Quality Control, Quality Assurance, and Fault Diagnosis

The intention of quality assurance and quality control is to recognize defects while the production is in progress or completed and also when defective parts are being or previously manufactured. Because of the inherent potential of neural networks for recognizing patterns and hidden relationships they are employed also for fulfilling different quality assurance, quality control and fault diagnostic tasks. For illustration, a neural network trained with the Back Propagation algorithm and learned to model the relationship in between process measurements or ultrasonic pulses and bond quality has been utilized for predicting wire bond superiority in microcircuit manufacturing.

The presentation of the system is causes and results illustrated the feasibility of neural network for a low cost quality and high reliability assurance systems. The neural networks have been represented to be a good choice to control charts and yet outperformed control charts in condition of small shifts of variance or and means and enhanced type II error rate.

Neural networks coupled along with rule-based expert systems or ES are utilized to diagnose and detect faults or and control difficult automated manufacturing procedures. The hybrid system assembles a more robust intelligent system quite than utilizing either expert systems or neural network alone by combining the strengths of expert systems and neural networks. A particularly helpful feature of the aforementioned system was its capability for self-organization throughout a feedback loop in between the neural network and the expert systems.

Here application of neural networks to quality assurance, quality control, and fault detect comprises composite floor detection, out of control prediction, surface roughness inspection, sampling plan finding, and machine and process fault diagnosis as shown in hierarchy following figure. The neural networks utilized were multilayer ART, feed-forward networks, and stochastic networks.

        Figure: Hierarchy of Neural Network Applications for Fault Diagnosis, Quality Assurance, and Quality Control