Week 7 State of charge estimation
1.
After downloading the specified file Battery_System.prj was opened in Simulink.Then 'Design,Prototype and implement BMS Algorithm' option was chosen.Now it is necessary to open BMS System and the Closed Loop harness Dashboard.BMS system is selected in MATLAB window and it was opened.It was observed that the system had two sub-systems,namely BMS_ECU and PLANT.At first BMS_ECU was opened and then State machine was opened.Next switchover was made to Harness Dashboard which was seen to include several Pushbuttons and Variant Subsystem.Green colored button indicates no error while red colored button indicates error.Test Case I was chosen by clicking 'BMS State Request' and then selecting 'BMS Test Sequence'.Simulation was done and six plots were obtained in Scope.The plots were :Cell Voltage,Cell Temperature,SoC,Pack Current,BMS_State and BalCmd.
BMS_State indicates that for 3000 secs the battery pack is in drive mode and discharge current is flowing.From 3000 secs to 4000 secs the module is in balancing standby mode and afer 4000 secs to 9000 secs it is in charging & balancing mode.After 9000 secs only post charge balancing mode operates upto end of simulation.Initially the SoC of the module is 80%.Yellow plot represents Soc estimation by coulomb counting technique.Blue plot represents SoC estimation by UKF technique,red plot by EKF technique.Initially SoC decreases for 3000 secs,then remains constant upto 4000 secs and thereafter increases and stabilizes at about 94%.During charging the battery pack voltage is increased upto 4.2 V and thereafter stabilizes at around 4.1 V.The cell temperature increases during the driving and charging cycle and starts falling off in the standby mode.The current supplied to the EV during the driving phase can be seen in the current plot.As it is clear from BalCmd balancing of cells start after 3000 sec and at the end of simulation all cell voltages try to balance equally.
Next simulation was done by selecting Test Case II in BMS_Test Sequence and clicking OK.The same set of six plots were obtained.But in this case the nature of plots were different.As seen from BMS_State this represents drive mode where SoC drops from 0.8 to 0.4.Balancing of cells is not performed.Cell voltage varies from 3.9 V maximum to about 3.0 V as cutoff voltage.The pack current discharge is maximum 70 A and slight charging occurs in the course of drive.The pattern of temperature variation is almost same as in the case of Test Drive I.
Again simulation was carried out after selecting Test Case III.As seen from BMS_State the battery pack is operating in charging mode.SoC value rises from 80% to 100% in 2000 secs.Cell balancing takes place during the whole period. Temperature increases upto full charge condition and once full charge is attained it reduces thereafter.Cell voltage rises to about 4.1 V and after full charge the current is cut off.
2.
Coulomb counting is a technique to track the State of Charge of a battery pack.It works by integrating the active flowing current measured in Amps over time to derive the total sum of energy entering or leaving the battery pack.This produces a capacity that is typically measured in Amp Hours. Thus
SoC=Q/Qn where Q is remaining capacity and Qn is rated capacity specified by the manufacturer.
Mathematically Soc(t)=Soc(t-1)+(I(t)/Qn)XDelta t
where I(t) is the discharge current and Delta t is time differential.
SoC generally is dependent on the following parameters:
A.Battery chemistry
B.Charging/Discharging Rate
C.Temperature
D.Impedance
E.Voltage
F.Current
G.Capacity
BMS model includes BMS_ECU and PLANT blocks.For SoC calculation we need to enter PLANT block.Here we find State Machine,SoC Estimation and Balancing Logic blocks.Then we need to select SoC Estimation block.Here again three methods have been outlined and these are: Coulomb Counting, Unscented Kalman Filter and Extended Kalman Filter.
We select Coulomb Counting Method block.It is seen from the block diagram that Coulomb Counting method takes two inputs which are current and temperature.While current input is fed to the Gain block,temperature input goes to a Look up Table for battery time and current.The outputs from these blocks are then fed to a Multiplier Cum Divider block from where the output goes to a Discrete Time Integrator block.This tracks charge by integrating current signal thereby estimating SoC by use of the standard formula.
Advantages of Coulomb Counting method are its simplicity and low Capital Cost.The main drawbacks are : accumulation of current sensor error and incorrect output resulting from wrong initial conditions.Calibration of initial SoC at regular intervals is a feature of this method and this has to be incorporated in BMS design.