A high-frequency response model predictive control method for dynamic systems

Abstract

This invention discloses a high-frequency response model predictive control method for a powertrain system, relating to the field of powertrain system control technology. The method includes real-time acquisition of various powertrain system data, establishment of a powertrain system model, real-time prediction of driving range based on the powertrain system model and display of driving range data, setting a driving range warning value for a kinetic energy alarm, comparing the warning value with the predicted driving range, issuing an alarm when the predicted driving range is lower than the warning value, issuing a prompt for refueling to supplement kinetic energy when the alarm is issued, having the driver confirm the refueling method, refueling the powertrain, predicting the health of the powertrain, and generating a health prediction report. This invention utilizes a high-frequency response model predictive control method for powertrain systems, supports dual modes of total driving range percentage threshold and user-defined quota threshold, distinguishes between independent warnings for pure electric and fuel-powered driving, meets the needs of different driving scenarios, provides more timely warning response, and supports automatic triggering of the refueling mode when the electric driving range is insufficient.

Description

Technical Field

[0001] This invention relates to the field of power system control technology, and in particular to a high-frequency response model predictive control method for power systems. Background Technology

[0002] New energy vehicles refer to new technologies and new structures of automobiles that use unconventional vehicle fuels as a power source or use new on-board power devices and integrate advanced power control and drive technologies. They are the core direction of the current low-carbon transformation of the automotive industry.

[0003] Series hybrid is a special technical route of hybrid power, commonly known as range-extended electric vehicle. The engine is only used as a generator and does not directly drive the wheels. The wheels are driven entirely by electric motors. The working process is as follows: the engine drives the generator to generate electricity, part of which is directly supplied to the drive motor, and the other part is used to charge the battery.

[0004] However, the fuel system of series hybrid vehicles only estimates the remaining energy based on the fluid level, without combining it with the engine's dynamic efficiency curve for correction. It only provides a single range value and does not distinguish between pure electric, fuel, and combined range types. It does not make accurate predictions based on the gradient and congestion of the navigation route, and there are no confidence level prompts. Users lack a basis for judging the reliability of the prediction results, thus having certain defects. Summary of the Invention

[0005] The purpose of this invention is to provide a high-frequency response model predictive control method for power systems to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a high-frequency response model predictive control method for a power system, comprising the following steps: Step 1: Collect various data from the power system in real time and establish a power system model; Step 2: Predict the driving range in real time based on the powertrain model and display the driving range data; Step 3: Set the kinetic energy alarm range warning value, compare the range warning value with the predicted range, and issue a reminder alarm when the predicted range is lower than the range warning value, and remind personnel to recharge. Step 4: When the alarm is issued, a prompt will be sent to replenish the power system with fuel. The driver should confirm the power replenishment method and confirm that the power system is being recharged. Step 5: Based on the real-time collected power system data, predict the health status of the power system and generate a health prediction report.

[0007] Preferably, the real-time acquisition of various data of the power system in step one is achieved by real-time acquisition and transmission of data by sensors arranged in the power system. The sensors arranged in the power system include, but are not limited to, one or more of the following: intake sensor, temperature sensor, speed sensor, current sensor, voltage sensor, pressure sensor, and wheel speed sensor.

[0008] Preferably, the step one of establishing the dynamic system model includes the following steps: S1.1: Process the collected power system data, including outlier removal, noise filtering and normalization, and extract features from the processed data; S1.2: A modeling framework that integrates mechanism and data-driven approaches is adopted, and the power source module, transmission execution module, and load characteristic module are modeled in layers according to the power system structure. S1.3: The model parameters are calibrated and validated using a multi-dimensional validation method, which includes static calibration, dynamic calibration, and robustness testing. S1.4: Set the model periodic update threshold so that after the system runs through the periodic update threshold, the model is incrementally updated. The actual running data collected is compared with the model prediction results. When the deviation exceeds the warning threshold, the model parameters are automatically adjusted to dynamically adapt to the characteristics of performance degradation and component aging of the model and power system.

[0009] Preferably, the real-time prediction of driving range in step two includes the following steps: S2.1: Calculate the remaining available energy in real time based on the data transmitted by the sensors in the power system; S2.2: Calculate the energy consumption per unit mileage under the current operating conditions using a fusion algorithm of mechanism model and data-driven approach; S2.3: Basic range calculation, route-related range prediction, and confidence level display provide multi-scenario calculation and confidence assessment of range; S2.4: Based on the update frequency, repeat steps S2.1 to S2.3 to dynamically update and visualize the battery life data.

[0010] Preferably, in step three, the kinetic energy alarm range warning value includes a pure electric range warning value and a fuel range warning value. The pure electric range warning value and the fuel range warning value are one of the values ​​of 10% of the total pure electric range and the total fuel range, and the fixed range value. The predicted range is the pure electric range and fuel range values ​​predicted in real time in step two. The predicted pure electric range value is compared with the pure electric range warning value. When the predicted pure electric range value is less than or equal to the pure electric range warning value, a pure electric range reminder alarm is issued. Similarly, the predicted fuel range value is compared with the fuel range warning value. When the predicted fuel range value is less than or equal to the fuel range warning value, a fuel range reminder alarm is issued.

[0011] Preferably, in step four, when the fuel-powered battery replenishment prompt is issued as an electric range reminder alarm, if the user confirms the fuel supply within 5 seconds or does not confirm within 5 seconds, the fuel-powered battery replenishment operation is executed. If the user cancels the fuel supply operation within 5 seconds, the fuel-powered battery replenishment operation is not executed. When the pure electric range value drops by 2%, the electric range reminder alarm step is repeated. When a fuel range reminder alarm is issued, a fuel alarm status is displayed on the instrument panel, and the fuel alarm display cannot be manually cleared. After fuel is replenished, the fuel alarm display is automatically cleared.

[0012] Preferably, step four, which involves replenishing the power system, also includes the automated generation of a charging scheme, comprising the following steps: S4.1: When the power system is guiding, the remaining mileage is calculated in real time, and the charging stations and gas stations along the route are counted online; S4.2: Calculate the comprehensive predicted range based on the sum of the real-time predicted pure electric range and fuel range values ​​in step two, and set the comprehensive range alarm threshold based on the comprehensive predicted range. S4.3: When the sum of the pure electric range and fuel range of the power system is less than or equal to the comprehensive range warning threshold, the remaining mileage in the navigation is compared with the comprehensive range warning threshold. When the remaining mileage in the navigation is less than the comprehensive range warning threshold, charging stations and gas stations are not recommended. S4.4: When the remaining mileage in the navigation is greater than or equal to the comprehensive range warning threshold, the system will automatically search for charging stations and gas stations along the navigation route and recommend them to the user. When the user selects a specific charging station or gas station, the system will automatically add that charging station or gas station as a waypoint.

[0013] Preferably, the prediction of the power system health in step five includes the following steps: S5.1: Based on the data of the power system collected in step one, extract the time-domain statistical features, performance degradation features and frequency-domain features; S5.2: An evaluation framework combining offline training and online updates is adopted to assess the health of the power system, and the health of the power system is divided into four levels based on the evaluation results: healthy state, sub-healthy state, deterioration state, and fault warning state.

[0014] Preferably, the prediction of the power system health in step five further includes the following steps: S5.3: When the system identifies degradation and fault warning states, it automatically performs fault location, associates abnormal features with the fault mode library based on the dynamic Bayesian network model, matches faulty components, and outputs fault impact analysis.

[0015] Preferably, the output of the health prediction report in step five includes the current health level, health status score of each component, remaining service life prediction, abnormal feature list and fault risk warning, and automatically generates maintenance suggestions, providing maintenance plans with different priorities according to the health level. The report supports local storage and cloud synchronization and can be exported in a standardized format for maintenance personnel and management platform to retrieve.

[0016] The technical effects and advantages of this invention are as follows: This invention utilizes a high-frequency response model predictive control method for power systems, supporting dual modes of total range percentage threshold and user-defined fixed threshold. It distinguishes between independent warnings for pure electric and fuel-powered driving, meeting the needs of different driving scenarios and providing more timely warning responses. When the electric range is insufficient, it supports automatic triggering of the fuel-powered generator mode and repeatedly reminds the user after the pure electric range drops by 2%. The fuel warning is forcibly displayed until refueling is completed, preventing user oversight. Combining the remaining navigation mileage and the comprehensive range threshold, it automatically searches for charging stations / gas stations along the route and recommends the optimal option. After the user selects an option, it is automatically added as a waypoint, realizing full automation of the refueling planning process and reducing the user's operational burden. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is the logic diagram for determining the supplementary kinetic energy of oil-fired power generation in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] This invention provides, for example Figures 1-2 The high-frequency response model predictive control method for a dynamic system, as shown, includes the following steps: Step 1: Collect various data from the power system in real time and establish a power system model; Step 2: Predict the driving range in real time based on the powertrain model and display the driving range data; Step 3: Set the kinetic energy alarm range warning value, compare the range warning value with the predicted range, and issue a reminder alarm when the predicted range is lower than the range warning value, and remind personnel to recharge. Step 4: When the alarm is issued, a prompt will be sent to replenish the power system with fuel. The driver should confirm the power replenishment method and confirm that the power system is being recharged. Step 5: Based on the real-time collected power system data, predict the health status of the power system and generate a health prediction report.

[0020] Specifically, in step one, the real-time acquisition of various data of the power system is achieved by using sensors deployed in the power system to collect and transmit data in real time. These sensors include, but are not limited to, one or more types of sensors such as intake sensors, temperature sensors, speed sensors, current sensors, voltage sensors, pressure sensors, and wheel speed sensors. These sensors can collect and transmit various data of the power system in real time to ensure the comprehensiveness of the data.

[0021] Furthermore, the establishment of the dynamic system model in step one includes the following steps: S1.1: Process the collected power system data. The data collection method is not limited to the sensors deployed in the power system in step one. It can also be based on historical data in the power system database. This includes outlier removal, noise filtering, and normalization. Outlier removal, noise filtering, and normalization are all routine data processing operations and will not be described in detail here. Then, feature extraction is performed on the processed data to facilitate the establishment of the subsequent power model based on the extracted features. S1.2: A modeling framework that integrates mechanism and data-driven approaches is adopted. The power source module, transmission execution module, and load characteristic module are modeled in layers according to the power system structure. The power source module is based on thermodynamic or electrochemical mechanisms to build input and output characteristic models of internal combustion engines, fuel cells, or power batteries, and fits the mapping relationship between power output, energy consumption, speed, temperature, and load. The transmission execution module establishes the transmission efficiency model of gearbox, drive shaft, and drive motor, and corrects the transmission loss coefficient by combining torque and speed sensor data. The load characteristic module is based on wheel speed and pressure sensor data to build a real-time calculation model of driving resistance and dynamically match the actual operating conditions. S1.3: A multi-dimensional verification method is used to calibrate and verify the model parameters. This method includes static calibration, dynamic calibration, and robustness testing. Static calibration involves comparing the model output with measured data under bench testing conditions and iteratively correcting the model parameters using a particle swarm optimization algorithm to keep the steady-state error within 3%. Dynamic verification simulates typical dynamic conditions such as acceleration, deceleration, and hill climbing, and uses similarity analysis to verify the consistency between the model output response and the actual system. The dynamic response delay is less than 50ms to meet the requirements of high-frequency response control. Robustness testing introduces disturbance conditions such as temperature fluctuations from -40℃ to 85℃ and a 10% power supply voltage deviation to verify the output stability of the model under extreme conditions, ensuring that the prediction deviation does not exceed 5%. S1.4: Set the model periodic update threshold, for example, perform incremental updates to the model every 100 running cycles. After the system runs through the periodic update threshold, perform incremental updates to the model, compare the collected actual running data with the model prediction results, and automatically trigger the adjustment of model parameters when the deviation exceeds the warning threshold, so as to dynamically adapt the model to the characteristics of power system performance degradation and component aging.

[0022] Furthermore, step two, which involves real-time prediction of driving range, includes the following steps: S2.1: Based on the data transmitted by the sensors in the power system, the remaining available energy is calculated in real time. For fuel or hybrid systems, the remaining fuel quantity can be read in real time by the fuel tank level sensor. Combined with the fuel calorific value and the engine dynamic efficiency curve, it is converted into the total output available energy. For pure electric or fuel cell systems, the remaining charge (SOC) and battery health status (SOH) can be collected by the battery management system (BMS). After deducting the battery's low-temperature degradation and the energy consumption reserve of high-voltage accessories, the actual dischargeable energy is calculated. For multi-energy system coupling, for hybrid electric vehicles, range-extended electric vehicles, and other configurations, the proportion of available fuel and electric energy can be dynamically allocated according to the energy management strategy to avoid double-counting cross-power supply parts. S2.2: A mechanistic model and data-driven fusion algorithm are used to calculate the energy consumption per unit mileage under the current operating conditions. Based on vehicle speed, acceleration, throttle opening, and braking frequency data from the last 30 seconds, a random forest algorithm is used to classify and identify the current operating condition type, such as urban congestion, suburban constant speed, high-speed driving, uphill driving, and low temperature. A baseline energy consumption is matched to the corresponding operating condition, and correction coefficients for ambient temperature, altitude, and drag coefficient are introduced. Energy consumption increases by 8%-12% for every 10°C decrease and by 4%-6% for every 1000m increase in altitude. In windy conditions, energy consumption increases by an additional 10%-20%. The vehicle's load is calculated in real time using suspension displacement sensors and acceleration recognition algorithms. For every 100kg increase in load, energy consumption increases by 5%-8%. The real-time power consumption of high-voltage or low-voltage accessories such as air conditioning, audio-visual systems, and intelligent driving sensors is also taken into account. This part of the energy consumption can account for up to 30% of the total energy consumption. The actual driving energy consumption data from the past 50km is retrieved, and the current energy consumption estimate is corrected using a sliding window weighted average method. The newer the driving data, the higher the weight, to adapt to the personalized differences in driving habits. S2.3: Basic range calculation, route-related range prediction, and confidence level display provide multi-scenario calculation and confidence assessment of driving range. Basic range calculation is performed according to the formula: It outputs three types of values: pure electric range, fuel range, and combined range. If the user has entered a navigation route, it calculates the energy consumption per kilometer of the route segment by combining the slope, speed limit, traffic light distribution, and congestion in the navigation data, and outputs an accurate range value based on the current planned route. It evaluates the error of the prediction result. When the operating conditions are stable and the data is sufficient, it is marked as: high confidence error ±5%. When the operating conditions fluctuate drastically and the remaining energy is less than 10%, it is marked as: low confidence, and the user is prompted that the range may fluctuate within a range of ±15%. S2.4: Based on the update frequency, the update frequency is once every 1 second under low-speed conditions and once every 500ms under high-speed conditions to ensure synchronization with actual energy consumption changes. Repeat steps S2.1 to S2.3 to dynamically update and visualize the battery life data.

[0023] Furthermore, in step three, the energy warning range alert value includes a pure electric range alert value and a fuel range alert value. These values ​​are either 10% of the total pure electric range or the total fuel range, or one of the fixed range values. The fixed range can be set by the driver according to their habits, such as 30 kilometers or 15 kilometers. The predicted range is obtained by predicting the pure electric range and fuel range in real time in step two. The predicted pure electric range value is compared with the pure electric range alert value. If the predicted pure electric range value is less than or equal to the pure electric range alert value, a pure electric range reminder alarm is issued. Similarly, the predicted fuel range value is compared with the fuel range alert value. If the predicted fuel range value is less than or equal to the fuel range alert value, a fuel range reminder alarm is issued.

[0024] Furthermore, in step four, when the fuel-powered charging reminder is issued as an electric range alert, if the driver confirms the fuel range within 5 seconds or does not confirm within 5 seconds, the fuel-powered charging will continue. If the driver cancels the fuel range operation within 5 seconds, the fuel-powered charging will not be performed. The electric range reminder will repeat once the pure electric range value drops by 2%. For example, when driving in the city, after the electric range reminder is issued, the driver can charge the battery upon reaching the destination according to their driving habits, thus eliminating the need for fuel-powered charging and reducing fuel consumption, thereby improving resource utilization efficiency. When the fuel range reminder is issued, the fuel warning status will be displayed on the instrument panel, and the fuel warning display cannot be manually cleared. The fuel warning display will be automatically cleared after refueling.

[0025] In particular, step four, which involves replenishing the power system, also includes the automated generation of a recharging scheme, including the following steps: S4.1: When the power system is being guided, such as when driving on a highway, the remaining mileage is calculated in real time, and the charging stations and gas stations along the route are connected to the network. S4.2: Calculate the comprehensive predicted range based on the sum of the real-time predicted pure electric range and fuel range values ​​in step two, and set the comprehensive range alarm threshold based on the comprehensive predicted range. The comprehensive range alarm threshold can be 1.5 times the sum of the pure electric range warning value and the fuel range warning value. S4.3: When the sum of the pure electric range and fuel range of the power system is less than or equal to the comprehensive range warning threshold, the remaining mileage in the navigation is compared with the comprehensive range warning threshold. When the remaining mileage in the navigation is less than the comprehensive range warning threshold, charging stations and gas stations are not recommended. Replenish energy according to the reminders in steps three and four. S4.4: When the remaining mileage in the navigation is greater than or equal to the comprehensive range warning threshold, the system will automatically search for charging stations and gas stations along the navigation route and recommend them to the user. When the user selects a specific charging station or gas station, the system will automatically add that charging station or gas station as a waypoint.

[0026] Furthermore, the prediction of the power system health in step five includes the following steps: S5.1: Based on the power system data collected in step one, extract time-domain statistical features, performance degradation features, and frequency-domain features. Time-domain statistical features calculate the mean, variance, kurtosis, and peak factor of dynamic signals such as speed, vibration, current, and voltage to characterize latent degradation features such as component wear and impact. Frequency-domain features are obtained by performing a Fast Fourier Transform (FFT) on vibration and noise signals. Fourier Transform is a common operation in vibration and noise signal processing and will not be discussed in detail here. Extract feature frequency amplitude changes to identify early fault characteristics of transmission components such as bearings and gears. Performance degradation features are obtained by comparing the deviation values ​​between the power system model output and the actual operating data, including engine efficiency deviation, battery charging and discharging efficiency deviation, and transmission system loss deviation, to quantify the overall performance degradation of the system. S5.2: An evaluation framework combining offline training and online updates is adopted to assess the health of the power system. Based on the evaluation results, the health of the power system is divided into four levels: healthy state, sub-healthy state, degraded state, and fault warning state. The healthy state is characterized by characteristic parameter fluctuations within ±5% of the factory reference value, with no signs of degradation. The sub-healthy state is characterized by characteristic parameter deviations within the range of 5%-15%, indicating early wear or attenuation, but not affecting normal operation. The degraded state is characterized by characteristic parameter deviations within the range of 15%-30%, indicating a significant decrease in performance, requiring preventive maintenance. The fault warning state is characterized by characteristic parameter deviations exceeding 30%, indicating a risk of failure, requiring immediate shutdown and inspection. S5.3: When the system identifies degradation and fault warning states, it automatically performs fault location. Based on the dynamic Bayesian network model, it associates abnormal features with the fault mode library, matches faulty components, and outputs fault impact analysis, explaining the degree of impact of the fault on power output, energy consumption, and safety, as well as the expected time of fault deterioration.

[0027] Furthermore, the output of the health prediction report in step five includes the current health level, health status scores of each component, remaining service life prediction, list of abnormal features and fault risk warnings, and automatically generates maintenance suggestions. It provides different priority maintenance plans based on the health level, such as suggestions for sub-health status (check transmission components at the next maintenance), suggestions for degradation status (replace worn bearings within 1000km), and the report supports local storage and cloud synchronization. It can be exported in a standardized format for maintenance personnel and management platforms to access.

[0028] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A high-frequency response model predictive control method for a dynamic system, characterized in that: Includes the following steps: Step 1: Collect various data from the power system in real time and establish a power system model; Step 2: Predict the driving range in real time based on the powertrain model and display the driving range data; Step 3: Set the kinetic energy alarm range warning value, compare the range warning value with the predicted range, and issue a reminder alarm when the predicted range is lower than the range warning value, and remind personnel to recharge. Step 4: When the alarm is issued, a prompt will be sent to replenish the power system with fuel. The driver should confirm the power replenishment method and confirm that the power system is being recharged. Step 5: Based on the real-time collected power system data, predict the health status of the power system and generate a health prediction report.

2. The high-frequency response model predictive control method for a power system according to claim 1, characterized in that: In step one, the real-time acquisition of various data of the power system is achieved by real-time acquisition and transmission of data by sensors arranged in the power system. The sensors arranged in the power system include, but are not limited to, one or more of the following: intake sensor, temperature sensor, speed sensor, current sensor, voltage sensor, pressure sensor, and wheel speed sensor.

3. The high-frequency response model predictive control method for a power system according to claim 2, characterized in that: The step one of establishing the dynamic system model includes the following steps: S1.1: Process the collected power system data, including outlier removal, noise filtering and normalization, and extract features from the processed data; S1.2: A modeling framework that integrates mechanism and data-driven approaches is adopted, and the power source module, transmission execution module, and load characteristic module are modeled in layers according to the power system structure. S1.3: The model parameters are calibrated and validated using a multi-dimensional validation method, which includes static calibration, dynamic calibration, and robustness testing. S1.4: Set the model periodic update threshold so that after the system runs through the periodic update threshold, the model is incrementally updated. The actual running data collected is compared with the model prediction results. When the deviation exceeds the warning threshold, the model parameters are automatically adjusted to dynamically adapt to the characteristics of performance degradation and component aging of the model and power system.

4. The high-frequency response model predictive control method for a power system according to claim 3, characterized in that: Step two, which involves real-time prediction of driving range, includes the following steps: S2.1: Calculate the remaining available energy in real time based on the data transmitted by the sensors in the power system; S2.2: Calculate the energy consumption per unit mileage under the current operating conditions using a fusion algorithm of mechanism model and data-driven approach; S2.3: Basic range calculation, route-related range prediction, and confidence level display provide multi-scenario calculation and confidence assessment of range; S2.4: Based on the update frequency, repeat steps S2.1 to S2.3 to dynamically update and visualize the battery life data.

5. The high-frequency response model predictive control method for a power system according to claim 4, characterized in that: In step three, the kinetic energy alarm range warning value includes a pure electric range warning value and a fuel range warning value. The pure electric range warning value and the fuel range warning value are one of the following: 10% of the total pure electric range and the total fuel range, and the fixed range value. The predicted range is the pure electric range and fuel range values ​​predicted in real time in step two. The predicted pure electric range value is compared with the pure electric range warning value. If the predicted pure electric range value is less than or equal to the pure electric range warning value, a pure electric range reminder alarm is issued. Similarly, the predicted fuel range value is compared with the fuel range warning value. If the predicted fuel range value is less than or equal to the fuel range warning value, a fuel range reminder alarm is issued.

6. The high-frequency response model predictive control method for a power system according to claim 5, characterized in that: In step four, when the fuel-powered generator is issued as a power range reminder alarm, if the user confirms the fuel range within 5 seconds or does not confirm within 5 seconds, the fuel-powered generator will be used to continue the power supply. If the user cancels the fuel range operation within 5 seconds, the fuel-powered generator will not be used. The power range reminder alarm will repeat after the pure electric range value drops by 2%. When a fuel range reminder alarm is issued, a fuel alarm status will be displayed on the instrument panel, and the fuel alarm display cannot be manually cleared. The fuel alarm display will be automatically cleared after fuel is replenished.

7. The high-frequency response model predictive control method for a power system according to claim 6, characterized in that: Step four, which involves replenishing the power system, also includes the automated generation of a charging scheme, comprising the following steps: S4.1: When the power system is guiding, the remaining mileage is calculated in real time, and the charging stations and gas stations along the route are counted online; S4.2: Calculate the comprehensive predicted range based on the sum of the real-time predicted pure electric range and fuel range values ​​in step two, and set the comprehensive range alarm threshold based on the comprehensive predicted range. S4.3: When the sum of the pure electric range and fuel range of the power system is less than or equal to the comprehensive range warning threshold, the remaining mileage in the navigation is compared with the comprehensive range warning threshold. When the remaining mileage in the navigation is less than the comprehensive range warning threshold, charging stations and gas stations are not recommended. S4.4: When the remaining mileage in the navigation is greater than or equal to the comprehensive range warning threshold, the system will automatically search for charging stations and gas stations along the navigation route and recommend them to the user. When the user selects a specific charging station or gas station, the system will automatically add that charging station or gas station as a waypoint.

8. The high-frequency response model predictive control method for a power system according to claim 7, characterized in that: The prediction of the power system health in step five includes the following steps: S5.1: Based on the data of the power system collected in step one, extract the time-domain statistical features, performance degradation features and frequency-domain features; S5.2: An evaluation framework combining offline training and online updates is adopted to assess the health of the power system, and the health of the power system is divided into four levels based on the evaluation results: healthy state, sub-healthy state, deterioration state, and fault warning state.

9. The high-frequency response model predictive control method for a power system according to claim 8, characterized in that: The prediction of the power system health in step five also includes the following steps: S5.3: When the system identifies degradation and fault warning states, it automatically performs fault location, associates abnormal features with the fault mode library based on the dynamic Bayesian network model, matches faulty components, and outputs fault impact analysis.

10. The high-frequency response model predictive control method for a power system according to claim 9, characterized in that: The output of the health prediction report in step five includes the current health level, health status score of each component, remaining service life prediction, abnormal feature list and fault risk warning, and automatically generates maintenance suggestions. It provides maintenance plans with different priorities according to the health level. The report supports local storage and cloud synchronization and can be exported in a standardized format for maintenance personnel and management platform to retrieve.

Images