RECENT DEVELOPMENTS OF INDUCTION MOTOR DRIVES FAULT DIAGNOSIS

RECENT DEVELOPMENTS OF INDUCTION MOTOR DRIVES FAULT DIAGNOSIS USING AI TECHNIQUES Oly Paz 1

ARTIFICIAL INTELLIGENCE • It is the science and engineering of making intelligent machines, specially intelligent computer programs. • It is important for AI is to have algorithms as capable as people at solving problems, and the identification of subdomains for which good algorithms exit.

Human involvement in the actual fault detection decision making is slowly being replaced by automated tools such as expert systems, neural networks and fuzzy logic based systems.

DATA ADQUISITION SYSTEM

The main step of a procedure can be classified as : • Signature extraction; • Fault identification; • Fault severity evaluation.

Basic stator current monitoring system configuration

Single-phase stator current monitoring scheme

Input current variation for a 5. 5 k. W machine with a load torque of 30 N starting at 0. 5 sec.

DATA RETRIEVING STRATEGIES:

SPECTRUM LINE SEARCH AND FAULT CLASSIFICATION

AI-BASED TECHNIQUES: • • Artificial Neural Networks (ANN), Fuzzy Logic, Fuzzy-NNs, Genetic Algorithms (GAs).

ANN based fault diagnosis

NN-Based Diagnosis Examples ANN architecture for stator short circuit diagnosis. In=negative sequence stator current Ip=positive sequence component of the healthy machine Ir=rated current fp= output fault percentage s= slip sr=rated slip

Fuzzy diagnostic system layout with feature extraction

Fuzzy-Logic-Based Diagnosis Examples Input variables fuzzy sets for I 1

Fuzzy rules for the detection of broken bars fault severity, using as input variables the fault components I 1 and I 2: 3 -D map of the input-output relationships between the sideband components I 1 and I 2

FUZZY NN-BASED DIAGNOSIS EXAMPLES Adaptative ANFIs architecture for rotor fault diagnosis based on the sideband components I 1 and I 2

FAULT DIAGNOSIS OF DRIVES Experimental spectra and instantaneous supply current and output converter current in (a), (b) healthy condition and (c), (d) fault condition.

Stator current Park’s vector pattern

GENETIC ALGORITHMS • GAs are stochastic optimization techniques inspired by laws of natural selection and genetics. They use the concept of Darwin’s theory of evolution, which is based on the ruled of the survival of the fittest. • These algorithms do not need functional derivative information to search for a set of parameters that minimize a given objective function.