多传感器数据融合在气固流化床流型识别中的应用

MULTI-SENSOR DATA FUSION IN IDENTIFYING DIFFERENT FLOW REGIMES IN GAS-SOLID FLUIDIZED BED

在模糊集理论的基础上给出了多传感器信息融合的一般化办法 ,并将这种办法用于气固流化床流型的识别中 .以压差、压力信号的算法复杂性C(n)、涨落复杂性Cf作为特征参数 ,根据特征参数建立了单个传感器判别流型的隶属度函数 .对多个传感器的识别结果进行数据融合 ,最后得到了多传感器对不同流型的识别结果 .实验结果表明 ,应用多参数。

Flow regime identification of gas-solid fluidized bed is a difficult problem in fluidized field. Up to now, it is almost impossible to recognize all the flow regimes with a single sensor or parameter because of the inner complexity of the pressure signal of gas-solid phase system and the existence of transition. A new regime-division method is proposed. Fuzzy language 'membership' is used to describe the degree of transition. Algorithmic complexity C(n) and fluctuation complexity Cf of the pressure signal of a separate sensor are used as nonlinear characteristic parameters to indicate the flow regimes. Simplified models of C(n) and Cf to identify the flow regimes are established according to observation and statistics of experiments. Membership functions are given to indicate the flow regimes according to the simplified models. Data fusion at the feature level is carried out through fuzzy transformation and the identification result of a separate sensor is obtained. Data fusion at the decision level is carried out in the same way and the initial identification results are input into the decision center as local decisions. Finally the identification result of multi-parameters and multi-sensors is obtained. The experimental results show that multi-sensor data fusion can well identify the fluidized states.