A multi-layer neural network model for detecting changes in the process mean
Computers and Industrial Engineering
Automated unnatural pattern recognition on control charts using correlation analysis techniques
Computers and Industrial Engineering
Hi-index | 0.00 |
The recent trends in optimisation of sustainability of production processes requires, amongst all the activities, a continuous detection and correction of process behaviours, monitoring those parameters critical to performance. Detection of special causes of variations is a basic task in manufacturing, that has to be performed continuously to maintain any process stable as well as predictable. In this paper, a contribution to automate performance control is presented, based on synthesizing Cellular Neural Networks as associative memories for pro-actively recognizing unnatural behaviours. As an example, a test case is developed by considering abnormal cyclic behaviours given by sinusoidal signals. For this purpose, a CNN is synthesized for an associative memory, to recognize these unnatural situations. A robustness analysis of the synthesized network is then developed in the presence of unnatural behaviours in the form of input noises. The behaviour of the designed circuit is illustrated in detail.