A Parameter Estimation Approach For Complex Industrial Systems By Human-computer Interaction

It is often difficult to estimate the facility performance parameters in a complex industrial system. The estimated parameters by only regression methods Iike contradictory to the people's experiences so they can't he accepted by the operators. We propose a new approach to estimate the parameters by human-computer interaction. At first the computer acquires the knowledge about the facility performance from experts then it cut out the unreasonable samplers and determines the above and below boundaries of parameters. The acceptable parameters call be obtained by a bounded parameter estimation algorithm. This approach is used to identify the boiler efficiency curves in a power plant, the results have proved that it is a very efficient and useful tool.