training data set, latent features are obtained;

[0012] The latent features are input into the prediction model to obtain the predicted values. Based on the predicted values ​​and the true values, an objective function is constructed. The adjustment parameters are obtained when the objective function's target value is minimized. The prediction model is then optimized based on the adjustment parameters to obtain the optimized prediction model.

[0013] The actual engine design parameters are input into the optimization prediction model to obtain the actual system weight.

[0014] Optionally, the step of inputting the original data set into a generative adversarial network (GAN) to obtain a generated data set, wherein the GAN includes a generator and a discriminator, setting constraints on the generator to enable the generator to evaluate the generated data set and obtain an evaluation value, the discriminator to discriminate the generated data set and obtain a discrimination value, and optimizing the GAN based on the discrimination value and the evaluation value to obtain an optimized GAN, includes:

[0015] The original data set is input into the generative adversarial network to obtain the generated data set;

[0016] The generator evaluates the generated data set and obtains an evaluation value, which is represented as:

[0017]

[0018] in, For generator, As a condition, Indicates the condition Below, generator The generated data set , To generate the prior distribution of the data set, express It is sampled from the generated data set. It is a discriminator. This indicates that for generators The generated data set Under given conditions The probability that the generated data set is derived from the original data set is given, and the range of values ​​is given. , This indicates that under random sampling, according to generated samples Input discriminator In the middle, the discriminator gives the probability of being true and takes the expected value of the logarithm;

[0019] The discriminator performs a discrimination operation on the generated data group to obtain a discrimination value, which is represented as follows:

[0020]

[0021] in, express It was sampled from the original dataset. For the discriminator pair The output of the probabilities from the original data set is then logarithmic. For discriminator right The expected value of the logarithm of the probability of being true given from the original data set;

[0022] Based on the discriminant value and the evaluation value, the generative adversarial network is optimized to obtain an optimized generative adversarial network, represented as:

[0023]

[0024] in, Indicates the distribution from the original data set The real samples obtained from sampling.

[0025] Optionally, the step of inputting the training data set into the neural network model, extracting hidden layer features, passing the hidden layer features through an activation function to obtain output values, calculating a loss value based on the output value and the true value, and optimizing the model parameters of the neural network model by minimizing the loss value to obtain an optimized neural network model includes:

[0026] The training data set is input into the neural network model to extract hidden layer features. , represented as:

[0027]

[0028] in, For activation function, To hide the number of layers, For the first The weight matrix of the hidden layer, For the first The bias vector of the hidden layer. For the training data set;

[0029] The hidden layer features are passed through an activation function to obtain the output value;

[0030] Based on the output value and the true value, the loss value is calculated and expressed as:

[0031]

[0032] in, This is the loss value. This is the weight matrix. For bias vectors, For the input sample, For real labels, For the output value, Denotes the Euclidean norm;

[0033] An optimized neural network model is obtained by optimizing the model parameters of the neural network model by minimizing the loss value.

[0034] Optionally, based on the optimized neural network model and the training data set, the latent feature representation is obtained as follows:

[0035]

[0036] in, For the first The weight matrix of the hidden layer, For the first Hidden layer output of the layer, and It is a constant. For the first The bias vector of the hidden layer. These are potential features.

[0037] Optionally, the step of inputting the latent features into the prediction model to obtain predicted values, constructing an objective function based on the predicted values ​​and the true values, obtaining adjustment parameters when the objective function's target value is minimized, and optimizing the prediction model based on the adjustment parameters to obtain an optimized prediction model includes:

[0038] The latent features are input into the prediction model to obtain the predicted value, which is expressed as:

[0039]

[0040] in, For predicted values, Indicates the number of regression trees. For the function space of the regression tree, For the first Tree samples The predicted value;

[0041] Based on the predicted and actual values, an objective function is constructed, expressed as:

[0042]

[0043] in, This represents the error between the predicted and actual values. This represents the total number of engine samples. for The sum of the regularization terms of the tree can be expanded as follows:

[0044]

[0045] in, For the first Round iteration, i.e. the first iteration Optimization of trees and To adjust the parameters, The number of leaves, For the first Output value of a leaf;

[0046] The adjustment parameters are obtained when the objective function is minimized. The prediction model is then optimized based on the adjustment parameters to obtain the optimized prediction model.

[0047] Optionally, obtaining the adjustment parameters when the objective function's target value is minimized, and optimizing the prediction model based on the adjustment parameters, includes:

[0048] No. The objective function of a tree can be expressed as:

[0049]

[0050] in, For the first The objective function value of each tree. For the first The predicted values ​​output by each tree. For the true value, For the first Tree pairs of samples The predicted value, For constant terms, This represents the total number of engine samples. For the first The regularization term of a tree, For the first The prediction error value for each tree;

[0051] The first The objective function expression of the tree is expanded using Taylor, resulting in the Taylor expansion expression, which is then simplified as follows:

[0052]

[0053] in, and Let these represent the first and second derivatives of the loss function, respectively. For belonging to the first The set of all samples of a leaf. For the first A leaf, For the first For each sample, take the partial derivative of the Taylor expansion and set it to 0. Then the optimal output value can be obtained. :

[0054]

[0055] At this point, the minimum objective value calculated using the objective function is... for:

[0056]

[0057] Obtain the adjustment parameters when the objective value of the objective function is minimized, and optimize the prediction model based on the adjustment parameters.

[0058] Optionally, the step of obtaining the adjustment parameter that minimizes the objective value of the objective function based on the Taylor expansion expression, and optimizing the prediction model based on the adjustment parameter to obtain the optimized prediction model includes:

[0059] Obtaining the segmentation evaluation value when the regression tree structure splits. A , represented as:

[0060]

[0061] in, and These are the sample sets of the left and right subtrees after the split, respectively. This represents the sample set before the split;

[0062] If the segmentation evaluation value is greater than 0, then continue to split the leaf nodes of the tree;

[0063] When the segmentation evaluation value is less than 0, the splitting of the logarithmic leaf nodes is stopped, and the number of leaves is obtained;

[0064] Based on the number of leaves and the Taylor expansion expression, we obtain the first... The optimal output value of the leaf nodes of the tree;

[0065] The minimum objective function value is calculated based on the optimal output value. The adjustment parameter is obtained when the objective function has the minimum objective value. The prediction model is then optimized based on the adjustment parameter to obtain the optimized prediction model.

[0066] An engine system weight prediction system, comprising:

[0067] The acquisition module is used to acquire engine design parameters and the corresponding system weight.

[0068] The association module is used to establish a mapping relationship between the engine design parameters and the system weights corresponding to the design parameters, and to obtain several sets of raw data.

[0069] A generative adversarial network (GAN) optimization module is used to input the original data set into a GAN to obtain a generated data set. The GAN includes a generator and a discriminator. Constraints are set on the generator to enable it to evaluate the generated data set and obtain an evaluation value. The discriminator discriminates the generated data set to obtain a discrimination value. Based on the discrimination value and the evaluation value, the GAN is optimized to obtain an optimized GAN.

[0070] The training data set generation module is used to obtain an optimized generated data set based on the optimized generative adversarial network and the original data set, and to combine the optimized generated data set and the original data set into a training data set.

[0071] The neural network model optimization module is used to input the training data set into the neural network model, extract hidden layer features, input the hidden layer features into the activation function to obtain the output value, calculate the loss value based on the output value and the true value, optimize the model parameters of the neural network model by minimizing the loss value, and obtain an optimized neural network model, wherein the true value is the system weight in the training data set;

[0072] The feature extraction module is used to obtain potential features based on the optimized neural network model and the training data set;

[0073] The prediction model optimization module is used to input the latent features into the prediction model to obtain the predicted value, construct an objective function based on the predicted value and the true value, obtain the adjustment parameter when the objective function's target value is minimized, and optimize the prediction model based on the adjustment parameter to obtain the optimized prediction model.

[0074] The prediction module is used to input the actual engine design parameters into the optimization prediction model to obtain the actual system weight.

[0075] A terminal device includes a memory and a processor. The memory stores a computer program that can run on the processor. When the processor loads and executes the computer program, it employs an engine system weight prediction method.

[0076] A computer-readable storage medium storing a computer program that, when loaded and executed by a processor, employs an engine system weight prediction method.

[0077] The beneficial effects of this invention are:

[0078] 1. By establishing a mapping relationship between engine design parameters and the corresponding system weight, an original data set is obtained. This data set is then input into a generative adversarial network (GAN) for training. The optimized GAN is obtained by optimizing the evaluation values ​​generated by the generator and the discriminant values ​​generated by the discriminator. A training data set is then generated and input into the neural network model to obtain a loss value. The optimized neural network model is obtained by minimizing the loss value, and latent features are extracted. These latent features are then input into the prediction model. By minimizing the objective function, the optimal prediction model is obtained. The actual engine design parameters are then input into the optimized prediction model to obtain the actual system weight. This application has the capability to generate a large amount of similar data (i.e., a generated data set) when data is insufficient. After feature extraction from the generated data set, the prediction model is trained to obtain an optimized training prediction model, thus improving prediction accuracy. Compared to traditional prediction model training, under conditions of scarce engine design data, the prediction accuracy of engine weight prediction can be significantly improved by enhancing training samples and extracting latent features.

[0079] 2. DNN (Neural Network Model) is used to capture complex feature interactions, while XGBoost (Prediction Model) provides high performance and accuracy in regression prediction, improving the accuracy and stability of prediction. Attached Figure Description

[0080] Figure 1 This is a schematic diagram of the overall structure of the engine system weight prediction method of the present invention;

[0081] Figure 2 This is a graph showing the evolution of the loss function during the training process of this invention.

[0082] Figure 3 This is a comparison chart of the actual values ​​and predicted values ​​of this invention;

[0083] Figure 4 This is a comparison chart showing the accuracy of predictions in a DNN model using the original dataset and the augmented dataset in this invention.

[0084] Figure 5 This invention compares the accuracy of predictions made in the XGBoost model using the original dataset and the augmented dataset.

[0085] Figure 6 This invention compares the accuracy of predictions made in the Int model using the original dataset and the augmented dataset.

[0086] Figure 7 This diagram illustrates the improvement in prediction performance brought about by the DNN and XGBoost modules of this invention.

[0087] Figure 8 This is a statistical comparison chart of the prediction accuracy of each model in this invention;

[0088] Figure 9 This is a schematic diagram comparing the prediction accuracy of the three models of this invention on the original dataset;

[0089] Figure 10 This is a comparison chart of the prediction accuracy of the three models of this invention on the expanded dataset;

[0090] Figure 11 This is a distribution of prediction accuracy for different models and datasets under different thresholds in this invention. Detailed Implementation

[0091] An engine system weight prediction method, such as Figure 1 As shown, it includes:

[0092] S1. Obtain the engine design parameters and the corresponding system weight;

[0093] Specifically, the engine design parameters are expressed as follows:

[0094]

[0095] in, This indicates the total number of engine samples, each Indicates the first One engine 3D feature vectors. This feature space contains There are several independent design parameters that may affect the system's weight characteristics. Accordingly, the system weight is represented as an objective vector. ,in Indicates the first The system weight is obtained from the measurements of each engine. Given the diverse factors influencing aircraft engine weight, each input sample can be formally represented as: Each input sample can be formally represented as:

[0096]

[0097] in The first term used for system weight prediction Engine design parameters This indicates the total number of design parameters of this type.

[0098] S2. Establish a mapping relationship between the engine design parameters and the system weight corresponding to the design parameters to obtain the original data set (real sample).

[0099] Specifically, based on the corresponding engine data The system weight is predicted using a learning function. This allows it to input the engine. Mapped to the corresponding system weight prediction value As shown in the following formula:

[0100]

[0101] in, Indicates the first The predicted weight of each engine is the predicted value.

[0102] Each And a corresponding real This is called a raw data set.

[0103] S3. Input the original data set into the generative adversarial network to obtain a generated data set. The generative adversarial network includes a generator and a discriminator. Set constraints on the generator to make the generator evaluate the generated data set and obtain an evaluation value. The discriminator discriminates the generated data set and obtains a discrimination value. Based on the discrimination value and the evaluation value, optimize the generative adversarial network to obtain an optimized generative adversarial network.

[0104] Specifically, due to the limited publicly available data on engine weight parameters, which is insufficient for model training, a Conditional Generative Adversarial Network (cGAN) model is used to augment the engine data. cGANs are considered a promising solution, approximating the distribution of real-world data and generating synthetic samples that retain the intrinsic characteristics of the original data, thereby expanding the training set and alleviating the data scarcity problem. Random noise and conditional information are input into the generator G, which then produces synthetic samples and feeds them into the discriminator D for evaluation. Unlike traditional GANs, the cGAN discriminator not only needs to distinguish between real and fake samples but also needs to determine whether the generated samples meet predetermined conditions.

[0105] The process involves inputting the original data set into a generative adversarial network (GAN) to obtain a generated data set. The GAN includes a generator and a discriminator. The generator evaluates the generated data set to obtain an evaluation value, and the discriminator discriminates the generated data to obtain a discrimination value. Based on the discrimination value and the evaluation value, the GAN is optimized to obtain an optimized GAN, comprising:

[0106] The original data set is input into the generative adversarial network to obtain the generated data set;

[0107] The generator evaluates the generated data set and obtains an evaluation value, which is represented as:

[0108]

[0109] in, For generator, As a condition, Indicates the condition Below, generator The generated data set , To generate the prior distribution of the data set, express It is sampled from the generated data set. It is a discriminator. This indicates that for generators The generated data set Under given conditions The probability that the generated data set is derived from the original data set is given, and the range of values ​​is given. , This indicates that under random sampling, according to generated samples Input discriminator In the middle, the discriminator gives the probability of being true and takes the expected value of the logarithm;

[0110] The discriminator performs a discrimination operation on the generated data group to obtain a discrimination value, which is represented as follows:

[0111]

[0112] in, express It was sampled from the original dataset. For the discriminator pair The output of the probabilities from the original data set is then logarithmic. For discriminator right The expected value of the logarithm of the probability of being true given from the original data set.

[0113] Based on the discriminant value and the evaluation value, the generative adversarial network is optimized to obtain an optimized generative adversarial network, represented as:

[0114]

[0115] in, Indicates the distribution from the original data set The real samples obtained from sampling.

[0116] S4. Based on the optimized generative adversarial network and the original data set, an optimized generated data set is obtained, and the optimized generated data set and the original data set are used as training data sets.

[0117] S5. Input the training data set into the neural network model, extract the hidden layer features, input the hidden layer features into the activation function to obtain the output value, calculate the loss value based on the output value and the true value, optimize the model parameters of the neural network model by minimizing the loss value, and obtain the optimized neural network model. The true value is the system weight in the training data set.

[0118] Specifically, to fully utilize the information contained in the input data, we employ a deep neural network (DNN) model for feature extraction from the engine data. DNNs possess structural advantages through the introduction of multiple layers of nonlinear transformations. By constructing a complex network architecture comprising an input layer, multiple hidden layers, and an output layer, DNNs can progressively extract features from the data. This multi-layered structure enables DNNs to learn various feature representations from the raw data, ranging from simple to complex.

[0119] Furthermore, DNNs possess powerful adaptive learning capabilities. They are trained on large-scale datasets using the backpropagation algorithm, which continuously adjusts the network parameters to minimize the error between the predicted output and the actual label. This process enables DNNs to sensitively capture subtle latent features in the data while fitting its underlying distribution characteristics well. Even with complex, high-dimensional, and noisy data, DNNs remain remarkably effective in extracting valuable latent features.

[0120] The steps of inputting the training data set into the neural network model, extracting hidden layer features, passing the hidden layer features through an activation function to obtain output values, calculating a loss value based on the output value and the true value, and optimizing the model parameters of the neural network model by minimizing the loss value to obtain an optimized neural network model include:

[0121] The training data set is input into the neural network model to extract hidden layer features. , represented as:

[0122]

[0123] in, For activation function, To hide the number of layers, For the first The weight matrix of the hidden layer, For the first The bias vector of the hidden layer. For the training data set;

[0124] Specifically, the constructed DNN model includes A hidden layer.

[0125] The hidden layer features are passed through an activation function to obtain the output value;

[0126] Specifically, in the hidden layer, ScaledExponentialLinearUnit (SELU) is used as the activation function, defined as follows:

[0127]

[0128] in, and These are constants; default values ​​are used. The above describes the forward propagation process of the DNN model. To find suitable weights... and bias DNN uses the backpropagation algorithm to minimize the error between the predicted output and the true label, and defines a loss function.

[0129] Based on the output value and the true value, the loss value is calculated and expressed as:

[0130]

[0131] in, The loss value. This is the weight matrix. For bias vectors, For the input sample, For real labels, For the output value, Denotes the Euclidean norm;

[0132] An optimized neural network model is obtained by optimizing the model parameters of the neural network model by minimizing the loss value.

[0133] Specifically, gradient descent is used to iteratively solve for the weights and biases of each layer, improving network performance and enabling it to learn. The trained DNN model can then effectively extract high-dimensional latent features from the original data.

[0134] S6. Based on the optimized neural network model and the training data set, obtain the potential features;

[0135] Based on the optimized neural network model and the training data set, the latent features are represented as follows:

[0136]

[0137] in, For the first The weight matrix of the hidden layer, For the first Hidden layer output of the layer, and It is a constant. For the first The bias vector of the hidden layer. These are potential features.

[0138] S7. Input the latent features into the prediction model to obtain the predicted value. Based on the predicted value and the true value, construct the objective function, obtain the adjustment parameter when the objective value of the objective function is minimized, and optimize the prediction model based on the adjustment parameter to obtain the optimized prediction model.

[0139] Specifically, the prediction model in this application employs XGBoost. XGBoost is an efficient and highly scalable gradient boosting framework that uses classification and regression trees as its base learners. These base learners are combined sequentially to form a strong learner, significantly improving the model's prediction accuracy and generalization ability.

[0140] The process of inputting the latent features into the prediction model to obtain predicted values, constructing an objective function based on the predicted values ​​and the true values, obtaining adjustment parameters when the objective function's target value is minimized, and optimizing the prediction model based on the adjustment parameters to obtain the optimized prediction model includes:

[0141] The latent features are input into the prediction model to obtain the predicted value, which is expressed as:

[0142]

[0143] in, For predicted values, Indicates the number of regression trees. For the function space of the regression tree, For the first Tree samples The predicted value;

[0144] Based on the predicted and actual values, an objective function is constructed, expressed as:

[0145]

[0146] in, This represents the error between the predicted and actual values. This represents the total number of engine samples. for The sum of the regularization terms of the tree can be expanded as follows:

[0147]

[0148] in, For the first Round iteration, i.e. the first iteration Optimization of trees and To adjust the parameters, The number of leaves, For the first Output value of a leaf;

[0149] The adjustment parameters are obtained when the objective function is minimized. The prediction model is then optimized based on the adjustment parameters to obtain the optimized prediction model.

[0150] The step of obtaining the adjustment parameter when the objective function is minimized, and optimizing the prediction model based on the adjustment parameter, includes:

[0151] No. The objective function of the tree is:

[0152]

[0153] in, For the first The objective function value of each tree. For the first The predicted values ​​output by each tree. For the true value, For the first Tree pairs of samples The predicted value, For constant terms, This represents the total number of engine samples. For the first The regularization term of a tree, For the first The prediction error value for each tree.

[0154] The first The objective function expression of the tree is expanded using Taylor, resulting in the Taylor expansion expression, which is then simplified as follows:

[0155]

[0156] in, and Let these represent the first and second derivatives of the loss function, respectively. For belonging to the first The set of all samples of a leaf. For the first A leaf, For the first Taking the partial derivative of the Taylor expansion with respect to a given sample and setting it to zero, the optimal output value is obtained. :

[0157]

[0158] At this point, the minimum objective value calculated using the objective function is... for:

[0159]

[0160] Obtain the adjustment parameters when the objective value of the objective function is minimized, and optimize the prediction model based on the adjustment parameters.

[0161] Specifically, to further optimize performance, XGBoost introduces a second-order Taylor expansion into the loss function to accelerate the convergence process. Simultaneously, to reduce the risk of overfitting, the complexity of the tree model is explicitly added to a regularization term. This approach improves the model's robustness, making it less prone to overfitting even on complex datasets.

[0162] Based on the Taylor expansion expression, the adjustment parameter is obtained when the objective value of the objective function is minimized. The prediction model is then optimized based on the adjustment parameter to obtain the optimized prediction model.

[0163] The step of obtaining the adjustment parameter when the objective function is minimized based on the Taylor expansion expression, and optimizing the prediction model based on the adjustment parameter to obtain the optimized prediction model includes:

[0164] Obtaining the segmentation evaluation value when the regression tree structure splits. A , represented as:

[0165]

[0166] in, and These are the sample sets of the left and right subtrees after the split, respectively. This represents the sample set before the split;

[0167] If the segmentation evaluation value is greater than 0, then continue to split the leaf nodes of the tree;

[0168] When the segmentation evaluation value is less than 0, the splitting of the logarithmic leaf nodes is stopped, and the number of leaves is obtained;

[0169] Based on the number of leaves and the Taylor expansion expression, we obtain the first... The optimal output value of the leaf nodes of the tree;

[0170] The minimum objective function value is calculated based on the optimal output value. The adjustment parameter is obtained when the objective function has the minimum objective value. The prediction model is then optimized based on the adjustment parameter to obtain the optimized prediction model.

[0171] Based on the number of leaves and the Taylor expansion expression, we obtain the first... Leaf nodes of a tree Output value Thus, a complete XGBoost prediction model is obtained.

[0172] S8. Input the actual engine design parameters into the prediction model to obtain the actual system weight. Specific Implementation

[0174] Dataset

[0175] The dataset used in this application includes design parameters for 144 commercially manufactured engines and 39 turbofan engines studied in multiple aviation projects. The production time of these commercial engines spans from the mid-1960s to the late 2010s, covering more than half a century of technological advancements and engineering experience, thus providing a realistic basis for predictive analysis.

[0176] This application focuses on 12 parameters of the engine, as shown in Table 1. In this dataset, turbofan engines can be further divided into three types based on system type and number of fans, as shown in Table 2. These are also the conditional labels used when generating the data using cGAN.

[0177] Table 1 Engine Dataset Parameters

[0178]

[0179] Table 2 Three types of turbofan systems

[0180]

[0181] The input features used in the prediction task correspond to the engine design parameters numbered 3 to 10 in Table 1. These symbols include: BPR (SLS), OPR (SLS), Thrust, IBS (SLS), CruiseMach, CruiseAlt.kft.YearCertified, SystemType, and No.ofSpools. The target variable for prediction (item 12) is the propulsion system weight.

[0182] Experimental setup

[0183] The dataset was divided into training and test sets in an 8:2 ratio. The algorithm used was implemented based on the TensorFlow framework. All experiments were run on a high-performance server configured with dual AMD EPYC740 224-core 2.8GHz processors, 256GB of memory, and eight GeForce RTX 3090 GPUs.

[0184] Evaluation metrics: We use three metrics to evaluate model performance: mean absolute error (MAE), coefficient of determination (R²R²R²), and accuracy (Acc).

[0185] The three are defined as follows:

[0186] MAE (Mean Absolute Error) is defined as:

[0187]

[0188] in, For the first The actual system weight of an engine. This is the predicted value (i.e., the predicted system weight). This represents the total number of engine samples.

[0189] Decision system Defined as:

[0190]

[0191] in This is the average of the true values.

[0192] Accuracy (Acc) is defined as:

[0193]

[0194] in It is the first The error between the predicted value and the reference value.

[0195] Experimental Results and Analysis

[0196] Data augmentation and prediction tasks are performed sequentially by the Int model (corresponding to feature extraction and prediction). During the cGAN-based data augmentation process, we used GridSearch to adjust model parameters and compared the convergence of the loss learning curves under various settings. Finally, we determined the optimal hyperparameters for the two sub-networks in cGAN; details are shown in Table 3.

[0197] Table 3 shows the parameter settings for the generator and discriminator in the cGAN model.

[0198]

[0199] Ideally, the discriminator should be able to accurately distinguish between all real samples (loss of 1) and fake samples (loss of 0), with the final discriminator loss being approximately 0.5. Related trends are as follows... Figure 2 As shown.

[0200] Subsequently, the original real samples and synthetic samples were merged and input into the DNN model for feature extraction. The extracted features were then input into the XGBoost model for prediction. Hyperparameter grid search was also performed on both the DNN and XGBoost models, and an EarlyStopping strategy was introduced to prevent overfitting and excessively long training time. The relevant hyperparameter settings are shown in Table 4.

[0201] Table 4 Parameter settings for DNN and XGBoost modules

[0202]

[0203] Table 5 Prediction results for each data point

[0204]

[0205] Table 5 shows the comparison results between the predicted values ​​and the test set. Of the 37 evaluated engines, 29 achieved a prediction accuracy exceeding 99%, with an overall accuracy exceeding 90%. However, there were some instances of inaccurate predictions; for example, the prediction accuracy for an engine with an actual weight of 5598 was only 92.06%. These outliers will be investigated further in future work.

[0206] Figure 3 The actual values ​​of 37 analyzed engines are shown. ) and predicted value ( A direct comparison between the predicted and actual data points. The chart clearly demonstrates the high degree of consistency between the predicted values ​​and the actual data points, using a perfect prediction line as a reference. Despite some minor deviations, most data points are close to this line, indicating high accuracy in the predictions across the entire dataset.

[0207] ablation experiment

[0208] The core innovation of this application framework lies in the application of data augmentation techniques and the combination of deep neural networks (DNNs) and the XGBoost model. To verify the effectiveness of these operations, ablation experiments are required. Two ablation models were designed for this study. The first experiment aimed to verify the improvement effect of data augmentation on weight prediction performance, while the second experiment evaluated the effectiveness of the Int model.

[0209] Experiment (1)

[0210] Two datasets were used: the original engine dataset (Raw_data) and the augmented engine dataset (Aug_data), along with three models: a deep neural network (DNN) model, an XGBoost model, and an ensemble model. First, Raw_data was fed into these three models to obtain baseline predictions. Next, Aug_data was fed into these models under the same settings. Finally, the prediction results of the two datasets were compared. The results show that, for each model, using augmented data significantly improved prediction accuracy. Figure 4 This diagram compares the prediction accuracy of three models (DNN, XGBoost, and an ensemble model) on the original dataset (Raw_data) and the augmented dataset (Aug_data). Each subplot shows the accuracy value for a single sample under both data settings, with dashed lines connecting the corresponding data points to highlight the enhancement effect. Figure 5 and Figure 6 As shown, DNN and XGBoost models exhibit significantly improved prediction accuracy when trained with augmented data, especially on samples with initially poor performance. Figure 5 and Figure 6In this model, one horizontal axis corresponds to one engine sample, with a total of 37 engine samples. This improvement is particularly evident in the XGBoost case, indicating its greater sensitivity to data richness. Figure 6 The effectiveness of the ensemble model is further emphasized, as it consistently outperforms individual models on most samples. When trained on the Aug_data dataset, the Int model not only achieves higher accuracy but also demonstrates greater stability, with almost all predictions exceeding 0.96. These results collectively demonstrate the effectiveness of the data augmentation method.

[0211] Furthermore, the improvement in prediction accuracy in both DNN and XGBoost models compared to Aug_data was quantified. The results are as follows: Figure 7 As shown in the bar chart, the positive change in prediction accuracy for 37 engine samples after switching from Raw_data to Aug_data when using DNN and XGBoost models is visually presented. The 37 groups of bars from bottom to top sequentially show the accuracy improvement for samples numbered 1 to 37. Blue bars represent the accuracy improvement achieved by the DNN model after using Aug_data, orange bars represent the accuracy improvement achieved by XGBoost after using Aug_data, and gray bars represent no accuracy improvement. It can be seen that the prediction accuracy of most samples improved, with some samples showing improvements exceeding 10%. Notably, in some cases, the improvement of the XGBoost model was more significant (e.g., samples 7, 12, and 23), while the DNN model showed a more consistent but moderate improvement across most data points. These results further demonstrate the effectiveness of data augmentation, especially in improving model performance for samples with initially low prediction accuracy.

[0212] The statistical characteristics of the prediction results from DNN and XGBoost models were further analyzed. For example... Figure 8 As shown, box plots compare the prediction accuracy distributions of the three models (DNN, XGBoost, and the ensemble model) on the original and augmented datasets. For each model, data augmentation significantly improves accuracy in terms of both central tendency and variability. Specifically, after augmentation, the median accuracy of all three models increases, and the interquartile range becomes more compact, which not only improves prediction performance but also enhances stability. The ensemble model (Int) outperforms the individual models, particularly when using augmented data, with its accuracy distribution concentrated above 0.99 and minimal variance.

[0213] Experiment (2)

[0214] To validate the effectiveness of the ensemble model component in the proposed framework, we compared the prediction performance of three models—with and without the ensemble model—on two datasets. When the ensemble model was not used, predictions were made using DNN and XGBoost models respectively. Figure 9 and Figure 10 The performance of the three models on both the original and augmented datasets is presented. These comparisons further demonstrate the contribution of the ensemble model to the overall prediction accuracy. The ensemble model consistently exhibits higher prediction accuracy than the individual DNN and XGBoost models on most data points. In particular, when using the augmented dataset (Aug_data), the ensemble model not only performs better overall, but its accuracy values ​​are also more tightly clustered around the upper limit (close to 1.0), showing better stability. In contrast, the individual models exhibit greater volatility in prediction accuracy, especially on the original dataset, with several noticeable outliers. These observations suggest that the ensemble of DNN and XGBoost effectively leverages their complementary strengths, thereby improving overall performance and reducing sensitivity to data variability.

[0215] We then analyzed the distribution of prediction accuracy within a predefined threshold range before and after introducing the ensemble model component. The results are presented as a heatmap. Figure 11 The heatmap displays the number and percentage of predictions exceeding each threshold for comparison. The five threshold ranges considered are >0.80, >0.85, >0.90, >0.95, and >0.99, which were used to evaluate all models on both the raw and augmented datasets (cGAN). Each cell in the heatmap shows the number of samples exceeding the corresponding threshold and its proportion. Figure 9 It is evident that the ensemble model outperforms individual models regardless of whether raw or augmented data is used. Particularly in the augmented data setting, the proportion of samples with a prediction accuracy exceeding 0.99 reached 78%, which not only improved precision but also ensured more stable and reliable predictions across the entire dataset. This result demonstrates that the proposed framework not only enhances prediction accuracy but also strengthens model reliability, providing a more robust foundation for practical applications requiring high-confidence predictions.

[0216] Finally, we used three evaluation metrics: mean absolute error (MAE), coefficient of determination (R²), and accuracy (ACC) to compare the predictive performance of the three models on the two datasets. The results are summarized in Table 6. The ensemble model trained on the augmented dataset, namely IA, consistently showed the best overall performance. It achieved the lowest mean absolute error of 110.656, the highest coefficient of determination of 0.998, and a narrow accuracy range between 0.921 and 1.000, with an average accuracy of 0.991. Compared to the models trained on the original dataset—namely DR, XR, and IR—all models benefiting from augmented data (DA, XA, and IA) showed significant improvements. This demonstrates the advantages of data augmentation in improving predictive accuracy and generalization ability. Although the XGBoost model performed well in terms of coefficient of determination, achieving 0.978 in both the XR and XA settings, it showed greater variability at the sample level. In particular, the lowest accuracy of XR dropped to 0.591, indicating a decrease in stability in some cases. In contrast, the ensemble models (IR and IA) not only achieved higher average accuracy but also demonstrated more consistent predictions across the entire dataset. These findings confirm that combining model ensemble with data augmentation provides significant benefits in terms of prediction accuracy and robustness. This approach offers a promising direction for improving the reliability of engine system weight estimation in spacecraft design.

[0217] Table 6 compares the prediction results of the two datasets and the three models. DR / XR / IR represent the Deep Neural Network (DNN), XGBoost, and ensemble model using the original dataset, respectively; DA / XA / IA represent the corresponding models trained on the augmented dataset.

[0218] Table 6. Comparison of Prediction Results for Different Models and Data Sets

[0219]

[0220] This application proposes a method for predicting the weight of aircraft engine systems, aiming to address the dual challenges of data scarcity and model complexity in the early design phase. By combining a cGAN model to generate synthetic data with an integrated DNN-XGBoost regression model, this framework significantly improves the accuracy and stability of predictions. Experimental results show that this framework not only reduces prediction errors but also improves inter-sample consistency. Specifically, the data augmentation module significantly improves the accuracy of all three baseline models, with several models showing improvements exceeding 10%. The integrated DNN-XGBoost model further enhances the stability of predictions, with over 90% of predictions achieving at least 95% accuracy and 78% achieving 99% accuracy. Therefore, this method has practical application value in aerospace engineering, especially in the preliminary design phase where empirical data is scarce but accurate estimation is crucial. It provides engineers with an effective tool to help them generate high-confidence weight predictions under uncertainty conditions, thereby supporting conceptual design studies, early risk assessments, and resource planning in aircraft engine system development. Although this method has proven effective, some limitations remain. While the generated data is diverse, it may not fully capture the subtle physical constraints inherent in engine architecture, and the current model does not explicitly incorporate domain knowledge or physical priors. Although the overall prediction accuracy is high, slight performance differences remain between individual samples, indicating room for improvement in the model's sensitivity to design-specific changes. Future research will focus on improving the physical realism of augmented data by introducing physics-based constraints and enhancing the model's sensitivity to design-specific changes to reduce performance differences between samples. These efforts may also include exploring adaptive feature selection or attention mechanisms to improve local prediction accuracy while maintaining computational efficiency.

[0221] An engine system weight prediction system, comprising:

[0222] The acquisition module is used to acquire engine design parameters and the corresponding system weight.

[0223] The association module is used to establish a mapping relationship between the engine design parameters and the system weights corresponding to the design parameters, and to obtain several sets of raw data.

[0224] A generative adversarial network (GAN) optimization module is used to input the original data set into a GAN to obtain a generated data set. The GAN includes a generator and a discriminator. Constraints are set on the generator to enable it to evaluate the generated data and obtain an evaluation value. The discriminator discriminates the generated data set to obtain a discrimination value. Based on the discrimination value and the evaluation value, the GAN is optimized to obtain an optimized GAN.

[0225] The training data set generation module is used to obtain an optimized generated data set based on the optimized generative adversarial network and the original data set, and to combine the optimized generated data set and the original data set into a training data set.

[0226] The neural network model optimization module is used to input the training data set into the neural network model, extract hidden layer features, input the hidden layer features into the activation function to obtain the output value, calculate the loss value based on the output value and the true value, optimize the model parameters of the neural network model by minimizing the loss value, and obtain an optimized neural network model, wherein the true value is the system weight in the training data set;

[0227] The feature extraction module is used to obtain potential features based on the optimized neural network model and the training data set;

[0228] The prediction model optimization module is used to input the latent features into the prediction model to obtain the predicted value, construct an objective function based on the predicted value and the true value, obtain the adjustment parameter when the objective function's target value is minimized, and optimize the prediction model based on the adjustment parameter to obtain the optimized prediction model.

[0229] The prediction module is used to input the actual engine design parameters into the optimization prediction model to obtain the actual system weight.

[0230] This application also discloses a terminal device, including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor loads and executes the computer program, it employs an engine system weight prediction method.

[0231] The terminal device can be a computer device such as a desktop computer, a laptop computer, or a cloud server. The terminal device includes, but is not limited to, a processor and a memory. For example, the terminal device may also include input / output devices, network access devices, and buses.

[0232] The processor can be a central processing unit (CPU). Of course, depending on the actual use, it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc., and this application does not limit it.

[0233] The memory can be an internal storage unit of the terminal device, such as a hard disk or RAM of the terminal device, or an external storage device of the terminal device, such as a plug-in hard disk, smart memory card (SMC), secure digital card (SD), or flash memory card (FC) equipped on the terminal device. Furthermore, the memory can be a combination of internal storage units and external storage devices of the terminal device. The memory is used to store computer programs and other programs and data required by the terminal device. The memory can also be used to temporarily store data that has been output or will be output. This application does not limit this.

[0234] In this terminal device, an engine system weight prediction method from the above embodiments is stored in the terminal device's memory and loaded and executed on the terminal device's processor for convenient use.

[0235] This application also discloses a computer-readable storage medium, which stores a computer program, wherein when the computer program is executed by a processor, it employs an engine system weight prediction method as described in the above embodiments.

[0236] The computer program can be stored in a computer-readable medium. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or certain middleware. The computer-readable medium includes any entity or device capable of carrying computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the computer-readable medium includes, but is not limited to, the above-mentioned components.

[0237] The engine system weight prediction method described in the above embodiments is stored in the computer-readable storage medium and loaded and executed on the processor to facilitate the storage and application of the above method.

[0238] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of protection of this application is limited to these examples; within the framework of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of one or more embodiments of this application as described above, which are not provided in detail for the sake of brevity.

[0239] One or more embodiments in this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments in this application should be included within the protection scope of this application.

Claims

1. A method for predicting the weight of an engine system, characterized in that, include: Obtain the engine design parameters and the corresponding system weight; A mapping relationship is established between the engine design parameters and the corresponding system weights to obtain the original data set. The original data set is input into a generative adversarial network (GAN) to obtain a generated data set. The GAN includes a generator and a discriminator. Constraints are set on the generator to enable it to evaluate the generated data set and obtain an evaluation value. The discriminator discriminates the generated data set and obtains a discrimination value. Based on the discrimination value and the evaluation value, the GAN is optimized to obtain an optimized GAN. Based on the optimized generative adversarial network and the original data set, an optimized generated data set is obtained, and the optimized generated data set and the original data set are combined into a training data set. The training data set is input into the neural network model, hidden layer features are extracted, the hidden layer features are input into the activation function to obtain the output value, the loss value is calculated based on the output value and the true value, and the model parameters of the neural network model are optimized by minimizing the loss value to obtain the optimized neural network model. The true value is the system weight in the training data set. Based on the optimized neural network model and the training data set, latent features are obtained; The latent features are input into the prediction model to obtain the predicted values. Based on the predicted values ​​and the true values, an objective function is constructed. The adjustment parameters are obtained when the objective function's target value is minimized. The prediction model is then optimized based on the adjustment parameters to obtain the optimized prediction model. The actual engine design parameters are input into the optimization prediction model to obtain the actual system weight.

2. The engine system weight prediction method as described in claim 1, characterized in that, The process involves inputting the original data set into a generative adversarial network (GAN) to obtain a generated data set. The GAN includes a generator and a discriminator. Constraints are set on the generator to evaluate the generated data set and obtain an evaluation value. The discriminator then discriminates the generated data set to obtain a discrimination value. Based on the discrimination value and the evaluation value, the GAN is optimized to obtain an optimized GAN, comprising: The original data set is input into the generative adversarial network to obtain the generated data set; The generator evaluates the generated data set and obtains an evaluation value, which is represented as: in, For generator, As a condition, Indicates the condition Below, generator The generated data set , To generate the prior distribution of the data set, express It is sampled from the generated data set. It is a discriminator. This indicates that for generators The generated data set Under given conditions The probability that the generated data set is derived from the original data set is given, and the range of values ​​is given. , This indicates that under random sampling, according to generated samples Input discriminator In the middle, the discriminator gives the probability of being true and takes the expected value of the logarithm; The discriminator performs a discrimination operation on the generated data group to obtain a discrimination value, which is represented as follows: in, express It was sampled from the original dataset. For the discriminator pair The output of the probabilities from the original data set is then logarithmic. For discriminator right The expected value of the logarithm of the probability of being true given from the original data set; Based on the discriminant value and the evaluation value, the generative adversarial network is optimized to obtain an optimized generative adversarial network, represented as: in, Indicates the distribution from the original data set The real samples obtained from sampling.

3. The engine system weight prediction method as described in claim 1, characterized in that, The steps of inputting the training data set into the neural network model, extracting hidden layer features, passing the hidden layer features through an activation function to obtain output values, calculating a loss value based on the output value and the true value, and optimizing the model parameters of the neural network model by minimizing the loss value to obtain an optimized neural network model include: The training data set is input into the neural network model to extract hidden layer features. , represented as: in, For activation function, To hide the number of layers, For the first The weight matrix of the hidden layer, For the first The bias vector of the hidden layer. For the training data set; The hidden layer features are passed through an activation function to obtain the output value; Based on the output value and the true value, the loss value is calculated and expressed as: in, The loss value. This is the weight matrix. For bias vectors, For the input sample, For real labels, For the output value, Denotes the Euclidean norm; An optimized neural network model is obtained by optimizing the model parameters of the neural network model by minimizing the loss value.

4. The engine system weight prediction method as described in claim 1, characterized in that, Based on the optimized neural network model and the training data set, the latent features are represented as follows: in, For the first The weight matrix of the hidden layer, For the first Hidden layer output of the layer, and It is a constant. For the first The bias vector of the hidden layer. These are potential features.

5. The engine system weight prediction method as described in claim 1, characterized in that, The process of inputting the latent features into the prediction model to obtain predicted values, constructing an objective function based on the predicted values ​​and the true values, obtaining adjustment parameters when the objective function's target value is minimized, and optimizing the prediction model based on the adjustment parameters to obtain the optimized prediction model includes: The latent features are input into the prediction model to obtain the predicted value, which is expressed as: in, For predicted values, Indicates the number of regression trees. For the function space of the regression tree, For the first Tree samples The predicted value; Based on the predicted and actual values, an objective function is constructed, expressed as: in, This represents the error between the predicted and actual values. This represents the total number of engine samples. for The sum of the regularization terms of the tree can be expanded as follows: in, For the first Round iteration, i.e. the first iteration Optimization of trees and To adjust the parameters, The number of leaves, For the first Output value of a leaf; The adjustment parameters are obtained when the objective function is minimized. The prediction model is then optimized based on the adjustment parameters to obtain the optimized prediction model.

6. The engine system weight prediction method as described in claim 5, characterized in that, The step of obtaining the adjustment parameter when the objective function is minimized, and optimizing the prediction model based on the adjustment parameter, includes: No. The objective function of the tree is expressed as: in, For the first The objective function value of each tree. For the first The predicted values ​​output by each tree. For the true value, For the first Tree pairs of samples The predicted value, For constant terms, This represents the total number of engine samples. For the first The regularization term of a tree, For the first The prediction error value for each tree; The first The objective function expression of the tree is expanded using Taylor, resulting in the Taylor expansion expression, which is then simplified as follows: in, and Let these represent the first and second derivatives of the loss function, respectively. For belonging to the first The set of all samples of a leaf. For the first A leaf, For the first Taking the partial derivative of the Taylor expansion with respect to a given sample and setting it to zero, the optimal output value is obtained. : At this point, the minimum objective value calculated using the objective function is... for: Obtain the adjustment parameters when the objective value of the objective function is minimized, and optimize the prediction model based on the adjustment parameters.

7. The engine system weight prediction method as described in claim 6, characterized in that, The step of obtaining the adjustment parameter when the objective function is minimized based on the Taylor expansion expression, and optimizing the prediction model based on the adjustment parameter to obtain the optimized prediction model includes: Obtaining the segmentation evaluation value when the regression tree structure splits. A , represented as: in, and These are the sample sets of the left and right subtrees after the split, respectively. This represents the sample set before the split; If the segmentation evaluation value is greater than 0, then continue to split the leaf nodes of the tree; When the segmentation evaluation value is less than 0, the splitting of the logarithmic leaf nodes is stopped, and the number of leaves is obtained; Based on the number of leaves and the Taylor expansion expression, we obtain the first... The optimal output value of the leaf nodes of the tree; The minimum objective function value is calculated based on the optimal output value. The adjustment parameter is obtained when the objective function has the minimum objective value. The prediction model is then optimized based on the adjustment parameter to obtain the optimized prediction model.

8. An engine system weight prediction system, characterized in that, include: The acquisition module is used to acquire engine design parameters and the corresponding system weight. The association module is used to establish a mapping relationship between the engine design parameters and the system weights corresponding to the design parameters, and to obtain several sets of raw data. A generative adversarial network (GAN) optimization module is used to input the original data set into a GAN to obtain a generated data set. The GAN includes a generator and a discriminator. Constraints are set on the generator to enable it to evaluate the generated data set and obtain an evaluation value. The discriminator discriminates the generated data set to obtain a discrimination value. Based on the discrimination value and the evaluation value, the GAN is optimized to obtain an optimized GAN. The training data set generation module is used to obtain an optimized generated data set based on the optimized generative adversarial network and the original data set, and to combine the optimized generated data set and the original data set into a training data set. The neural network model optimization module is used to input the training data set into the neural network model, extract hidden layer features, input the hidden layer features into the activation function to obtain the output value, calculate the loss value based on the output value and the true value, optimize the model parameters of the neural network model by minimizing the loss value, and obtain an optimized neural network model, wherein the true value is the system weight in the training data set; The feature extraction module is used to obtain potential features based on the optimized neural network model and the training data set; The prediction model optimization module is used to input the latent features into the prediction model to obtain the predicted value, construct an objective function based on the predicted value and the true value, obtain the adjustment parameter when the objective function's target value is minimized, and optimize the prediction model based on the adjustment parameter to obtain the optimized prediction model. The prediction module is used to input the actual engine design parameters into the optimization prediction model to obtain the actual system weight.

9. A terminal device, comprising a memory and a processor, characterized in that, The memory stores a computer program that can run on a processor, and when the processor loads and executes the computer program, it employs the method described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is loaded and executed by the processor, it employs the method described in any one of claims 1 to 7.

Images