Study On Health Condition Evaluation Method Of High Pressure Common Rail Injector

The reliability of Marine diesel engines is reduced due to their complex mechanical structure and harsh working environment.A high-pressure common rail injector is the core component of the diesel engine fuel system.To detect the performance degradation of highpressure common rail injectors in time,it is necessary to carry out health status evaluation research on this component.Given the problems of inaccurate model evaluation and difficult division of state space in existing research,this paper carried out a study on the health state evaluation method of high-pressure common rail fuel injectors.The main contents are as follows:Firstly,through the typical failure mode analysis of a high-pressure common rail fuel system,the research object of this paper is determined to be a high-pressure common rail fuel injector.Based on the common failure modes of high-pressure common rail injectors,this paper studied two degradation paths,needle valve wear,and nozzle blockage.To obtain signals that can reflect the health status of high-pressure common rail injectors,fault experiment schemes were designed and oil pressure signals were collected under different health statuses of highpressure common rail injectors.Secondly,aiming at the problem that the fault information is easily covered by the strong noise in the working environment of a high-pressure common rail injector,a new adaptive signal denoising method based on the ensemble empirical mode decomposition algorithm is proposed in this paper.Firstly,a signal extension method based on an improved whale optimization algorithm and support vector regression algorithm is proposed to solve the endpoint effect problem of the ensemble empirical mode decomposition algorithm.The improved whale optimization algorithm uses information entropy to realize the adaptive adjustment of inertia weight and sets the iteration termination condition,which greatly reduces the operation time on the premise of guaranteeing iteration accuracy.In this paper,the proposed signal continuation method is compared with other common continuation methods,which can better suppress the end effects of the ensemble empirical mode decomposition algorithm.In this paper,the proposed adaptive signal denoising method is applied to simulation and real signals.The experimental results show that the proposed method can effectively remove the noise in the original signal.Then,to extract typical fault features of high-pressure common rail injectors,a feature extraction method based on hierarchical weighted permutation entropy is introduced in this paper.The hierarchical weighted prearrangement entropy is compared with common feature extraction methods.The results show that,compared with multi-scale weighted permutation entropy and multi-scale permutation entropy,the feature extraction method based on hierarchical weighted permutation entropy can effectively distinguish fault categories.Moreover,the fault diagnosis method based on hierarchical weighted permutation entropy improves the fault recognition accuracy by 3.9% and 11.67%,respectively.Considering that the degradation path of needle valve wear fault is different from that of nozzle blockage fault,this paper uses a support vector machine algorithm to identify the failure mode of a highpressure common rail injector before health status evaluation.Compared with the fault diagnosis method based on the hidden Markov model,this model can effectively distinguish the fault categories of high-pressure common rail injectors,and the calculation time is less.Finally,to solve the problems of difficult and costly life-cycle signal acquisition of highpressure common rail fuel injectors,and subjective artificially defined health status of highpressure common rail fuel injectors,this paper proposes a health status evaluation method based on the hidden Markov model and Mahalanobis distance.In the model training step,this method only needs oil pressure data of high-pressure common rail fuel injectors working under health status.Reduced data integrity requirements.In addition,based on the working data of the highpressure common rail injector under the health condition and the oil pressure data under individual fault nodes of the high-pressure common rail injector,the Mahalanobis distance and hidden Markov model were combined to construct the performance evaluation index of the high-pressure common rail injector,and the health condition evaluation experiment was conducted by using the measurement signals under different working conditions.The experimental results show that the accuracy of state evaluation under different conditions reaches 100%,which proves that this method is suitable for small sample data and can weaken the negative effects caused by the artificial division of state space.