Ioannis D. Chasiotis

An Intelligent Multi-Agent System Framework for Fault Diagnosis of Squirrel-Cage Induction Motor Broken Bars

Decision making in fault diagnosis is a critical factor in industrial production maintenance. Artificial intelligence (AI) techniques are widely used to accurately identify faults in induction motors. Multi-agent systems (MAS) as a distributed AI method are efficient in representing intelligent manufacturing systems to achieve decision robustness. In this paper, an intelligent MAS is developed for the decision making in fault diagnosis of three-phase squirrel cage induction motor rotor bars. Agents in the proposed MAS represent induction motor in different health conditions, i.e. healthy motor and motor with 1, 2 and 3 broken bars and also a supervisor agent. Each agent is embedded with an artificial neural network and trained with measurement data taken from a motor in the corresponding health condition. Measurement data are obtained with the classical motor current signature analysis (MSCA) method. Each agent makes local decision making and communicates its output to the supervisor agent that makes the final fault diagnosis based on a threshold value.

Fault diagnosis of squirrel-cage induction motor broken bars based on a model identification method with subtractive clustering

Fault diagnosis in electric motors is a field that evolves and grows constantly, aiming at their effective maintenance and protection scenarios under the lowest possible cost. Especially for induction motors, since they are of fundamental importance to the industry worldwide, many techniques and methodologies for the early fault detection-diagnosis have been proposed so far. In this paper, an attempt is made to develop a mechanism in order to diagnose faults in a three-phase squirrel cage induction motor rotor bars. The concept is implemented by primarily taking into account the information extracted from the classical motor current signature analysis (MSCA) and then a model identification method approach is formulated using data set manipulation known as subtractive clustering. The method is based on adaptive neuro fuzzy inference system (ANFIS). An investigation on the validity of the proposed method is performed, through experimental data taken from a healthy motor operation as well as those from the same motor with 1, 2 and 3 broken bars. From the derived results it is shown that they present satisfactory sensitivity and accuracy characteristics and thus the proposed method may be a suitable candidate mechanism in the early rotor bar fault detection phase of induction motors.

Cooling system design and thermal analysis of an electric vehicle's in-wheel PMSM

This paper proposes a simple though effective cooling system, which has been designed for a light permanent magnet synchronous motor (PMSM) with outer rotor topology, for an in-wheel electric vehicle (EV) application. Various thermal considerations, found in the recent literature and a large number of variables are taken into account in order to ensure that the cooling system meets the special features for this kind of motor. In this context, numerous simulations using Finite Elements Method (FEM) are conducted first, for the determination of temperature distribution at the various parts of the machine under different operating conditions. Next, an investigation of the proposed cooling system performance is done by examining some parameters, such as the coolants inlet temperature and the corresponding flow rate. Based on these results the final specifications of the cooling system are determined. Finally, the cooling systems efficiency according to the motors output power range is examined.

PMDC coreless micro-motor parameters estimation through Grey Wolf Optimizer

Compared to conventional iron-core brush-type dc motors, the coreless types have no iron losses, low friction, and acceptable thermal dissipation, which make them extremely efficient. The design of their low-inertia rotor is the key to rapid acceleration and fast reaction time characteristics. In addition, their linear behavior fit well with simple drive circuits. A sub-category, namely coreless micro-motor (CMM), is especially considered in low-power equipment, as well as many medical and industrial precision tools. For an analytical control system design and optimization though, their precise model used in the control system is often needed. In this case, the values for reference of the motor parameters should be estimated accurately, regardless the given motor specifications provided by the motor manufacturer. In this context, the application of a newly introduced meta-heuristic optimization method called “Grey Wolf Optimizer” (GWO) is studied here. The effectiveness of the method in estimating accurately the PMDC CMM parameters by using speed step responses only, is then evaluated. The obtained results of the applied method reveal very satisfactory performance when compared with those of another similar meta-heuristic technique.

A Simple Knowledge Base Software Architecture for Industrial Electrical Machine Design: Application to Electric Vehicle’s In-Wheel Motor

The paper presents the application of a knowledge based software architecture (KBSA) scheme which has been developed and implemented in order to be used as a tool in the electrical machines design industrial process. The proposed scheme’s layers are introduced, considering several impact factors from many points of view (i.e. technical, material, algorithmic, economic etc), as well as their interference. It is evident that the specific engineering design problem poses inherent demand for a knowledge representation framework that could support the entire life cycle: requirements, specification, coding, as well as the software process itself. In this context, the work continues by presenting design results of the implemented KBSA for a certain type of permanent magnet motor currently under research in electric vehicle industry, for an in-wheel electric vehicle (EV) application. The KBSA employs evolutionary algorithms for the systematic optimization and the results reveal the effectiveness of the aforementioned procedure followed.

Optimization of Standard PMDC Motors Used in Automotive Applications for Higher Power Density

Aim of this work is the development of a competitive alternate design topology (in terms of power density) of a small permanent magnet dc (PMDC) motor found in automotive applications. Initially, a real industrial motor is measured, designed and simulated, while its measurements and the relevant manufacturer data are considered as a benchmark. In turn, through custom developed software, a redesigned configuration is proposed regarding the structural (stator, rotor, magnets) geometry and magnet material. The resulting geometry was obtained through a constrained optimization algorithm having as goal the minimization of the overall volume and it was further verified by commercial finite element method (FEM) analysis software. Also, the new model is compared with the benchmark motor. Last but not least, FEM analysis was used for thermal behavior evaluation. The overall results reveal that the energy density and the performance of the proposed topology were substantially increased, while the cost was remained low.

On the neural network single phase induction motor efficiency estimation as a design tool

This paper presents a modified design approach as well as a practical and effective neural network efficiency estimation procedure of a permanent capacitor single-phase induction motor (SPIM). The standard industrial motor frame sizes along with the current design trends of larger lengths and smaller diameters are taken into account, which are not likely presented in literature. In this context, a computer-aided design approach of a SPIM -based mainly on the classical output coefficient-is proposed first. Numerous simulations using Finite Elements Method (FEM) are conducted in order to verify the proposed procedure and investigations regarding the number of stator and rotor slots are performed. Secondly, based on the previous results, a neural network (NN) scheme is proposed in order to estimate the efficiency of SPIMs with various stator/rotor slots combinations and different output power. It is seen that the proposed methodology is verified satisfactorily and could be of great use as an aid tool to industrial SPIM designers.