Week 7 State of charge estimation
AIM:
1. To simulate the model for 3test cases from harness dashboard and write a detialed report on the results.
2.To explain the coulomb counting referring to the above model and explain how BMS implements coulomb counting for SOC estimation
DESCRIPTION:
1.TO SIMULATE THE MODEL FOR 3TEST CASES FROM HARNESS DASHBOARD AND REPORT ON RESULTS:
A BMS file is extracted and file folder is opened in MATLAB and harness model is opened from the file downloaded from the link given https://in.mathworks.com/matlabcentral/fileexchange/72865-design-and-test-lithium-ion-battery-management-algorithms and selected the option of open project and model as shown below
It opens the two models as below
This 1st model of BMS closed loop consists of two subsystems: ECU and Plant unit.
The plant output from plant unit is given as input to BMS ECU as BMS_input feedback loop. The state request is given to BMS ECU unit.
The BMS output command given as input to plant unit as From_BMS.Alsp the state request is given to plant unit.
The final outputs from closed loop model are BMS_input, BMS_Output, BMS_info.
The BMS ECU unit consist of current power limit calculations carried out by using state machine block, SOC Estimations, Cell balancing logics are as shown above
The above plant unit consist of the battery unit pack, pre charge unit, charger/load units.
In battery packs there are two types of configurations as module1 (6cells) and 16modules (96cells).
Here we are choosing module1 for lesser simulation running time.
The second model of test harness consists of the following
1.Test Sequence varient (for selecting the test case)
2.Manual varient(for selecting the type of varient)
3.BMS_Test sequence state request i.e a sub-system to provide state request commands.
4.Lamps to indicate fault state, High temperature, over voltage, over current, charging, voltage sensor operation, inverter operation, low temperature, under voltage. If any fault interrupts these lamps glow in red and all good operations are indicated by green lamps.
5.A rotating switch for manual varient changes while running simulation.
6.The scope results the SOC of module, Module current, Individual cell voltages, Individual cell temperature, Battery state, cell balancing states are seen as outputs.
Now, running the model for 3 test cases by selecting the test sequence varient from dialog box and to click apply as shown
CASE1:
This above selection selects the 1st sequence from the BMS_state request sub-system as shown below
The test case consists of the following parameters
The parameters for 1st sequence is set as
1.Driving for first 3000secs
2.After 3000sec starts balancing operations occurs for next 1000 sec
3.Then charging and balancing occurs for next 5000secs.
4.After that post charge sequence occus till end of simulation perios.
CASE1 RESULTS:
After running the simulation the results for Case 1 are as follows
All lamps are glowing in green which shows that there are no faults in the simulation and the plot generated is as shown below
This shows that for first 3000sec are in balance mode and the voltage varies from 3.1 to 3.9v. From 3000 to 4000secs cell balancing occurs and in this stage the module is in standby mode and voltage is constant at nearly 3.7v. After 4000sec charging occurs, voltage raised from 3.7v to 4v and from 4v to 4.2v at constant current at nearly 30A and after 4.2v as constant voltage, current reduces from 30A to 2.6A and the cell balancing occurs upto 9000seconds. After 9000sec post charge cell balancing starts and all cell voltages are balancing to a voltage with respect to cell 1 to 4.1v as showin.
Temperature for intial 3000sec increases from 288K to 302K. The temperature of cell 1 is lesser and cell 6 is higher for all stages. After 3000 to 4000sec the cell temperatures gets slightly reduced as no current flows through in standby mode. After 4000sec temperatures starts increases exponentially upto CC and then reduces in CV and till cell balancing upto 9000sec.
For 3000sec the battery module is in drive state (discharging). From 3000 to 4000sec the battery module is in standby.After 4000 to 9000 sec the battery module in charging and in balancing mode.
SOC is initially at 80%. Yellow plot reperesents SOC estimation done by coulomb counting.Blue plot represents SOC estimation by Unscented Kalman filter. Red plot represents SOC estimation by Extended Kalman filter.Among the 3 UKF and EKF estimate accurately the coulomb counting. So, initially for 3000sec SOC decreases nearly to 45% in drive mode then after 3000 to 4000sec the SOC is constant at 45% as cell balance occurs in standby mode. After 4000 to 9000sec the module is in charging mode (4000 to 7000 in CC and 7000 to 9000 in CV). After 9000sec the SOC becomes almost approx 94-95%.
Cell balancing was not performed for initial 3000sec. After 3000sec cell balancing starts with on cell6 and cell 2 as they raise to 1 as shown in plot since its voltage is higher than all other cell voltages. After this charging and balancing, post charge balancing occurs and cell 6 and cell2 still remains in state1 till end of simulation. As cell 1 i.e in yellow colour has lesser voltage than other cells it remains in 0 state through out the simulation. At the end of the simulation all cell voltages try to balance equally.
CASE2 RESULTS:
After running the simulation the results for Case 2 are as follows
This test consists of driving for 10000sec and after that the charging occurs till end of simulation
The plots for the above case is as follows
The cell voltages varies from maximum 3.9V and reduces to 3.05v as cutoff voltage at 9000sec. The entire capacity is drained at 9000sec.
The pack current discharges to maximum of 70A.
Here cell1 has lesser temperature and varies from 288K to 302.5K. Cell 6 has high temperature and varies from 288K to 322.5K. The cell temperature slightly reduces between drive and standby mode.
For the battery module state entire sequence is performing in drive mode as in plot.
Here the SOC gets reduced from 80% to 0% as current drain is for longer duration as in entire drive mode. At stand by mode in between the driving the SOC remains constant.
Here cell balancing is not performed as balance of states is 0 for entire drive modes.
CASE3 RESULTS:
After running the simulation the results for Case 3 are as follows
The case consist of charging for entire duration of simulation time
The plot results are as follows
For cell voltages from above plot, initially cell2 voltage raises from 3.9v to 4.2v and stabilizes at 4.15v. cell1 raises from 3.85v to 4.1v and stabilized to 4.08v.
The pack current raises from 0A to 30A and starts to reduce to 2.5A at the end of full charging time of 2000sec. At 2000sec the cell pack is fully charger ti 100%.After full charge the current flow will be cut-off as shown in plot.
The cell temperatures are in increased state upto full charge and after that the temperature starts reducing.
The BMS state is operating in charging mode entirely.
The SOC gets fully charged at 2000sec and remains at full charge for rest of the sequence.The actual SOC is approx 93% by UKF and EKF techniques.
Cell balance starts from 0sec for cell 6 and 2 w.r.t cell1 and all cella are trying o balance at end of simulation.
2.TO EXPLAIN COULOMB COUNTING FOR FINDING SOC ESTIMATION IN BMS
COULOMB COUNTING:
SOC is defined as the rate of available capacity to its maximum capacity when a battery is fully charges and describes the remaining % of the battery capacity.
To find the SOC % we are using coulomb counting method.
The coulomb counting method measures the discharging cureent of the battery and integrates the discharging current over time in order to estimate the SOC.
It can be gives by SOC(t) = SOC(t-1) + (I(t)/Qn) delta t
where I(t)= discharging current
SOC(t-1) = previously estimated SOC values
MODIFIED COULOMB COUNTING:
In this technique discharge current is connected with a function and corrected current is found out for accurate SOC estimation.
The function is given as Ic(t) = k2I(t)^2 + k1I(t) + k0
Where k0,k1,k2 are constant values obtained from the experimental data
The SOC estimation is given as follows
SOC(t) = SOC(t-1) + (Ic(t)/Qn)delta t
SOC ESTIMATION IN BMS MODEL
To observe the coulomb counting technique in BMS model click the sub-system of SOC estimation and click coulomb counting
In the coulomb counting it is developed as follows
Here discharge current from battery is given to gain block with value 1/3600 to convert to Ah. The value is then divided by capacity due to change in temperature then integrated with a gain value of 1 with sample time -1 as we are using discrete integrator. The saturation limits are from 0 to 1. We will get the SOC values from 0 to 1 multiplied with 100 to get SOC %.
CONCLUSION:
Thus the challenge for the tasks of week 7 for SOC estimation and to study the results of 3 cases of test harness has been completed successfully.