Study on least square modeling of heat engineering object based on typical signal response

This paper proposes a kind of least square modeling method based on typical signal response to enhance modeling accuracy of heat engineering process and adapts the environment that modeling experiment conditions are limited. The principle of this method is, under the condition of known typical pulse, step and slope signal response and model structure, to give algorithm of model parameters of identified continuous system by least square mode through derivation. The method is applied to the identification of heat exchange process for a consumer substation, and identification result obtained is compared with that of other conventional methods. After the comparison the result shows that identification accuracy is improved obviously. In addition to the good identification accuracy, this method has the characteristics such as it can identify directly continuous system model, pure lagging time, and is not sensitive to data length in the identification process. All these characteristics show that this method is simple, easy to implement and has good practicability.