Subspace Predictive Control And Its Research On Gasification Of Coal In Dual Fluidized-bed System

Coal is the main energy resource in our country, mainly used for direct combustion.With low utilization and serious pollution, the way of direct combustion has broughtserious resources and environmental problems. Finding comprehensive methods anddeveloping high efficiency and environmental technology for coal utilization is animportant research direction to achieve the goal of energy-saving and emissions reduction.The technology of gasification of coal in dual fluidized-bed(DFB) is newly developed atthe end of the last century, realizing the stage-by-stage coal utilization and with theadvantages of mild operating condition, wide adaptability and low pollution emission.However, the process characteristics of dual fluidized-bed have also brought newproblems for the control system design, which makes it too hard for the conventionalcontrol method to achieve good effect. In this paper, a recursive subspace predictivecontrol strategy based on forgetting factor has been proposed, and supposed to be appliedto the automatic control system of the gasification system. The main works are as follows:This paper systematically analysed the process principle and the dynamicperformance of the gasification technology in dual fluidized-bed. On the basis of itscharacteristic and experimental research, the simplified control architecture was proposed.The system inputs are air/coal ratio and fuel and the output are the gasifier temperatureand gas calorific value.The principles and basic arithmetic of Subspace Identification Methods(SIM) andGeneralized Predictive Control(GPC) was deeply studied, and then deduced the subspacepredictive control arithmetic. For solving the problem of control in dual fluidized beds,which is a kind of complex industrial multi-input, multi-output process with tight couplingand large changes, a recursive subspace identification method was proposed. A forgettingfactor was introduced to increase the convergent rate and improve the performance intracking the time-varying information. The arithmetic proved to be effective through asimulation and combined with the GPC method, the recursive subspace predictive controlarithmetic was put forward. The system model of gasification of coal in dual fluidized-bed was identified byusing the recursive subspace identification arithmetic. Then designed the gasifiertemperature and gas calorific value subspace predictive system. And the simulation resultshows that the system not only meet the requirements of gasifier temperature and gascalorific value, but also assured the stability of air/coal ratio and fuel. The simulationproves that the recursive subspace predictive control arithmetic is completely feasible forthe gasification of coal in dual fluidized-bed system.