Multi-Mode Modeling And Predictive Control Method For High Speed Multiple Units

High speed multiple units operation is covered with a large number of random, un-certain modal information and so on, involving a series of science problems such as multi–objective optimal control of complex fast giant system in dynamic unknown environme-nt. However, the control of high speed multiple units is different from that of middle-lowspeed trains and it’s difficluty to meet the requirements of modeling and control with suchcomplicated nonlinear dynamical system based on conventional modeling and optimizati-on methods.According to the characteristics of high speed multiple units operation process,multi-mode modeling is used to describe the operation and the corresponding predictivecontrol method is also proposed. Multi-model, adaptive multi-model and T-S adaptivemulti-mode for high speed multiple units are set up, and the corresponding predictivecontroller is designed. Main contents of the paper are summarized as following:1. Train operation is a typical nonlinear process, when the train speed enhances, thenonlinear characteristics become more and more stronger, which results in higher require-ments for the optimal control of the operation process. Firstly, on the basis of the nonline-ar characteristics, the numbers of multi-model are determined by subtraction clusteringalgorithm; then near each working point, the corresponding linear model is bulit with therecursive least squares method; the optimal control model is selected online based onmodel switching scheme with smallest model accumulative error and multi-modelpredictive controller is derived to ensure high precision tracking in both velocity andposition.2. With the continuously high speed operation, the higher demand is needed forreliability and anti-interference ability of high speed multiple units control system. Com-bined with nonlinear dynamics and system operation data, the multiple models set forhigh speed multiple units is built by subtractive clustering and pattern clssificationalgorithm; high speed multiple units adaptive models are employed to adapt to the changeof object and disturb characteristic; the optimal control model is selected online by modelswitching scheme and the corresponding active fault tolerant predictive controller isdesigned to realize the secure and efficient operation of high speed multiple units.3. High speed multiple units operation needs a large consumption of energy resour-ce and different manipulative methods contribute to obvious differences in energyconsumption. How to simulate experienced train drivers experience, expert control know-ledge and ensure energy-saving optimal operation has become a realistic problem to scientific workers. So, an integrated modeling composed of T-S knowledge model andadaptive models is proposed for high speed multiple units operation description, thepredictive controller considering energy consumption indexes and model switching sche-me is comed up with its operation control method.To verify the effectiveness of the methods above, simulation study has been carriedbased on the new generation of CRH2-300high speed multiple units, simulation resultsshow that the proposed methods can better satisfy the needs of high presicion trackingcontrol, fault tolerant control and energy-saving.