Font Size: a A A

Optimization Design And Implementation Of State Estimation And Balancing Management System For Lithium-Ion Batteries In Electric Vehicles

Posted on:2018-05-06Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y L ShangFull Text:PDF
GTID:1362330572953613Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
In today's world,energy crisis and environmental pollution have seriously restricted sustainable development of social economy,which are the severe problems each country has to face.Large-scale development of electric vehicles(EVs)is the primary way to solve the energy and environmental crisis.Vehicle power batteries are the power sources of EVs.Their performances have significant impacts on the power performance,fuel economy,and safety for vehicles,and have became the bottleneck in large-scale development of EVs.Battery management system(BMS)is the key to operate reliably vehicle power batteries.However,BMS still have many key scientific and technological problems,which need to be overcome.Statistics indicate that improper management of BMS is a key factor leading to great reduction in battery capacity and cycle life or even battery failure.Therefore,aiming at the key technical bottlenecks of BMS industrialization development,this dissertation has developed a series of theories and techniques of modeling,estimating,balancing,and heating for vehicle power batteries.This dissertation has mainly finishesd the following research works.Due to the nonlinear characteristic of lithium-ion batteries,i.e.,the very broad voltage plain at the middle of charge or discharge and the serious polarization on both ends,it is difficult to solve the contradiction between the accuracy and practicability by a fixed-structured RC model.Therefore,a variable-integral-order RC equivalent circuit model based on Akaike information criterion(AIC)was firstly proposed.In the exponential area where the battery voltage changes acutely,the high-order RC model is chosen to ensure the model accuracy.In the flat area where the battery voltage changes slowly,the low order RC model is chosen to reduce the model complexity.The simulation and experimental results show that by slightly increasing the model complexity,the proposed variable-integral-order RC equivalent circuit model can simulate precisely the battery nonlinear characteristics with the error of less than 1.5%,which achieves the balance between the accuracy and practicality of the battery model.However,there are abrupt changes in the mode output when the integer-order model is switched.Therfore,a variable-fractional-order equivalent circuit model is proposed to achieve the continuous change in the model order,which further improves the stability and precision of the model output.The proposed variable-order RC equivalent circuit models can be applied to battery state estimation,BMS design,battery test/simulation,and so on,which have important academic and engineering values to ascertain the non-linear characteristics of batteries,the main factors affecting the accuracy of battery model,and to estimate rapidly and accurately battery states.The batteries have the characteristics of strong non-linear,time varying,and noise interference.Moreover,for the traditional SOC(State of Charge)estimation methods based on EKF,the SOC estimation precision is highly dependent on an accurate battery model.To solve the above these problems,a SOC estimation method for Lithium-Ion batteries is proposed based on EKF optimized by fuzzy neural networks(FNN).An error prediction model is built based on fuzzy neural network,by which the measurement noise covariance of EKF is real-time revised.When the predicted model error is smaller,the measurement model is updated,otherwise,the process model is updated only.Therefore,the proposed algorithm can effectively eliminate the SOC estimation error caused by the model error and the uncertain noise statistical properties,with the maximum error of less than 1.2%.Moreover,it has good convergence and robustness.The proposed SOC estimation method has great importance and significance to develop the theory and method system of battery state estimation,to maximize the power and energy of batteries,and to improve the driving range of EVs.The existing active equalizers cannot realize the zero-voltage-gap balancing among cells and have high switching loss.To overcome these difficulties,a direct cell-to-cell battery equalizer based on boost converter and resonant LC converter is proposed to improve the balancing speed and efficiency,and achieve the full balancing among cells.Due to the ohmic internal resistances of batteries,the voltage-based balancing methods have the serious problems of the over-equalization or under-equalization.Therefore,a self-learning fuzzy logic control strategy is proposed,not only greatly reducing the balancing time and the switching cycles but also effectively improving the consistency of the battery string.In order to overcome the disadvantages of the existing active equalizers,including the large size,complex control,and low reliability,two automatic equalizers are skillfully proposed based on switched coupling capacitors and forward-flyback converters,respectively.The proposed equalizers have simple control,which only need one pair of complementary PWM signals.The proposed equalizers achieve the any-cells-to-any-cells automatic equalization,leading to a high balancing speed and efficiency independent of the cell number and the initial cell voltages.Contrary to the conventional equalizers,the proposed equalizers achieve the simultaneous equalization among cells and modules by a simple connection,resulting in smaller size and lower cost.Especially,the balancing operation can be carried out regardless of the battery working state of charging,discharging or rest,achieving the all-time balancing.The proposed equalizers can effectively improve the consistency of the battery pack,maximize the capacity and energy of the battery pack,prolong the battery cycle life,and has high practical value in series-connected battery packs of EVs.Aiming at that lithium-ion batteries suffer severe power/energy loss,reduced life cycle,and bad safety in cold climates,based on the balancing technologies,several automotive on-board AC heaters are proposed to heat lithium-ion batteries at low temperatures without the requirement of external power supplies.Firstly,three internal heating topologies are introduced,i.e.,the basic heater,the interleaved parallel heater,and the integrated heater-equalizer.The basic heating topology has the simplest structure,but only takes half time to heat batteries.The interleaved parallel structure can effectively heat batteries all the time,which not only achieves a higher heating speed and efficiency but also has no more damage to batteries compared with the basic method.The integrated heater-equalizer can heat batteries at lower frequencies and balance cell or module volatges at higher frequencies without the need of additional balancing circuits,consequently increasing the power density of BMS.Furthermore,by utilizing the heat generated in MOSFETs working at the high-frequency hard-switching mode,an internal and external heater is proposed to take full advantage of the energy in batteries,leading to a high heating speed and efficiency.Moreover,in order to provide a guidance for the optimized design of the proposed heater,a thermoelectric model for the internal and external combined heating is developed to accurately predict the battery temperature-rises at different switching frequencies and duty cycles.In summary,the proposed on-board heaters have the advantges of small size,low cost,fast heating speed,high efficiency,ease of control,good uniformity,and high reliability.Particularlly,the heating speed can be regulated online by controlling the switching frequency with a good flexibility,which can meet different application requirements.The proposed heaters can be easily integrated into battery packs,by which "all-climate" and"all-voltage" batteries are immediately achieved without the need of changing battery structures and electrolytes.The proposed heaters are of great significance in engineering practices to improve the available capacity and useful life of batteries,reduce the cost,and increase the driving range of EVs.To sum up,this dissertation has achieved innovative research results in regard to battery modeling,state estimation,balancing,and AC heating,which have been published in the top journals,e.g.,IEEE Trans.on Power Electronics,Industrial Electronics,and Vehicular Technology,etc.,followed by peer experts.This dissertation provides important references and design schemes for high-efficiency,safety,and reliable operation of vehicle power batteries.
Keywords/Search Tags:Variable-order RC models, SOC estimation, battery equalizers, low-temperature heating, battery management systems(BMSs), electric vehicles(EVs)
PDF Full Text Request
Related items