The Soft Measurement Method Study Of Thermal Parameter Based On State Observer

Soft measurement method based on mechanism modeling and data fitting has beenwidely applied in thermal parameter measurement of coal-fired power plant, whichsolved problem that some important thermal parameters are hard to be measured on-linereal-time with conventional sensor, such as oxygen, air volume, the ash content of coal,coal calorific value and moisture, etc. The model established by mechanism method is akind of static relationship, which fully considered the various instrumental variables tothe measure static effect. The result under the calibration working condition of thestability has high precision, and can satisfy experiment requirement. However, from theangle of time domain analysis, the model relies too much on the established staticfunction relationship, not considering the dynamic characteristics of each variable. Thereis no problem under steady state conditions, but a huge dynamic error in dynamiccondition. For example of the coal calorific soft measurement based on the mechanismanalysis, because the boiler is big inertia and large delay, when the load was up, the coalflew first increased while the total quantity of boiler heat was changeless, coal calorificvalue simply by the total quantity of heat of boiler divided by the coal flew got lower; Onthe contrary, when the load was down, it got higher.The small inertial states in the thermal object or constructed by compensation areahead of the system output greatly. Most of these states cannot be measured directly.They are estimated by state observer as secondary variables of soft sensor to improve thedynamic characteristics of soft measurement based on mechanism analysis. The coalcalorific value as an example, dynamic compensation method of thermal parameter softmeasurement was studied. According to the load-pressure model of unit, the dynamicrelationship between the boiler fuel flow and the total heat is described by a high orderinertia link. By designing a state observer for the high order inertia link, the intermediatestate with smaller inertia is obtained, which is advanced compared with the total heat andcan be use to compensate the lag caused by the fuel total quantity of heat with big inertiaand long delay. The soft measurement model is established by using the observedintermediate state, which can reflect the coal calorific value change quickly. A speed limitmedian filter was added after the model to make the dynamic characteristics of resultbetter. The simulation experiment of variable load and variable calorific value of coal inMatlab environment, the simulation curve showed that dynamic compensation resultremain unchanged under variable load and response to the change of coal quickly.The furnace simulation experiment was carried out through the actual operation ofthe offline data on a660MW pulverized coal. The experiment results show that: The coal calorific value, calculated by dynamic compensation method for parameters softmeasurement based on state observer, has small dynamic error under varying loadconditions and can quickly response to the new coal calorific value under variable coalquality conditions. It was validated effectively that the presented method has betterbehaviors on load side and boiler side disturbance restraint and dynamic response, andmeets the demands of precision and response time in practical industrial application.