CN-121981009-A - High-fidelity state evaluation method for high-end equipment pump motor

CN121981009ACN 121981009 ACN121981009 ACN 121981009ACN-121981009-A

Abstract

The invention discloses a high-fidelity state evaluation method of a high-end equipment pump motor, which comprises the steps of sequentially constructing an ideal mathematical model, an internal leakage amount mechanism model and an actual output torque calculation model to obtain theoretical performance parameters, dynamic volumetric efficiency and a torque reference, inverting model parameters to be determined by adopting a multi-objective optimization algorithm based on health state test data to establish an individual health reference model, inputting real-time performance signals into the model, calculating residual errors of actual measurement data and model response, extracting dimensionless health feature vectors through normalization processing, obtaining comprehensive health indexes through weighting fusion, and realizing accurate quantitative evaluation of the health state of the pump motor.

Inventors

YU LIANG
ZHENG CHANGSONG
JIA RAN
WU WEI
QIN JINPING
ZHANG XIAOPENG
Min Sijie
HUANG JIAHAI

Assignees

北京理工大学

Dates

Publication Date: 20260505
Application Date: 20260127

Claims (10)

1. A high-end equipped pump motor high-fidelity state assessment method, comprising: According to the structural parameters and the operation condition data of the pump motor, an ideal pump motor mathematical model is constructed, and theoretical performance parameters of the pump motor are obtained; according to the theoretical performance parameters, constructing a pump motor internal leakage amount mechanism model to obtain dynamic volumetric efficiency; According to the dynamic volumetric efficiency and the friction loss mechanism, constructing a pump motor actual output torque calculation model to obtain a theoretical output torque reference value; Inverting undetermined parameters in the pump motor internal leakage amount mechanism model and the pump motor actual output torque calculation model by adopting a multi-objective optimization algorithm according to test data of the pump motor in a healthy state to obtain a pump motor individuation health reference model; inputting the pump motor performance signals acquired in real time into the pump motor individuation health reference model to obtain model response output; According to the residual error between the pump motor operation data acquired in real time and the model response output, carrying out normalization processing to obtain a dimensionless health feature vector; And carrying out weighted fusion on the dimensionless health feature vectors to obtain a comprehensive health index, and evaluating the health state of the pump motor according to the comprehensive health index to obtain an evaluation result of the health state of the pump motor.
2. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of constructing an ideal pump motor mathematical model to obtain theoretical performance parameters of the pump motor includes: constructing an input flow equation and an output flow equation of the pump according to the maximum displacement, the inclined angle of the inclined plate and the input rotating speed of the variable pump; Constructing an input flow equation and an output flow equation of the motor according to the motor displacement and the input rotating speed; constructing a motor rotation speed calculation equation according to the volumetric efficiency and the rotation speed transmission ratio of the pump; constructing a motor torque calculation equation according to the pressure difference at two sides of the motor and the motor displacement; constructing an oil supplementing flow equation according to the flow pressure gradient of the check valve and the oil supplementing pressure; Respectively constructing flow balance differential equations according to the flow continuity principles of the high pressure side and the low pressure side; Establishing an ideal pump motor mathematical model by combining an input flow equation and an output flow equation of the pump, an input flow equation and an output flow equation of the motor, a motor rotating speed calculation equation, a motor torque calculation equation, an oil supplementing flow equation and a flow balance differential equation, and obtaining theoretical performance parameters of the pump motor; Theoretical performance parameters of the pump motor include system pressure distribution, dynamic flow characteristics, output rotational speed and torque.
3. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of constructing a pump motor internal leakage mechanism model to obtain dynamic volumetric efficiency includes: Constructing an oil dynamic equivalent viscosity calculation model according to the oil temperature, the oil pressure, the viscosity-temperature characteristic parameters and the viscosity-pressure coefficient to obtain real-time dynamic viscosity; According to geometric conductance, shear conductance, effective clearance, inlet pressure and outlet pressure of the matched pair, constructing a matched pair leakage amount calculation model based on a gap flow theory; and constructing a total leakage quantity model in the pump motor according to the sum of leakage quantities of the plunger piston and cylinder hole matching pair, the slipper and sloping cam plate matching pair and the valve plate and cylinder end face matching pair, and calculating dynamic volumetric efficiency according to the total leakage quantity and theoretical flow.
4. The method of claim 3, wherein the step of, The process for constructing the oil dynamic equivalent viscosity calculation model comprises the following steps: Determining a viscosity-temperature characteristic parameter and a viscosity-temperature coefficient according to the oil type; determining a viscosity coefficient according to the oil pressure, and correcting the viscosity increasing effect under high load; And calculating the real-time dynamic viscosity according to the oil temperature, the high-pressure side oil pressure, the viscosity-temperature characteristic parameter, the viscosity-pressure coefficient and the viscosity-temperature coefficient which are acquired in real time.
5. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of constructing a pump motor actual output torque calculation model and obtaining a theoretical output torque reference value comprises the following steps: constructing an ideal output torque equation according to the pressure difference at two sides of the motor, the motor displacement and the dynamic volumetric efficiency; constructing a mixed friction moment model containing viscous friction, coulomb friction and static friction according to the real-time dynamic viscosity, the motor rotating speed, the coulomb friction coefficient and the static friction parameter; according to the equivalent rotational inertia and angular acceleration of the rotating component of the pump motor, an inertia moment compensation model is constructed; and constructing a theoretical output torque reference value calculation equation according to the ideal output torque, the mixed friction torque and the inertia torque to obtain a theoretical output torque reference value.
6. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of obtaining a pump motor personalized health benchmark model includes: Under the health state of the pump motor, collecting multi-working condition operation data covering different rotating speeds, different load pressures and different oil temperatures, and constructing a health sample data set; Constructing an oil reference viscosity, a viscosity pressing coefficient, an equivalent effective gap reference value and a geometric flow passage as parameter vectors to be identified; driving a simulation model containing a parameter vector to be identified according to working condition data in the health sample data set, and constructing a multi-objective optimization function according to the degree of flow fitting errors, torque fitting errors and parameter deviation priori knowledge; adopting an NSGA-III framework to carry out multi-objective optimization, predicting an optimization target value by utilizing a Gaussian process regression model, and adopting expected super-volume improvement as an acquisition function to select a candidate solution; And after the pareto optimal solution set is obtained, calculating a parameter confidence interval, outputting a global optimal parameter vector, solidifying the global optimal parameter vector to the pump motor internal leakage amount mechanism model and the pump motor actual output torque calculation model, and constructing and obtaining a pump motor individuation health reference model.
7. The method of claim 6, wherein the step of providing the first layer comprises, The process of multi-objective optimization using the NSGA-III framework includes: uniformly sampling an initial population in a parameter space, evaluating multiple target values of individuals of the initial population by using a simulation model, and constructing an initial pareto front; Generating offspring according to the selection, crossing and mutation mechanisms of NSGA-III; respectively predicting a flow fitting precision target, a torque fitting precision target and a model simplicity target by utilizing an estimated individual training Gaussian process regression model; selecting an individual improving the current pareto front from the candidate solutions by adopting the expected super-volume improvement as an acquisition function to carry out real simulation evaluation; repeating the above process until the algorithm converges, and outputting the pareto optimal solution set with uniform distribution.
8. The method of claim 1, wherein the step of determining the position of the substrate comprises, The process of estimating the health state of the pump motor according to the comprehensive health index and obtaining the estimation result of the health state of the pump motor comprises the following steps: calculating motor output rotation speed residual errors and motor output torque residual errors according to the real-time operation data of the pump motor and the output data of the pump motor individuation health reference model; Performing dynamic standardization processing on the torque residual error by adopting a Z-Score method to obtain dimensionless rotation speed abnormal times and torque abnormal times; Performing root mean square processing or moving average processing on the abnormal multiples by adopting a sliding time window, extracting robust trend characteristics, and constructing a multidimensional health characteristic vector; Calculating a weighted sum according to the multidimensional health feature vector and a preset weight coefficient vector, and mapping the weighted sum result to a [0,1] interval by adopting a mapping function; And introducing a short board correction term, multiplying the mapping result by the short board correction term to obtain a comprehensive health index, and dividing the health grade according to the numerical value of the comprehensive health index.
9. The method of claim 8, wherein the step of determining the position of the first electrode is performed, Introducing a short-board correction term, multiplying the mapping result by the short-board correction term, and obtaining the comprehensive health index comprises the following steps: constructing a multidimensional health feature vector according to the abnormal rotation speed multiple and the abnormal torque multiple; distributing weight coefficients according to the sensitivity of each dimensional characteristic to the fault mode, wherein the weight coefficient corresponding to the abnormal torque multiple for representing the increase of leakage is larger than the weight coefficient corresponding to the abnormal rotation speed multiple; Calculating the health degree of each subsystem and the average health degree of all subsystems, and constructing a short-board correction term according to the ratio of the health degree of each subsystem to the average health degree, wherein the correction intensity coefficient is greater than or equal to 1; And multiplying the weighted sum of the multidimensional health feature vectors with a short-board correction term to obtain a comprehensive health index.
10. The method of claim 8, wherein the step of determining the position of the first electrode is performed, The process of classifying the health grade according to the value of the comprehensive health index comprises the following steps: When the integrated health index is greater than 0.8 and less than or equal to 1, determining a health grade; when the integrated health index is greater than 0.6 and less than or equal to 0.8, determining a good grade; determining a degradation level when the integrated health index is greater than 0.4 and less than or equal to 0.6; and when the integrated health index is greater than or equal to 0 and less than or equal to 0.4, determining that the integrated health index is a fault level.

Description

High-fidelity state evaluation method for high-end equipment pump motor Technical Field The invention belongs to the field of state monitoring of fluid transmission equipment, and particularly relates to a high-fidelity state evaluation method for a high-end equipment pump motor. Background The pump motor is used as a core power unit of the hydrostatic transmission system, and is widely applied to the national strategic fields of aerospace, polar scientific investigation equipment, high-end intelligent manufacturing, military special equipment and the like by virtue of the advantages of high power density, compact structure, remarkable efficiency and the like. Aiming at the health state evaluation of a pump motor, the prior art mainly comprises three types, namely a monitoring method based on vibration spectrum analysis or fixed efficiency threshold, an abnormal judgment method through setting an experience threshold, a pure data driving method based on deep learning, a state recognition method realized by using a large amount of historical data training models, and an evaluation method based on a fixed parameter physical mechanism model, wherein the performance benchmark is calculated through a theoretical formula. The method can realize the basic monitoring function under the conventional working condition, but has obviously limited applicability under the working conditions of multiple physical field intensity couplings such as high-low temperature circulation, variable rotation speed-variable load alternation, long-term low-temperature heavy load and the like faced by high-end equipment, and time-varying non-stationary working conditions. However, the prior art has the outstanding problems that the traditional monitoring method based on vibration frequency spectrum or fixed threshold cannot distinguish normal parameter drift and real physical degradation induced by working conditions under variable working conditions, fluid pulsation noise and normal leakage fluctuation are misjudged as fault signals, false alarms are frequent, severe requirements of zero misjudgment in the scenes such as aerospace on-orbit operation and polar scientific investigation are difficult to meet, the pure data driving method based on deep learning relies on massive full life cycle sample data, a high-end equipment pump motor is mostly a customized product, the full life data is scarce, the working conditions are not covered fully, the model generalization capability is seriously insufficient, meanwhile, a black box model lacks physical interpretability, a degradation source cannot be positioned, accurate operation and maintenance decision is difficult to support, the traditional physical mechanism adopts fixed experience parameters, and cannot dynamically follow actual working condition evolution such as viscosity change caused by oil temperature change, gap change caused by abrasion, so that a theoretical health standard has significant deviation from an actual physical process, and the value as a reliable evaluation reference is lost. The problems cause that the evaluation accuracy, reliability and engineering practicability of the existing method under the complex working condition can not meet the requirement of high-end equipment on the evaluation of the state of the pump motor. Disclosure of Invention In order to solve the technical problems, the invention provides a high-fidelity state evaluation method of a high-end equipment pump motor, which comprises the following steps: According to the structural parameters and the operation condition data of the pump motor, an ideal pump motor mathematical model is constructed, and theoretical performance parameters of the pump motor are obtained; according to the theoretical performance parameters, constructing a pump motor internal leakage amount mechanism model to obtain dynamic volumetric efficiency; According to the dynamic volumetric efficiency and the friction loss mechanism, constructing a pump motor actual output torque calculation model to obtain a theoretical output torque reference value; Inverting undetermined parameters in the pump motor internal leakage amount mechanism model and the pump motor actual output torque calculation model by adopting a multi-objective optimization algorithm according to test data of the pump motor in a healthy state to obtain a pump motor individuation health reference model; inputting the pump motor performance signals acquired in real time into the pump motor individuation health reference model to obtain model response output; According to the residual error between the pump motor operation data acquired in real time and the model response output, carrying out normalization processing to obtain a dimensionless health feature vector; And carrying out weighted fusion on the dimensionless health feature vectors to obtain a comprehensive health index, and evaluating the health state of the pump motor according to the comprehensive health