Pump Fault Research Articles

Review of Machine Learning Based Fault Detection for Centrifugal Pump Induction Motors

Centrifugal pumps are an integral part of many industrial processes and are used extensively in water supply, sewage, heating and cooling systems. While there are several review papers on machine learning-based fault diagnosis on induction motors, its application to centrifugal pumps has received relatively little attention. This work attempts to summarize and review recent research and development in machine learning-based pump condition monitoring and fault diagnosis. The paper starts with a brief explanation of pump operation including common pump faults and the main principles of the motor current signature analysis (MCSA) method. This is followed by a detailed explanation of various machine learning-based methods including the types of detected faults, experimental details and reported accuracies. The performances of different approaches are then presented systematically in a unified table. Finally, the authors discuss practical aspects and challenges related to data collection, storage and real-world implementation.

Intelligent fault diagnosis system design and implementation of diaphragm pump based on artificial intelligence

This article designs and implements a fault diagnosis system for diaphragm pump. The key components of the diaphragm pump are equipped with three-axis vibration intelligent sensor to collect vibration data, the edge side completes the data pre-processing, and then transmits it to the cloud platform through NBIoT. For the non-linear, non-smooth complex system, the time frequency representation is obtained by continuous wavelet transformation, after compressing the time frequency representation as the input of the two-dimensional neural network model, using the two-dimensional convolutional neural network model for training and monitoring, to complete the identification and positioning of diaphragm pump fault.

Fuzzy entropy assisted singular spectrum decomposition to detect bearing faults in axial piston pump

The fault feature of vibration signal is always submerged in heavy interference noise and strong pumping impulses in axial piston pump bearings. A fuzzy entropy assisted singular spectrum decomposition denoising method is introduced to extract bearing fault in axial piston pump. The raw measured data is adaptively decomposed to singular spectrum components. The fuzzy entropy of singular spectrum components is used to screen the component with critical information and remove the unrelated components. The remaining singular spectrum components are conducted for the secondary screening to decrease impulse interference based on the prior knowledge. Simulation results of signals with different noise intensity thoroughly validate performance. This method is applied to process practical engineering signals and the characteristic frequency intensity coefficient is taken as the evaluation index in the field. The experimental results show that the evaluation index of the presented technique are increased by 80%, 56% and 40%, respectively.

Fault Prediction of Centrifugal Pump Based on Improved KNN

To effectively predict the faults of centrifugal pumps, the idea of machine learning k-nearest neighbor algorithm (KNN) was introduced into the traditional Mahalanobis distance fault discrimination, and an improved centrifugal pump fault prediction model of KNN based on the Mahalanobis distance is proposed. In this method, the Mahalanobis distance is used to replace the distance function in the conventional KNN algorithm. Grid search and cross-validation are used to determine the optimal K value of the prediction model. A centrifugal pump test rig was established to solve three common faults of centrifugal pumps: cavitation, impeller damage, and machine seal damage, and the method was verified. The results show that this method can effectively distinguish the specific fault types of centrifugal pumps based on vibration signals, and the fault prediction accuracy of the off-balance condition is up to 82%. This study provides a novel idea and method for centrifugal pump fault prediction and diagnosis and avoids the interaction between parameters when monitoring multiple parameters.

A physics-based parametric regression approach for feedwater pump system diagnosis

In this paper, we assessed the performance of a model-based approach for fault diagnosis of a nuclear power plant feedwater pump system. Physics-based models were constructed to monitor the performance of the system components. Plant data were used for the calibration of the models and subsequently in the diagnosis of abnormal events. We considered two real-time events representing scenarios of a pump fault and component performance degradation. The two events were correctly diagnosed, and the results demonstrated the high detection sensitivity of the physics-based models. Various sensor fault scenarios were simulated to show the capability of the approach to detect and uniquely identify sensor faults. Results for a scenario in which the plant operated in flexible power mode also showed that the diagnostic approach is insensitive to changes of operating conditions, which is one of the advantages of the model-based approach using physics-based models over purely data-driven approach.

Fault Detection of Circulation Pumps on the Basis of Motor Current Evaluation

Motor current signature analysis (MCSA) for fault detection has found widespread application, especially for induction motors. The basis of MCSA is the evaluation of a motor's current. This analysis is now also used for other motor types and can be used to detect faults of the coupled load. The purpose of this article is to examine whether MCSA can be used to detect faults in a wet-rotor pump. A total of three faults are examined. The results show that, compared with a healthy pump, all faults could be detected. However, a detailed analysis of frequency components has to be carried out to differentiate the faults. A circulation pump with a maximum power consumption of 1.1 kW was used as the test item.

Fault Diagnosis in Centrifugal Pump using Support Vector Machine and Artificial Neural Network

Centrifugal pumps are commonly utilized in thermo-fluidic systems in the industry. Being a rotating machinery, they are prone to vibrations and their premature failure may affect the system predictability and reliability. To avoid their premature breakdown during operation, it is necessary to diagnose the faults in a pump at their initial stage. This study presents the methodology to diagnose fault of a cent rifugal pump using two distinct machine learning techniques, namely, Support vector machine (SVM) and Artificial neural network (ANN). Different statistical features are extracted in the time and the frequency domain of the vibration signal for different working conditions of the pump. Furthermore, to decrease the dimensionality of the obtained features different feature ranking (FR) methods, namely, Chi-square, ReliefF and XGBoost are employed. ANN technique is found to be more efficient in classifying faults in a centrifugal pump as compared to the SVM, and Chi-square and XGBoost ranking techniques are better than ReliefF at sorting more relevant features. The results presented in thus study demonstrate that an ANN based machine learning approach with Chi-square and XGBoost feature ranking techniques can be used effectively for the fault diagnosis of a centrifugal pump.

Machine Learning–Based Fault Detection and Diagnosis of Organic Rankine Cycle System for Waste-Heat Recovery

Utilizing the organic Rankine cycle (ORC) for waste heat recovery is an important energy conversion method. Some faults may occur in the ORC in actual operation, but few studies have focused on the fault detection and diagnosis of the whole ORC system. Fault detection detects whether a fault occurs in the system and fault diagnosis diagnoses where the fault is. This paper investigated a fault detection and diagnosis scheme of the ORC system for waste heat recovery based on machine learning. First, a thermodynamic ORC model was established. Three kinds of faults (expander fault, pump fault, and heat exchanger fault) and three kinds of algorithms [logistic regression, softmax regression, and support vector machines (SVMs)] were described. The data of four major important faults (fouling fault of the evaporator and of the condenser, looseness of the mechanical moving parts in the expander, and blocking of the pump) were generated from the thermodynamic ORC model and used to train the fault detection and diagnosis schemes. To evaluate the accuracy of the fault detection and diagnosis schemes, a set of experimental data was employed to test the schemes. The accuracy scores of fault detection using logistic regression and support vector machines were 77.42% and 96.77%, respectively. The accuracy scores of fault diagnosis using softmax regression and SVM were 91.78% and 94.52%, respectively. The test times of fault diagnosis using softmax regression and SVM were 0.0099 and 0.0085 s, respectively. The results demonstrated that machine learning–based fault detection and diagnosis schemes for the ORC have high accuracy and immediacy. Therefore, the proposed schemes are promising tools for fault detection and diagnosis of the ORC system for waste heat recovery.

Two Current-Based Methods for the Detection of Bearing and Impeller Faults in Variable Speed Pumps

The growing number of variable speed drives (VSDs) in industry has an impact on the future development of condition monitoring methods. In research, more and more attention is being paid to condition monitoring based on motor current evaluation. However, there are currently only a few contributions to current-based pump diagnosis. In this paper, two current-based methods for the detection of bearing defects, impeller clogging, and cracked impellers are presented. The first approach, load point-dependent fault indicator analysis (LoPoFIA), is an approach that was derived from motor current signature analysis (MCSA). Compared to MCSA, the novelty of LoPoFIA is that only amplitudes at typical fault frequencies in the current spectrum are considered as a function of the hydraulic load point. The second approach is advanced transient current signature analysis (ATCSA), which represents a time-frequency analysis of a current signal during start-up. According to the literature, ATCSA is mainly used for motor diagnosis. As a test item, a VSD-driven circulation pump was measured in a pump test bench. Compared to MCSA, both LoPoFIA and ATCSA showed improvements in terms of minimizing false alarms. However, LoPoFIA simplifies the separation of bearing defects and impeller defects, as impeller defects especially influence higher flow ranges. Compared to LoPoFIA, ATCSA represents a more efficient method in terms of minimizing measurement effort. In summary, both LoPoFIA and ATCSA provide important insights into the behavior of faulty pumps and can be advantageous compared to MCSA in terms of false alarms and fault separation.

Intelligent fault diagnosis of machinery using digital twin-assisted deep transfer learning

Digital twin (DT) is emerging as a key technology for smart manufacturing. The high fidelity DT model of the physical assets can produce system performance data that is close to reality, which provides remarkable opportunities for machine fault diagnosis when the measured fault condition data are insufficient. This paper presents an intelligent fault diagnosis framework for machinery based on DT and deep transfer learning. First, the DT model of the machine is built by establishing the simulation model and with further updating through continuously measured data from the physical asset. Second, all important machine conditions can be simulated from the built DT. Third, a new-type deep structure based on novel sparse de-noising auto-encoder (NSDAE) is developed and pre-trained with condition data from the source domain, as generated from the DT. Then, to achieve accurate machine fault diagnosis with possible variations in working conditions and system characteristics, the pre-trained NSDAE is fine-tuned using parameter transfer with only one sample from the target domain. The presented method is validated through a case study of triplex pump fault diagnosis. The experimental results demonstrate that the proposed method achieves intelligent fault diagnosis with a limited amount of measured data and outperforms other state-of-the-art data-driven methods.

A Non-Intrusive Signal-Based Fault Diagnosis Method for Proton Exchange Membrane Water Electrolyzer Using Empirical Mode Decomposition

This work focuses on a signal-based diagnosis approach dedicated to proton exchange membrane water electrolyzer (PEM WE) anode pump fault. The PEM WE cell measurements are performed with an experimental test bench to highlight the impact of water flow rate in the anode compartment. This approach is non-intrusive, and it can detect anode flow rate variation during the electrolysis and is designed to fulfill online diagnosis requirements. Contrary to electrochemical impedance spectroscopy-based approaches (EIS), this method stands out from existing procedures as a result of its few requirements, excluding any signal with perturbing amplitude. Therefore, the electrolyzer remains continuously available, even while the analysis is performed. The empirical mode decomposition (EMD) is used to decompose the signal variation into a sum of amplitude modulation and frequency modulation (AM-FM) components, called intrinsic mode functions (IMFs). In this work, the PEM WE current signal is decomposed into several IMFs using EMD. Then, the energetic contribution of each IMF is calculated. Experimental results exhibited that the energetic contribution of IMFs can be used as relevant criteria for fault diagnosis in PEM WE systems. This process only requires monitoring of the PEM WE current and has a low computational cost, which is a significant economic and technical advantage.

Application of Spectral Kurtosis on vibration signals for the detection of cavitation in centrifugal pumps

The detection of cavitation formation in hydraulic turbomachinery has been widely investigated due to its significant impact on their steady and dynamic operation. The aim of this study is the application of Spectral Kurtosis tool in order to effectively detect the impulsive shock waves generated during the implosion of vapour bubbles. The methodology is applied and evaluated on the vibration signals obtained from two different semi-open impellers. The results under initial cavitating conditions show that the high frequency implosions of vapour bubbles interact with the low frequency passing of the rotating blades and compose part of the vibration signal. The application of the band pass filter, with central frequency and bandwidth estimated from the Fast Kurtogram, to the original signal allows to extract this information both in time and in frequency domain, and to correlate the periodic impulsive behaviour with the blade passing frequency of the impeller. The present results support the establishment of a robust detection cavitation criterion in centrifugal pumps and show that Spectral Kurtosis is a useful tool for the prevention of related faults in centrifugal pumps.

Using MLP‐GABP and SVM with wavelet packet transform‐based feature extraction for fault diagnosis of a centrifugal pump

AbstractThis paper explores artificial intelligent training schemes based on multilayer perceptron, considering back propagation and genetic algorithm (GA). The hybrid scheme is compared with the traditional support vector machine approach in the literature to analyze both fault and normal scenarios of a centrifugal pump. A comparative analysis of the performance of the variables was carried out using both schemes. The study used features extracted for three decomposition levels based on wavelet packet transform. In order to investigate the effectiveness of the extracted features, two mother wavelets were investigated. The salient part of this work is the optimization of the hidden layers numbers using GA. Furthermore, this optimization process was extended to the multilayer perceptron neurons. The evaluation of the model system performance used for the study shows better response of the extracted features, and hidden layers variables including the selected neurons. Moreover, the applied training algorithm used in the work was able to enhance the classifications obtained considering the hybrid artificial intelligent scheme been proposed. This work has achieved a number of contributions like GA‐based selection of hidden layers and neuron, applied in neural network of centrifugal pump condition classification. Furthermore, a hybrid training method combining GA and back propagation (BP) algorithms has been applied for condition classification of a centrifugal pump. The obtained results have shown the good ability of the proposed methods and algorithms.

Fault Detection Algorithm for Multiple-Simultaneous Refrigerant Charge and Secondary Fluid Flow Rate Faults in Heat Pumps

The detection and diagnosis of faults is becoming necessary in ensuring energy savings in heat pump units. Faults can exist independently or simultaneously in heat pumps at the refrigerant side and secondary fluid flow loops. In this work, we discuss the effects that simultaneous refrigerant charge faults and faults associated with the flow rate of secondary fluids have on the performance of a heat pump operating in summer season and we developed a correlation to detect and diagnose these faults using multiple linear regression. The faults considered include simultaneous refrigerant charge and indoor heat exchanger secondary fluid flow rate faults (IFRFs), simultaneous refrigerant charge and outdoor heat exchanger secondary fluid flow rate faults (OFRFs) and simultaneous refrigerant charge, IFRF and OFRF. The occurrence of simultaneous refrigerant charge fault, IFRF and OFRF caused up to a 5.7% and 8% decrease in cooling capacity compared to simultaneous refrigerant charge and indoor heat exchanger secondary fluid flow rate faults, and simultaneous refrigerant charge and outdoor heat exchanger secondary fluid flow rate faults, respectively. Simultaneous refrigerant charge fault, IFRF and OFRF resulted in up to an 11.6% and 5.9% decrease in COP of the heat pump unit compared to simultaneous refrigerant charge fault and IFRF, and simultaneous refrigerant charge fault and OFRF, respectively. The developed FDD correlations accurately predicted the simultaneous refrigerant charge and faults in the flow rate of the secondary fluid within an error margin of 7.7%.

Hydraulic pump fault diagnosis based on chaotic characteristics of speed signals under non-stationary conditions

Pumps are vital components in hydraulic equipment, and their malfunction is directly related to the normal operation of the whole system. The hydraulic system is accompanied by the conversion of multi-domain energy during operation. In particular, it exhibits non-stationarity and nonlinearity under variable operating conditions, which brings difficulties to fault diagnosis. In this paper, the instantaneous angular speed (IAS) signal obtained by the equal-angle measurement is studied to diagnose pump faults under non-stationary conditions. For this purpose, a synchroextracting normal S transform method is proposed to extract the time-frequency feature components properly. Then, the extracted signal is reconstructed into a high-dimensional phase space based on the improved C-C method. On this basis, the quantitative analysis indicators of pump state are obtained by the G-P algorithm and small-data method, including correlation dimension, Kolmogorov entropy, and largest Lyapunov exponent. Pump faults can be classified in a chaotic space by using these sensitive features. The results show that the proposed method is capable of diagnosing different states of pumps and robust to the variation of load values.

Recent Advancement of Deep Learning Applications to Machine Condition Monitoring Part 1: A Critical Review

With the huge success of applying deep learning (DL) methodologies to image recognition and natural language processing in recent years, researchers are now keen to use them in the machine condition monitoring (MCM) context. There are numerous papers in applying various DL techniques, such as auto-encoder, restricted Boltzmann machine, convolutional neural network and recurrent neural network, etc., to MCM problems ranging from component-level condition monitoring (machine tool wear prediction, bearing fault diagnosis and classification and hydraulic pump fault diagnosis) to system-level health management (aircraft and spacecraft diagnosis). In this paper, we give a brief overview in the area of DL for MCM with a focus on reviewing the most recent papers published since 2019. In Part 1, we present some critical views regarding whether any breakthrough has been achieved from an MCM domain expert perspective, with the main conclusion that DL has great potential for MCM applications, and a major breakthrough could come soon since the shortfalls lie more in data than in the DL methodologies. Our overall impression is that (a) DL models are not really showing their great potentials with only a small training data; (b) faulty-condition data is hard to come by for training DL, but normal condition data is abundant, so anomaly detection makes more sense; (c) applying DL only to the Case Western Reserve University (CWRU) bearing fault dataset is not sufficient for real world industrial applications as it was from a very simple test rig, and applying DL to data from complex systems like helicopter gearbox data may deliver much more convincing results. In Part 2, we enhance the main conclusion of the critical review with supplement views and a case study on analysing Bell-206B helicopter main gearbox planet bearing failure data using some traditional MCM techniques in contrast to applying the long short-term memory (LSTM) DL method. We can conclude from the case study that the DL-based methods are not necessarily always superior to the traditional MCM techniques for dataset from moderately complex machinery.

A Hydraulic Pump Fault Diagnosis Method Based on the Modified Ensemble Empirical Mode Decomposition and Wavelet Kernel Extreme Learning Machine Methods.

To address the problem that the faults in axial piston pumps are complex and difficult to effectively diagnose, an integrated hydraulic pump fault diagnosis method based on the modified ensemble empirical mode decomposition (MEEMD), autoregressive (AR) spectrum energy, and wavelet kernel extreme learning machine (WKELM) methods is presented in this paper. First, the non-linear and non-stationary hydraulic pump vibration signals are decomposed into several intrinsic mode function (IMF) components by the MEEMD method. Next, AR spectrum analysis is performed for each IMF component, in order to extract the AR spectrum energy of each component as fault characteristics. Then, a hydraulic pump fault diagnosis model based on WKELM is built, in order to extract the features and diagnose faults of hydraulic pump vibration signals, for which the recognition accuracy reached 100%. Finally, the fault diagnosis effect of the hydraulic pump fault diagnosis method proposed in this paper is compared with BP neural network, support vector machine (SVM), and extreme learning machine (ELM) methods. The hydraulic pump fault diagnosis method presented in this paper can diagnose faults of single slipper wear, single slipper loosing and center spring wear type with 100% accuracy, and the fault diagnosis time is only 0.002 s. The results demonstrate that the integrated hydraulic pump fault diagnosis method based on MEEMD, AR spectrum, and WKELM methods has higher fault recognition accuracy and faster speed than existing alternatives.

Hydraulic Pump Fault Diagnosis Method Based on EWT Decomposition Denoising and Deep Learning on Cloud Platform

An axial piston pump fault diagnosis algorithm based on empirical wavelet transform (EWT) and one-dimensional convolutional neural network (1D-CNN) is presented. The fault vibration signals and pressure signals of axial piston pump are taken as the analysis objects. Firstly, the original signals are decomposed by EWT, and each signal component is screened and reconstructed according to the energy characteristics. Then, the time-domain features and the frequency-domain features of the denoised signal are extracted, and features of time domain and frequency domain are fused. Finally, the 1D-CNN model was deployed to the WISE-Platform as a Service (WISE-PaaS) cloud platform to realize the real-time fault diagnosis of axial piston pump based on the cloud platform. Compared with ensemble empirical mode decomposition (EEMD) and complementary ensemble empirical mode decomposition (CEEMD), the results show that the axial piston pump fault diagnosis algorithm based on EWT and 1D-CNN has higher fault identification accuracy.

Axial Piston Pump Fault Diagnosis Method Based on Symmetrical Polar Coordinate Image and Fuzzy C-Means Clustering Algorithm

In this paper, a fault diagnosis method based on symmetric polar coordinate image and Fuzzy C-Means clustering algorithm is proposed to solve the problem that the fault signal of axial piston pump is not intuitive under the time-domain waveform diagram. In this paper, the sampled vibration signals of axial piston pump were denoised firstly by the combination of ensemble empirical mode decomposition and Pearson correlation coefficient. Secondly, the data, after noise reduction, was converted into images, called snowflake images, according to symmetric polar coordinate method. Different fault types of axial piston pump can be identified by observing the snowflake images. After that, in order to evaluate the research results objectively, the obtained images were converted into Gray-Level Cooccurrence Matrixes. Their multiple eigenvalues were extracted, and the eigenvectors consisting of multiple eigenvalues were classified by Fuzzy C-Means clustering algorithm. Finally, according to the accuracy of classification results, the feasibility of applying the symmetric polar coordinate method to axial piston pump fault diagnosis has been validated.

Global and Local Feature Extraction Using a Convolutional Autoencoder and Neural Networks for Diagnosing Centrifugal Pump Mechanical Faults

Centrifugal pumps are important types of electro-mechanical machines used for fluid and energy conveyance. Mechanical faults in centrifugal pumps lead to abnormal impacts in the vibration signal of the system. Those impacts induce nonstationarity in vibration signals and hence complex time-frequency domain signal analysis techniques are required to investigate the mechanical fault features of centrifugal pumps. In this paper, an end-to-end pipeline for diagnosing faults in centrifugal pumps is proposed. To create a two-dimensional representation of the transients that appear in the vibration signals due to centrifugal pump operating conditions, first, a 1/3-binary tree fast kurtogram is computed. Next, a convolutional autoencoder and convolutional neural network are trained to autonomously extract global and local features from the kurtograms. Then, global, and local features are merged to form a joined feature vector that contains different visual features that are extracted using convolutional deep architectures using their specific loss functions during the training. Finally, this feature vector is propagated to a shallow-structured artificial neural network to accomplish fault identification. The proposed framework has been validated by the dataset collected from a real industrial centrifugal pump test rig. The results obtained during the series of experimental trials demonstrated that the introduced method achieved high classification accuracies when diagnosing faults based on signals collected under 3.0 and 4.0 bars of pressure.