An improved neural network method for the prediction of comprehensive production indices in coking process

The coking process, which involves coke oven heating-combustion process, gas-collector collecting process and coke-pushing process, is a very complex industrial process with the feature of multi input and multi output.

On the basis of analyzing coking production technology and system requirements, a web based system realization scheme is proposed to implement real-time monitoring for coking process, and a three-layer B/S architecture is designed which contains data service layer, business logic layer and user interaction layer.

The determination of the end of coking process is an important problem for coke oven plant managers, The aim of this study is to apply on-line GG analysis of raw coke oven gas to determine the end-point and compare its results with criteria based on temperature control at the centre of the charge.

Aiming at the requirement on complexity,security and timeliness for fault diagnosis of coking process,the multi-agentfault diagnosis model for coking system was proposed by the hiberarchy analysis on coking system fault.

However, coke production process is complicated and the production environment is very bad, which leads that the automation level of coke production in China is not high and a series of automations are hard to achieve.

Application of steam turbine in coking production

During the coking production process,the coke pushing with frequently high electric current would cause serious failures to the coking oven body,would highly reduce the service life of the coking oven and would increase the coking production cost.