Armored vehicle diesel engine digital twin model reduction method and system

Abstract

The invention discloses an armored vehicle diesel engine digital twin model reduction method and system, and the method comprises the steps: building a diesel engine twin model which is a high-dimensional nonlinear system and comprises a thermodynamic equation of a diesel engine; operating the diesel engine twin model under different working conditions, collecting input and output data of variables, and performing standardization processing on the input and output data of the variables to obtain simulation data; a neural network is used for constructing a nerve differential equation model, simulation data is used for training the nerve differential equation model to obtain a reduced-order model, and the output of the diesel engine is predicted based on the obtained reduced-order model. Through Simulink modeling, neural network training, incremental learning and real-time data updating, a dynamic model of the diesel engine is effectively simplified, and it is ensured that high-precision prediction capacity is kept under different working conditions and operation environments. The system architecture supports real-time monitoring, online learning and self-adaptive adjustment, and high efficiency and stability of the diesel engine control system in practical application are guaranteed.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent control technology, and specifically relates to a digital twin model order reduction method and system for an armored vehicle diesel engine. Background Art

[0002] Diesel engine performance directly determines the mobility, durability, and reliability of armored vehicles. To ensure efficient operation of armored vehicles in complex battlefield environments, diesel engine condition monitoring and fault prediction are critical tasks. Diesel engine performance analysis and optimization typically rely on complex mathematical models. However, with increasing system complexity and computing resources, traditional modeling methods face challenges such as low computational efficiency and poor real-time performance. To improve the accuracy and computational efficiency of diesel engine performance prediction, digital twin technology has been proposed in recent years and is increasingly being applied to diesel engine condition monitoring and fault prediction.

[0003] However, existing digital twin models often involve extensive computation during their construction and solution, making model order reduction crucial. While traditional order reduction methods (such as POD and Krylov methods) can reduce computational complexity, they struggle to simultaneously maintain high accuracy and real-time performance. With advances in artificial intelligence (AI), the combination of neural networks and ordinary differential equations (ODEs) has emerged as a new approach to improving model accuracy and computational efficiency. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to address the deficiencies in the above-mentioned prior art and provide a method and system for reducing the order of a digital twin model of an armored vehicle diesel engine. The method uses a neural network to learn the dynamic behavior of the diesel engine system and obtains the predicted output of the system at any time through ordinary differential equations. The method can significantly improve the computational efficiency while ensuring the accuracy of the model, and provide a new method for the optimization design, performance evaluation and fault diagnosis of diesel engines, thereby solving the technical problems of low computational efficiency, poor real-time performance and difficulty in balancing accuracy in traditional diesel engine digital twin models.

[0005] The present invention adopts the following technical solutions: A digital twin model reduction method for an armored vehicle diesel engine includes the following steps: Establish a diesel engine twin model, which is a high-dimensional nonlinear system including the thermodynamic equations of the diesel engine; Run the diesel engine twin model under different operating conditions, collect the input and output data of the variables, and standardize the input and output data of the variables to obtain simulation data; A neural network is used to construct a neural network ordinary differential equation model, which is then trained with simulation data to obtain a reduced-order model. The output of the diesel engine is then predicted based on the obtained reduced-order model.

[0006] Preferably, the variable input and output data include rotational speed, cylinder pressure, intake pressure, exhaust pressure, intake temperature, exhaust temperature, intake flow rate, and exhaust flow rate.

[0007] Preferably, the input and output data of the variables are standardized to obtain simulation data, specifically: The input and output data of the variables are divided into 80% training set and 20% test set, the value of each variable is adjusted to zero mean and unit variance, and the value range of the variable is limited to -1 to 1.

[0008] Preferably, a neural network is used to construct a neural ordinary differential equation model, and a low-dimensional neural network architecture is used to fit the dynamic behavior of the diesel engine twin model. The neural network structure includes an activation function and a network depth.

[0009] Preferably, the feature matrix and target variable are normalized and converted into PyTorch tensors, which are then batch processed and loaded using TensorDataset and DataLoader. The neural network module defined by the ODEFunc class uses a three-layer fully connected network with a Tanh activation function in each layer; The ODEBlock class uses the odeint function to solve ordinary differential equations and return the prediction results of the system at different time steps; The NeuralODE class combines the neural ordinary differential equation model, including the input layer, the ODE solver module, and the output layer; the input layer maps the input features to the hidden layer dimensions, and after ODEBlock processing, the final prediction result is obtained through the output layer.

[0010] Preferably, the training process of the Neural ODE model uses the Adam optimizer and the mean squared error as the loss function, and traverses the training data 100 times; in each epoch, the data is loaded in batches through train_loader and sent to the multilayer perceptron for forward propagation; The calculated MSE loss between the predicted results and the actual target values is back-propagated to optimize the parameters of the neural ordinary differential equation network; the predicted results and the actual results are visualized using matplotlib.

[0011] Preferably, key parameters include the dimension of the input features, the dimension of the hidden layer, the batch size, the learning rate, and the number of epochs for training.

[0012] Preferably, the Kalman filter algorithm is used to train the neural ordinary differential equation network, and the Simulink simulation data is used for optimization to adjust the weights of the neural ordinary differential equation network to obtain a low-order reduced-order model that reflects the dynamic behavior of the Simulink model.

[0013] Preferably, Simulink is used to build a diesel engine digital twin model.

[0014] In a second aspect, an embodiment of the present invention provides an armored vehicle diesel engine digital twin model reduction system, comprising: Constructing modules and establishing a diesel engine twin model. The diesel engine twin model is a high-dimensional nonlinear system that includes the thermodynamic equations of the diesel engine. The data module runs the diesel engine twin model under different working conditions, collects the input and output data of the variables, and standardizes the input and output data of the variables to obtain simulation data; The order reduction module uses a neural network to build a neural network ordinary differential equation model, uses simulation data to train the neural network ordinary differential equation model to obtain a reduced-order model, and predicts the output of the diesel engine based on the obtained reduced-order model.

[0015] In a third aspect, a computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned armored vehicle diesel engine digital twin model reduction method when executing the computer program.

[0016] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, comprising a computer program, which, when executed by a processor, implements the steps of the above-mentioned armored vehicle diesel engine digital twin model reduction method.

[0017] In the fifth aspect, a chip includes a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, the steps of the above-mentioned armored vehicle diesel engine digital twin model reduction method are implemented.

[0018] In a sixth aspect, an embodiment of the present invention provides an electronic device, including a computer program, which, when executed by the electronic device, implements the steps of the above-mentioned armored vehicle diesel engine digital twin model reduction method.

[0019] Compared with the prior art, the present invention has at least the following beneficial effects: A digital twin model reduction method for armored vehicle diesel engines reduces the order of a high-dimensional diesel engine Simulink model using a Neural Ordinary Differential Equation (NODE) model, reducing the computational effort, optimizing memory usage, and thus improving computational efficiency. The reduced-order model can effectively preserve the nonlinear and time-varying characteristics of the original diesel engine system. Through incremental learning, the reduced-order model can continuously adapt to different operating conditions and environments, ensuring high-precision prediction and control in real-time control systems. The present invention is not only suitable for diesel engine performance prediction, but can also be widely used in multiple fields such as fault diagnosis and control optimization, thus possessing important industrial application value.

[0020] Furthermore, the core state variables of diesel engine thermodynamics are covered to ensure data completeness. By screening highly relevant variables, redundant calculations are reduced, and irrelevant noise is prevented from interfering with model training, the physical rationality of the reduced-order model is improved from the source of the data.

[0021] Furthermore, normalization is used to eliminate the negative impact of dimensional differences on neural network training, which in principle makes the gradient descent optimization direction consistent and accelerates convergence; the test set is divided to verify the generalization ability of the model and prevent overfitting.

[0022] Furthermore, the neural network is limited to a low-dimensional architecture with Tanh activation and three-layer full connectivity. This enforces dimensionality reduction through low-dimensional hidden layers, and the Tanh function smoothes nonlinear responses, preventing parameter explosion in deep networks. In the PyTorch implementation, DataLoader batch loading is combined with GPU parallel computing to improve training efficiency. The ODEBlock module encapsulates differential equation solving, enabling flexible time step prediction to meet real-time requirements.

[0023] Furthermore, the Adam optimizer + MSE loss is used, and adaptive learning rate adjustment is combined with mean square error minimization to balance convergence speed and stability in principle; the training process is visually monitored to facilitate dynamic optimization of hyperparameters and ensure that the model quickly reaches the optimal state.

[0024] Furthermore, Kalman filtering and Simulink simulation are introduced. Kalman filtering performs state estimation on noisy data to improve data quality; Simulink provides a high-fidelity simulation environment to ensure that the training data is dynamically consistent with the real system, thereby enhancing the practical applicability of the reduced-order model.

[0025] It can be understood that the beneficial effects of the second to sixth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here.

[0026] In summary, the present invention uses the dual mechanisms of data-driven dimensionality reduction and physical constraint modeling to reduce computational complexity while embedding the thermodynamic laws of diesel engines, ultimately achieving high-precision real-time prediction and providing reliable technical support for online monitoring and fault diagnosis of armored vehicle diesel engines.

[0027] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0029] Figure 1 Flow chart of the method of the present invention; Figure 2 This is the Simulink model diagram of the diesel engine; Figure 3 This is a diagram showing the training effect of the Neural Ordinary Differential Equation algorithm; Figure 4 A schematic diagram of a computer device provided in accordance with an embodiment of the present invention; Figure 5 The present invention is a block diagram of an electronic device according to an embodiment of the present invention.

[0030] Among them, 60. Computer device; 61. Processor; 62. Memory; 63. Computer program; 600. Electronic device; 610. Processing unit; 620. Storage unit; 6201. Random access memory unit; 6202. Cache memory unit; 6203. Read-only memory unit; 6204. Program / Utility; 6205. Program module; 630. Bus; 640. Display unit; 650. Input / output interface; 660. Network adapter; 700. External device. DETAILED DESCRIPTION

[0031] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0032] In the description of the present invention, it is to be understood that the terms “include” and “comprise” indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0033] It should also be understood that the terms used in the present specification are only for the purpose of describing particular embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0034] It should be further understood that the term "and / or" as used in the present specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in the present invention generally indicates that the associated objects are in an "or" relationship.

[0035] It should be understood that although the terms "first," "second," and "third" may be used to describe preset ranges in embodiments of the present invention, these preset ranges should not be limited to these terms. These terms are merely used to distinguish one preset range from another. For example, without departing from the scope of embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0036] The word "if," as used herein, may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to the determination" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)," depending on the context.

[0037] The accompanying drawings illustrate various schematic diagrams of structures according to embodiments disclosed herein. These figures are not drawn to scale; for clarity, some details are exaggerated and some details may be omitted. The shapes of the various regions and layers shown in the figures, as well as their relative sizes and positional relationships, are merely exemplary and may deviate in practice due to manufacturing tolerances or technical limitations. Those skilled in the art may design regions / layers with different shapes, sizes, and relative positions as needed.

[0038] This paper provides a digital twin model reduction method for armored vehicle diesel engines. Through Simulink modeling, neural network training, incremental learning, and real-time data updates, it effectively simplifies the engine's dynamic model and ensures high-precision prediction capabilities under various operating conditions and environments. Furthermore, the system architecture supports real-time monitoring, online learning, and adaptive adjustment, ensuring the efficiency and stability of the diesel engine control system in practical applications.

[0039] Example 1 See also Figure 1 The present invention provides an armored vehicle diesel engine digital twin model reduction method, comprising the following steps: S1. Use Simulink modeling tools to create a complete diesel engine digital twin model. The diesel engine digital twin model includes the main thermodynamic equations of the diesel engine and can reflect the complex dynamic behavior of the diesel engine under different operating conditions. First, the interaction between the diesel engine's thermodynamic equations and dynamic behavior is comprehensively considered. Based on the basic thermodynamic principles of the diesel engine, such as compression, combustion, and expansion processes, the energy conversion and loss in its working cycle are accurately described. By establishing an appropriate thermodynamic model to simulate the mixing of fuel and air, the compression ignition process, the combustion reaction and the dynamic changes of exhaust, the performance of the diesel engine under different working conditions can be accurately reflected.

[0040] In addition, in order to improve the reliability and complexity of the diesel engine digital twin model, it is necessary to consider the diesel engine behavior under different load and speed conditions, covering the entire process from engine intake, combustion to exhaust. Figure 2 As shown, it includes a single-cylinder model, a multi-cylinder superposition model, a two-stage turbocharger model, a two-stage intercooler model, an intake pipe model, and an exhaust pipe model.

[0041] S2. Run the diesel engine digital twin model under different operating conditions to collect input and output data of key variables; Specifically including: speed, cylinder pressure, intake pressure, exhaust pressure, intake temperature, exhaust temperature, intake flow, exhaust flow; these data will be used as input to the Neural Ordinary Differential Equation (NODE) model for neural network training. The details are as follows: Set up multiple typical operating conditions to cover the diesel engine's behavior under different operating conditions. For each operating condition, adjust input parameters such as load, speed, temperature, and pressure to ensure that the complex dynamic characteristics of the diesel engine in actual operation can be simulated; Next, the diesel engine digital twin model is run, and during the simulation, the input and output data of key variables are recorded in real time. These key variables include speed, cylinder pressure, intake pressure, exhaust pressure, intake temperature, exhaust temperature, intake flow rate, and exhaust flow rate. The changes in these variables at each time step will reflect the thermodynamic, aerodynamic, and mechanical processes within the diesel engine. This data will serve as input to the Neural Ordinary Differential Equation (NODE) model and enter the subsequent neural network training phase. During the training process, the NODE model learns these time series data to identify and capture the dynamic evolution laws of diesel engines under different operating conditions.

[0042] S3. Before building the neural network model, extract the input and output data of key variables from the diesel engine digital twin model and standardize the data. This data normalization ensures that the neural network can learn efficiently and reduces the impact of different dimensions. The data is divided into an 80% training set and a 20% test set. To ensure that the neural network can learn efficiently and avoid adverse effects on model training due to differences in data scale, the collected data must be standardized. The numerical value of each variable is adjusted to zero mean and unit variance, and the value range of the variable is limited to between -1 and 1; this can eliminate the influence of different dimensions between different physical quantities, so that the neural network can focus more on learning the intrinsic relationship between the variables during training, rather than being disturbed by dimensional differences, thereby improving the convergence speed and prediction accuracy of the neural network.

[0043] S4. Based on the simulation data characteristics of the diesel engine digital twin model, a neural network is used to construct a neural ordinary differential equation (NODE) model. This model can fit the differential equations of the diesel engine dynamic system through the neural network structure, capturing the nonlinear and time-varying characteristics of the system. It uses a deep neural network architecture with an appropriate number of network layers and neurons per layer, and adopts activation functions suitable for dynamic system modeling to ensure that the complex behavior of the diesel engine system can be effectively captured. The specific Python implementation process is as follows: After normalizing the feature matrix and target variables, they are converted into PyTorch tensors, which are then batch processed and loaded using TensorDataset and DataLoader for more efficient training and testing. The neural network module defined by the ODEFunc class uses a three-layer fully connected network with a Tanh activation function in each layer; The ODEBlock class uses the odeint function to solve ordinary differential equations and return the prediction results of the system at different time steps; The NeuralODE class combines the entire neural ordinary differential equation model, including the input layer, the ODE solver, and the output layer. The input layer maps the input features to the hidden layer dimensions, and after ODEBlock processing, the final prediction result is obtained through the output layer. The output of the Neural ODE Reduced Order model is the value at the last time step, which is based on the result of solving the differential equation.

[0044] The training process of the Neural ODE reduction model uses the Adam optimizer and the mean squared error (MSE) as the loss function. During training, the Neural ODE model iterates over the training data 100 times (i.e., 100 training epochs). In each epoch, the data is loaded in batches using train_loader and fed into the network for forward propagation. The calculated MSE loss between the predicted result and the actual target value will be back-propagated to optimize the network parameters; During the training process, the performance of the neural ordinary differential equation reduction model is monitored by testing the results on the validation set, and the MSE, RMSE, MAE and relative error of the test set are calculated; Finally, the predicted results and actual results of the neural ordinary differential equation reduction model are visualized using matplotlib. The visualization effect is as follows: Figure 3 to compare the difference between the predicted and actual values.

[0045] The key parameters in the whole process include the dimension of the input features (input_dim=7), the dimension of the hidden layer (hidden_dim=20), the batch size (batch_size=8), the learning rate (learning_rate=0.001), and the number of epochs for training (epochs=100).

[0046] S5. Use the Kalman filter algorithm to adjust the neural network parameters, optimize using Simulink simulation data, adjust the network weights, and reduce fitting errors, thereby obtaining a low-order reduced-order model that can accurately reflect the dynamic behavior of the diesel engine. The details are as follows: Use the Kalman filter algorithm to estimate the optimal values of neural network parameters in real time, and gradually correct the network weights by comparing the output of the neural network with Simulink simulation data; The simulation data provides the real dynamic response of the diesel engine system under different operating conditions, and these data are used as the target output of the neural network; To achieve effective reduced-order modeling, we focused on optimizing several key neural network parameters, including the input feature dimension, hidden layer dimension, batch size, learning rate, and the number of training epochs. By adjusting these parameters, the Kalman filter algorithm can dynamically adjust network weights, reducing the fitting error of the reduced-order neural ordinary differential equation model. This allows the neural network to efficiently learn and reflect the complex dynamic behavior of the diesel engine system within limited computing resources. Ultimately, the optimized reduced-order neural ordinary differential equation model can not only accurately predict the output of the diesel engine, but also has a lower model order, thus achieving a balance between computational efficiency and prediction accuracy.

[0047] This process makes the neural network model simpler and has better generalization ability, which can stably predict the dynamic characteristics of the diesel engine under different operating conditions.

[0048] S6. Verification and adjustment of reduced-order model.

[0049] S601: Verify the trained reduced-order model and check the output differences between it and the original Simulink model under different operating conditions. Adjust the reduced-order model based on the difference results to ensure that the reduced-order model accurately reflects the dynamic behavior of the original model.

[0050] S602: Evaluate the accuracy and reliability of the model by comparing the reduced-order model simulation data with the experimental data. If the prediction error is large, adjust the network parameters and retrain to further optimize the model's accuracy and generalization ability.

[0051] Those skilled in the art will appreciate that various aspects of the present invention may be implemented as systems, methods, or program products. Accordingly, various aspects of the present invention may be implemented in the following forms: entirely in hardware, entirely in software (including firmware, microcode, etc.), or in a combination of hardware and software, collectively referred to herein as "circuits," "modules," or "platforms."

[0052] Example 2 The present invention provides an armored vehicle diesel engine digital twin model order reduction system, which can be used to implement the above-mentioned armored vehicle diesel engine digital twin model order reduction method. Specifically, the armored vehicle diesel engine digital twin model order reduction system includes a construction module, a data module and a reduction module.

[0053] Among them, the construction module establishes the diesel engine twin model, which is a high-dimensional nonlinear system including the thermodynamic equations of the diesel engine; The data module runs the diesel engine twin model under different working conditions, collects the input and output data of the variables, and standardizes the input and output data of the variables to obtain simulation data; The order reduction module uses a neural network to build a neural network ordinary differential equation model, uses simulation data to train the neural network ordinary differential equation model to obtain a reduced-order model, and predicts the output of the diesel engine based on the obtained reduced-order model.

[0054] Example 3 The present invention provides a terminal device, which includes a processor and a memory, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or may also be other general-purpose processors, graphics processing units (GPUs), tensor processing units (TPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to implement corresponding method processes or corresponding functions; the processor described in the embodiment of the present invention can be used for the operation of the digital twin model reduction method of armored vehicle diesel engines, including: A diesel engine twin model is established. The diesel engine twin model is a high-dimensional nonlinear system that includes the thermodynamic equations of the diesel engine. The diesel engine twin model is operated under different operating conditions to collect the input and output data of the variables, and the input and output data of the variables are standardized to obtain simulation data. A neural network is used to construct a neural network ordinary differential equation model, and the neural network ordinary differential equation model is trained with simulation data to obtain a reduced-order model. The output of the diesel engine is predicted based on the obtained reduced-order model.

[0055] See also Figure 4 The terminal device is a computer device. The computer device 60 of this embodiment includes: a processor 61, a memory 62, and a computer program 63 stored in the memory 62 and executable on the processor 61. When executed by the processor 61, the computer program 63 implements the armored vehicle diesel engine digital twin model order reduction method of the embodiment. To avoid repetition, a detailed description is not given here. Alternatively, when executed by the processor 61, the computer program 63 implements the functions of each model / unit in the armored vehicle diesel engine digital twin model order reduction system of the embodiment. To avoid repetition, a detailed description is not given here.

[0056] The computer device 60 may be a desktop computer, a notebook computer, a PDA, a cloud server, or other computing devices. The computer device 60 may include, but is not limited to, a processor 61 and a memory 62. It will be understood by those skilled in the art that Figure 4 This is merely an example of the computer device 60 and does not constitute a limitation of the computer device 60 . The computer device 60 may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the computer device may also include input and output devices, network access devices, buses, etc.

[0057] The processor 61 may be a central processing unit (CPU), or other general-purpose processors, a graphics processing unit (GPU), a tensor processing unit (TPU), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0058] The memory 62 may be an internal storage unit of the computer device 60, such as a hard disk or memory of the computer device 60. The memory 62 may also be an external storage device of the computer device 60, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device 60.

[0059] Furthermore, the memory 62 may include both an internal storage unit of the computer device 60 and an external storage device. The memory 62 is used to store computer programs and other programs and data required by the computer device. The memory 62 may also be used to temporarily store data that has been output or is about to be output.

[0060] See also Figure 5The terminal device is an electronic device 600, which is implemented as a general-purpose computing device. The components of the electronic device may include, but are not limited to, at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including the storage unit 620 and the processing unit 610), and a display unit 640.

[0061] The storage unit stores program codes, which can be executed by the processing unit 610, so that the processing unit 610 performs the steps according to various exemplary embodiments of the present invention described in the above method section of this specification. For example, the processing unit 610 can perform the following steps: Figure 1 Follow the steps shown in .

[0062] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache memory unit 6202 , and may further include a read-only memory unit (ROM) 6203 .

[0063] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0064] Bus 630 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0065] The electronic device 600 may also communicate with one or more external devices 700 (e.g., a keyboard, a pointing device, a Bluetooth device, etc.), one or more devices that enable a user to interact with the electronic device 600, and / or any device that enables the electronic device 600 to communicate with one or more other computing devices (e.g., a router, a modem). Such communication may occur via an input / output interface 650. Furthermore, the electronic device 600 may also communicate with one or more networks (e.g., a local area network, a wide area network, and / or a public network, such as the Internet) via a network adapter 660. The network adapter 660 may communicate with other modules of the electronic device 600 via a bus 630. It should be understood that, although not shown in the figures, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms.

[0066] Example 4 The present invention also provides a storage medium, specifically a computer-readable storage medium. The computer-readable storage medium is a memory device in a terminal device, used to store programs and data. It is understood that the computer-readable storage medium herein may include both the built-in storage medium in the terminal device and, of course, the extended storage medium supported by the terminal device. It may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for being loaded and executed by a processor. These instructions may be one or more computer programs (including program code). It should be noted that more specific examples of the computer-readable storage medium herein include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical fiber, a portable compact disk read-only memory, an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0067] Computer-readable storage media also include data signals propagated in baseband or as part of a carrier wave, which carry readable program code. Such propagated data signals can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable storage medium can also be any readable medium other than a readable storage medium, which can send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, device, or device. The program code contained on the readable storage medium can be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, radio frequency, etc., or any suitable combination of the above.

[0068] The program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, and the like, as well as conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a stand-alone software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device via any type of network, including a local area network or a wide area network, or may be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0069] The processor may load and execute one or more instructions stored in a computer-readable storage medium to implement the corresponding steps of the armored vehicle diesel engine digital twin model order reduction method in the above embodiment; the processor may load and execute the following steps: A diesel engine twin model is established. The diesel engine twin model is a high-dimensional nonlinear system that includes the thermodynamic equations of the diesel engine. The diesel engine twin model is operated under different operating conditions to collect the input and output data of the variables, and the input and output data of the variables are standardized to obtain simulation data. A neural network is used to construct a neural network ordinary differential equation model, and the neural network ordinary differential equation model is trained with simulation data to obtain a reduced-order model. The output of the diesel engine is predicted based on the obtained reduced-order model.

[0070] The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may include, but are not limited to, general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic units, data processing logic units based on quantum computing, and the like.

[0071] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0072] Example The real-time application and online update method of the armored vehicle diesel engine digital twin model order reduction includes the following steps: Applying the reduced-order neural ordinary differential equation model to the diesel engine control system. Real-time sensor data from the diesel engine, such as speed, cylinder pressure, intake flow, and exhaust flow, is collected and provided to the reduced-order model for prediction and analysis. The reduced-order model is updated through an incremental learning algorithm. Whenever new data arrives, the incremental update method adjusts the weights of the neural network so that the model always reflects the current dynamic behavior of the diesel engine. By comparing the real-time prediction results of the reduced-order model with the output of the actual diesel engine system, the model parameters are corrected to ensure that the reduced-order model is highly consistent with the actual diesel engine dynamic behavior; The system continuously monitors the prediction accuracy of the reduced-order model and makes adaptive adjustments when necessary. When the model accuracy deviates, it adjusts the learning rate and update frequency of incremental learning, or adjusts the structure of the neural network, to further improve the model's accuracy and generalization ability.

[0073] 1. Real-time Prediction Performance Verification (Based on a 12V150 Diesel Engine Bench Test) In full-operation simulation tests, the reduced-order model completed state prediction in a real-time calculation cycle of 0.8ms, a 550% improvement in efficiency compared to the traditional POD reduced-order model (5.2ms). Key performance comparison data is as follows:

[0074] Through the time series modeling capabilities of neural ordinary differential equations (Neural ODEs), the computational speed bottleneck of traditional reduced-order models is broken through while ensuring accuracy, meeting the real-time requirements of transient working conditions such as rapid acceleration and emergency braking of armored vehicles.

[0075] 2. Verification of Incremental Learning Effects (Based on Plateau Combat Environment Data) At an altitude of 4,500 meters, by deploying the incremental learning module, the model achieved the following adaptive adjustment results for key parameters after receiving 200 hours of continuous operating data: Air-fuel ratio control accuracy: The initial model error was 12.7%, which was reduced to 4.1% after incremental learning (optimization of 67.7%). Turbo lag compensation: After adding the altitude-intake density correlation parameter, the turbo response time is shortened from 1.8s to 0.9s; Model reconstruction efficiency: Only the weights of the first three layers of the neural network need to be updated (accounting for 21% of the total network parameters) to complete more than 90% of feature migration.

[0076] The elastic network regularized incremental learning algorithm is adopted to achieve efficient integration of new working condition features while retaining the original thermodynamic constraints, avoiding the catastrophic forgetting problem.

[0077] 3. Fault Diagnosis Application Case (A Certain Armored Brigade's Actual Combat Training Data) For diesel engine injector partial blockage failures, the system triggers a three-level early warning mechanism through model prediction deviation:

[0078] Fault sensitivity grading is achieved through a dynamic residual threshold algorithm, which reduces the false alarm rate by 42% compared to the traditional threshold method and increases the advance warning time by 3-5 times.

[0079] 4. Hardware-in-the-Loop Simulation Verification (Based on the dSPACE SCALEXIO Platform) In the hardware-in-the-loop test, the model was deployed on a domestically produced in-vehicle computing unit (Huawei MDC 610). The performance indicators are as follows: Resource usage: Memory usage is compressed from 1.2GB (original model) to 180MB, meeting the requirements of in-vehicle embedded systems; Multitasking: Maintaining a 10ms control cycle while simultaneously running fault diagnosis (20% CPU) and status prediction (45% CPU); Communication load: CAN bus data throughput is reduced by 72% (only feature vectors are transmitted instead of full-dimensional data).

[0080] Through feature space dimensional compression technology (Feature Space Compression Rate = 85.3%), efficient deployment in edge computing scenarios is achieved.

[0081] Experimental data validates this method's significant advantages in computational efficiency (response speed increased by 5.5 times), environmental adaptability (error reduction in plateau conditions by 67.7%), and cost-effectiveness (maintenance costs reduced by 81%). In particular, its 0.6-second fault warning capability provides revolutionary technical support for the reliable operation of armored vehicle diesel engines in battlefield environments.

[0082] In summary, the present invention provides a method and system for reducing the order of a digital twin model of an armored vehicle diesel engine. The system has a millisecond-level response to support reliable power output for tactical actions (such as rapid assaults); the fault warning time is advanced by 3-5 times, significantly improving the equipment's availability; the annual maintenance cost is saved by over 300,000 yuan, and the fuel economy is optimized to extend the combat radius; the model's self-update capability ensures mission continuity under extreme working conditions such as plateaus and deserts; it provides systematic technical support for the status of armored vehicle diesel engines to be known, faults to be prevented, and performance to be controlled, and is a key innovation for improving the combat effectiveness of modern armored forces.

[0083] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0084] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0085] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians 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 present invention.

[0086] In the embodiments provided by the present invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For example, the division of the modules or units is merely a logical functional division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of devices or units, and can be electrical, mechanical, or other forms.

[0087] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0088] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0089] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of the above-mentioned various method embodiments. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0090] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices, and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0091] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0092] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0093] The above content is only for explaining the technical idea of the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A digital twin model reduction method for an armored vehicle diesel engine, characterized in that: The following steps are involved: Use Simulink to build a preliminary diesel engine twin model; Run the diesel engine twin model under different operating conditions, collect the input and output data of the variables, and standardize the input and output data of the variables to obtain simulation data; Use neural networks to build a neural network ordinary differential equation model, and use simulation data to train the neural network ordinary differential equation model to obtain a reduced-order model; The Kalman filter algorithm is used to adjust the neural network parameters, optimize the model, and obtain the final reduced-order model.

2. The armored vehicle diesel engine digital twin model order reduction method according to claim 1 is characterized in that: The variable input and output data include speed, cylinder pressure, intake pressure, exhaust pressure, intake temperature, exhaust temperature, intake flow rate, and exhaust flow rate.

3. The digital twin model order reduction method for an armored vehicle diesel engine according to claim 1 is characterized in that: The input and output data of the variables are standardized to obtain simulation data, specifically: The input and output data of the variables are divided into 80% training set and 20% test set, the value of each variable is adjusted to zero mean and unit variance, and the value range of the variable is limited to -1 to 1.

4. The digital twin model order reduction method for an armored vehicle diesel engine according to claim 1, characterized in that: A neural network is used to construct a neural ordinary differential equation model, and a low-dimensional neural network architecture is used to fit the dynamic behavior of the diesel engine twin model. The neural network structure includes activation function and network depth.

5. The digital twin model order reduction method for an armored vehicle diesel engine according to claim 4 is characterized in that: After normalizing the feature matrix and target variables, they are converted into PyTorch tensors and batch processed and loaded using TensorDataset and DataLoader. The neural network module defined by the ODEFunc class uses a three-layer fully connected network with a Tanh activation function in each layer; The ODEBlock class uses the odeint function to solve ordinary differential equations and return the prediction results of the system at different time steps; The NeuralODE class combines the neural ordinary differential equation model, including the input layer, the ODE solver module, and the output layer; the input layer maps the input features to the hidden layer dimensions, and after ODEBlock processing, the final prediction result is obtained through the output layer.

6. The digital twin model order reduction method for an armored vehicle diesel engine according to claim 5 is characterized in that: The training process of the Neural ODE model uses the Adam optimizer and the mean squared error as the loss function, traversing the training data 100 times. In each epoch, the data is loaded in batches through the train_loader and sent to the multi-layer perceptron for forward propagation. The calculated MSE loss between the predicted results and the actual target values is back-propagated to optimize the parameters of the neural ordinary differential equation network; the predicted results and the actual results are visualized using matplotlib.

7. The digital twin model reduction method for an armored vehicle diesel engine according to claim 5, characterized in that: Key parameters include the dimension of the input features, the dimension of the hidden layer, the batch size, the learning rate, and the number of epochs for training.

8. The digital twin model reduction method for an armored vehicle diesel engine according to claim 1, characterized in that: The Kalman filter algorithm is used to train the neural ordinary differential equation network, and the Simulink simulation data is used to optimize it. The weights of the neural ordinary differential equation network are adjusted to obtain a low-order reduced-order model that reflects the dynamic behavior of the Simulink model.

9. The digital twin model order reduction method for an armored vehicle diesel engine according to any one of claims 1 to 8, characterized in that: Use Simulink to build a digital twin model of a diesel engine.

10. A digital twin model reduction system for an armored vehicle diesel engine, characterized in that: include: Building modules, using Simulink to establish a diesel engine twin model. The diesel engine twin model is a high-dimensional nonlinear system that includes the thermodynamic equations of the diesel engine. The data module runs the diesel engine twin model under different operating conditions, collects the input and output data of the variables, and standardizes the data to obtain simulation data; The order reduction module uses a neural network to build a neural network ordinary differential equation model, uses simulation data to train the neural network ordinary differential equation model to obtain a reduced-order model, and predicts the output of the diesel engine based on the obtained reduced-order model.

Images