Abstract

Fault detection in chemical production processes is difficult because of the amount of data that needs to be analysed and the absence of fault data. This paper uses support vector machines (SVMs) to diagnose faults, as it fits the small sample problem. To eliminate system disturbances and noise from the high levels of data, it then combines SVMs with a novel quantum ant colony optimization (QACO) algorithm to select the fault features. The proposed method is then tested using the benchmark Tennessee Eastman Process and shown to be effective. In particular, it is demonstrated that QACO and SVMs can find the essential fault variables exactly, and will greatly improve the fault diagnosis performance of SVMs for a complex chemical process.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call