A higher order statistics based pneumatic regulating valve friction detection and diagnosis of non-invasive method. Nonlinear exist using non Gauss index and nonlinear index detection control valve, to determine whether the fault caused by friction with the controlled variable and the output of the controller of the coordinate chart, through the simulation example shows the effectiveness of the proposed method.
A typical chemical plant have hundreds or even thousands of control loops, performance and plant control loop is closely related to the economic benefit. Concussion is main reason caused control loop control performance degradation. The controlled variable in the loop of shock caused the decline in the quality of products, defective rate, increases energy consumption, increase production efficiency and reduce. The only activity control loop part is the control valve. If the control valve includes nonlinear, for example: friction, backlash and dead zone, valve output may shock, this will cause the process output shocks. In many kinds of nonlinear control valve, friction is the most widespread and long-term problems, it not only reduces the performance of control valve, but also leads to a decrease in control loop performance. The use of invasive method (control valve in the non-working state, detection and diagnosis of the fault (1)), such as stroke detection, can detect friction easily. But will this invasive method is applied to the detection of plant in the whole factory in hundreds or more control valve which is time-consuming, it is not feasible.
Although there are a lot of invasive method to analysis performance of the control valve (2 ~ 5), analysis and research for non-invasive methods rarely appeared in the literature. The Horch method successfully detect friction flow in the circuit, but it can not be applied to compressible fluid (6). Time trend method proposed by Ren2gaswamy is dependent on the data, but the impact of this often affected by noise or interference. The trend curve data affected by the process and the controller to a great extent (7). Stenman proposed a model based approach to the detection of friction control valve (8), this method needs to know the process model and a large number of tuning parameters, and obtain closed loop model from the daily operation of the data is very difficult.
A kind of data based on the non invasive method can effectively reduce the cost to maintain the desired control performance. This paper introduces a kind of data model not based on non-invasive methods, this method is not imposed on the system, additional excitation or experiment, as long as the system input and output data under normal operation condition to estimate the dynamics of the process, so the application is simple, it is these advantages make it useful the control system performance testing tool
The article explains:
1, this is the network editor reproduced or the release, the release of the article aims to convey more information to visitors.
2, involves the content of the works, copyright and other issues, please contact with the network in 30 days, we will make the appropriate treatment for the first time! The relevant copyright matters please contact: 13515194999
3, this article please indicate are from: Jiangyin OuLeiShi valve valve repair factory www.olschina.com
4 other links:
Valve repair valve repair valve repair equipment valve grinding machine import valve repair valve repair valve repair
5, the article source:
|
