Intelligent Control Method for Inverters Based on Adaptive Algorithms

Abstract

This invention discloses an intelligent inverter control method based on an adaptive algorithm, belonging to the field of intelligent inverter control technology. It employs a five-step approach: multi-source temperature acquisition and environmental identification, machine learning-based electro-thermal coupling prediction, adaptive inverter parameter optimization, collaborative thermal management, and fault diagnosis and self-repair. Addressing the risk of high or low temperature failures in motors and inverters under extreme climates, it acquires real-time temperature distribution and load information, predicts temperature rise trends in advance, and proactively adjusts inverter current, voltage, and modulation strategies to achieve safe derating or torque compensation. Furthermore, it performs multi-sensor cross-validation and observer fusion when sensors drift or components age, maintaining stable system operation and efficient energy utilization, significantly improving driving performance and vehicle reliability in extreme environments. During this process, online model updates further refine fault diagnosis and self-repair, enhancing the durability and economy of the electric drive system.

Description

Technical Field

[0001] This invention relates to the field of inverter intelligent control technology, specifically to an inverter intelligent control method based on an adaptive algorithm. Background Technology

[0002] When electric vehicles operate in extremely cold environments (such as high-latitude regions with temperatures below -10°C) or extremely hot climates (such as deserts with temperatures exceeding 40°C), the temperatures of key components such as the motor stator, permanent magnets, inverter power modules, and battery packs differ significantly from the ambient temperature. Furthermore, the accumulation and dissipation of heat are accelerated during long-distance high-speed driving or under heavy loads. At these conditions, power electronic devices (such as IGBTs and MOSFETs) are prone to a sharp increase in switching losses or overheating protection at high temperatures, while in extremely cold conditions, insufficient output power can result from poor lubrication and increased battery internal resistance. Simultaneously, vehicles have high demands for power response and driving range; slow starts at low temperatures or power derating at high temperatures can lead to a decline in user experience and even sudden safety hazards. Under these harsh conditions, not only is the onboard thermal management system's ability to quickly adapt to changes in the external environment and load tested, but the inverter and motor also require highly reliable and sophisticated temperature monitoring strategies to balance power performance and safety.

[0003] Chinese invention patent CN106240341B discloses a cooling system and control method for a permanent magnet synchronous motor (PMSM) in electric vehicles. The cooling system includes a control unit, a motor winding ammeter, a cooling water pump, a cooling water jacket, a cooling water temperature sensor, a coolant flow meter, a radiator, and a cooling fan. The control method detects the current flowing through the motor stator windings to determine whether the PMSM is trending towards heating or cooling under certain operating conditions, adjusting the cooling capacity of the cooling system accordingly to achieve active cooling system regulation. It also introduces auxiliary feedback regulation, using a motor efficiency map to query the motor's operating point, obtaining the motor's efficiency under the corresponding operating conditions, calculating the motor's heat generation, and then feeding back the difference between the motor's heat generation and the radiator's heat dissipation to the control unit to correct the cooling system's adjustment. This invention provides active cooling control of the cooling system before the motor temperature changes, ensuring sufficient motor heat dissipation while saving energy.

[0004] However, considering the above practical application scenarios and existing technologies:

[0005] Currently, under such extreme temperature conditions, the control methods for vehicle inverters and motors often adopt fixed safety thresholds or simple temperature switching strategies, limiting power only when significant over-temperature is detected. In extremely cold conditions, there is also a lack of forward-looking compensation and adaptive regulation of temperature dynamic characteristics, resulting in protection being triggered only when the temperature has seriously exceeded the design range. This causes excessive stress accumulation in power devices or motor magnets, accelerating aging and even causing failures.

[0006] Meanwhile, sensors are prone to drift or failure under extreme temperatures, leading to inaccurate identification of the temperature field and component performance status. Traditional solutions, relying solely on single-point temperature measurement and shallow temperature threshold judgment, struggle to accurately grasp the vehicle's temperature distribution and trends. These issues can cause significant vehicle deceleration, sudden torque drops, or unstable power output during actual operation, posing serious challenges to overall vehicle reliability and safety.

[0007] To address this, the present invention provides an intelligent control method for inverters based on an adaptive algorithm. Summary of the Invention

[0008] (a) Technical problems to be solved

[0009] To address the shortcomings of existing technologies, this invention provides an intelligent inverter control method based on adaptive algorithms. This method integrates five steps: multi-source temperature acquisition and environmental identification, machine learning-based electro-thermal coupling prediction, adaptive inverter parameter optimization, collaborative thermal management, and fault diagnosis and self-repair. Addressing the risk of high or low temperature failures in motors and inverters under extreme climates, it acquires real-time temperature distribution and load information, predicts temperature rise trends in advance, and proactively adjusts inverter current, voltage, and modulation strategies to achieve safe derating or torque compensation. Furthermore, it performs multi-sensor cross-verification and observer fusion when sensors drift or components age, maintaining stable system operation and efficient energy utilization, significantly improving driving performance and vehicle reliability in extreme environments; thus solving the technical problems described in the background art.

[0010] (II) Technical Solution

[0011] To achieve the above objectives, the present invention provides the following technical solution:

[0012] The inverter intelligent control method based on adaptive algorithm includes, when the ambient temperature is detected to be significantly higher than the preset range, integrating multi-source sensor readings, vehicle operating parameters and external meteorological data and using kernel function interpolation to generate a multi-source fused temperature distribution field;

[0013] After collecting the multi-source fusion temperature distribution field and vehicle operating conditions, the electro-thermal coupling prediction model is invoked and error constraints are introduced to generate a prediction quantity that includes at least the predicted temperature distribution field and the predicted heat load value.

[0014] If the predicted temperature distribution field is detected to be about to exceed the temperature margin, the inverter control parameters are adaptively adjusted to actively limit and reduce power loss before the critical temperature, so as to maintain power output and device safety.

[0015] When the inverter output power and actual temperature distribution field are transmitted back to the vehicle thermal management module in real time, the temperature and heat dissipation information of the BMS side are compared to predict the temperature distribution field and correct the electro-thermal coupling model. Then, the fan or liquid cooling rate is dynamically scheduled and the correction data is fed back to the previous steps.

[0016] When a sensor reading deviates significantly from the predicted temperature distribution field, the fault type is determined and a metric function is calculated. Then, a degraded operating value is initiated, or redundant signals are enabled to facilitate self-repair of power devices or temperature measurement failures.

[0017] Preferably, a multi-point temperature sensing network is deployed in the vehicle power system to record and mark the readings of each sensor. Based on the spatial location of each measuring point, the measurement results are mapped to the original temperature field. Based on the external environmental information obtained during vehicle operation, an external temperature curve is collected and an environmental influence coefficient is constructed.

[0018] Preferably, the environmental impact coefficient is coupled with the original temperature field to obtain the environmentally corrected multi-source fused temperature distribution field;

[0019] The environmentally corrected multi-source fused temperature distribution field and external environment identification results are stored in a real-time database and interact with a historical operating condition database.

[0020] Preferably, the acquired multi-source fusion temperature distribution field, external environment identification results, vehicle operating status, and relevant records in the historical operating condition database are synchronized to the feature construction module;

[0021] Feature extraction is performed using a temporal fusion method with an exponentially decaying kernel to generate comprehensive feature terms for subsequent models. The comprehensive feature term sequence between two historical moments is used in conjunction with the target value to construct the feature-label pairs required for supervised training.

[0022] Preferably, based on the thermophysical characteristics of electric vehicle inverters and motors under extreme temperatures, a hybrid electro-thermal coupling model with comprehensive feature terms as input is introduced into the machine learning framework, and the predicted temperature distribution field, heat load, and magnetic flux attenuation at future moments are output. A loss function that integrates physical constraints is introduced to train the electro-thermal coupling prediction model.

[0023] The prediction performance is evaluated through both offline verification and online update modes, and fine-tuning training is performed periodically. After the electro-thermal coupling prediction model passes the evaluation, its prediction output is transmitted to the inverter control parameter optimization module.

[0024] Preferably, the obtained predicted output is compared with the designed upper limit of temperature safety and lower limit of magnetic flux to form temperature margin and magnetic flux margin; the driving intention of the vehicle in the current and next time period is collected, and the control parameter set of the existing inverter control is extracted at the same time. If the risk of overheating or insufficient magnetic flux is predicted, the corresponding power limiting strategy is selected according to the temperature margin and magnetic flux margin.

[0025] The dynamic power limit value is combined with the original control parameter set to obtain a parameter mapping with upper and lower limit constraints; if temperature or magnetic flux risk is detected, the control parameters are dynamically limited or adjusted according to the dynamic power limit value.

[0026] Preferably, based on the determined parameter mapping, an adaptive algorithm is built into the inverter controller to solve for each control parameter online, obtaining the optimal control parameters finally applied to the actuator, wherein:

[0027] The predicted output is incorporated into the objective function to optimize the control parameters for future time periods, thereby obtaining the optimal control parameters that change over time. A dynamic control strategy is then constructed and finally executed in conjunction with parameter mapping. If the actual temperature or magnetic flux demagnetization rate is still found to deviate from the expected value, the online prediction update mechanism is automatically invoked in the next control cycle for online fine-tuning or incremental training to correct the upper limits of the input and control parameters of the electro-thermal coupling prediction model.

[0028] Preferably, the actual temperature distribution field and power flow information of the inverter after the actual execution of the dynamic control strategy are collected, as well as the updated optimal control parameters; the inverter-side data are compared with the predicted temperature distribution field, and the deviation between the prediction and reality is evaluated through the control error term;

[0029] If the control error term value is greater than expected, the machine learning model or sensor calibration should be corrected.

[0030] The optimal control parameters, actual temperature distribution field, and control error terms are transmitted to the vehicle thermal management module. At the same time, the battery status and the operating status of the cooling system are obtained from the BMS side and marked as battery management system status information. After alignment, a multi-terminal integrated data packet is formed.

[0031] Preferably, after acquiring the merged multi-terminal integrated data packet, the vehicle thermal management module and the BMS jointly execute a global scheduling; after calculating the optimal solution of the introduced global objective function, the vehicle thermal management module sends a scheduling instruction to the cooling system and guides battery preheating or equalization operations on the battery management system (BMS) side.

[0032] Preferably, the actual action feedback generated by the final execution of the scheduling command is fed back in real time to the sensor fusion parameter correction and environmental identification update, as well as the input correction and fine-tuning of the electro-thermal coupling prediction model; if it is found that the actual temperature response after the implementation of the adaptive control strategy deviates significantly from the prediction, the scheduling strategy is dynamically adjusted, the global objective function is re-evaluated, and local corrections are made.

[0033] Preferably, after obtaining the actual temperature distribution field and other key sensor readings and corresponding predicted values, a metric function based on the fusion of observer and residual is defined. When the metric function increases sharply in a short period of time and exceeds the threshold, the specific fault type is determined by multidimensional residual pattern matching for the detected suspicious signals.

[0034] Preferably, the degraded operating value is defined based on the identified fault type and degree of abnormality, indicating the extent to which partial degrading or backup parameter activation is required;

[0035] When the fault is severe, the degraded operation value increases. If the fault has limited impact, only a minor temporary replacement is performed. For sensor drift or partial failure, redundant sensor data or the correction value based on the predicted output is temporarily used as the replacement input. When the degraded operation value exceeds the preset warning line, degraded operation is automatically executed.

[0036] (III) Beneficial Effects

[0037] This invention provides an intelligent control method for inverters based on an adaptive algorithm, which has the following advantages:

[0038] Through five closely linked steps, a complete process optimization for motors and inverters in extreme temperature environments is provided, from real-time monitoring to fault diagnosis and self-repair. The overall beneficial effects are mainly reflected in the following aspects:

[0039] Through multi-source temperature measurement and environmental recognition, the quality of sensor data in extremely cold and hot regions has been significantly improved, and dynamic adaptation to external conditions such as wind speed and humidity has been achieved, thus providing high-precision input for subsequent machine learning models.

[0040] Based on predicted temperature distribution field and power heat load Electro-thermal coupling prediction can identify overheating and magnetic flux attenuation risks in advance, thus improving the safety margin ΔT. safe ,ΔΨ safe Quantitative evaluation is conducted, thereby reducing fatigue accumulation and failure probability of components under large temperature differences;

[0041] Adaptive inverter control uses prediction results to flexibly adjust control parameters before the temperature is about to reach the high-risk range, breaking away from the traditional passive mode of reducing power only when the temperature is too high, and significantly enhancing the vehicle's power and efficiency in extreme environments.

[0042] Utilizing the actual temperature distribution field of the inverter Real-time feedback information is used to couple the inverter's execution results with the BMS and vehicle cooling system, creating a unified optimization from battery preheating to fan liquid cooling regulation. This collaborative thermal management can allocate more heat dissipation resources to protect core power devices at high temperatures, and can moderately preheat the battery and motor at low temperatures, balancing energy consumption and performance.

[0043] To address sensor drift, power component aging, or sudden failures, a multidimensional residual vector R is used. s (x,t) and the metric function Φ obs (t) and other multidimensional residual fusion methods are used to diagnose anomalies in a timely manner and to reduce the operating value Γ. degrade (t) Trigger self-repair or graded throttling. In this way, it can maintain safe operation even in extreme cold or heat, and the collected abnormal data feeds back into the machine learning model, forming self-learning and adaptive capabilities;

[0044] In summary, this solution not only ensures the safety and stability of the powertrain system in terms of temperature monitoring, thermal prediction, inverter control, vehicle thermal management, and fault diagnosis and self-repair, but also maintains superior efficiency and range in various extreme environments, achieving synergistic gains that balance multiple objectives. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the inverter intelligent control method based on adaptive algorithm of the present invention. Detailed Implementation

[0046] 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.

[0047] Please see Figure 1 This invention provides an intelligent control method for inverters based on an adaptive algorithm, including:

[0048] Step 1: When the vehicle detects that the ambient temperature exceeds the preset range, it automatically integrates the outputs of multiple temperature sensors, data from external weather stations, historical operating condition information, and the vehicle's current operating parameters, and constructs a multi-source fused temperature distribution field through a kernel function weighted interpolation strategy. Simultaneously integrate the environmental impact coefficient Θ ext (t) to capture the thermal coupling effect of key components;

[0049] Step one includes the following:

[0050] Step 101: Sensor Deployment and Temperature Data Field Construction

[0051] A series of sensors are deployed in the vehicle's powertrain system (such as the motor stator, motor end cover, and inverter power module surface) to form a multi-point temperature sensing network. These sensors include thermistor sensors, platinum resistance temperature sensors, thermocouples, and integrated temperature / humidity sensors. The readings of each sensor at time t are recorded and marked. (Where i = 1, 2, ..., N represents the sensor number, and N is the total number of sensors); based on the spatial position x of each measuring point (i) (This can be represented as three-dimensional coordinates), and its measurement results are mapped to the original temperature field T(x,t). To avoid local distortion caused by differences in measurements from various sensors, the following fusion formula with a weighted kernel function is introduced:

[0052]

[0053] In the formula: Ω represents the overall position domain of the sensor; Ψ(||xx) (i) ||) is the kernel function, used to calculate the spatial distance ||xx. (i) ||Weighting the sensor measurements can generally be achieved using a Gaussian kernel or other positive definite kernels to achieve smooth interpolation; Let be the measurement value of the i-th sensor at time t; where the kernel function is a Gaussian kernel.

[0054] K(x i ,x)=exp(-||x i -x|| 2 / (2σ 2 )), where σ=0.12m was obtained through actual vehicle calibration;

[0055] When used, the kernel function Φ is used for weighting, which can significantly reduce the local deviation caused by single-point measurement, realize more reliable temperature field reconstruction after multi-sensor fusion, and improve the accuracy of temperature distribution field. By defining the original temperature field T(x,t) as a continuous distribution field function, it can be directly output to the machine learning model required in the second step for training or real-time prediction input, avoiding the secondary error of data format conversion.

[0056] Step 102: External Environment Identification and Multi-Source Information Fusion

[0057] Based on real-time weather data acquired during vehicle operation, cloud-based weather APIs, or other external environmental information, the external temperature curve is marked as T. env (t), and further define the environmental impact coefficient Θ based on the vehicle's current location and possible altitude, wind speed, humidity, etc. ext (t), which can be referred to in the following form:

[0058]

[0059] In the formula: h(t) is the vehicle's current altitude, w(t) is the wind speed, and RH(t) is the relative humidity; h0, w0, and RH0 are reference values ​​for altitude, wind speed, and humidity, respectively. To form a column vector by taking the natural logarithm of x(τ) / x0 (element-by-element ratio). There is no dimensionlessness; A is a 3×3 matrix, which can be expanded or reduced according to the number of environmental elements; ω(τ) is a scalar function, which is unweighted by time; or exponential decay, Gaussian kernel, etc. can be used; A = diag(1 / h0, 1 / v0, 1 / u0), where h0 = 1000m, v0 = 5m / s, u0 = 60%, is established based on the average value of the national meteorological reference station;

[0060] If a new environmental element (such as light intensity S) is added, the column expansion method is used while maintaining the unitized diagonal: A' = diag(A, 1 / S0);

[0061] The environmental impact coefficient Θ ext The temperature field T(x,t) is coupled with the original temperature field T(x,t) to obtain the environmentally corrected multi-source fused temperature distribution field. Its form can be expressed as:

[0062]

[0063] Where Λ(·) is a mapping function coupled with the external environment, used to characterize the additional temperature compensation effect under wind cooling, heat dissipation, or extreme cold conditions. That is, the dimensionless environmental influence coefficient Θ can be implemented through two paths: a physical-parameterized linear combination or a data-driven neural network. ext (t) is mapped to the temperature increment ΔT ext (x,t); For example, in extremely cold conditions, there may be a situation where the sensor location is not completely cooled down but the external temperature is extremely low. In this case, Λ(Θ) ext (t) can provide a negative offset based on simulation or data-driven methods to more accurately characterize the actual cooling effect; where Λ(·) has the following form:

[0064]

[0065] For the dimensionless index of forced convection; A dimensionless index for natural convection; f is a dimensionless index for radiant heat; c (x) represents the forced convection spatial weights; f r (x) represents the spatial weight of natural convection; fs (x) represents the radiation space weight; γ c To force the global weights of convection, values ​​are taken between 0.3 and 0.6; γ r γ represents the global weight for natural convection, with values ​​ranging from 0.1 to 0.4. s The global weights are set to a value between 0.1 and 0.3; the constraint is γ. c +γ r +γ s =1;T scale It is a temperature scaling measure used to restore dimensionless terms to temperature increments;

[0066] The environmentally corrected multi-source fusion temperature distribution field The results of external environment identification (such as classification labels for extreme cold, extreme heat, and humid heat) are uniformly stored in a real-time database and interact with the historical operating condition database, and are marked as...

[0067] When using it, the environmental impact factor Θ is used. ext (t) and the environmentally corrected multi-source fused temperature distribution field The integration of these technologies enables timely capture of heat flow differences caused by extreme temperatures, avoiding the lag or bias that can occur with relying solely on internal temperature measurements. Furthermore, it mitigates the impact of significant changes in external climate or altitude, which can affect the environmental impact coefficient Θ. ext (t) will change accordingly and be reflected in the mapping function Λ(·), thus enhancing the system's adaptability to the external environment; the environmental influence coefficient Θ is introduced. ext Using (t) and the mapping function Λ(·) as environmental correction units, compared with the traditional method of using only the vehicle sensor itself for temperature limiting protection, it increases the quantitative characterization of the external dynamic environment, so that thermal management decisions no longer rely on internal measurements in isolation, but achieve a comprehensive evaluation of the climate scenario in which the vehicle is located.

[0068] Step 2: Receive the multi-source fused temperature distribution field After determining the vehicle's operating status, an electro-thermal coupling model combining machine learning and physical priors is invoked to analyze the inverter and predict the temperature distribution field. and power heat load Predictions are made, and potential overheating or extreme cold risks are identified through high-order error constraints and online updates, providing a quantitative reference for the adaptive control parameter Π.

[0069] Step two includes the following:

[0070] Step 201: Multidimensional Feature Engineering and Data Reconstruction

[0071] The acquired multi-source fusion temperature distribution field External environment identification results Θ ext(t), vehicle operating status u(τ) (vehicle speed, load, acceleration), and relevant records in the historical operating condition database are synchronized to the feature construction module;

[0072] Align all data on the timeline to have a uniform sampling point t. k Let k = 1, 2, ..., K, where K is the total number of samples. A time fusion method with an exponentially decaying kernel is used for feature extraction to emphasize the impact of recent temperature changes on the system's thermal response. The following mapping is defined to generate a comprehensive feature term F(t) for subsequent models. k ):

[0073]

[0074] In the formula: Δt represents the fusion time, which is used to comprehensively consider the influence of temperature and external environment over a recent period;

[0075] α is the time decay factor, ranging from 0.01 to 0.25, used to give greater weight to recent data during integration; G(·) is the feature combination function, used to incorporate the environmental impact coefficient Θ. ext (τ) is superimposed with the vehicle operating state u(τ) to form a temperature distribution field that can be fused with multiple sources. The corresponding comprehensive characteristic term F(t) k For example, two feasible implementation schemes can be provided: a linear-physical hybrid method and a lightweight neural network mapping. The input features, normalization methods, weight / network structures and safety clamping strategies are given in detail to ensure that the output is consistent with the temperature field dimensions and can be directly used in subsequent steps.

[0076] Comprehensive characteristic term F(t) k The result dimension can be simplified by discretizing or projecting the spatial location x. Specifically, the spatial location x can be divided into several key regions and integrals can be calculated separately. Then, the results can be combined into a vector or matrix, which is convenient for subsequent machine learning algorithms.

[0077] In real-time prediction scenarios, historical time points t are utilized. k-j to t k The comprehensive characteristic term F(t) sequence between them, combined with the corresponding future temperature distribution field or the heat load H(t) of power devices k+δ ) and other target values ​​to construct the feature-label pairs required for supervised training;

[0078] For key indicators such as motor flux attenuation or inverter overheating, corresponding tag sequences Y(t) can also be generated. k This includes factors such as the rate of temperature rise or the potential performance degradation, which facilitates the construction of multi-task learning models.

[0079] In use, it can more sensitively capture rapid temperature changes and, combined with external environment and vehicle operating conditions, improve the perception accuracy of sudden thermal shocks or rapid cooling, highlighting the impact of recent conditions; the comprehensive characteristic term F(t) k It also integrates multi-source fusion temperature distribution fields. Environmental impact coefficient Θ ext The values ​​of (t) and vehicle operating state u(τ) help machine learning models to more comprehensively characterize the electro-thermal coupling process, retaining historical information while highlighting the importance of recent changes. This allows the combined effects of external climate and vehicle power demand to be reflected in the model input stage.

[0080] Step 202: Training and Dynamic Evaluation of the Electro-Thermal Coupling Prediction Model

[0081] Based on the thermophysical characteristics of electric vehicle inverters and motors under extreme temperatures, a hybrid electro-thermal coupling model is introduced into a machine learning framework: on the one hand, it utilizes neural networks (such as multilayer perceptrons, denoted as Φ) NN On the one hand, it captures the complex nonlinearity of the temperature distribution field; on the other hand, it introduces a small number of physical priors (such as thermal conductivity and stator copper loss characteristics in the heat transfer equation), integrates them through the structure of the neural network, and constrains them in the loss function.

[0082] The input comprehensive feature term F(t) of the neural network k Output the predicted temperature distribution field at future times. Inverter switching device heat load Magnetic flux attenuation Etc. The symbols here. This represents the model's predicted values, compared with actual measured or validated data;

[0083] To reflect the mechanism of electro-thermal coupling and reduce extrapolation inaccuracies that may arise from purely data-driven approaches, a loss function incorporating physical constraints is introduced. Where W is the model weight vector, including neural network weights and related physical hyperparameters, and the loss function can be defined as:

[0084]

[0085] In the formula: For actual observed or labeled temperature distribution fields; To predict the temperature distribution field for the model;

[0086] Γ represents the spatiotemporal domain (covering the training period and the spatial range of key components), Γ f This is an adjustable factor, ranging from 1 to 20, used to focus on key areas; This is a physical residual operator used to measure the degree to which the prediction results violate fundamental thermodynamic equations (such as heat conduction equations and heat dissipation models) or inverter power loss mechanisms. A typical form is:

[0087]

[0088] ρ, c p Here, k represents the material density and specific heat capacity; k is the thermal conductivity (which may differ between copper windings and aluminum shell sections).

[0089] q gen (x,t) represents the local heat generation rate per unit volume, calculated from the current density and device switching losses; h conv (x) is the local convective heat transfer coefficient; T ∞ For ambient temperature, β is the predicted temperature; β is the physical constraint weighting coefficient, which takes a value greater than 0, and is used to balance the loss function between prediction accuracy and physical consistency.

[0090] After the electro-thermal coupling prediction model is trained, its prediction performance is evaluated through two modes: offline validation and online updating. Offline validation involves inputting complete test data (including extreme cold and extreme heat conditions) into the model and comparing it with measured temperatures. Online updating involves continuously collecting the latest multi-source fused temperature distribution field during actual vehicle operation. And periodically fine-tune the training to correct the bias of the model under different environments and aging conditions;

[0091] After the model passes evaluation, its predicted output (i.e., the predicted future temperature distribution field) will be... Predicted thermal load values ​​of power devices Magnetic flux attenuation prediction The parameters (etc.) will be transmitted to the inverter control parameter optimization module in step three, guiding the latter to actively adjust before overheating or performance degradation may occur.

[0092] When in use, it integrates physical mechanisms and machine learning, and can maintain high prediction accuracy even in scenarios with missing or distributed data (such as extreme cold and high altitude or prolonged heat). The dynamic evaluation and iterative training mechanism ensures that the model can be continuously corrected according to the constant changes in vehicle operating conditions (such as component aging and sensor offset), thus maintaining high reliability throughout the vehicle's life cycle. To provide necessary temperature and performance evolution references, it helps the inverter controller to achieve proactive adaptive regulation in extreme environments, and can maintain the model's efficiency and accuracy during real vehicle operation through online updates.

[0093] Step 3: When overheating or undertemperature risks are detected, an adaptive optimization algorithm is automatically loaded into the inverter controller based on the predicted temperature distribution field. and temperature margin ΔT safe(t) Adjust control parameters in real time, dynamically limit or advance power output before the temperature critical value, and adjust the magnetic flux margin ΔΨ. safe (t) Prevent excessive demagnetization of the magnet;

[0094] Step three includes the following:

[0095] Step 301: Control Requirements Extraction and Key Parameter Mapping

[0096] Read the future time period [t] from the obtained prediction results k , t k+δ Predicted future temperature distribution field of internal motor windings and power modules And the predicted heat load of the inverter and flux attenuation prediction wait;

[0097] Compare the above predicted values ​​with the designed upper limit of temperature safety T. safe and lower limit of magnetic flux Ψ safe Comparisons are made to form the temperature margin ΔT safe (t) and flux margin ΔΨ safe (t) is used to determine whether there is a potential risk of overheating or severe demagnetization; the driving intentions of the vehicle in the current and next time periods (such as acceleration, constant speed cruising, slope conditions, etc.) are collected and labeled as D(t); at the same time, the set of control parameters of the existing inverter control Π={ω switch i max v dc …},in:

[0098] ω switch Indicates the inverter modulation switching frequency; i max This indicates the peak output current allowed by the inverter; v dc Indicates bus voltage, etc.;

[0099] If the prediction is made at time t k+δ If there is a risk of overheating or insufficient magnetic flux in the vicinity, then the temperature margin ΔT should be considered. safe (t) and flux margin ΔΨ safe (t) Pre-calculate appropriate power limiting strategies, such as limiting peak current amplitude or slightly reducing the upper limit of bus voltage, to achieve early cooling or reduce losses. The specific mapping relationship can be expressed as:

[0100] Γ limit (t)=Θ ctrl [ΔT safe (t),ΔΨ safe (t),D(t)]

[0101] Where Γ limit(t) represents the dynamic power limit imposed on the system at a future time t, Θ ctrl (·) is a function that combines safety margins with vehicle requirements to generate control limits, for example:

[0102]

[0103] Where: Θ ext (t) represents the environmental impact coefficient; u(t) represents the vehicle operating status; T dyn (t) represents the real-time temperature data of the vehicle's power system; To predict the temperature distribution field in the future; This is the predicted value for magnetic flux attenuation;

[0104] The dynamic power limit value Γ limit (t) is combined with the original control parameter set Π to obtain the parameter mapping Π with upper and lower limit constraints. bounded (t); When it is determined that no reduction is needed, the parameter mapping Π bounded (t) is the same as the default value of the control parameter set Π;

[0105] If a temperature or magnetic flux risk is detected, then the dynamic power limit value Γ is applied. limit (t) with respect to switching frequency ω switch Maximum current i max or DC bus voltage v dc Dynamically limit or appropriately adjust control parameters;

[0106] In practice, by using the predicted values ​​output from step two, power scheduling preparation can be completed before the arrival of extreme high / low temperatures, rather than waiting until the system issues a temperature alarm before urgently derating; parameter mapping Π bounded (t) not only includes simple peak current limiting, but also allows for comprehensive adjustment of inverter switching frequency or bus voltage, improving the ability to withstand extreme temperatures; a temperature margin ΔT is introduced. safe (t) and flux margin ΔΨ safe The mapping mechanism of (t) can accurately indicate when and to what extent power limiting is performed, avoiding excessive or insufficient protection operations.

[0107] Step 302: Adaptive control algorithm implementation and closed-loop execution

[0108] Map Π based on the determined parameters bounded (t) An adaptive algorithm (which can be nonlinear model predictive control or Lyapunov basis control) is built into the inverter controller to solve for each control parameter online, thus obtaining the optimal control parameter Π that is finally applied to the actuator. * (t);

[0109] The predicted heat load value output from step two And flux attenuation prediction These factors are incorporated into the objective function to ensure that driving requirements are met to the greatest extent possible while taking into account motor temperature and thermal stress of inverter power devices. Below is an example of a nonlinear cost function for the future time period ΔT. ctrl Optimize the control parameters:

[0110]

[0111] In the formula: Ω1, Ω2, and Ω3 are weighting coefficients, all ranging from 0 to 1, used to balance vehicle power requirements and temperature safety; Θ drive (D(τ)) measures the power demand for a vehicle's dynamic response; This is a constraint penalty term for the inverter's heat load; Π margin (ΔΨ safe (τ) represents the allowable variation range of the control parameters after considering the magnetic flux safety margin;

[0112] By analyzing the objective function Solving for the maximum (or minimum) value yields the optimal control parameter Π that varies with time. * (t), and construct a dynamic control strategy and combine it with parameter mapping Π bounded (t) will be executed finally;

[0113] Inverter controller maps parameters ∏ bounded (t) Real-time updates of modulation strategies (such as carrier frequency and duty cycle limit of Space, Vector, and PWM), current command upper limit, and bus voltage control signals, etc., and comparison with the actual output feedback of the motor. If the actual temperature or magnetic flux demagnetization rate is detected to still deviate from the expectation, the online prediction update mechanism of step two is automatically called in the next control cycle. The online fine-tuning or incremental training is performed using real-time feedback from actual operation to correct the upper limit of the input and control parameters of the electro-thermal coupling prediction model.

[0114] The execution data of this closed-loop strategy (including actual temperature, real-time power output, and optimal control parameters) * The temporal distribution field of (t) will be passed to step four, and will work with the BMS and vehicle thermal management module to improve thermal management and performance optimization at the whole vehicle level.

[0115] In use, an adaptive algorithm incorporating heat load prediction is employed, enabling dynamic and precise control strategies under different temperature conditions. Through closed-loop interaction with the prediction model, it can quickly correct discrepancies between actual operation and prediction, maintaining the system's sustained performance. In the objective function... Simultaneously considering vehicle power requirements Θ drive(D(τ)) and inverter heat load prediction and magnetic flux safety margin ΔΨ safe (τ) Achieve multi-objective nonlinear optimization, rather than simply pursuing temperature not exceeding the limit or a single maximum torque output;

[0116] Through the changing space Π margin (·) Incorporate the risk of magnetic flux attenuation into the control constraints, and combine it with a derating strategy based on thermal prediction to achieve synergy between magnet protection and inverter safety.

[0117] Step 4: After the inverter completes adaptive control and outputs the actual temperature distribution field... After combining power data, the vehicle thermal management module, along with battery temperature and equalization information from the BMS, predicts the temperature distribution field. With respect to the actual temperature distribution field The differences are evaluated for residuals and the fan or liquid cooling schedule is activated to preheat the battery and motor at low temperatures. The correction data is then fed back to the machine learning model for continuous updates.

[0118] Step four includes the following:

[0119] Step 401: Real-time feedback and dynamic multi-terminal linkage

[0120] The actual temperature distribution field of the inverter after the actual execution of the dynamic control strategy is collected. With power flow information (including output current, voltage, etc.) and the updated optimal control parameters Π * (t);

[0121] Inverter-side data and future predicted temperature distribution field For comparison, the following error term Λ is controlled. err The formula (t) is used to evaluate the deviation between the prediction and reality, and is defined as follows:

[0122]

[0123] In the formula: This represents the actual temperature distribution field of the inverter and related components at location x. The predicted temperature distribution field is obtained based on thermal models and machine learning; W loc (x) is a local weighting function used to emphasize the importance of specific parts (such as IGBT modules, DC bus terminals, etc.), and is usually set to W. loc (x)>1, to amplify the deviation in critical areas; for minor parts, W can be taken. loc (x)<1;

[0124] α is the global adjustment factor. Used to control the proportion of the cubic error term in the overall metric, it can be set according to the sensitivity to the risk of temperature peaks. Cubic calculation is more sensitive to large deviations and can impose a higher penalty when there are extreme temperature changes.

[0125] The gradient of the difference field between the actual temperature and the predicted temperature in the spatial domain reflects the direction and rate of local temperature change.

[0126] ||·|| 2 For vectors The squared Euclidean norm is used to measure the magnitude of the gradient difference; β is the gradient difference weight. Ω inv This refers to the space area where key components such as inverters and motors are located.

[0127] If the control error term Λ err The larger the value of (t), the more serious the deviation between the prediction and the actual situation, requiring timely correction of the machine learning model or sensor calibration in the second step; if the control error term Λ err (t) If it remains within a preset small range, it is considered that the current prediction model and control measures are basically matched;

[0128] The optimal control parameters Π * (t), actual temperature distribution field Control error term Λ err (t) Control feedback information is transmitted to the vehicle thermal management module. At the same time, the battery status (temperature, remaining capacity, internal resistance, etc.) and the operating status of the cooling system (fan, water pump or liquid cooling device) are obtained from the battery management system (BMS) and marked as battery management system status information (BMS). info (t k );

[0129] To ensure a unified time base, this information is aligned on the same time axis t to form a multi-terminal integrated data packet:

[0130]

[0131] The data packet will serve as the basic input for global thermal management optimization decisions, providing complete status information;

[0132] During use, the real-time data from the inverter control terminal, battery management system (BMS), and cooling system are unified and encapsulated, facilitating comprehensive decision-making by the vehicle's thermal management module. This is achieved by comparing the actual temperature distribution field of the inverter. With the predicted temperature distribution field This provides new corrected data for the model in the second step, gradually reducing prediction errors, and allows for differentiated monitoring of different sensitive areas of the inverter (such as IGBT modules and DC bus contacts), greatly improving accuracy.

[0133] Step 402: Global Thermal Management Scheduling and Comprehensive Performance Optimization

[0134] Obtain the merged multi-terminal integrated data packet M(t) k Subsequently, the vehicle thermal management module and the battery management system (BMS) collaboratively execute a global schedule to maximize the vehicle's power performance and energy utilization efficiency while ensuring safe temperature limits. A global objective function is introduced as follows:

[0135]

[0136] Where: ΔT plan For the planning timeframe of thermal management; Ω A With Ω B These represent the weights for measuring vehicle driving performance and thermal safety, with values ​​ranging from 0 to 1;

[0137] C drive (D(τ)) represents the contribution to meeting driving needs (such as torque output and acceleration requirements);

[0138] The safety penalty function for overheating risk or low temperature degradation may also include factors such as heat dissipation power consumption and BMS preheating loss;

[0139] By analyzing the global objective function Π global Maximization allows for finding the optimal balance between vehicle power and temperature safety under different operating conditions; the global objective function Π is calculated. global After finding the optimal solution, the vehicle thermal management module sends scheduling instructions (such as fan speed and liquid cooling rate) to the cooling system and guides battery preheating or equalization operations on the battery management system (BMS) side.

[0140] For example, when the actual temperature distribution field of the inverter If the battery is close to the safety threshold but is at a relatively low temperature, the inverter's heat dissipation intensity can be increased first, and the battery can be preheated appropriately to avoid a double temperature bottleneck caused by the upcoming high load conditions. If the inverter temperature is low but the motor demand is high during the early start-up phase of the vehicle in a cold region, the thermal management module can reduce the power consumption of the heat dissipation device accordingly, or even use the small amount of heat generated by the motor to help accelerate the system to enter the operating temperature range, thereby avoiding efficiency loss caused by the motor and inverter temperatures being too low.

[0141] When using it, it is compared with the optimal control parameter Π *(t) Coupling can reduce the inverter output power and battery discharge power in the same direction when necessary, or reduce the peak torque within the vehicle speed limit, thereby completing the integrated design of vehicle energy consumption optimization and hardware safety protection;

[0142] The actual action feedback generated by the final execution of the scheduling command (including fan speed curve, liquid cooling flow rate, BMS heating power, etc.) is fed back in real time to the sensor fusion parameter correction and environmental identification update in step one, as well as the input correction and fine-tuning of the electro-thermal coupling prediction model in step two. This ensures that subsequent machine learning model updates can fully incorporate the temperature variation effects caused by the vehicle's cooling and preheating actions. If a significant deviation is found between the actual temperature response and the prediction after implementing the adaptive control strategy (which can be determined by the control error term Λ), the system will take action accordingly. err (t) If the monitoring results show that the scheduling strategy is dynamically adjusted, the global objective function Π is re-evaluated. global And make local corrections to maintain continuous system optimization;

[0143] When in use, the thermal states of the inverter and battery terminals are included in the same optimization objective to avoid excessive heat dissipation or preheating of a single system, which could lead to increased total energy consumption or insufficient power. The scheduling time range is ΔT. plan The objective function form can be adapted to different vehicle operation scenarios (long-distance highways, urban commuting) to enhance flexibility; by feeding back actual execution and temperature results to previous steps, the cyclical system from data acquisition to vehicle-wide scheduling is further improved, enabling the system to have self-iterative and self-correcting characteristics; in the global objective function Π global The contribution of C in meeting driving needs is also considered. drive (D(t)) and the safety penalty function A multi-objective optimization framework was constructed; through the bidirectional coordination of heat dissipation and preheating, proactive scheduling can be carried out under both high and low temperature extreme conditions, rather than simply managing the heat dissipation system or the preheating system separately.

[0144] Taking the deviation between the actual operation at the inverter and the preceding prediction as a starting point, a real-time multi-terminal linkage feedback mechanism was constructed, enabling the system to quickly grasp the real situation of temperature and power changes. This feedback is combined with the scheduling requirements of the BMS and cooling system, utilizing the global objective function Π global Multi-objective optimization is performed to coordinate vehicle thermal management and power output under extreme temperatures. Through this closed-loop collaboration, a complete technology chain is formed, which not only ensures the safety of the inverter and motor, but also provides the driver with better power and range performance in extreme environments, demonstrating the systematic advantages of this solution in efficient integration, proactive adaptation and precise control at the vehicle level.

[0145] Step 5: When the sensor or power device data matches the predicted temperature distribution field When a significant deviation occurs, a fault diagnosis and self-repair process is triggered, which fuses the multidimensional residual vector R of redundant measurement points through the observer. s (x,t) and the metric function Φ obs (t) Quantify the degree of anomaly and locate the failure site. If it is determined to be high-risk, then proceed according to the downgraded operating value Γ. degrade (t) Execute graded derating or enable temporary alternative signals to feed back fault information to the inverter controller and thermal management module in order to work together to reduce the impact of the fault;

[0146] Step five includes the following:

[0147] Step 501: Fault Identification and Anomaly Location

[0148] From the actual temperature distribution field obtained Extract redundant dataset S for cross-validation from readings of other key sensors. snes (t), for example, temperature observations at multiple measuring points around the same component or cross information from different signal channels;

[0149] Simultaneously, predicted values ​​for the same physical quantities, such as the predicted temperature distribution field, are obtained from the electro-thermal coupling prediction model.

[0150] R s (x,t) represents the vector form of the difference between the sensor's measured value and the predicted value. Specifically, it can include multi-dimensional components such as temperature, power, and magnetic flux attenuation rate, forming a residual signal.

[0151] To identify the root cause of the fault, a metric function Φ is defined that combines the observer (a state estimator based on the system model) and the residual. obs (t):

[0152]

[0153] Where: Ω sens Represents the sensor spatial domain; R s (x,t) is a multidimensional residual vector (e.g., temperature difference, power difference, magnetic flux difference) at position x and time t, which aggregates the multidimensional errors between the actual measured values ​​and the model or predicted values; ||·|| F The Frobenius norm of the matrix is ​​used to measure R. s (x),t)R s (x,t) T The overall amplitude; γ>1 is a higher-order amplification factor used to improve sensitivity to significant deviations;

[0154] When the metric function Φ obs(t) If it increases sharply and exceeds the threshold in a short period of time, it indicates that the corresponding sensor or power module data is suspected to be faulty or drifting; if it is significantly abnormal only in a few areas, it may indicate aging of local devices or loose connection.

[0155] For detected suspicious signals, the specific fault type is determined by comparing the cross-redundancy between different sensors and the prior knowledge of the physical model (such as the stator temperature should be below a certain threshold, the possible point of irreversible demagnetization of the magnet, etc.). That is, by using multi-dimensional residual pattern matching (rule base method) or machine learning classification, such as temperature sensor drift, IGBT gate fault, poor DC bus contact, etc. The preliminary judgment result is combined with the metric function Φ. obs (t) is passed to the next step to help subsequent execution of self-repair or degradation mechanisms;

[0156] When using it, through the multidimensional residual vector R s The vectorized structure of (x,t) can not only capture single temperature deviations but also identify inconsistencies between multiple signals such as power and magnetic flux, improving the comprehensiveness of fault detection. Based on higher-order operations, sudden or severe local deviations are not masked by average values, making it more sensitive to high-temperature hotspots or rapid aging. The observer is fused with the residual matrix R. s (x,t)R s (x,t) T By combining the Frobenius norm and higher-order power operations, a metric function Φ is obtained. obs Simultaneously evaluating the differences between multiple sensors and multiple physical quantities in (t) can maintain high recognition of complex failure modes.

[0157] Step 502: Self-repair and Degradation Operation Strategy

[0158] Based on the identified fault type and anomaly severity, a degraded operation value Γ is defined. degrade (t) indicates the extent to which the system may require partial de-rating or activation of backup parameters:

[0159] Γ degrade (t)=f deg (Φ obs (t), fault type, over-temperature / over-current margin)

[0160] Where f deg It is a function that converts the observer output and fault characteristics into executable scheduling instructions;

[0161] When the fault is severe, the operating value is downgraded to Γ. degrade If (t) increases, and the impact of the fault is limited, then only a minor temporary replacement should be performed;

[0162] For sensor drift or partial failure, redundant sensor data or the corrected value based on the prediction model in the second step can be temporarily used as a substitute input, so that the adaptive control logic in the third step can still operate smoothly.

[0163] When the downgraded operating value Γ degrade (t) If the preset warning line is exceeded, automatic downgrade operation will be executed;

[0164] The optimal control parameter Π in the third step * (t) adds a more conservative limit or lowers the power upper limit to reduce device thermal stress; in conjunction with the global thermal management in step four, if a fault causes a decrease in heat dissipation efficiency, the battery cooling or current diversion on the BMS side can be strengthened in advance; if a local high-temperature sensor fails, other measuring points and predictive models are used to maintain overall temperature monitoring; more specifically, based on the degraded operating value Γ degrade α classifies the faults and obtains:

[0165] Minor fault: Only the inverter peak current i max Reduce the frequency by 5% to 10% to decrease IGBT switching frequency fluctuations;

[0166] Moderate fault: Switch to backup sensor channel or enable predictive value replacement, and adjust bus voltage V. dc The upper limit is lowered by a certain percentage to protect key components;

[0167] Severe Fault: If a critical component is detected to be endangering safety, an emergency safety mode can be triggered, significantly reducing power output and prompting the driver to limit the speed or have it repaired as soon as possible;

[0168] Meanwhile, by recording the self-healing process (such as failure time, sensor replacement rate, aging characteristics, etc.), new abnormal data samples are incorporated when updating the machine learning model, continuously improving the ability to learn and warn of similar failure scenarios.

[0169] When using it, the downgraded running value Γ is used. degrade (t) Automatically adjust the operating mode so that even if some sensors or power devices fail, the core functions of the vehicle can still be maintained and secondary damage can be avoided. The degradation strategy is linked with the inverter adaptive control, BMS battery management and vehicle thermal management to ensure that a single point of failure will not quickly evolve into a systemic failure. The abnormal samples collected during the failure will feed back into the machine learning prediction model and sensor calibration model to gradually accumulate fault characteristics and repair experience and build a more complete closed loop.

[0170] A dual mechanism of self-repair and degraded operation is proposed: in the case of minor faults, prediction or redundant data is used for correction to avoid unnecessary overall power reduction; in the case of severe faults, the vehicle's thermal management and inverter power reduction are linked to balance safety and sustainable use. Through fault identification and anomaly localization in sub-step 501 and the self-repair and degraded operation strategy in sub-step 502, timely responses can be made to sensor drift, power device aging, or sudden faults under extreme temperature conditions. On the one hand, early warning and accurate localization are achieved by using observer and residual fusion methods; on the other hand, the basic functions of the inverter and the vehicle are maintained during the fault through degraded operation or backup data channels, and the fault characteristic data is fed back to the electro-thermal coupling prediction model in step two and the vehicle's thermal management in step four to continuously improve fault handling experience in subsequent operation. Thus, it can be organically combined with the previous four steps to enable the vehicle to balance temperature safety, system performance, and fault tolerance in extreme environments, further enhancing the robustness and practical value of the entire solution.

[0171] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0172] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0173] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0174] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0175] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An intelligent control method for inverters based on adaptive algorithm, characterized in that: Comprising, When detecting that the ambient temperature is significantly out of the preset range, integrating multi-source sensor readings, vehicle operating parameters and external weather data, generating a multi-source fusion temperature distribution field using a weighted positive definite kernel function with weight based on the spatial distance between sensors; After collecting the multi-source fusion temperature distribution field and vehicle operating conditions, constructing comprehensive feature items including the multi-source fusion temperature distribution field, external environment recognition results and vehicle operating state, calling a hybrid electric-thermal coupling prediction model with the comprehensive feature items as input and output, predicting temperature distribution field and power device thermal load prediction value at future time, and introducing physical residual constraint based on heat conduction equation and heat dissipation mechanism into the loss function of the prediction model to generate prediction quantities including at least predicted temperature distribution field and thermal load prediction value; If it is detected that the predicted temperature distribution field is about to exceed the temperature margin, the inverter control parameters are adaptively scheduled to actively limit the amplitude and reduce the power loss before the critical temperature, so as to maintain the power output and device safety; When the inverter output power and the actual temperature distribution field are transmitted back to the thermal management module, the BMS side temperature and heat dissipation information are compared with the predicted temperature distribution field and the electric-thermal coupling model is corrected, and then the fan or liquid cooling rate is dynamically scheduled, and the corrected data is fed back to the previous step; When it is monitored that the sensor values and the predicted temperature distribution field deviate beyond the expectation, the fault type is determined and the metric function is calculated, and then the hierarchical derating control and / or self-repairing processing are performed based on the degraded operating value, wherein the hierarchical derating control includes limiting or reducing the output current, switching frequency and / or bus voltage of the power device, and the self-repairing processing includes enabling redundant sensor data or a modified value based on the predicted output as an alternative input when a temperature sensor drift or local failure is detected, so that the inverter control system continues to operate in the degraded mode.

2. The adaptive algorithm-based intelligent inverter control method of claim 1, wherein: A multi-point temperature sensing network is laid out in the vehicle power system, the readings of each sensor are recorded and marked; According to the spatial position of each measurement point, the measurement results are mapped to the original temperature field, and based on the external environment information obtained during vehicle operation, the external temperature curve is collected and the environmental influence coefficient is constructed.

3. The adaptive algorithm-based intelligent inverter control method of claim 2, wherein: The environmental influence coefficient is coupled with the original temperature field to obtain an environment-corrected multi-source fusion temperature distribution field; The environment-corrected multi-source fusion temperature distribution field and the external environment recognition results are stored in a real-time database and interacted with a historical operating condition database.

4. The adaptive algorithm-based intelligent inverter control method of claim 2, wherein: The obtained multi-source fusion temperature distribution field, external environment recognition results, vehicle operating state, and related records in the historical operating condition database are synchronized to a feature construction module; The time fusion method with exponential decay kernel is used for feature extraction to generate comprehensive feature items for subsequent model; the sequence of comprehensive feature items between two historical time points is used to construct feature-label pairs required for supervised training together with target values.

5. The adaptive algorithm-based intelligent control method for inverters according to claim 4, characterized in that: According to the thermal physical properties of the electric vehicle inverter and the motor under extreme temperature, a hybrid electro-thermal coupling model with comprehensive feature items as input is introduced into the machine learning framework, which outputs the predicted temperature distribution field, thermal load and magnetic flux decay at future time points, and a loss function with physical constraints is introduced to train the electro-thermal coupling prediction model; The prediction effect is evaluated through offline verification and online updating modes, and periodic fine-tuning training is performed. After the electro-thermal coupling prediction model is evaluated and qualified, its prediction output is transmitted to the inverter control parameter optimization module.

6. The adaptive algorithm-based intelligent control method for inverters according to claim 5, characterized in that: The predicted output is compared with the designed upper limit of temperature safety and lower limit of magnetic flux to form temperature margin and magnetic flux margin; the current and next period driving intention of the vehicle is collected, and the existing inverter control parameter set is extracted at the same time. If the risk of over-temperature or insufficient magnetic flux is predicted, the corresponding power limiting strategy is selected according to the temperature margin and magnetic flux margin; The dynamic power limiting value is combined with the original control parameter set to obtain a parameter mapping with upper and lower limit constraints. If the temperature or magnetic flux risk is detected, the control parameters are dynamically limited or adjusted according to the dynamic power limiting value.

7. The adaptive algorithm-based intelligent control method for inverters according to claim 6, characterized in that: According to the determined parameter mapping, the online solution of each control parameter is realized in the adaptive algorithm built in the inverter controller to obtain the optimal control parameters for the execution end, wherein: The prediction output is included in the objective function to optimize the control parameters for the future period, obtain the optimal control parameters varying with time, and construct a dynamic control strategy and finally execute it in combination with the parameter mapping. If the actual temperature or magnetic flux demagnetization rate deviates from the expected value, the online prediction update mechanism is automatically called for online fine-tuning or incremental training in the next control period to correct the input of the electro-thermal coupling prediction model and the upper limit of the control parameters.

8. The adaptive algorithm-based intelligent control method for inverters according to claim 7, characterized in that: The actual temperature distribution field and power flow information of the inverter after executing the dynamic control strategy and the updated optimal control parameters are collected, and the inverter end data is compared with the predicted temperature distribution field to evaluate the deviation between prediction and reality through the control error term; If the control error term value is greater than expected, the machine learning model or sensor calibration is corrected; The optimal control parameters, actual temperature distribution field and control error term are transmitted to the vehicle thermal management module, and the battery state and operating state of the heat dissipation system are obtained from the BMS side, which are marked as battery management system state information. After alignment, a comprehensive data package is formed.

9. The adaptive algorithm-based intelligent control method for inverters according to claim 8, characterized in that: After obtaining the combined multi-terminal comprehensive data packet, the vehicle thermal management module cooperates with the BMS to perform a global scheduling; After calculating the optimal solution of the introduced global objective function, the vehicle thermal management module sends a scheduling instruction to the heat dissipation system and guides the battery preheating or equalization operation on the BMS side.

10. The adaptive algorithm-based intelligent control method for inverters according to claim 9, characterized in that: The actual action resulting from the final execution of the scheduling instruction is fed back to the real-time feedback of the sensor fusion parameter correction and environment recognition update, as well as the input correction and fine-tuning of the electric-thermal coupling prediction model; If it is found that there is a significant deviation between the actual temperature response and the prediction of the adaptive control strategy, the scheduling strategy is dynamically adjusted, the global objective function is re-evaluated and locally corrected.

11. The adaptive algorithm-based intelligent control method for inverters according to claim 10, characterized in that: After obtaining the actual temperature distribution field and other key sensor readings and corresponding predicted values, a metric function based on observer and residual fusion is defined, and when the metric function sharply increases and exceeds the threshold value within a short time, the specific fault type is determined by multi-dimensional residual pattern matching for the detected suspicious signal.

12. The adaptive algorithm-based intelligent control method for inverters according to claim 10, characterized in that: According to the identified fault type and abnormality degree, a degradation operation value representing the amplitude of partial derating or backup parameter activation is defined; When the fault is serious, the degradation operation value increases, and if the fault impact is limited, only a slight temporary replacement treatment is performed; for sensor drift or local failure, temporarily enable redundant sensor data or correction values based on predicted output as alternative input; when the degradation operation value exceeds the preset warning line, automatic degradation operation is performed.

Images