Fault analogue simulation method for ship dual-fuel electronic injection diesel engine
By acquiring rail pressure variation data of dual-fuel electronically controlled diesel engines and performing empirical mode decomposition, the simulation data is adjusted based on similarity calculations. This solves the problem of low accuracy of simulation signals in existing technologies and enables more efficient fault diagnosis and control strategy optimization.
Patent Information
- Application Number
- CN202511878761.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-12
- Publication Date
- 2026-03-17
AI Technical Summary
Existing methods for simulating faults in marine dual-fuel electronically controlled diesel engines rely on mechanistic models or semi-empirical models, which leads to discrepancies between the generated simulation signals and the actual measurement data at the time of the fault. This results in low simulation accuracy and affects the effectiveness of fault diagnosis algorithm verification and control strategy optimization.
By acquiring rail pressure change data of a dual-fuel electronically controlled diesel engine during navigation failure, a simulation model is constructed and empirical mode decomposition is performed to obtain intrinsic mode function components. The similarity is calculated and the simulation data is adjusted to improve the simulation accuracy, ensuring that the corrected data conforms to the fault mechanism and closely matches the real fault details.
It significantly improves the effectiveness of fault diagnosis algorithm verification and control strategy optimization, reduces the risk and cost of physical equipment fault testing, and improves the accuracy of simulation signals.
Smart Images
Figure CN121683259A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of ship fault detection technology, and specifically to a fault simulation method for ship dual-fuel electronically controlled diesel engines. Background Technology
[0002] In related technologies, in order to reduce the failure risk of dual-fuel electronically controlled diesel engines in actual operation and improve maintenance efficiency, the research and development, debugging and training of dual-fuel electronically controlled diesel engines are simulated by model simulation. By simulating typical faults in a virtual environment, control strategies, evaluation and diagnostic algorithms and fault drills can be verified without damaging the equipment, thereby significantly reducing test costs and safety risks.
[0003] However, existing methods for simulating faults in marine dual-fuel electronically controlled diesel engines typically rely on mechanistic models or semi-empirical models, generating corresponding fault signals by artificially setting changes in fault parameters. While this approach can reflect fault trends to some extent, errors in model design and parameter settings lead to discrepancies between the generated signals and actual measured data during fault occurrence. Therefore, improving the simulation accuracy of the generated signals has become a pressing technical challenge. Summary of the Invention
[0004] To address the problem of low simulation accuracy in existing technologies, the present invention aims to provide a fault simulation method for marine dual-fuel electronically controlled diesel engines. The specific technical solution adopted is as follows: Acquire multiple rail pressure change data of dual-fuel electronically controlled diesel engines when a ship experiences a navigation malfunction; A simulation model of a dual-fuel electronically controlled diesel engine was constructed, and the adjustable fault parameters of each fault type in the simulation model were adjusted to perform rail pressure simulation, thus obtaining the first rail pressure simulation data. Empirical mode decomposition (EMD) is performed on each rail pressure change data to obtain multiple first eigenmode functions and eigenmode function components for each rail pressure change data. Empirical mode decomposition is also performed on the first rail pressure simulation data to obtain multiple second eigenmode function components. Based on the first and second eigenmode function components, the similarity between multiple rail pressure variation data and the first rail pressure simulation data is calculated respectively. The first rail pressure simulation data is adjusted based on the K rail pressure change data with the highest similarity to obtain the second rail pressure simulation data, where K is a positive integer.
[0005] In one possible implementation, the process of determining the similarity between any rail pressure variation data and the first rail pressure simulation data includes: Calculate the difference between each first eigenmode function component and the second eigenmode function component of any rail pressure variation data; The correlation between each adjustable fault parameter and each second intrinsic mode function component is calculated, and multiple fault severity values corresponding to the first rail pressure simulation data are calculated. The fault severity values are used to characterize the degree of adjustment when adjusting the adjustable fault parameters of each fault type in the simulation model. Based on the difference value, correlation value, and fault degree value, the similarity between any rail pressure change data and the first rail pressure simulation data is calculated.
[0006] In one possible implementation, the similarity is determined by a weighted sum of multiple first products determined during the rail pressure simulation process, each first product being a product of a fault severity value, a correlation value, and the reciprocal of a difference value.
[0007] In one possible implementation, the correlation is the product of the reciprocal of the square of the number of simulations of the adjustable fault parameter and a preset weighted summation result. The preset weighted summation result is a weighted summation of multiple first ratios in the rail pressure simulation process. The first ratio is the ratio of the difference value of the same second eigenmode function component under different simulations to the absolute value of the difference value of the same adjustable fault parameter under different simulations.
[0008] In one possible implementation, the fault severity value is positively correlated with the absolute value of a first difference, which is the difference between the actual adjusted value and the optimal adjusted value of the corresponding adjustable fault parameter, and the fault severity value is negatively correlated with the adjustable range of the corresponding adjustable fault parameter.
[0009] In one possible implementation, the method includes: Based on the correlation between each second intrinsic mode function component and the adjustable fault parameters, and the corresponding fault severity value, the correction coefficient of each second intrinsic mode function component is calculated. Based on the correction coefficient and the first eigenmode function component corresponding to each second eigenmode function component, each second eigenmode function component is corrected to obtain the third eigenmode function component corresponding to each second eigenmode function component. The third eigenmode function components corresponding to each second eigenmode function component are superimposed to obtain the second rail pressure simulation data.
[0010] In one possible implementation, the correction coefficient of the second intrinsic mode function component is positively correlated with the frequency of the second intrinsic mode function component; the correction coefficient of the second intrinsic mode function component is negatively correlated with a preset product, which is the product of the correlation between the second intrinsic mode function component and the adjustable fault parameter and the corresponding fault severity value.
[0011] In one possible implementation, the third intrinsic mode function component is proportional to the corresponding second intrinsic mode function component, proportional to the correction coefficient, and proportional to the average of multiple target values. The target value is the difference between the average amplitude of the same intrinsic mode function component in the rail pressure change data and the first rail pressure simulation data, and the ratio of the average amplitude of the same intrinsic mode function component in the first rail pressure simulation data.
[0012] In one possible implementation, the simulation model includes: a low-pressure fuel supply module, a high-pressure pump module, a common rail module, an injector module, and a controller module. The low-pressure fuel supply module is used to simulate the process of supplying fuel to the high-pressure pump from the low-pressure pump, filter element, and low-pressure pipeline. The input parameters of the low-pressure fuel supply module include: low-pressure pump speed or fuel supply pressure, and fuel temperature; the output parameters include high-pressure pump inlet pressure and inlet flow rate. The high-pressure pump module is used to simulate the process of pressurizing low-pressure fuel to the high pressure required for the common rail. The input parameters of the high-pressure pump module include: pump speed, geometric displacement, volumetric efficiency, mechanical efficiency, and drive throttle signal or pump control valve opening; the output parameters include: outlet flow rate and outlet pressure. The common rail module is used to simulate the high-pressure energy storage and distribution process of fuel. The input parameters of the common rail module include: high-pressure pump outlet flow rate, injector return flow rate, pressure regulating valve opening and common rail geometric volume; the output parameters include: common rail pressure. The fuel injector module is used to simulate receiving control signals and complete the fuel injection process. The input parameters of the fuel injector module include the pulse width and number of injections, common rail pressure, nozzle flow coefficient, and return path leakage coefficient; the output parameters include the injection quantity and return flow rate. The controller module is used to simulate the closed-loop regulation of common rail pressure and injection timing. The input parameters of the controller module include: rail pressure target, common rail pressure, engine speed, and load; the output parameters include: high-pressure pump control signal and injection pulse signal.
[0013] In one possible implementation, the intrinsic mode function components include high-frequency components, mid-frequency components, and low-frequency components. High-frequency components are used to characterize fuel injection pulsation, sensor noise, and inter-tooth pressure fluctuations. The intermediate frequency component is used to characterize the dynamic response information of the pressure regulating valve and the pump; Low-frequency components are used to characterize load changes and rail pressure slow drift information.
[0014] This invention offers the following advantages: It obtains rail pressure variation data (i.e., real data) of a dual-fuel electronically controlled diesel engine during a ship malfunction and generates first rail pressure simulation data by constructing a simulation model. Then, empirical mode decomposition (EMD) is performed on both the rail pressure variation data and the first rail pressure simulation data to obtain corresponding intrinsic mode function (IMF) components. This allows for frequency-based stratification of the rail pressure signal, enabling precise capture of fault characteristics across different frequency bands. Finally, the most relevant real data is selected by calculating the similarity between the two types of component data to correct the simulation data, ensuring that the corrected second rail pressure simulation data both conforms to the fault mechanism and closely approximates the detailed characteristics of the actual fault. Compared to traditional methods, this embodiment achieves simulation data correction through the above method, thereby increasing the simulation accuracy of the generated signal. This significantly improves the effectiveness of fault diagnosis algorithm verification and control strategy optimization, while reducing the risk and cost of physical equipment fault testing. Attached Figure Description
[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] Figure 1 This is a flowchart of a fault simulation method for a marine dual-fuel electronically controlled diesel engine, provided in one embodiment of the present invention. Figure 2 This is a structural diagram of a simulation model provided in one embodiment of the present invention; Figure 3 A flowchart of a method for determining the similarity between any rail pressure variation data and first rail pressure simulation data, provided in an embodiment of the present invention; Figure 4 This is a flowchart illustrating a method for simulating faults in a marine dual-fuel electronically controlled diesel engine, as provided in one embodiment of the present invention. Detailed Implementation
[0017] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a fault simulation method for a marine dual-fuel electronically controlled diesel engine proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.
[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0019] Dual-fuel electronically controlled diesel engines for ships are crucial power units in modern marine propulsion systems. Capable of using multiple fuels, including diesel and natural gas, they offer advantages such as high fuel economy, low emissions, and strong adaptability. During operation, the fuel injection system, fuel switching system, and combustion control system are core components, and their operational status directly impacts the overall safety and economy of the engine. To reduce the risk of failure in actual operation and improve maintenance efficiency, fault simulation technology is increasingly being introduced into the research, development, commissioning, and training phases of dual-fuel electronically controlled diesel engines. By simulating typical faults in a virtual environment, control strategies can be verified, diagnostic algorithms evaluated, and fault drills conducted without damaging the equipment, thereby significantly reducing testing costs and safety risks.
[0020] Existing methods for simulating faults in marine dual-fuel electronic fuel injection diesel engines typically rely on mechanistic or semi-empirical models, generating corresponding fault signals by artificially setting changes in fault parameters. While this approach can reflect fault trends to some extent, errors in model design and parameter settings lead to discrepancies between the generated signals and actual measured data during fault occurrence. For example, when the sealing surface of the electronic fuel injection nozzle wears, fuel leakage occurs during non-injection cycles, resulting in persistently low and abnormally fluctuating high-pressure common rail pressure. Common simplifications in simulation models include treating leakage as a constant flow rate and ignoring inter-tooth pulsation and transient injection suction. This causes the instantaneous pressure drop curve to deviate from reality (the actual pressure drop rate is faster, not simply a reduction in pressure based on normal pressure changes), and the high-frequency common rail pressure becomes overly smooth. These shortcomings result in simulation signals that "appear reasonable but lack realism," reducing the effectiveness of simulation data in training and validating fault diagnosis algorithms. Therefore, increasing the accuracy of generated signals is a pressing technical problem that needs to be addressed.
[0021] In view of this, this application provides a fault simulation method for dual-fuel electronically controlled diesel engines for ships. It acquires rail pressure change data (i.e., real data) of the dual-fuel electronically controlled diesel engine during a ship malfunction and obtains first rail pressure simulation data by constructing a simulation model. Then, empirical mode decomposition (EMD) is performed on both the rail pressure change data and the first rail pressure simulation data to obtain corresponding intrinsic mode function (IMF) components. This allows the rail pressure signal to be stratified by frequency, thereby accurately capturing fault characteristics in different frequency bands. Finally, the most relevant real data is selected by calculating the similarity between the two types of component data to correct the simulation data, ensuring that the corrected second rail pressure simulation data both conforms to the fault mechanism and closely approximates the detailed characteristics of the real fault. Compared with traditional methods, the embodiments of this application can correct the simulation data in the above manner, thereby increasing the simulation accuracy of the generated signal. This significantly improves the effectiveness of fault diagnosis algorithm verification and control strategy optimization, while reducing the risk and cost of physical equipment fault testing.
[0022] The specific scheme of the fault simulation method for marine dual-fuel electronically controlled diesel engines provided by the present invention will be described in detail below with reference to the accompanying drawings.
[0023] Please see Figure 1 Flowchart, such as Figure 1 As shown, the method includes the following steps: Step 101: Obtain multiple rail pressure change data of the dual-fuel electronically controlled diesel engine when a ship malfunctions during navigation.
[0024] In marine dual-fuel electronically controlled diesel engines, the fuel injection subsystem can suppress knocking, reduce NOx emissions, and improve combustion efficiency by adjusting the fuel injection quantity, injection timing, and injection mode. These adjustment functions rely on the rapid response and high-precision control of the fuel injection subsystem.
[0025] To make the simulation of rail pressure data more closely resemble real-world conditions, this application collects common rail pressure change data under historical fault conditions by recording fault events during actual navigation and verifying them through maintenance, and uses this data as reference data for common rail pressure under actual fault conditions.
[0026] For example, rail pressure change data refers to the historical data of common rail pressure changes over time, recorded by the ship's monitoring system and verified by maintenance personnel when a dual-fuel electronically controlled diesel engine malfunctions during actual navigation. This data can truly reflect the dynamic characteristics of rail pressure under fault conditions (such as pressure fluctuation amplitude, pressure drop rate, etc.).
[0027] Step 102: Construct a simulation model of a dual-fuel electronically controlled diesel engine, and adjust the adjustable fault parameters of each fault type in the simulation model to perform rail pressure simulation and obtain the first rail pressure simulation data.
[0028] Adjustable fault parameters refer to the parameters used to simulate various fault types in dual-fuel electronically controlled diesel engines. The first rail pressure simulation data is the uncorrected rail pressure simulation data output by the simulation model after adjusting the adjustable fault parameters. In the simulation model, the rail pressure curve of the fuel injection subsystem is often used as a core monitoring indicator, which can not only detect abnormalities in the fuel injection subsystem, but also help to judge the health status of other components of the fuel system.
[0029] In the simulation model, abnormalities in fuel supply flow, fuel consumption flow, system compressibility, and control strategy can all lead to abnormal common rail pressure in the diesel engine injection system. Furthermore, during fault simulation, this simulation model is based on the original simulation model, with adjustments made to some model parameters to simulate fault signals. For example, the following are some parameter simulation adjustment methods: High-pressure pump maximum flow rate: The oil pump's oil supply capacity at rated speed can be reduced to 70-90% of normal conditions to simulate pump wear. Lower pump flow rate will cause changes in the output common rail pressure, resulting in a decrease in steady-state rail pressure and increased under-injection during fuel injection. Nozzle internal leakage channel diameter: As the fuel injector ages, the corresponding fuel injector diameter will increase, resulting in the rail pressure being different from the normal data during subsequent fuel injection, the steady-state rail pressure will decrease, and the rail pressure fluctuation will increase. Mechanical efficiency and volumetric efficiency of the oil pump: When the oil pump ages and its efficiency decreases, the rail pressure will also decrease as a result.
[0030] Step 103: Perform empirical mode decomposition on each rail pressure change data to obtain multiple first eigenmode function components of each rail pressure change data, and perform empirical mode decomposition on the first rail pressure simulation data to obtain multiple second eigenmode function components.
[0031] It should be noted that during the fault simulation process, some adjustable fault parameters change with the actual operation of the diesel engine. For example, the volumetric efficiency of the high-pressure pump changes with the pressure difference and temperature. In traditional fault simulation, in order to reproduce a certain fault (such as pump wear), the volumetric efficiency is directly set to a fixed value, such as changing it from 0.95 to 0.85, and it is impossible to adjust the actual high-pressure pump volumetric efficiency and other parameters in real time.
[0032] Therefore, there is still a certain difference between the rail pressure data generated under a certain fault condition and the real data. In order to further analyze the difference between the generated simulated rail pressure data and the real data, this application can decompose the two types of data by means of empirical mode decomposition, so as to improve the component with poor simulation effect of the simulation data in the future, thereby improving the simulation effect.
[0033] In some embodiments, empirical mode decomposition can decompose the rail pressure signal in the rail pressure variation data (or the first rail pressure simulation data) into several intrinsic mode function components, with different intrinsic mode function components corresponding to different frequency components. After decomposing the rail pressure variation data (or the first rail pressure simulation data), the frequency ranges of its different intrinsic mode function components are different, and the information data they contain also have certain differences.
[0034] In some embodiments, the intrinsic mode function components include high-frequency components, mid-frequency components, and low-frequency components; the high-frequency components are used to characterize fuel injection pulsation, sensor noise, and inter-tooth pressure fluctuation information; the mid-frequency components are used to characterize pressure regulating valve adjustment and pump dynamic response information; and the low-frequency components are used to characterize load changes and rail pressure slow drift information.
[0035] For example, taking the intrinsic mode function components as divided into IMF1~IMF2 (high frequency part), IMF3~IMF4 (medium frequency part), and MF5 and above (low frequency part) according to frequency as an example, the high frequency part of IMF1~IMF2 often includes information on fuel injection pulsation, sensor noise and inter-tooth pressure fluctuation; while the medium frequency part of IMF3~IMF4 includes pressure regulating valve adjustment and pump dynamic response; the low frequency part of MF5 and above includes information on load changes, rail pressure slow drift, etc.
[0036] Step 104: Based on the first and second intrinsic mode function components, calculate the similarity between multiple rail pressure variation data and the first rail pressure simulation data.
[0037] The similarity is used to measure the degree of similarity between a certain rail pressure change data (i.e., real data) and the first rail pressure simulation data (i.e., simulation data before correction). The higher the similarity, the closer the real data is to the fault state corresponding to the initial simulation data, and the more suitable it is for correction of subsequent simulation data.
[0038] It should be noted that, since diesel engine monitoring systems primarily measure direct parameters such as rail pressure, engine speed, fuel injection quantity, and exhaust temperature, many parameters are not directly measured using sensors and can only be indirectly estimated. Therefore, it is impossible to directly determine whether the simulated fault and the real fault are the same based on the degree of fault. Therefore, this application can divide the two types of data into multiple components based on empirical mode decomposition. Since different intrinsic mode function components correspond to different frequency components, and different frequency components are often affected by different equipment anomaly information, this application can evaluate the similarity between multiple rail pressure change data and the first rail pressure simulation data based on the comparison of differences between the components.
[0039] Step 105: Adjust the first rail pressure simulation data based on the K rail pressure change data with the highest similarity to obtain the second rail pressure simulation data.
[0040] Where K is a positive integer. The second rail pressure simulation data refers to the rail pressure simulation data that is closer to the real fault characteristics after being corrected by the top K real rail pressure data with the highest similarity; the value of K can be set according to the quantity and quality of real data. For example, K is 10, that is, the 10 rail pressure change data with the highest similarity to the first rail pressure simulation data are selected for correction.
[0041] As can be seen from step 104 above, the similarity between the rail pressure change data calculated by the intrinsic mode function components and the first rail pressure simulation data can characterize the fault state corresponding to the real data and the initial simulation data. Therefore, the top K rail pressure change data with the highest similarity not only maintain the original mechanism consistency in the overall trend, but also are closer to the real fault data in high-frequency dynamic characteristics. Therefore, by adjusting the first rail pressure simulation data based on the top K rail pressure change data with the highest similarity, this application can effectively make up for the shortcomings of traditional simulation models in transient fitting and noise characteristics, thereby improving the simulation effect.
[0042] Based on the above technical solution, this application can obtain rail pressure change data (i.e., real data) of a dual-fuel electronically injected diesel engine when a ship experiences a navigation malfunction, and obtain first rail pressure simulation data by constructing a simulation model. Then, empirical mode decomposition is performed on both the rail pressure change data and the first rail pressure simulation data to obtain the corresponding intrinsic mode function components. This allows the rail pressure signal to be layered by frequency, thereby accurately capturing fault characteristics in different frequency bands. Finally, the most relevant real data is selected by calculating the similarity between the two types of component data to correct the simulation data, ensuring that the corrected second rail pressure simulation data both conforms to the fault mechanism and closely approximates the detailed characteristics of the real fault. Compared with traditional methods, the embodiments of this application can correct the simulation data in the above manner, thereby increasing the simulation accuracy of the generated signal. Therefore, it significantly improves the effectiveness of fault diagnosis algorithm verification and control strategy optimization, while reducing the risk and cost of physical equipment fault testing.
[0043] As one possible embodiment of this application, such as Figure 2 As shown, the simulation model constructed in this embodiment includes: a low-pressure fuel supply module, a high-pressure pump module, a common rail module, an injector module, and a controller module.
[0044] The low-pressure fuel supply module is used to simulate the process of supplying fuel to the high-pressure pump from the low-pressure pump, filter element, and low-pressure pipeline. The input parameters of the low-pressure fuel supply module include: low-pressure pump speed or fuel supply pressure, and fuel temperature; the output parameters include high-pressure pump inlet pressure and inlet flow rate.
[0045] The high-pressure pump module is used to simulate the process of pressurizing low-pressure fuel to the high pressure required for the common rail. The input parameters of the high-pressure pump module include: pump speed, geometric displacement, volumetric efficiency, mechanical efficiency, and drive throttle signal or pump control valve opening; the output parameters include: outlet flow rate and outlet pressure.
[0046] The common rail module is used to simulate the high-pressure energy storage and distribution process of fuel. The input parameters of the common rail module include: high-pressure pump outlet flow rate, injector return flow rate, pressure regulating valve opening and common rail geometric volume; the output parameter includes: common rail pressure.
[0047] The injector module is used to simulate receiving control signals and complete the fuel injection process. The input parameters of the injector module include the pulse width and number of injections, common rail pressure, nozzle flow coefficient, and return path leakage coefficient; the output parameters include the injection quantity and return flow rate.
[0048] The controller module is used to simulate the closed-loop regulation of common rail pressure and injection timing. The input parameters of the controller module include: rail pressure target, common rail pressure, engine speed, and load; the output parameters include: high-pressure pump control signal and injection pulse signal.
[0049] In some embodiments, the simulation model may further include an exhaust / combustion feedback module, which is used to monitor the combustion process status and exhaust parameters in the simulation model of a marine dual-fuel electronically controlled diesel engine in real time; the input parameters of the exhaust / combustion feedback module include combustion chamber temperature, combustion chamber pressure, exhaust temperature, and exhaust pressure; the output parameters of the exhaust / combustion feedback module include combustion state feedback signal and exhaust parameter feedback signal.
[0050] Based on the above model modules, the simulation model can simulate the normal operation and various fault states of the fuel injection subsystem (such as high-pressure pump wear, nozzle internal leakage, etc.), providing basic model support for subsequent fault simulation.
[0051] As one possible embodiment of this application, such as Figure 3 As shown, the process of determining the similarity between any of the above rail pressure variation data and the first rail pressure simulation data includes the following steps: Step 301: Calculate the difference between each first intrinsic mode function component and the second intrinsic mode function component of any rail pressure change data.
[0052] The difference value is used to measure the waveform difference between the first intrinsic mode function component and the second intrinsic mode function component. For example, this application can use dynamic time warping (DTW) distance as the calculation index for the difference value, which can be expressed as follows: ,in, Indicates the first The first rail pressure simulation data under the second simulation. One second eigenmode function component, Indicates the first The first data point of the rail pressure change. For each first intrinsic mode function component, the smaller the DTW distance, the more similar the waveforms of the two intrinsic mode function components are, i.e., the smaller the difference value.
[0053] Step 302: Calculate the correlation between each adjustable fault parameter and each second intrinsic mode function component, and calculate multiple fault severity values corresponding to the first rail pressure simulation data.
[0054] Among them, the fault severity value is used to characterize the degree of adjustment when adjusting the adjustable fault parameters of each fault type in the simulation model.
[0055] It should be noted that correlation is used to measure the degree of influence of a certain adjustable fault parameter on the second intrinsic mode function component. The correlation between different second intrinsic mode function components and each adjustable fault parameter varies. If the correlation between a certain adjustable fault parameter and a certain second intrinsic mode function component is strong, the corresponding component will change more under different fault severity values. That is to say, the higher the correlation, the more significant the impact of the change of the adjustable fault parameter on the dynamic characteristics of the second intrinsic mode function component.
[0056] In some embodiments, the correlation is negatively correlated with the number of simulations of the adjustable fault parameter and the absolute value of the difference between the same adjustable fault parameter in different simulations; the correlation is positively correlated with the magnitude of the difference between the same second eigenmode function component in different simulations. Optionally, the correlation is the product of the reciprocal of the square of the number of simulations of the adjustable fault parameter and a preset weighted summation result, where the preset weighted summation result is a weighted summation of multiple first ratios in the rail pressure simulation process, and the first ratio is the ratio of the absolute value of the difference between the same second eigenmode function component in different simulations and the absolute value of the difference between the same adjustable fault parameter in different simulations.
[0057] For example, the correlation between the adjustable fault parameters and each second intrinsic mode function component satisfies the following formula: , in, Indicates the first Adjustable fault parameters and the first The correlation between the components of the second eigenmode function Indicates the first The number of simulations for adjustable fault parameters; Indicates the first The first rail pressure simulation data under the second simulation. One second eigenmode function component, Indicates the first The first rail pressure simulation data under the second simulation. One second eigenmode function component; This indicates the difference between two sequences, and the DTW distance can be selected. Indicates the first The simulation in the second year The fault severity value of the adjustable fault parameters. Indicates the first The simulation in the second year The fault severity value of the adjustable fault parameters. It is a constant greater than 0, and in the embodiments of the present invention, it can be taken as 0.01.
[0058] Thus, by observing the differences in the components under a unit change in fault severity, the influence of each adjustable fault parameter on the rail pressure data can be determined. When the differences between the second intrinsic mode function components are large, it indicates that the second intrinsic mode function component is greatly affected by the corresponding adjustable fault parameters.
[0059] In some other embodiments, the fault severity value is positively correlated with the absolute value of a first difference, which is the difference between the actual adjusted value and the optimal adjusted value of the corresponding adjustable fault parameter, and the fault severity value is negatively correlated with the adjustable range of the corresponding adjustable fault parameter.
[0060] For example, the fault severity value satisfies the following formula: , in, Indicates the first The fault severity value of the adjustable fault parameters. Indicating the simulation time The values of the control parameters; Indicates the first The optimal values for adjustable fault parameters (i.e., the values when the equipment is operating normally), such as the values when the maximum flow rate of the high-pressure pump is operating normally. Indicates the first The fluctuation range of adjustable fault parameters (for example, it can be determined by the upper and lower limits of the adjustable fault parameters within a preset time when the equipment is operating normally).
[0061] Step 303: Based on the difference value, correlation value, and fault degree value, calculate the similarity between any rail pressure change data and the first rail pressure simulation data.
[0062] In some embodiments, similarity is positively correlated with correlation and fault severity values; similarity is negatively correlated with difference values. Optionally, similarity is determined by a weighted sum of multiple first products determined during the rail pressure simulation process, each first product being the product of a fault severity value, a correlation value, and the reciprocal of a difference value.
[0063] For example, the similarity between the rail pressure variation data and the first rail pressure simulation data satisfies the following formula: , in, Indicates the first The rail pressure change data and the first Similarity of the first rail pressure simulation data under the second simulation. Indicates the number of components; Indicates the number of adjustable fault parameters; Indicates the i-th adjustable fault parameter and the i-th Correlation between the components of the second intrinsic mode function; Indicates the first The simulation in the second year The fault severity value of the adjustable fault parameters; Indicates the first The first rail pressure simulation data under the second simulation. One second eigenmode function component, Then it means the first The first data point of the rail pressure change. One first eigenmode function component; This indicates calculating the difference between two sequences. It is a constant greater than 0.
[0064] Based on the above technical solution, this application measures the differences between intrinsic mode function components and quantifies the impact of adjustable fault parameters on intrinsic mode function components at different frequencies, enabling similarity calculations to reflect the correlation of fault mechanisms. Simultaneously, by introducing a fault severity value, it ensures that the similarity matches the severity of the fault. Thus, the similarity calculated in this way can accurately select the real data that most closely matches the fault state of the initial simulation data, providing a reliable basis for correcting subsequent simulation data and further improving the realism of the final simulation signal.
[0065] As one possible embodiment of this application, combined with Figure 1 ,like Figure 4 As shown, step 105 above can be achieved through the following steps: Step 401: Based on the correlation between each second intrinsic mode function component and the adjustable fault parameters, and the corresponding fault severity value, calculate the correction coefficient for each second intrinsic mode function component.
[0066] It should be noted that after obtaining the K rail pressure change data with the highest similarity, this application can correct the signal components of each analog signal. Since the degree of influence of each second intrinsic mode function component is different under the adjustable fault parameters corresponding to each analog signal, the similarity judgment is mainly based on the difference between the second intrinsic mode function components under similar faults. However, for the components with a lower degree of fault influence, the difference between them and the real data is the reason for the large difference between the analog data and the real data.
[0067] In other words, for components that are less affected by faults but have significant differences between the two, the differences mainly come from "simulation defects". For example, the low-frequency component heavily bears the mechanism dynamics of the pump-valve-rail volume and control strategy. In each simulation, if the fault parameters mainly affect the high-frequency component, and the simulation effect of the high-frequency component is good, the low-frequency component cannot be modified or adjusted too much. Using real rail pressure data to modify the simulated low frequency will destroy the consistency of the mechanism.
[0068] For example, high frequencies include fuel injection transients, pipeline acoustics, sensor chains, and other textures and color noise. This type of data is easy to simplify during simulation and can be corrected using real data. It is less likely to violate energy conservation and control logic, and the simulation effect of real data will be more realistic.
[0069] In some embodiments, the correction coefficient of the second intrinsic mode function component is positively correlated with the frequency of the second intrinsic mode function component; the correction coefficient of the second intrinsic mode function component is negatively correlated with a preset product, which is the product of the correlation between the second intrinsic mode function component and the adjustable fault parameter and the corresponding fault severity value.
[0070] For example, the correction coefficients for the second eigenmode function components satisfy the following formula: , Wherein, it represents the first The first rail pressure simulation data under the second simulation. Correction coefficients for each of the second eigenmode function components Represents the normalization function; This represents the frequency of the second eigenmode function component, which varies with the component index of the second eigenmode function component. As the value increases, the frequency of the corresponding component decreases, and the degree of correction should be reduced accordingly. Indicates the number of adjustable fault parameters; Indicates the first Adjustable fault parameters and the first Correlation between the components of the second intrinsic mode function; Indicates the first The simulation in the second year The fault severity value of the adjustable fault parameters.
[0071] Step 402: Based on the correction coefficient and the first eigenmode function component corresponding to each second eigenmode function component, correct each second eigenmode function component to obtain the third eigenmode function component corresponding to each second eigenmode function component.
[0072] It should be noted that for each second intrinsic mode function component, this application can compare the amplitude difference between the simulation data and the corresponding component in the historical real fault data. For second intrinsic mode function components with significant amplitude deviation, the simulation accuracy of this second intrinsic mode function component is insufficient in this frequency band. Therefore, the correction coefficients mentioned above, as well as the differences between each second intrinsic mode function component and the first IMF classification, can be combined to correct each second intrinsic mode function component.
[0073] In some embodiments, the third intrinsic mode function component is proportional to the corresponding second intrinsic mode function component, proportional to the correction coefficient, and proportional to the average of multiple target values. The target value is the difference between the average amplitude of the same intrinsic mode function component in the rail pressure change data and the first rail pressure simulation data, and the ratio of the average amplitude of the same intrinsic mode function component in the first rail pressure simulation data.
[0074] For example, the third eigenmode function component can satisfy the following formula: , in, Indicates the first The first rail pressure simulation data under the second simulation. The third eigenmode function component is obtained by correcting the second eigenmode function component. Indicates the first The first rail pressure simulation data under the second simulation. One second eigenmode function component; Indicates the first The first rail pressure simulation data under the second simulation. Correction coefficients for each second eigenmode function component; This represents the mean function; This indicates the number of rail pressure change data points with the highest similarity obtained; Indicates the first The first data point of the rail pressure change. The average magnitude of the first eigenmode function components; Indicates the first The first rail pressure simulation data under the second simulation. The average magnitude of each second eigenmode function component It is a constant greater than 0.
[0075] In this way, the present application can indirectly correct the amplitude based on the actual signal performance, rather than directly replacing the amplitude, thus ensuring the continuity and physical rationality of the signal waveform.
[0076] Step 403: Superimpose the third eigenmode function components corresponding to each second eigenmode function component to obtain the second rail pressure simulation data.
[0077] For example, after correcting the amplitude of each second intrinsic mode function component, this application can superimpose the corrected third intrinsic mode function components according to the reconstruction formula of empirical mode decomposition to obtain improved second rail pressure simulation data. This second rail pressure simulation data not only maintains the original mechanism consistency in the overall trend, but also more closely resembles real fault data in high-frequency dynamic characteristics, thereby effectively making up for the shortcomings of traditional simulation models in transient fitting and noise characteristics.
[0078] This simulation improvement method based on component amplitude correction and signal reconstruction can achieve higher fidelity reproduction of typical fault rail pressure changes in the fault simulation of marine dual-fuel electronically controlled diesel engines. It provides more realistic and reliable signal input for subsequent fault diagnosis algorithm verification, control strategy optimization and fault prediction, and significantly improves the practical value of the simulation system in engineering applications.
[0079] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0080] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.
Claims
1. A method for fault simulation modeling of a dual-fuel electrically injected diesel engine of a marine vessel, characterized in that, The method comprises: acquiring multiple rail pressure change data of a dual-fuel electric injection diesel engine when a ship navigation failure occurs; constructing a simulation model of the dual-fuel electric injection diesel engine, and adjusting adjustable fault parameters of each fault type in the simulation model to perform rail pressure simulation to obtain first rail pressure simulation data; performing empirical mode decomposition on each rail pressure change data to obtain multiple first intrinsic mode function components of each rail pressure change data, and performing empirical mode decomposition on the first rail pressure simulation data to obtain multiple second intrinsic mode function components; based on the first intrinsic mode function components and the second intrinsic mode function components, calculating the similarity of the multiple rail pressure change data and the first rail pressure simulation data respectively; based on the top K rail pressure change data with the highest similarity, adjusting the first rail pressure simulation data to obtain second rail pressure simulation data, K being a positive integer.
2. The method for simulating the malfunction of a marine dual-fuel electric injection diesel engine according to claim 1, characterized in that, The determination process of the similarity of any rail pressure change data and the first rail pressure simulation data comprises: calculating the difference value between each first intrinsic mode function component of the any rail pressure change data and the second intrinsic mode function component; calculating the correlation between each adjustable fault parameter and each second intrinsic mode function component, and calculating multiple fault degree values corresponding to the first rail pressure simulation data, the fault degree values being used to represent the degree of adjustment of the adjustable fault parameters of each fault type in the simulation model; based on the difference value, the correlation and the fault degree values, calculating the similarity of the any rail pressure change data and the first rail pressure simulation data.
3. The method for simulating the malfunction of a marine dual-fuel electric injection diesel engine according to claim 2, characterized in that, The similarity is determined by multiple first product weight sums in the rail pressure simulation process, each first product being the product of one fault degree value, one correlation and the reciprocal of one difference value.
4. The method for simulating the malfunction of a marine dual-fuel electric injection diesel engine according to claim 2, characterized in that, The correlation is the product of the reciprocal of the square of the number of simulations of the adjustable fault parameter and a preset weight sum result, the preset weight sum result being the weighted sum of multiple first ratios in the rail pressure simulation process, the first ratio being the ratio of the difference value of one second intrinsic mode function component under different simulations to the absolute value of the difference of one adjustable fault parameter under different simulations.
5. The method for simulating the malfunction of a marine dual-fuel electric injection diesel engine according to claim 2, wherein The fault degree value is positively correlated with the absolute value of a first difference value, the first difference value being the difference between the actual adjustment value and the optimal adjustment value of the corresponding adjustable fault parameter, and the fault degree value is negatively correlated with the adjustable range of the corresponding adjustable fault parameter.
6. The method for simulating the malfunction of a marine dual-fuel electric injection diesel engine according to claim 2, wherein The adjustment of the first rail pressure simulation data based on the top K rail pressure change data with the highest similarity to obtain the second rail pressure simulation data comprises: based on the correlation between each second intrinsic mode function component and the adjustable fault parameter and the corresponding fault degree value, calculating a correction coefficient of each second intrinsic mode function component; based on the correction coefficient and the first intrinsic mode function component corresponding to each second intrinsic mode function component, correcting each second intrinsic mode function component to obtain a third intrinsic mode function component corresponding to each second intrinsic mode function component. Superimpose the third eigenmode function component corresponding to each second eigenmode function component to obtain second rail pressure simulation data.
7. The method for simulating the malfunction of a marine dual-fuel electric injection diesel engine according to claim 6, characterized in that, The correction coefficient of the second eigenmode function component is positively correlated with the frequency of the second eigenmode function component; and the correction coefficient of the second eigenmode function component is positively correlated with a preset product, which is the product of the correlation between the second eigenmode function component and the adjustable fault parameter and the corresponding fault degree value.
8. The method for simulating the malfunction of a marine dual-fuel electric injection diesel engine according to claim 6, characterized in that, The third eigenmode function component is proportional to the corresponding second eigenmode function component, the correction coefficient and the average value of a plurality of target values, the target value being the difference between the rail pressure change data and the average amplitude of the same eigenmode function component in the first rail pressure simulation data, and the ratio of the average amplitude of the same eigenmode function component in the first rail pressure simulation data.
9. A method of simulating faults in a marine dual-fuel electrically fuelled diesel engine according to any one of claims 1-8, characterized in that, The simulation model comprises a low-pressure oil supply module, a high-pressure pump module, a common rail module, an injector module, and a controller module. The low-pressure oil supply module is configured to simulate the process of supplying oil to the high-pressure pump by the low-pressure pump, filter element and low-pressure pipeline, and the input parameters of the low-pressure oil supply module include the low-pressure pump speed or oil supply pressure and fuel temperature, and the output parameters include the high-pressure pump inlet pressure and inlet flow rate. The high-pressure pump module is configured to simulate the process of pressurizing the low-pressure fuel to the high pressure required by the common rail, and the input parameters of the high-pressure pump module include the pump speed, geometric displacement, volumetric efficiency, mechanical efficiency, and drive throttle signal or pump control valve opening, and the output parameters include the outlet flow rate and outlet pressure. The common rail module is configured to simulate the high-pressure energy storage and distribution process of fuel, and the input parameters of the common rail module include the high-pressure pump outlet flow rate, injector backflow flow rate, pressure regulating valve opening and common rail geometric volume, and the output parameter includes the common rail pressure. The injector module is configured to simulate the process of receiving a control signal and completing fuel injection, and the input parameters of the injector module include the pulse width and frequency of the control injection, common rail pressure, nozzle flow coefficient and oil return path leakage coefficient, and the output parameters include the injection amount and oil return flow rate. The controller module is configured to simulate the process of closed-loop regulating the common rail pressure and injection timing, and the input parameters of the controller module include the rail pressure target, common rail pressure and engine speed, and load, and the output parameters include the high-pressure pump control signal and injection pulse signal.
10. A method of simulating a fault in a marine dual-fuel electrically fuelled diesel engine according to any one of claims 1-8, characterized in that, The eigenmode function component includes a high-frequency component, a medium-frequency component and a low-frequency component. The high-frequency component is used to represent the information of injection pulsation, sensor noise and inter-tooth pressure fluctuation. The medium-frequency component is used to represent the information of pressure regulating valve adjustment and pump dynamic response. The low-frequency component is used to represent the information of load change and rail pressure slow drift.