Condition monitoring and protection of induction reduced device considering reliability aspects

Abstract

Induction machines are the most frequently used electrical machines in domestic and industrial processes. Around 85% of motors used in industrial appliances are induction machines. The main reason behind it is lower cost, ruggedness, robust in structure, lower maintenance requirement, easiness in availability and capability to work under severe working atmosphere. The fault in the induction motor distracts the overall production of the industry, which may lead to increase the idle time and losses of revenue. In order to decrease the down time and for reliable and safe operation, fault recognition in early stage is desirable which necessitates condition-based monitoring of the induction motor. The basic principle of condition monitoring (CM) lies on investigating the running characteristics of the machine such that prediction for maintenance is done prior to breakdown or deterioration to occur in order to introspect the health monitoring of the machine. In this context, the individual part’s life or the life of the whole machine is critically analyzed. In this direction the correct data acquiring process and the data analysis is done in order to capture the trends that might occur. The maintenance based on time investigate the machines repair in offline mode in accordance with time schedule are working hour that leads to avoid the probability of failure. However unwanted shutdown or sudden accidents that may occur in the stipulated period should be taken into account in order to explore the health of the equipments. Thus fast fault detection in early stage can improve the performance of the motor and reduce the consequential harms, breakdown repairs, decrease the cost of maintenance and unpredicted failure risk is remarkably reduced with the availability of the machine. Accurate identification scheme urges the methodologies to be implemented in the direction of condition-based preventive and predictive maintenance rather than conventional time-based maintenance. In this context, the focus of condition based repair is to illustrate the evaluation accurately and identifies the fault a prior. Condition based maintenance leads to the set of information about the machine’s state and focuses CM approach followed by efficacy of the type of maintenance needed in order to reduce the manpower. The said scheme would not lead to halt the machine accidentally. Further protection issues with prognostics of condition monitoring is not been adequately discussed in the recent literatures. Subsequently to address the protection issues thermal relay has been considered and its detailed modelling followed by the operational aspect has been discussed in this thesis. Reliability is an important indicator associated with condition monitoring as a matter of fact different reliability indices and its impact on condition monitoring is discussed which would obviously augment the reliability aspects of condition based maintenance. Further for inter-turn fault analysis pattern classification scheme is explored for accurate detection and comparative assessment of different classification scheme such as ANN and SVM is carried out. The thesis organization and brief focus of different chapters are illustrated as mentioned below. Chapter 1 presents needs of induction motor condition monitoring, various faults and their root causes, various condition monitoring techniques employed for induction motor condition assessment, major challenges. Chapter 2 presents the thermal protection theory of induction motor, it’s first-order thermal model, and identification of parameters’ by using particle swarm optimization technique. Chapter 3 illustrates the theoretical background of mechanical overload and voltage unbalance on induction motors, motor protection with NEMA norms, and implementation of protection schemes under overload and unbalance voltage condition by using MATLAB™ Simulink for induction motor. Chapter 4 presents the analysis of 1hp, 400Volts, induction motor by ANSYS Maxwell electromagnetic field simulation software for inter-turn fault, experimental set up has been made for inter-turn fault, signature of various current signal has been taken at no load as well as different loading condition and their THD has been calculated, finally the ANN (Artificial Neural Network) and SVM (Support Vector Machine) have been used for detection of fault in three phase induction motor. It has been observed that SVM gives better accuracy in comparison to ANN. Chapter 5 describe the hazard model of induction motor in brief, the attempt is made to comprehensively discuss the reliability, MTBF, and the failure rate of induction motors have been evaluated with the help of the industrial data. It has also been described mathematically how the purchase of a standby machine increases the reliability of system operation. For various failure modes, preventive and suggestive methods are demonstrated in order to reduce the faults in induction machines. The reliability indices (reliability, MTBF and failure rate) v/s operational time curve based on probabilistic evaluation has been demonstrated successfully. This curve is very helpful to provide information about the planning maintenance schedules to obtain reliable operation without interruption.Chapter 6 presents the main conclusions and recommendations for future work.