Control Strategy Optimization Of Hybrid Electric Bus

The control strategy optimization of JieFang Hybrid Electric Bus is studied in thispaper, which is sponsored by the state “863” high-tech program (No.2008AA11A140)“Development of New Vehicle Technologies for the JieFang Hybrid Electric Bus ofFAW”and the jointed action project among Jilin province, FAW, and Jilin university“Pre-research on the Key Technologies for the SecondGeneration Hybrid Electric BUS ofFAW”. There are five parts of contents were proposed in this paper.1. A driving intention fuzzy identification method was proposed.Driving intentionswere classified and driving intention identification parameters were selected.Thetheoretical basis of driving intention identification was elaborated. The membershipfunctions, fuzzy inference rules and driving intention fuzzy identification model based onSIMULINK were built.To test and verify the driving intention identification method, theexperimental verification was done. According to identification result, someparameters’membership functions were optimized by fuzzy-neural network, whichincreases the identification accuracy of driving intention fuzzy identification model.2. A demand torque calculation method of hybrid electric bus was proposed based ondriving intention identification result, which makes demand torque calculation resultaccord with drivers’ demand perfectly.Aiming at drivers of different driving styles,different control strategies of hybrid electric bus were designed which lets control strategyadjust at real time according with the driver’s driving style and makes control strategyhave adaptability.Control strategy simulation was done, which indicates that economycontrol strategy makes the hybrid electric bus’s economy performance better with nodynamic performance decline and dynamic control strategy makes the hybrid electricbus’s dynamic performance better with no economy performance. Comparing withoriginal control strategy, the economy and dynamic performance are better because ofadaptive control strategy based on driving style identification.3. Driving intention prediction method based on Markov chain is proposed. Byanalysis of massive experimental data, driving intention has been analyzed statistically inthe process of driving and one-step transition probability matrix of driving intentionMarkov chain has been calculated. A hybrid vehicle control algorithm based on drivingintention prediction stochastic dynamic programming is proposed. Take the average ofdynamic programming cost function, according to driving intention one-step transition probability matrix and adopt weighted average method to calculate the cost function,which make every driving intention has an effect on next stage cost function based ondriving intention one-step transition probability. This algorithm is independent on drivingintentions at every moment and takes driving intention transition as a stochastic process.On the condition that driving intention one-step transition probability matrix is known, agroup of optimal control variables can be calculated. The control strategy based ondriving intention prediction and other control strategies have been contrasted bysimulation. From the fuel economy aspect, control strategy based on driving intentionprediction cannot obtain optimal control, but it is an approximate optimal control strategy.This control strategy makes dynamic programming algorithm don’t need knowing drivingintention at every moment and lay a solid foundation for the real-time using of dynamicprogramming algorithm in hybrid vehicle control strategy.4. Braking characteristics of hybrid electric bus were analyzed. The minimal value ofbraking distribution coefficient and the maximal value of motor regenerative brakingforce were calculated.Computing method of braking demand severity was optimizedbased on braking intention identification result. A regenerative braking control strategybased on braking intention identification was proposed. Motor brake is mainly used in thebrake process according to the brake emergency degree, so energy recovery ratio isincreased.5. Bench test was done to prove that the bench has functions and operation modes ofhybrid electric vehicles under the control strategy. Dynamic and economic performancetests of prototype vehicle were done on the heavy-duty vehicle chassis dynamometer ofFAW R&D center to prove that prototype vehicle has functions and operation modes thatare preconceived.At dynamic and economic performance aspect, prototype vehicle testresult is not better than simulation result, but is better than FAW first generation hybridelectric bus.