Study and application on multi-level process monitoring and predictive fault diagnosis software system framework
Journal Title: Science Paper Online - Year 2007, Vol 2, Issue 12
Abstract
The rapid development of automation and computer technologies makes it possible for process industry to retrieve a huge amount of process data, from which complete information on process operation can be obtained, but its fast increasing amount makes process operators difficult to handle in real-time, especially when more than one alarm are triggered in the operation. The software system framework proposed in this paper realize sensor validation, fault diagnosis and predication in real-time by incorporating the operator’s experience, basic engineering model, and statistical analysis. The case study on active oxidation process of Tonawanda, New York, USA, shows this software can effectively capture sensor failures, and operation faults, therefore, reduce unscheduled downtime more than 50%, save operation cost 2%～3%, and insure the operation safety.
Authors and Affiliations
Wei SUN , Zhongqi REN, Douglas H. Lenz
Keywords
Related Articles
- EP ID EP113062
- DOI -
- Views 157
- Downloads 0
ZnO p-type doping and room-temperature electro-luminescence from ZnO homojunction
This article reports on several technique for p-type doping in ZnO thin films, including N-doping, co-doping, Li-doping, nominally undoped method, large-size-mismatched group-Ⅴ doping. In addition, room-temperature elect...
Functional inequalities for fractional Dirichlet forms
Based on the functional inequality for a given Dirichlet form, the same type inequality is established for the quadric form associated to the generator with a power. As applications, some concrete functional inequalities...
Study on crystallizing phosphorus removal system based on back-propagation neural network
A BP neural network model was developed for the crystallizing phosphorus removal process. Results indicated that the BP model showed a good convergence performance during training, and the model predictions of outflow ph...
一种股骨头假体设计方法
为解决人工髋关节置换手术后假体无菌松动等并发症，适应假体个性化快速设计和制造的需求，采用医用CT技术、计算机辅助设计和有限元计算分析相结合的方法，依据患者股骨髓腔解剖形状和置换要求设计髋关节假体。用有限元软件ANSYS建立股骨...
Study on recognition method based on RBF neural network for <br /> intrusion events of oil and gas pipeline<br />
A recognition method for intrusion events is studied, which is used in the distributed optical fiber pre-warning system for the safety of oil and gas pipeline. In this pre-warning system, which is based on the principle...