scholarly journals A Cost-Aware DNN-Based FDI Technology for Solenoid Pumps

Electronics ◽  
2021 ◽  
Vol 10 (19) ◽  
pp. 2323
Author(s):  
Suju Kim ◽  
Ugochukwu Ejike Akpudo ◽  
Jang-Wook Hur
Keyword(s):  
Activation Function ◽  
Operating Conditions ◽  
Fault Detection And Isolation ◽  
Vibration Monitoring ◽  
Diagnostic Model ◽  
Test Accuracy ◽  
Critical Function ◽  
Hand Signal ◽  
Hyperparameter Selection ◽  
Filter Clogging

Fluid Pumps serve a critical function in hydraulic and thermodynamic systems, and this often exposes them to prolonged use, leading to fatigue, stress, contamination, filter clogging, etc. On one hand, vibration monitoring for hydraulic components has shown reliable efficiencies in fault detection and isolation (FDI) practices. On the other hand, signal processing techniques provide reliable FDI parameters for artificial intelligence (AI)-based data-driven diagnostics (and prognostics) and have recently attracted global interest across different disciplines and applications. Particularly for cost-aware systems, the choice of diagnostic parameters determines the reliability of an FDI/diagnostic model. By extracting (and selecting) discriminative spectral and transient features from solenoid pump vibration signals, accurate diagnostics across operating conditions can be achieved using AI-based FDI algorithms. This study employs a deep neural network (DNN) for fault diagnosis after a correlation-based selection of discriminative spectral and transient features. To solve the problem of hyperparameter selection for the proposed model, a grid search technique was employed for optimal search for parameters (number of layers, neurons, activation function, weight optimizer, etc.) on different network architectures.The results reveal the high accuracy of a three-layer DNN with ReLU activation function, with a test accuracy of 99.23% and a minimal false alarm rate on a case study.

A New Model-Based Tool for Fault Detection and Isolation in Machine- and Rotordynamics

1995 ◽  
Author(s):  
Dirk Söffker
Keyword(s):  
Decision Making ◽  
Fault Detection ◽  
Signal Analysis ◽  
Power Plants ◽  
Crack Detection ◽  
Fault Detection And Isolation ◽  
Vibration Monitoring ◽  
Monitoring Systems ◽  
Implicit Assumption ◽  
Signal Parameters

Abstract Reliability and safety aspects are becoming much more important due to higher quality requirements, complicated and/or connected processes. The fault monitoring systems to be commonly used in machine- and rotordynamics are based on signal analysis methods. Furthermore, various kinds of fault detection and isolation (FDI)-schemes are already applied to a lot of technical applications of detecting and isolating sensor and actuator failures (Isermann, 1994; van Schrick, 1994) and also to fault detection in power plants (in general) or in manufacturing machines. An implicit assumption is that process or machine changes due to faults lead to changes in calculated parameters, which are unique and unambiguous. In the case of applying methods of signal analysis this means spectrums etc. the vibration behaviour will be monitored very well but have to be interpreted. On the other hand signal parameters usually only describe the system by analyzing output signals without use of known and unknown inner parameters and/or inputs. These parameters are available, and normally this knowledge is used by the operating staff interpreting the resulting signal parameters. In this way a decision-making problem appears so that questions about the physical character of faults, about the existence of special faults and also about the location of failures/faults has to be answered. In this way the experience and knowledge of the interpreting persons are very important. In this contribution the problems of the decision-making process are tried to defuse: • The available knowledge about the unfaulty system parameters is used to built up beside a nominal system model an unambiguous fault-specific ratio. Inner states of the structure are estimated by an PI-observer. • The developed robust PI-observer (Söffker et al., 1993a; Söffker et al., 1995a) estimates inner states and unknown inputs. In (Söffker et al., 1993b) this new method is applied to the crack detection of a rotor, but not proved. In this paper the proof is given and a generalization is described. The advantages in contrast to usual signal based vibration monitoring systems and also modern FDI-schemes are shown.

Analysis of Fin-Tube Evaporator Performance With Limited Experimental Data Using Artificial Neural Networks

2000 ◽  
Author(s):  
Arturo Pacheco-Vega ◽  
Mihir Sen ◽  
Rodney L. McClain
Keyword(s):  
Neural Network ◽  
Heat Rate ◽  
Network Models ◽  
Activation Function ◽  
Operating Conditions ◽  
Training Data ◽  
Neural Network Models ◽  
The Neural Network ◽  
Artificial Neural ◽  
Fin Tube

Abstract In the current study we consider the problem of accuracy in heat rate estimations from artificial neural network models of heat exchangers used for refrigeration applications. The network configuration is of the feedforward type with a sigmoid activation function and a backpropagation algorithm. Limited experimental measurements from a manufacturer are used to show the capability of the neural network technique in modeling the heat transfer in these systems. Results from this exercise show that a well-trained network correlates the data with errors of the same order as the uncertainty of the measurements. It is also shown that the number and distribution of the training data are linked to the performance of the network when estimating the heat rates under different operating conditions, and that networks trained from few tests may give large errors. A methodology based on the cross-validation technique is presented to find regions where not enough data are available to construct a reliable neural network. The results from three tests show that the proposed methodology gives an upper bound of the estimated error in the heat rates.

On Possibilities of Using Relative Shaft Vibration Signals for Rotating Blades Monitoring

2018 ◽  
Author(s):  
Jindrich Liska ◽  
Vojtech Vasicek ◽  
Jan Jakl
Keyword(s):  
Steam Turbine ◽  
Operating Conditions ◽  
Vibration Monitoring ◽  
Monitoring Systems ◽  
Blade Vibration ◽  
Vibration Signals ◽  
Rotating Blades ◽  
Timing Measurement ◽  
Blade Tip ◽  
Shaft Vibration

Ensuring the reliability of the steam turbine is the key for its long life. For this purpose monitoring systems are standardly used. Early detection of any failure can avoid possible economical and material losses. A monitoring of rotating blades vibration belongs to the very important tasks of the turbomachinery state assessment. Especially in terms of the last stages of low-pressure part, where the longest blades are vibrating at most. Commonly used methods for blade vibration monitoring are based on contact measurement using strain gauges or non-contact approach based on blade tip timing measurement. Rising demand for low-cost monitoring systems has initiated development of a new approach in blade vibration monitoring task. The presented approach is based on usage of relative rotor vibration signals. Its advantage is in using of standardly installed sensors making this approach economically interesting for the turbine operators compared to the traditionally used methods, mentioned above. This paper summarizes the symptoms of blade vibration phenomenon in relative shaft vibration signals, the impact of operating conditions on the blade vibration amplitude and its comparison to blade tip-timing measurement results. In addition of several examples, the article also describes an evaluation of proposed method in operation of steam turbine with power of 170MW.

scholarly journals Intelligent Fault Diagnosis Techniques Applied to an Offshore Wind Turbine System

2019 ◽  
Vol 9 (4) ◽  
pp. 783 ◽  
Author(s):  
Silvio Simani ◽  
Paolo Castaldi
Keyword(s):  
Fault Diagnosis ◽  
Wind Turbine ◽  
Operating Conditions ◽  
Fault Detection And Isolation ◽  
Offshore Wind ◽  
Data Driven ◽  
Offshore Wind Turbine ◽  
Wind Turbine System ◽  
High Degree ◽  
Offshore Plants

Fault diagnosis of wind turbine systems is a challenging process, especially for offshore plants, and the search for solutions motivates the research discussed in this paper. In fact, these systems must have a high degree of reliability and availability to remain functional in specified operating conditions without needing expensive maintenance works. Especially for offshore plants, a clear conflict exists between ensuring a high degree of availability and reducing costly maintenance. Therefore, this paper presents viable fault detection and isolation techniques applied to a wind turbine system. The design of the so-called fault indicator relies on an estimate of the fault using data-driven methods and effective tools for managing partial knowledge of system dynamics, as well as noise and disturbance effects. In particular, the suggested data-driven strategies exploit fuzzy systems and neural networks that are used to determine nonlinear links between measurements and faults. The selected architectures are based on nonlinear autoregressive with exogenous input prototypes, which approximate dynamic relations with arbitrary accuracy. The designed fault diagnosis schemes were verified and validated using a high-fidelity simulator that describes the normal and faulty behavior of a realistic offshore wind turbine plant. Finally, by accounting for the uncertainty and disturbance in the wind turbine simulator, a hardware-in-the-loop test rig was used to assess the proposed methods for robustness and reliability. These aspects are fundamental when the developed fault diagnosis methods are applied to real offshore wind turbines.

Diagnosis for Engine Misfire Fault Based on Torsional Vibration and Neural-Network Analysis

2012 ◽  
Vol 433-440 ◽  
pp. 7240-7246
Author(s):  
Can Yi Du ◽  
Kang Ding ◽  
Zhi Jian Yang ◽  
Cui Li Yang
Keyword(s):  
Neural Network ◽  
Torsional Vibration ◽  
Bp Neural Network ◽  
Signal Analysis ◽  
Signal To Noise Ratio ◽  
Vibration Signal ◽  
Correction Method ◽  
Operating Conditions ◽  
Diagnostic Model ◽  
Vibration Signal Analysis

Misfire is a common fault which affects the engine performances. Because the signal-to-noise ratio of torsional vibration signal is high, torsional vibration test and analysis for the engine were performed in a variety of operating conditions, including healthy condition and single-cylinder misfire condition. In order to improve the accuracy of analysis, energy centrobaric correction method was used to correct the amplitude. Taking the corrected amplitude of main order as the fault feature, and then a BP neural-network diagnostic model can be established for misfire diagnosis. The result shows that the method of combining torsional vibration signal analysis and neural-network can diagnose engine misfire fault correctly.

Mahalanobis Taguchi System (MTS) as a Prognostics Tool for Rolling Element Bearing Failures

2010 ◽  
Vol 132 (5) ◽  
Author(s):  
Ahmet Soylemezoglu ◽  
S. Jagannathan ◽  
Can Saygin
Keyword(s):  
Fault Detection ◽  
Real Time ◽  
Operating Conditions ◽  
Fault Detection And Isolation ◽  
Numerical Control ◽  
Rolling Element Bearing ◽  
Time To Failure ◽  
Cnc Machine ◽  
Bearing Failures ◽  
Rolling Element

In this paper, a novel Mahalanobis–Taguchi system (MTS)-based fault detection, isolation, and prognostics scheme is presented. The proposed data-driven scheme utilizes the Mahalanobis distance (MD)-based fault clustering and the progression of MD values over time. MD thresholds derived from the clustering analysis are used for fault detection and isolation. When a fault is detected, the prognostics scheme, which monitors the progression of the MD values, is initiated. Then, using a linear approximation, time to failure is estimated. The performance of the scheme has been validated via experiments performed on rolling element bearings inside the spindle headstock of a microcomputer numerical control (CNC) machine testbed. The bearings have been instrumented with vibration and temperature sensors and experiments involving healthy and various types of faulty operating conditions have been performed. The experiments show that the proposed approach renders satisfactory results for bearing fault detection, isolation, and prognostics. Overall, the proposed solution provides a reliable multivariate analysis and real-time decision making tool that (1) presents a single tool for fault detection, isolation, and prognosis, eliminating the need to develop each separately and (2) offers a systematic way to determine the key features, thus reducing analysis overhead. In addition, the MTS-based scheme is process independent and can easily be implemented on wireless motes and deployed for real-time monitoring, diagnostics, and prognostics in a wide variety of industrial environments.

Nonlinear Fault Detection and Isolation for a Lithium-Ion Battery Management System

2010 ◽  
Author(s):  
Jim Marcicki ◽  
Simona Onori ◽  
Giorgio Rizzoni
Keyword(s):  
Fault Detection ◽  
Hybrid Electric Vehicle ◽  
Sliding Mode ◽  
Estimation Error ◽  
Lithium Ion ◽  
Open Loop ◽  
Operating Conditions ◽  
Fault Detection And Isolation ◽  
Battery Management System ◽  
Loop Model

Lithium-ion batteries are a growing source for electric power, but must be maintained within acceptable operating conditions to ensure efficiency and reliability. Therefore, a robust fault detection and isolation scheme is required that is sensitive enough to determine when sensor or actuator faults present a threat to the health of the battery. A scheme suitable for a hybrid electric vehicle battery application is presented in this work. The diagnostic problem is formulated as a nonlinear parity equation approach, but is modified for the considered application. Sliding mode observers are designed for input estimation, while the output voltage estimation is performed using an open loop model. The selection of optimal thresholds given a maximum allowable probability of error is also considered. An assessment of the design using real-world driving-cycle data leads to the conclusion that the estimation error of the observers determines a lower bound on the minimum detectable fault magnitude.

Fault Detection and Lubricant Health Monitoring for Slow Bearings in Extreme Operating Conditions

2005 ◽  
Author(s):  
Pattada Kallappa ◽  
Carl Byington ◽  
Bryan Donovan
Keyword(s):  
Fault Detection ◽  
Remaining Useful Life ◽  
Operating Conditions ◽  
Vibration Monitoring ◽  
Time To Failure ◽  
Grease Lubrication ◽  
Detection And Identification ◽  
Fault Detection And Identification ◽  
System Data ◽  
Actuation Systems

Slow rotating bearings are an integral part of aerospace and turbomachinery actuation systems. These actuation systems may be driven by electric, hydraulic or fueldraulic power and often operate under high loads and extreme temperatures. This makes these actuation systems and their slow rotating bearings highly susceptible to degradation and failure. Vibration monitoring techniques are not applicable to the PHM of these bearings, because their slow speeds are unable to produce a measureable vibration signature. Furthermore, the slow bearings are sealed and use grease lubrication, thus eliminating traditional oil debris monitoring. To address these problems, Impact Technologies, LLC has developed a PHM system that relies on system identification and uses available control system data and sensor measurements. This PHM system consists of algorithms and models that perform fault detection and identification for the bearings and its actuation train components like valves, pumps, motors, gears and bearings. The PHM process is divided into two stages — diagnostics and prognostics. Diagnostics is the process of detecting and isolating faults, while prognostics is the process of predicting remaining useful life (RUL) or time to failure. The authors demonstrate the PHM system through simulation on a dynamic model that is representative of hydraulic-mechanical actuation systems used in new and existing manned aircrafts, UAVs and Short Take-off and Vertical Landing aircrafts.

Sign in / Sign up

Export Citation Format

Share Document