Electronic water pump service life prediction method and electronic water pump

Abstract

The invention provides a life prediction method of an electronic water pump and the electronic water pump. The life prediction method comprises the steps of obtaining real-time operation data of the electronic water pump; processing the real-time operation data according to a preset feature extraction rule to generate a health feature vector; inputting the health feature vector into a pre-trained fault diagnosis model, and outputting a fault detection result of the electronic water pump; determining a target performance degradation algorithm corresponding to the fault detection result; and calculating the health feature vector and the historical operation data of the electronic water pump by using a pre-trained life prediction model based on a target decline algorithm, and outputting a life prediction result of the electronic water pump. In the mode, the fault state of the electronic water pump can be identified, and the residual life of the electronic water pump can be predicted based on the determined fault state, so that the calculation logic of the residual life is consistent with the current actual physical degradation process of the electronic water pump, and the accuracy of predicting the residual useful life of the electronic water pump in a complex fault mode is improved.

Description

Technical Field

[0001] This application relates to the field of electronic water pump technology, and in particular to a method for predicting the lifespan of an electronic water pump and an electronic water pump. Background Technology

[0002] With the development of new energy vehicle technology, the electronic water pump, as a core power component of the thermal management system, directly determines the operating temperature of the engine, battery, and motor control system, as well as the overall vehicle safety, based on its operational reliability. Currently, the electronic water pump mainly adjusts the motor speed through a controller to meet the vehicle's cooling circulation requirements.

[0003] However, existing maintenance strategies for electric water pumps primarily rely on reactive repairs or preventative replacements based on fixed mileage / time intervals, lacking the ability to perceive the actual health status of components in real time. Regarding fault monitoring, current technologies typically only detect serious faults such as motor stall or open circuits based on single electrical parameters, making it difficult to detect hidden mechanical and hydraulic performance degradation in its early stages, such as impeller cavitation, bearing wear, or flow channel blockage. Furthermore, electric water pumps cannot differentiate the varying impacts of different failure modes on their lifespan, making accurate prediction of remaining lifespan difficult. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a method for predicting the lifespan of an electronic water pump and an electronic water pump. By fusing the collection of electrical and hydraulic parameters and using fault detection results to match the target performance degradation algorithm, the fault state of the electronic water pump can be identified, and the remaining lifespan of the electronic water pump can be accurately predicted based on the determined fault state. This ensures that the calculation logic of the remaining lifespan is consistent with the actual physical degradation process of the electronic water pump, thereby improving the accuracy of the remaining useful life prediction of the electronic water pump under complex fault modes.

[0005] In a first aspect, the present invention provides a method for predicting the lifespan of an electronic water pump, comprising: Acquire real-time operating data of the electronic water pump; real-time operating data includes electrical parameters and hydraulic parameters.

[0006] The real-time operating data is processed according to the preset feature extraction rules to generate a health feature vector; the health feature vector is used to characterize the current operating status of the electronic water pump.

[0007] The health feature vector is input into a pre-trained fault diagnosis model, and the fault detection result of the electronic water pump is output; the fault detection result indicates whether the electronic water pump is in a fault-free state or a faulty state.

[0008] Based on the pre-defined mapping relationship between fault detection results and performance degradation algorithms, the target performance degradation algorithm corresponding to the fault detection results is determined.

[0009] Using a pre-trained lifespan prediction model and based on the target decay algorithm, the health feature vector and historical operating data of the electronic water pump are calculated to output the lifespan prediction results of the electronic water pump.

[0010] In an optional implementation, the electronic water pump includes a current sensor, a position sensor, and a miniature differential pressure sensor; the step of acquiring real-time operating data of the electronic water pump includes: The phase current signal of the motor of the electronic water pump is acquired by a current sensor.

[0011] The rotor position signal of the motor is acquired by a position sensor, and the real-time speed of the motor is calculated.

[0012] The pressure difference signal between the inlet and outlet of the electronic water pump is measured in real time using a miniature differential pressure sensor.

[0013] The phase current signal and differential pressure signal are synchronously sampled and denoised to obtain real-time operating data.

[0014] In an optional implementation, the step of processing real-time running data according to preset feature extraction rules to generate a health feature vector includes: The phase current signal is transformed in the frequency domain to extract the current harmonic distortion rate and spectral energy value as electrical characteristic components.

[0015] Based on the real-time rotational speed, the pre-stored standard differential pressure-flow characteristic curve is invoked to calculate the deviation between the measured value of the differential pressure signal and the corresponding theoretical value in the standard differential pressure-flow characteristic curve, which is then used as a hydraulic characteristic component.

[0016] The electrical and hydraulic feature components are normalized and their data are concatenated to generate a health feature vector.

[0017] In an optional implementation, the fault diagnosis model is trained in the following manner: The sample health feature vectors corresponding to various preset operating states are used as input, and the true state category labels corresponding to the sample health feature vectors are used as output. The preset initial classification network model is trained until the error between the output of the initial classification network model and the true state category label meets the preset convergence condition, thus obtaining the trained fault diagnosis model.

[0018] In an optional implementation, the fault state includes at least one of the following: impeller cavitation state, flow channel blockage state, bearing wear state, and stator winding aging state.

[0019] The steps for determining the target performance degradation algorithm corresponding to the fault detection result based on the preset mapping relationship between fault detection results and performance degradation algorithms include: When the fault detection result is impeller cavitation or flow channel blockage, the corresponding target performance degradation algorithm is determined to be the preset performance degradation algorithm.

[0020] When the fault detection result indicates bearing wear, the corresponding target performance degradation algorithm is determined to be the preset mechanical fatigue wear algorithm.

[0021] When the fault detection result is that the stator winding is in an aging state or has no fault, the corresponding target performance degradation algorithm is determined to be the preset aging algorithm.

[0022] In an optional implementation, the step of using a pre-trained lifespan prediction model, based on a target decay algorithm, to calculate the health feature vector and historical operating data of the electric water pump, and outputting the lifespan prediction result of the electric water pump includes: Acquire historical operating data of the electronic water pump; historical operating data includes the health feature vector sequence and cumulative operating time at historical moments.

[0023] The health degradation index at the current moment is calculated based on the target degradation algorithm.

[0024] By using health degradation indicators as constraints to input into the life prediction model, the model fits and extrapolates the time-series evolution trend of historical operating data, calculates the remaining time required for the performance indicators of the electronic water pump to decline to the preset failure threshold, and determines the remaining time as the life prediction result.

[0025] In an optional implementation, the lifetime prediction model is trained in the following manner: The pre-acquired sample feature vector sequence containing historical decay trend information is used as input, and the actual remaining useful life corresponding to the sample feature vector sequence is used as output. The preset initial time series prediction model is trained until the loss function of the initial time series prediction model satisfies the preset minimization condition, and the trained life prediction model is obtained.

[0026] In an optional implementation, the method further includes: The fault detection results and life prediction results are encapsulated into status messages.

[0027] Status messages are sent to the vehicle controller via the vehicle's communication bus.

[0028] In an optional implementation, the method further includes: When the fault detection result indicates impeller cavitation and the life prediction result is lower than the preset warning value, the target speed of the motor controlling the electronic water pump is lower than the preset safe speed threshold.

[0029] In a second aspect, the present invention provides an electronic water pump, including a controller, and further including a pumping component and a sensing component connected to the controller; the controller is used to execute the life prediction method of the electronic water pump according to any of the foregoing embodiments.

[0030] The pumping assembly includes a pump casing, an impeller housed within the pump casing, and a brushless motor that drives the impeller to rotate.

[0031] The sensing components include current sensors and position sensors for acquiring electrical parameters, and miniature differential pressure sensors for acquiring hydraulic parameters.

[0032] In an optional embodiment, the pump casing is provided with a first pressure-inlet hole and a second pressure-inlet hole. The first pressure-inlet hole is connected to the low-pressure area at the inlet of the impeller, and the second pressure-inlet hole is connected to the high-pressure area at the outlet of the impeller.

[0033] The pump casing is equipped with microchannels, and a miniature differential pressure sensor is airtightly connected to the first pressure inlet and the second pressure inlet through the microchannels to measure the pressure difference before and after the impeller.

[0034] In an optional implementation, the electric water pump also includes a communication interface.

[0035] The controller is configured to connect to the vehicle's onboard network system via a communication interface.

[0036] This application provides a method for predicting the lifespan of an electronic water pump and an electronic water pump. By acquiring real-time operating data including electrical and hydraulic parameters and matching a target performance degradation algorithm based on fault detection results, the current operating state of the electronic water pump can be accurately identified and a calculation strategy that conforms to the current physical degradation mechanism can be invoked. This enables high-precision prediction of the remaining useful life based on the evolution law of specific fault modes and triggers active protection when necessary, thereby improving the operational reliability of the electronic water pump and ensuring the continuous and stable operation of the vehicle thermal management system.

[0037] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the application. The objectives and other advantages of this application are realized and obtained through the structures particularly pointed out in the description, claims and drawings.

[0038] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0040] Figure 1 A flowchart illustrating the lifespan prediction method for an electronic water pump provided in this application embodiment; Figure 2 A flowchart of a real-time runtime data acquisition method provided in this application embodiment; Figure 3 A flowchart of the health feature vector generation method provided in this application embodiment; Figure 4 A flowchart of the lifetime prediction result generation method provided in the embodiments of this application; Figure 5 A schematic diagram of an electronic water pump provided in an embodiment of this application.

[0041] Diagram: 1-Controller; 2-Pumping assembly; 3-Sensing assembly. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0043] To help those skilled in the art better understand this application, a brief introduction to its application scenarios and design concepts is provided.

[0044] In the thermal management system of new energy vehicles, the electronic water pump, as the core power component of the cooling cycle, directly affects the safety of the engine and the three-electric system (battery, motor, and electronic control). However, existing brushless electronic water pumps typically function only as actuators, possessing basic speed regulation capabilities but lacking a deep understanding of their own health status. During long-term operation, water pumps are highly susceptible to failures such as impeller cavitation, bearing wear, mechanical seal leakage, or stator winding aging. These failures are highly insidious in their early stages, difficult to detect using only current monitoring, and are often only discovered after complete failure (such as jamming or failure to pump water) leading to system overheating, resulting in significant economic losses. Furthermore, existing maintenance strategies often employ high-risk reactive repairs or uneconomical fixed-mileage preventative replacements, failing to accurately assess the remaining value of components and hindering scientific predictive maintenance.

[0045] Based on this, embodiments of this application provide a life prediction method and an electronic water pump, which can effectively distinguish between failure modes with different mechanisms and achieve early and accurate warnings. Furthermore, it can utilize a built-in life prediction model to quantify and calculate the remaining useful life based on the decay patterns of specific faults, thereby transforming traditional passive maintenance into state-based proactive predictive maintenance, improving the operational reliability and energy efficiency of the vehicle's thermal management system.

[0046] To facilitate understanding of this embodiment, the embodiments of this application will be described in detail below.

[0047] This application provides a method for predicting the lifespan of an electronic water pump, referring to... Figure 1 Methods for predicting the lifespan of electronic water pumps include: Step S101: Obtain real-time operating data of the electronic water pump; real-time operating data includes electrical parameters and hydraulic parameters.

[0048] Here, the electronic water pump collects raw signals during the pump's operation at a preset sampling frequency through built-in or external sensing components.

[0049] Electrical parameters primarily reflect the electromagnetic drive state of the motor. Specific electrical parameters may include phase currents (such as three-phase currents or bus currents), bus voltages, motor power, and the temperature of power devices in the motor control circuit. In one implementation, phase current signals can be acquired using a sampling resistor or Hall sensor connected in series in the drive circuit, and voltage signals can be acquired using a voltage divider circuit.

[0050] Hydraulic parameters primarily reflect the load characteristics and working state of a water pump as a fluid machine. Specific hydraulic parameters may include the pressure difference between the inlet and outlet of the electric water pump, the flow rate of the coolant flowing through the pump, the fluid temperature, or the absolute pressure within the pipeline.

[0051] A miniature differential pressure sensor can be integrated inside the electronic water pump to directly measure the pressure difference before and after the impeller as a key hydraulic parameter.

[0052] If no direct fluid sensor is installed, the current flow rate or pressure difference can be estimated based on the motor's speed, current, and power using sensorless observer algorithms (such as Kalman filtering or flux linkage observers), and the estimated value can be used as hydraulic parameters.

[0053] In addition, real-time operating data may also include mechanical parameters, such as rotor position signals collected by position sensors (such as Hall position sensors or magnetic encoders), and the real-time speed, angular acceleration or moment of inertia of the motor calculated accordingly.

[0054] Step S102: Process the real-time operating data according to the preset feature extraction rules to generate a health feature vector; the health feature vector is used to characterize the current operating status of the electronic water pump.

[0055] Here, the preset feature extraction rules may include signal preprocessing, such as denoising the original signal, filtering (e.g., low-pass filtering to remove high-frequency interference), and time axis alignment.

[0056] Frequency domain analysis methods can be used to extract features from electrical parameters. For example, fast Fourier transform or wavelet packet decomposition can be performed on the phase current signal to extract the fundamental amplitude, total harmonic distortion (THD), and the amplitude or energy percentage of specific harmonics (such as the 3rd, 5th, or 7th harmonics). These features can effectively reflect whether there is an inter-turn short circuit in the motor windings, whether the magnets are demagnetized, or whether the bearings are experiencing eccentric vibration.

[0057] For feature extraction of hydraulic parameters, correlation analysis methods can be used. For example, by combining the current real-time rotational speed with a pre-calibrated speed-standard pressure difference characteristic curve, the deviation between the current measured pressure difference value and the theoretical standard value, as well as the rate of change of the deviation, can be calculated. The deviation can intuitively characterize whether the impeller has experienced cavitation, corrosion damage, or whether the flow channel is blocked by foreign objects.

[0058] To eliminate the influence between different units, it is usually necessary to normalize or standardize each feature component before generating the vector. Redundant information can also be removed by dimensionality reduction algorithms such as principal component analysis, and finally, a high-dimensional or low-dimensional health feature vector containing multiple state information is generated.

[0059] Step S103: Input the health feature vector into the pre-trained fault diagnosis model and output the fault detection result of the electronic water pump; the fault detection result indicates whether the electronic water pump is in a fault-free state or a faulty state.

[0060] Here, fault detection results not only include simple binary judgments of normal / abnormal, but also specific fault mode classifications. For example, the results can characterize the electric water pump as being in one of the following states: fault-free (healthy), impeller cavitation, flow channel blockage, bearing wear, stator winding insulation aging, or rotor demagnetization. Fault detection results can also include the confidence probability of the fault occurring (e.g., 90% probability of cavitation).

[0061] Step S104: Determine the target performance degradation algorithm corresponding to the fault detection result based on the preset mapping relationship between the fault detection result and the performance degradation algorithm.

[0062] Here, the preset mapping relationship is a logical lookup table or rule base used to match the most suitable performance degradation algorithm based on the diagnosed fault.

[0063] Step S105: Using a pre-trained lifespan prediction model and based on the target decay algorithm, calculate the health feature vector and the historical operating data of the electronic water pump, and output the lifespan prediction result of the electronic water pump.

[0064] Here, lifetime prediction models typically employ algorithmic architectures for processing time series data, such as Long Short-Term Memory Networks, Gated Recurrent Units, Recurrent Neural Networks, or various variants based on particle filtering and Kalman filtering.

[0065] The input historical operating data refers to the health characteristic vector sequence, cumulative operating time, cumulative number of starts, or historical load spectrum of the electronic water pump over a period of time. This data records the historical trajectory of the water pump's performance evolution over time.

[0066] The method of calculation based on the target degradation algorithm can be to use the theoretical degradation rate (or physical damage factor) calculated by the determined target performance degradation algorithm as an additional feature input to the neural network model such as the long short-term memory network model, so as to guide the neural network to fit the degradation curve more accurately.

[0067] The lifetime prediction model itself contains multiple sub-models (such as cavitation prediction sub-network and wear prediction sub-network), and can also directly activate the sub-model corresponding to the target performance degradation algorithm to extrapolate from historical data.

[0068] Life prediction results are quantitative indicators and can be at least one of the following: remaining useful life, health index, and failure probability curve.

[0069] The remaining useful life is the estimated time (in hours or days) remaining from the current moment until the electric pump fails (such as completely jammed or under-headed).

[0070] The health index is a normalized value (such as 0-100 or 0-1), which intuitively displays the current health percentage.

[0071] The failure probability curve represents the probability distribution of failure occurring at different points in the future.

[0072] In one embodiment, the electronic water pump includes a current sensor, a position sensor, and a miniature differential pressure sensor.

[0073] Reference Figure 2 Step S101 includes the following steps S201-S204.

[0074] Step S201: Acquire the phase current signal of the motor of the electronic water pump through a current sensor.

[0075] Here, the current sensor is integrated into the power drive circuit within the intelligent controller of the electric water pump. The current sensor can be a sampling resistor connected in series in the lower arm of a three-phase inverter bridge, using Ohm's law to measure the current flowing through the motor windings. Alternatively, a non-contact Hall effect current sensor can be used to reduce circuit losses.

[0076] The data collected is typically the three-phase current or DC bus current of a brushless DC motor. The controller reads the instantaneous current values ​​at a high-frequency sampling rate (e.g., 10kHz-20kHz) to capture dynamic load changes during motor operation. The phase current signal contains rich fault characteristic information; for example, when impeller cavitation or mechanical jamming occurs, the current waveform will exhibit specific distortions or amplitude fluctuations. When an inter-turn short circuit occurs in the winding, the imbalance of the three-phase current will increase significantly.

[0077] In step S202, the rotor position signal of the motor is acquired by the position sensor, and the real-time speed of the motor is calculated.

[0078] Here, a position sensor is used to monitor the physical angular position of the motor rotor in real time. In specific implementations, the position sensor can consist of three switching-type Hall elements evenly distributed on the stator circumference, or it can employ a high-precision magnetic encoder or rotary transformer. By reading the pulse signal or absolute angle value from the position sensor, the rotor phase is determined, thereby enabling precise electronic commutation in conjunction with the power drive unit.

[0079] Based on the obtained rotor position signal, the real-time speed of the motor is obtained by calculating the rate of change of position per unit time (i.e., taking the derivative of the position signal).

[0080] Step S203: The pressure difference signal between the inlet and outlet of the electronic water pump is measured in real time using a miniature differential pressure sensor.

[0081] Here, the miniature differential pressure sensor is preferably manufactured using MEMS (Micro-Electro-Mechanical Systems) technology.

[0082] In terms of physical structure, the electric water pump has a dedicated pressure-guiding microchannel designed inside the pump casing or controller housing. One end of the microchannel is connected to the low-pressure chamber at the impeller inlet, and the other end is connected to the high-pressure chamber at the impeller outlet. A miniature differential pressure sensor is airtightly connected between the inlet and outlet through these microchannels, directly measuring and outputting a differential pressure signal representing the actual head capacity.

[0083] By directly measuring the differential pressure signal, it is possible to distinguish between load changes caused by mechanical wear and load changes caused by changes in fluid system resistance (such as pipeline blockage).

[0084] Step S204: Synchronously sample and denoise filter the phase current signal and differential pressure signal to obtain real-time operating data.

[0085] Here, synchronous sampling means that the controller ensures that the current signal, position signal, and differential pressure signal are aligned on the time axis.

[0086] Noise reduction and filtering include hardware filtering and software filtering. Hardware filtering uses an RC (resistor-capacitor) low-pass filter to remove high-frequency spikes. Software filtering can employ moving average filtering, Kalman filtering, or wavelet thresholding algorithms to eliminate measurement noise and outliers. The processed real-time operating data is smoother and cleaner, more accurately reflecting the physical operating status of the electronic water pump.

[0087] In one embodiment, reference is made to Figure 3 Step S102 includes the following steps S301-S303.

[0088] Step S301: Perform frequency domain transformation on the phase current signal to extract the current harmonic distortion rate and spectral energy value as electrical characteristic components.

[0089] Here, since motor faults often introduce noise or distortion at specific frequencies into the current signal, simple time-domain analysis is insufficient to capture these subtle changes; therefore, frequency-domain analysis is employed. The controller utilizes its built-in digital signal processing unit or hardware accelerator to perform a Fast Fourier Transform on the acquired time-domain phase current signal.

[0090] Frequency domain transformation can decompose the current signal into fundamental and harmonic components. Based on the transformation results, the current harmonic distortion rate is first calculated. The current harmonic distortion rate reflects the overall distortion degree of the current waveform and is used to assess whether there are electrical faults such as insulation aging or inter-turn short circuits in the stator winding.

[0091] Simultaneously, by extracting the spectral energy values ​​within specific frequency bands, different mechanical faults can excite different characteristic frequencies in the current spectrum. For example, when impeller cavitation occurs, the impact of bubble collapse significantly increases the energy of specific high-frequency components in the current spectrum. When bearing wear occurs, it may lead to an increase in low-frequency vibration energy in the current spectrum.

[0092] Step S302: Based on the real-time rotational speed, call the pre-stored standard differential pressure-flow characteristic curve, calculate the deviation between the measured value of the differential pressure signal and the corresponding theoretical value in the standard differential pressure-flow characteristic curve, and use it as a hydraulic characteristic component.

[0093] Here, standard differential pressure-flow characteristic curves or differential pressure-speed characteristic data of the electronic water pump under ideal healthy conditions are pre-stored. The standard differential pressure-flow characteristic curve describes the standard head or differential pressure range that the water pump should have at different speeds.

[0094] During operation, based on the calculated real-time rotational speed, the corresponding theoretical standard differential pressure value at that speed is retrieved from a table or through function fitting. Subsequently, the measured value of the differential pressure signal obtained by the miniature differential pressure sensor is compared with the theoretical value, and the difference between the two, i.e., the deviation, is calculated.

[0095] Deviation is a core indicator for judging hydraulic faults. For example, when impeller cavitation or mechanical seal leakage occurs, the measured differential pressure value is usually significantly lower than the range of the standard curve, resulting in a negative deviation. When the flow channel is severely blocked, the differential pressure characteristics will also deviate significantly. Using the deviation and its rate of change as hydraulic characteristic components, the degree of pump damage can be intuitively quantified.

[0096] Step S303: Normalize the electrical feature components and hydraulic feature components and stitch the data together to generate a health feature vector.

[0097] Here, the normalization method can employ min-max normalization to map the data to the [0,1] interval, or Z-score standardization (zero-mean standardization) to transform it into a distribution with a mean of 0 and a variance of 1. After processing, all standardized feature components are concatenated in a preset order to form a multidimensional health feature vector 10101010. The health feature vector integrates the electrical, mechanical, and hydraulic status information of the electric water pump.

[0098] In one embodiment, the fault diagnosis model in step S103 is trained in the following manner: The sample health feature vectors corresponding to various preset operating states are used as input, and the true state category labels corresponding to the sample health feature vectors are used as output. The preset initial classification network model is trained until the error between the output of the initial classification network model and the true state category label meets the preset convergence condition, thus obtaining the trained fault diagnosis model.

[0099] Here, a sample dataset is constructed, and the data in the sample dataset typically comes from bench tests of electric water pumps or high-precision simulation environments. In the laboratory environment, technicians simulate various preset operating states that the electric water pump may encounter throughout its entire life cycle by controlling the operating parameters of the test bench.

[0100] Multiple preset operating states cover the normal operating range of the electric water pump and various typical failure modes. Specifically, these include: 1. Fault-free state: Normal operating state under standard voltage, temperature and fluid load.

[0101] 2. Impeller cavitation: Cavitation phenomena of varying degrees induced by reducing inlet pressure or adjusting fluid temperature.

[0102] 3. Flow channel blockage state: The situation of foreign objects blocking the flow channel is simulated by setting obstacles in the pipeline.

[0103] 4. Bearing wear condition: Mechanical fatigue is simulated by installing bearing components with different degrees of wear.

[0104] 5. Stator winding aging condition: Electrical aging is reproduced by simulating inter-turn short circuits or reducing the insulation level.

[0105] In each of the above states, electrical and hydraulic parameters are simultaneously collected and processed according to feature extraction rules to generate corresponding sample health feature vectors. At the same time, each set of sample health feature vectors is labeled with a corresponding real-state category label (e.g., label 0 represents normal, label 1 represents cavitation, label 2 represents wear, etc.), thus forming a supervised training dataset containing feature-label pairs.

[0106] Depending on actual computing resources and accuracy requirements, the preset initial classification network model can be a random forest model, a support vector machine, or a one-dimensional convolutional neural network. A lightweight one-dimensional convolutional neural network or a simplified random forest algorithm is preferred. The number of input layer nodes in the initial classification network model matches the dimension of the healthy feature vector, and the number of output layer nodes corresponds to the preset number of running state categories (e.g., 5 categories).

[0107] The prepared sample health feature vectors are input in batches into the initial classification network model. After forward propagation, the model outputs the predicted classification result (i.e., the predicted probability distribution). Subsequently, the error or loss function value (such as cross-entropy loss) between the predicted result and the true state class label is calculated.

[0108] Based on this error, the weight parameters within the initial classification network model are updated and adjusted using backpropagation or gradient descent algorithms to minimize the difference between the predicted results and the true labels. This process is repeated, continuously refining the model parameters using new sample data.

[0109] Training continues until the error between the output of the initial classification network model and the true state class label meets a preset convergence condition. The preset convergence condition may include the model's classification accuracy on the validation set reaching a preset threshold (e.g., above 95%), or the loss function value no longer significantly decreasing after several consecutive iterations (i.e., model convergence).

[0110] When the above conditions are met, training stops, and the model at this point is the trained fault diagnosis model. The fault diagnosis model has the ability to automatically summarize fault patterns from multi-dimensional features and can accurately distinguish between the high-frequency energy characteristics of cavitation and the low-frequency vibration characteristics of wear.

[0111] In one embodiment, the fault state includes at least one of impeller cavitation, flow channel blockage, bearing wear, and stator winding aging.

[0112] Step S104 includes: When the fault detection result is impeller cavitation or flow channel blockage, the corresponding target performance degradation algorithm is determined to be the preset performance degradation algorithm.

[0113] Here, when the fault detection result indicates impeller cavitation or flow channel blockage, the main driving force for failure originates from fluid dynamics factors. Impeller cavitation refers to the shock waves generated by the collapse of bubbles in the fluid continuously eroding the impeller surface, causing microscopic damage to the material. Flow channel blockage alters the system's resistance characteristics, causing the pump to operate under off-rated conditions for extended periods.

[0114] The pre-defined performance degradation algorithm focuses on calculations using hydraulic characteristic components. For example, it considers the deviation between the pump inlet / outlet pressure difference and the flow rate. For cavitation faults, the algorithm calculates the intensity and duration of cavitation, using cumulative damage models (such as variants based on the Palmgren-Miner theorem) to assess the impeller's mass loss or the rate of structural strength reduction. For blockage faults, the algorithm assesses the additional motor load due to increased resistance, and then estimates the time point at which performance degrades to the point where cooling requirements cannot be met.

[0115] When the fault detection result indicates bearing wear, the corresponding target performance degradation algorithm is determined to be the preset mechanical fatigue wear algorithm.

[0116] Here, when the fault detection result indicates bearing wear, the main driving force for failure originates from mechanical friction and contact fatigue. As the core component supporting the rotor, bearing wear can lead to rotor eccentricity, increased vibration, and ultimately, seizure.

[0117] The pre-defined mechanical fatigue wear algorithm focuses on calculations using the spectral energy values ​​in electrical characteristic components. Since bearing wear introduces specific low- or mid-frequency vibration signals into the current spectrum, the algorithm employs a physics-based fatigue life model (such as the Weibull distribution model or the Paris fatigue crack propagation model). The algorithm uses the current vibration energy level as stress input to extrapolate the remaining fatigue life of the bearing balls or raceways, predicting the operating time required for the wear to progress from the current wear state to complete seizure.

[0118] When the fault detection result is that the stator winding is in an aging state or has no fault, the corresponding target performance degradation algorithm is determined to be the preset aging algorithm.

[0119] Here, when the fault detection result indicates that the stator winding is in an aging state or a fault-free state, the main driving force for failure comes from thermal stress and time. Even if the water pump is in a fault-free normal operating state, or only shows slight stator winding aging, its insulation system (enameled wire, insulating paper, etc.) will gradually degrade over time and with cyclical temperature changes until insulation breakdown or short circuit occurs.

[0120] Preset aging algorithms are typically built upon the Arrhenius equation. These algorithms focus on the cumulative operating time and operating temperature load of the electric water pump. They calculate the cumulative thermal aging losses by statistically analyzing the pump's residence time in different temperature ranges. For a fault-free state, the algorithm calculates the electric water pump's natural theoretical lifespan. For winding aging, the algorithm accelerates the thermal aging rate based on detected harmonic distortion rates (reflecting inter-turn short circuits or insulation degradation), thus determining the remaining safe operating time under the current thermal load.

[0121] In one embodiment, reference is made to Figure 4 Step S105 includes the following steps S401-S403.

[0122] Step S401: Obtain historical operating data of the electronic water pump; historical operating data includes the health feature vector sequence and cumulative working time at historical moments.

[0123] Here, historical operating data of the electronic water pump since it was put into use is stored in advance.

[0124] A historical health feature vector sequence refers to a set of health feature vectors arranged in chronological order. The current health feature vector is recorded every preset time interval (e.g., every hour or at the end of each trip), thus forming time series data reflecting the evolution of performance over time.

[0125] Cumulative operating time refers to the total effective operating time of the electric water pump since it left the factory. In addition, historical operating data may also include statistical information such as the cumulative number of start-stop cycles and the percentage of operating time under extreme temperatures.

[0126] Step S402: Calculate the health degradation index at the current moment based on the target degradation algorithm.

[0127] Here, based on the determined fault type, the corresponding target decay algorithm is invoked to evaluate the current state.

[0128] Health degradation indexes are dimensionless values ​​used to quantify the health status of electronic water pumps. For example, a brand-new condition at the time of manufacture can be defined as 1 (or 100%), and a completely failed condition can be defined as 0 (or 0%).

[0129] If the target degradation algorithm is the cavitation cumulative damage algorithm, it will calculate the current mass loss rate or hydraulic efficiency reduction percentage of the impeller using the physical damage formula based on the current pressure difference offset and cavitation duration, and convert it into the current health degradation index (e.g., 0.85).

[0130] If the target degradation algorithm is a natural aging algorithm, it will calculate the theoretical life loss of the insulation material based on the cumulative working time and the Arrhenius model, and obtain the current health degradation index.

[0131] Health degradation indicators reflect how far an electronic water pump is from becoming obsolete at the current moment.

[0132] Step S403: Input the health degradation index as a constraint into the life prediction model, fit and extrapolate the time-series evolution trend of historical operating data, calculate the remaining time required for the performance index of the electronic water pump to decline to the preset failure threshold, and determine the remaining time as the life prediction result.

[0133] Here, the lifetime prediction model preferably uses a time series prediction network with memory function, such as a long short-term memory network or a gated recurrent unit.

[0134] The acquired health feature vector sequence at historical moments is input into the lifespan prediction model. By learning from historical data, the lifespan prediction model fits the temporal evolution trend of the electric water pump's performance degradation (i.e., the past degradation curve).

[0135] The health degradation index is input as a constraint into the model. This constraint can be used as the initial state input of the model or as part of the loss function to correct the model's prediction of future trends, ensuring that the predicted degradation trajectory conforms to the actual physical laws of cavitation, wear, or aging.

[0136] After fitting historical trends and being constrained by the current physical state, the lifetime prediction model extrapolates to the future timeline to simulate the future trajectory of the electric water pump's performance indicators.

[0137] Set a preset failure threshold (e.g., when the health index drops to 0.3 or 0.2). When the performance index trajectory curve predicted by the model reaches the preset failure threshold, the time difference between that moment and the current moment is the remaining useful life.

[0138] The remaining useful life is determined as a life prediction result and output to the vehicle control system or user in order to schedule maintenance.

[0139] In one embodiment, the lifetime prediction model in step S105 is trained in the following manner: The pre-acquired sample feature vector sequence containing historical decay trend information is used as input, and the actual remaining useful life corresponding to the sample feature vector sequence is used as output. The preset initial time series prediction model is trained until the loss function of the initial time series prediction model satisfies the preset minimization condition, and the trained life prediction model is obtained.

[0140] Here, the sample feature vector sequence containing historical degradation trend information comes from the full life cycle bench test or accelerated aging test of the electronic water pump. In the test environment, technicians simulate the long-term operation of the electronic water pump under different operating conditions (such as high temperature, high load, frequent start and stop) until the water pump experiences functional failure (such as the head falling below the standard value, bearing seizure, or coil burnout).

[0141] During this process, health feature vectors are continuously collected and recorded at each time point, thus forming a pre-acquired sequence of sample feature vectors containing historical decay trend information. These sequences record the entire process of the electric water pump from start to scrap (failure), including trend information of gradually increasing electrical parameters and gradually expanding hydraulic parameters.

[0142] For each sample feature vector sequence, since the experiment has recorded the final failure time of the water pump, the true remaining useful life value corresponding to each time step in the sequence can be calculated accurately. These sample feature vector sequences are used as the input to the model, and the corresponding true remaining useful life values ​​are used as the expected output of the model, constructing a training dataset for supervised learning.

[0143] To handle time-dependent feature sequences, the preferred initial time series prediction model is a recurrent neural network architecture in deep learning, especially a long short-term memory network or a gated recurrent unit.

[0144] The initial model's input layer dimension matches the dimension of the health feature vector, and the output layer typically consists of a single neuron that outputs the predicted lifespan value. Before training begins, the model's weight parameters are randomly initialized.

[0145] The constructed training dataset is input into the initial time series prediction model for training. The model receives a sequence of sample feature vectors (e.g., feature changes over the past 100 hours) and outputs a predicted remaining lifespan value.

[0146] The difference between the predicted value and the actual remaining useful life recorded in the sample is calculated, and this prediction error is quantified by a loss function.

[0147] Based on the value of the loss function, the backpropagation algorithm and optimizer are used to adjust the weight parameters inside the model, so that the model's prediction results get closer and closer to the true values.

[0148] This training process will be repeated multiple times. As training progresses, the value of the loss function will gradually decrease. The changes in the loss function will be monitored in real time until the loss function of the initial time series prediction model meets the preset minimization condition.

[0149] The preset minimization condition can be set as the loss function value being lower than a very small preset threshold, or the rate of change of the loss function over multiple consecutive training epochs being less than a preset value (i.e., the model converges and there is no longer a significant improvement).

[0150] When this condition is met, training stops, and the model parameters are fixed, resulting in a trained lifetime prediction model. This lifetime prediction model solidifies the understanding of the performance degradation trajectory of the electric water pump, enabling it to accurately infer, based on historical trend information, which stage of its life cycle the current performance is in and how far it is from failure.

[0151] In one embodiment, the method further includes the following steps S501-S502.

[0152] Step S501: Encapsulate the fault detection results and life prediction results into a status message.

[0153] Here, after completing fault diagnosis and life prediction calculations, this information is not stored locally in isolation, but rather encapsulated according to a predefined communication protocol.

[0154] The encapsulation process transforms abstract computation results into digital messages that conform to vehicular network standards. The data field of the status message contains several key information fields: 1. Fault Detection Result Field: Carries specific fault codes. For example, specific hexadecimal codes can be used to identify normal, impeller cavitation, flow channel blockage, bearing wear, or stator winding aging, respectively.

[0155] 2. Life Prediction Results Field: Carries a quantified value of remaining useful life (in hours) or a health index percentage (0-100%).

[0156] 3. Original alarm signal field: includes a flag indicating whether immediate maintenance is required.

[0157] In step S502, the status message is sent to the vehicle controller via the vehicle's communication bus.

[0158] Here, after encapsulation, the status message is sent to the vehicle controller or thermal management system controller via the physically connected vehicle communication bus. Upon receiving the status message, the vehicle controller can optimize the vehicle's energy management strategy accordingly. For example, when it detects a decline in water pump performance, it can adjust the cooling fan speed appropriately to compensate. Alternatively, it can forward this information to the driver's dashboard, displaying a message in text or on indicator lights indicating "Cooling system efficiency has decreased; please check promptly."

[0159] In one embodiment, the method further includes: When the fault detection result indicates impeller cavitation and the life prediction result is lower than the preset warning value, the target speed of the motor controlling the electronic water pump is lower than the preset safe speed threshold.

[0160] Here, fault detection results and life prediction results are monitored in real time. When the fault detection results clearly indicate impeller cavitation, and the calculated life prediction results (such as remaining hours or health index) are lower than the preset warning value (e.g., remaining life less than 200 hours or health index less than 30%), the water pump is determined to be in a high-risk stage.

[0161] At this point, to prevent further cavitation damage, the speed limiting logic is automatically triggered.

[0162] Because cavitation is closely related to flow velocity and pressure, the higher the rotational speed, the lower the local pressure at the impeller inlet, resulting in more intense cavitation bubble formation and collapse, and stronger erosion damage to the impeller surface. Therefore, the target rotational speed of the electric water pump motor should be less than a preset safe rotational speed threshold (usually set below the cavitation initiation speed or 70%-80% of the rated speed). By actively reducing the rotational speed, the static pressure at the impeller inlet can be effectively increased, suppressing the formation and collapse of bubbles. This reduces the rate of physical damage propagation while ensuring basic cooling circulation, thus extending the remaining usable time of the faulty pump as much as possible.

[0163] Based on the above embodiments, this application provides an electronic water pump, referring to... Figure 5 An electronic water pump includes: a controller 1, a pumping assembly 2 and a sensing assembly 3 connected to the controller 1; the controller 1 is used to execute the life prediction method of the electronic water pump according to any of the foregoing embodiments.

[0164] The pumping assembly 2 includes a pump casing, an impeller disposed inside the pump casing, and a brushless motor that drives the impeller to rotate.

[0165] The sensing component 3 includes a current sensor and a position sensor for acquiring electrical parameters, and a miniature differential pressure sensor for acquiring hydraulic parameters.

[0166] Here, the pumping assembly 2 includes a pump casing, an impeller disposed inside the pump casing, and a brushless motor for driving the impeller to rotate. The pump casing is typically made of high-temperature resistant and corrosion-resistant engineering plastics or aluminum alloys, and its interior is designed with a fluid cavity. The impeller is located within the fluid cavity and is connected to the brushless motor via a rotor shaft. The brushless motor is preferably a brushless DC motor, whose stator windings generate a rotating magnetic field that drives the rotor to rotate the impeller at high speed, thereby drawing coolant in from the inlet and pumping it out from the outlet, thus achieving energy conversion and fluid transport.

[0167] Sensing component 3 is used to acquire multi-dimensional physical parameters during the operation of the electronic water pump in real time. Sensing component 3 is electrically connected to controller 1 and specifically includes a current sensor, a position sensor, and a miniature differential pressure sensor. The current sensor and position sensor are mainly used to acquire electrical and mechanical parameters. The current sensor can be integrated into the power drive circuit of controller 1 to detect the phase current of the brushless motor; the position sensor (such as a Hall effect sensor) is installed at the stator end of the motor to detect the real-time position of the rotor. The miniature differential pressure sensor is mainly used to acquire hydraulic parameters and is preferably manufactured using microelectromechanical systems (MEMS) technology, resulting in a small size and high integration.

[0168] Controller 1 can be an MCU. Controller 1 is installed inside the electric water pump, integrated with the brushless motor and various sensors in a single package. Controller 1 is configured to execute the aforementioned lifetime prediction method. Specifically, controller 1 controls the operation of the brushless motor through a power drive unit, while simultaneously receiving current, position, and differential pressure signals from sensing components 3 through a signal processing unit. Controller 1 internally stores pre-trained fault diagnosis and lifetime prediction models, enabling it to identify faults such as cavitation and wear based on the acquired hardware signals, and calculate the remaining useful life. Controller 1 also incorporates a heat dissipation design, utilizing coolant within the pump housing to cool the power devices, ensuring operational stability.

[0169] In one embodiment, the pump casing is provided with a first pressure-inlet hole and a second pressure-inlet hole. The first pressure-inlet hole is connected to the low-pressure area at the inlet of the impeller, and the second pressure-inlet hole is connected to the high-pressure area at the outlet of the impeller.

[0170] The pump casing is equipped with microchannels, and a miniature differential pressure sensor is airtightly connected to the first pressure inlet and the second pressure inlet through the microchannels to measure the pressure difference before and after the impeller.

[0171] Here, to directly measure the impeller's working state, a first pressure-sensing hole and a second pressure-sensing hole are provided on the pump casing. The first pressure-sensing hole is located in the low-pressure zone (i.e., the suction chamber) connecting to the impeller inlet, and the second pressure-sensing hole is located in the high-pressure zone (i.e., the discharge chamber) connecting to the impeller outlet. A microchannel is provided inside the pump casing or at the junction of the controller 1 and the pump casing. A miniature differential pressure sensor is airtightly connected to the first and second pressure-sensing holes through the microchannel. The sensing surface of the miniature differential pressure sensor can directly withstand the pressure difference before and after the impeller, thereby outputting a precise differential pressure signal without the need for an additional sensor installed on an external pipeline.

[0172] In one embodiment, the electronic water pump also includes a communication interface.

[0173] Controller 1 is configured to connect to the vehicle's in-vehicle network system via a communication interface.

[0174] Here, the electric water pump is also equipped with a standard physical communication interface. Controller 1 is connected to the vehicle's onboard network system (such as vehicle controller 1 or thermal management controller 1) via the communication interface. The communication interface supports CAN bus or LIN bus protocols and is used to report status information such as health index, fault codes, and remaining lifespan calculated by controller 1 to the vehicle, and to receive control commands issued by the vehicle.

[0175] The computer program product provided in this application includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the preceding method embodiments. For specific implementation details, please refer to the method embodiments, which will not be repeated here.

[0176] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system and apparatus described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0177] Furthermore, in the description of the embodiments of this application, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0178] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0179] In the description of this application, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0180] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The scope of protection of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of protection of the claims.

Claims

1. A method for predicting the lifespan of an electronic water pump, characterized in that, include: Obtain the real-time operating data of the electronic water pump; The real-time operating data includes electrical parameters and hydraulic parameters; The real-time operating data is processed according to preset feature extraction rules to generate a health feature vector; the health feature vector is used to characterize the current operating status of the electronic water pump. The health feature vector is input into a pre-trained fault diagnosis model, and the fault detection result of the electronic water pump is output. The fault detection result indicates whether the electronic water pump is in a fault-free state or a faulty state. Based on the preset mapping relationship between fault detection results and performance degradation algorithms, the target performance degradation algorithm corresponding to the fault detection results is determined. Using a pre-trained lifespan prediction model and based on the target degradation algorithm, the health feature vector and the historical operating data of the electronic water pump are calculated, and the lifespan prediction result of the electronic water pump is output.

2. The method for predicting the lifespan of an electronic water pump according to claim 1, characterized in that, The electronic water pump includes a current sensor, a position sensor, and a miniature differential pressure sensor; the step of acquiring the real-time operating data of the electronic water pump includes: The phase current signal of the motor of the electronic water pump is acquired by the current sensor. The rotor position signal of the motor is acquired by the position sensor, and the real-time speed of the motor is calculated. The pressure difference signal between the inlet and outlet of the electronic water pump is measured in real time using the miniature differential pressure sensor. The phase current signal and the differential pressure signal are synchronously sampled and denoised to obtain the real-time operating data.

3. The method for predicting the lifespan of an electronic water pump according to claim 2, characterized in that, The step of processing the real-time running data according to preset feature extraction rules to generate a health feature vector includes: The phase current signal is subjected to frequency domain transformation to extract the current harmonic distortion rate and spectral energy value as electrical characteristic components; Based on the real-time rotational speed, the pre-stored standard differential pressure-flow characteristic curve is invoked to calculate the deviation between the measured value of the differential pressure signal and the corresponding theoretical value in the standard differential pressure-flow characteristic curve, which is used as a hydraulic characteristic component. The electrical feature components and the hydraulic feature components are normalized and concatenated to generate the health feature vector.

4. The method for predicting the lifespan of an electronic water pump according to claim 1, characterized in that, The fault diagnosis model is trained in the following manner: The sample health feature vectors corresponding to various preset operating states are used as input, and the true state category labels corresponding to the sample health feature vectors are used as output. The preset initial classification network model is trained until the error between the output of the initial classification network model and the true state category label meets the preset convergence condition, thus obtaining the trained fault diagnosis model.

5. The method for predicting the lifespan of an electronic water pump according to claim 1, characterized in that, The fault conditions include at least one of the following: impeller cavitation, flow channel blockage, bearing wear, and stator winding aging. The step of determining the target performance degradation algorithm corresponding to the fault detection result based on the preset mapping relationship between the fault detection result and the performance degradation algorithm includes: When the fault detection result is the impeller cavitation state or the flow channel blockage state, the corresponding target performance degradation algorithm is determined to be the preset performance degradation algorithm. When the fault detection result is the bearing wear state, the corresponding target performance degradation algorithm is determined to be the preset mechanical fatigue wear algorithm; When the fault detection result is the aging state of the stator winding or the fault-free state, the corresponding target performance degradation algorithm is determined to be the preset aging algorithm.

6. The method for predicting the lifespan of an electronic water pump according to claim 1, characterized in that, The step of using a pre-trained lifespan prediction model, based on the target degradation algorithm, to calculate the health feature vector and the historical operating data of the electronic water pump, and outputting the lifespan prediction result of the electronic water pump, includes: Acquire the historical operating data of the electronic water pump; the historical operating data includes the health feature vector sequence and cumulative operating time at historical moments. Calculate the current health degradation index based on the target degradation algorithm; The health degradation index is used as a constraint input into the life prediction model. The time-series evolution trend of the historical operating data is fitted and extrapolated to calculate the remaining time required for the performance index of the electronic water pump to decline to a preset failure threshold. The remaining time is then determined as the life prediction result.

7. The method for predicting the lifespan of an electronic water pump according to claim 1, characterized in that, The lifetime prediction model is trained in the following manner: The preset initial time series prediction model is trained by taking a pre-acquired sample feature vector sequence containing historical decay trend information as input and the actual remaining useful life corresponding to the sample feature vector sequence as output, until the loss function of the preset time series prediction model satisfies the preset minimization condition, thus obtaining the trained life prediction model.

8. The method for predicting the lifespan of an electronic water pump according to claim 1, characterized in that, The method further includes: The fault detection results and the lifetime prediction results are encapsulated into a status message; The status message is sent to the vehicle controller via the vehicle's communication bus.

9. The method for predicting the lifespan of an electronic water pump according to claim 1, characterized in that, The method further includes: When the fault detection result indicates impeller cavitation and the life prediction result is lower than the preset warning value, the target speed of the motor of the electronic water pump is controlled to be less than the preset safe speed threshold.

10. An electronic water pump, characterized in that, The device includes a controller, and further includes a pumping assembly and a sensing assembly connected to the controller; the controller is used to perform the life prediction method for the electronic water pump according to any one of claims 1-9. The pumping assembly includes a pump casing, an impeller disposed within the pump casing, and a brushless motor that drives the impeller to rotate. The sensing components include a current sensor and a position sensor for acquiring electrical parameters, and a miniature differential pressure sensor for acquiring hydraulic parameters.

11. The electronic water pump according to claim 10, characterized in that, The pump casing is provided with a first pressure-inlet hole and a second pressure-inlet hole. The first pressure-inlet hole is connected to the low-pressure area at the inlet of the impeller, and the second pressure-inlet hole is connected to the high-pressure area at the outlet of the impeller. The pump casing is provided with a microchannel, and the micro differential pressure sensor is airtightly connected to the first pressure inlet and the second pressure inlet through the microchannel to measure the pressure difference before and after the impeller.

12. The electronic water pump according to claim 10, characterized in that, The electronic water pump also includes a communication interface; The controller is configured to connect to the vehicle's in-vehicle network system via the communication interface.

Images