Study On The Cavitation Characteristics And Intelligent Monitoring Of Automotive Electronic Water Pumps Under Thermodynamic Effect

With the continuous breakthrough of energy revolution,new materials and new generation of information technology,automotive products are developing towards the direction of new energy,lightweight,intelligent and network-linked acceleratively.Automotive electronic pump is a high-tech product that has been spawned with the development of automotive industry technology and market demand.It is a kind of mini high-speed centrifugal pump,it has the characteristics of wide range of speed variation,complex geometry structure and high operating environment temperature.During operation,it is prone to cavitation,which cause performance degradation,severe vibration and noise,and even damage to the flow passage components in severe cases,or will affect traffic safety,etc.Cavitation is a multi-phase flow phenomenon involving complex phase transitions.The multi-scale spatiotemporal collapse of cavitation has always been a hot issue in hydraulic machinery.This study intends to use numerical simulation and experimental research methods to investigate the cavitation characteristics of automotive electronic water pumps under the influence of thermodynamic effects,and then intelligent monitoring of its operating status based on the support vector machine method,which is of great significance to improve the economy and reliability of the vehicle.The main research work and innovations of this thesis are as follows:1.In-depth analysis of the hydraulic machinery cavitation numerical simulation and condition monitoring methods are carried out.Through comparison,it is concluded that the thermodynamic effect has an obvious effect on the occurrence and development of automotive electronic water pump cavitation,and the corresponding model modification should be considered in the numerical simulation.The vibration method can reflect the cavitation state of the pump very well,the use of vibration characteristics makes the intelligent monitoring based on support vector machines has the advantages of simple and accurate.According to actual demand,the vibration method is used to collect cavitation signals.2.Aiming at the cavitation problem of automotive electronic pumps under different temperature working fluids,the cavitation model considering the influence of thermodynamic effects was modified,and the unsteady numerical simulation of cavitation under three temperature conditions was carried out,and performance test verification was carried out.The research results show that: the simulation value of the external characteristics and the actual test value basically have the same change trend at full flow,and the calculation accuracy meets the subsequent cavitation needs.With the increase of the temperature,the area of low pressure zone inside the impeller of the automotive electronic pump increases,and the critical cavitation point will increase,the number of cavitation bubbles increases and the cavitation becomes more serious.From no cavitation to cavitation stage,the maximum amplitude value of the pressure pulsation in the impeller flow-path increases gradually.The generation and collapse of the bubbles causes the pressure pulsation of the internal flow field variation,which in turn causes the pump body to vibrate,and the vibration signal can be collected to reflect the operation state of the pump.3.An automotive electronic pump cavitation monitoring test rig considering the effects of thermodynamics was built,and the cavitation performance and vibration tests of automotive electronic pumps at three temperatures under corresponding conditions were completed.The research results show that the dimensionless external characteristic curve basically conforms to the similarity theorem.The temperature has an obvious promotion effect on cavitation,the higher the temperature,the greater the cavitation number,cavitation is more easier to occur under the same other conditions.The main frequency of vibration frequency domain is the shaft frequency and its multiplier frequency.With the development of cavitation,the amplitude of the main frequency increases gradually.4.Aiming at the cavitation characteristics of automotive electronic pumps under different working conditions considering the influence of thermodynamic effects,methods such as time domain,frequency domain and time-frequency domain are used to extract the characteristics of vibration signals,and the cavitation intelligent monitoring database is formed.The research results show that due to the different amplitudes of the time domain signals,the mean value,standard deviation,peak factor,kurtosis factor and skewness factor of the time domain signal are selected as features characteristics.The frequency domain signal feature extraction selects the barycentric frequency,mean square frequency and frequency variance of the power spectrum as features according to the different amplitude and dispersion degree of the power spectrum.The feature value selection for time-frequency domain feature extraction is the ratio of the energy of the first six layers to the total amount decomposed by the empirical mode decomposition method.5.The support vector machine is used to intelligently monitor the cavitation state.By introducing the penalty factor C and the kernel function parameter g,the optimal parameter combination (C,g) is selected to improve the recognition accuracy rate of the support vector machine.The research results show that when performing the two classification of cavitation occurrence,the recognition rate of the time domain signal is the highest in the recognition accuracy rate of single features,followed by the timefrequency domain signal,and finally the frequency domain signal.When performing multi-class recognition of cavitation state,it is found that the characteristics of the timedomain signal at the four monitoring points and the time-frequency domain signal in the X direction and Y directions of the inlet pipe are combined.The multi-classification recognition rate of the combined eigenvalues can achieve more than 94%.It is proved that support vector machine has high accuracy in intelligent monitoring of cavitation state of automobile electronic pump.