Week 7 State of charge estimation

1.Simulate the 3 test cases from harness dashboard and write a detailed report on the results

ANS:

 Downloading & opening the model (Closedloop BMS harness dashboard model)

From the link provided in the question, we open the document of 'Design & Test Lithium Ion Battery Management Algorithms'.

Download the zip file & extract it to the current folder of MATLAB. The folder consists of following files.

 

Then open the MATLAB software & opening the simulink model file of 'BMS_Closedloop_Harness_Dashboard' from Test==> System==> BMS_Closedloop_Harness_Dashboard model file.

Then give open project and model in pop-up dialog box to install all the model related library files into the current folders and workspace.

 

When we open the model file, it looks like:

 

BLOCKS PRESENT IN MODEL:

1) Dashboard

• Manual Varient(Selector)
• BMS_Testsequence state request(a Subsystem) which provide state request commands
• Test Sequence Varient(Selector)
• Lamps to show High temperature, Fault State, Over current, Over voltage, Voltage sensor operation, Charging, Low temperature, Inverter operation, Under voltage. If any fault arises the lamp will indicate it by red colour and if all is good, then indicated by green in colour.
• While running the simulation, for manual variant changes, rotating switch is used.
• In scope results the Module current, SOC of module, Individual cell temperature,Individual cell voltages, Cell balancing states, Battery states are seen as outputs.

 

2) BMS_ClosedLoop

• This block consists of 2 main subsystems, BMS ECU and PLANT units.
• The output from the PLANT unit is given as input BMS ECU as BMS_Input feedback loop. Also the BMS ECU unit receives the state request.
• The input to the plant unit as 'From_BMS' is given by the BMS output commands (3MS_output) and also the PLANT unit receives the state request.
• The end/final outputs from the 'BMS_ClosedLoop' model are BMS_info, BMS_output, BMS_input.

 

In the BMS ECU subsystem/unit, the 'Current_Power_Limits_Calc' does the the power & current limitation calculation which are used in 'State machine block', 'Cell Balancing Logic' & 'SOC estimations' as shown below.

 

In the PLANT unit/subsystem we have 'Charger-Load', 'Battery_Pack' & the 'PreChargeCircuit' block.

The battery_Pack subsystem consists of two types of configurations namely, 1 module (6 cells series) and 16 module (96 cells series). 

We are using the 1 module configuartion of Battery_Pack to reduce the overall simulation running time.

 

 

Test case Discussion:

For this model to test the all 3 cases we need to select the Test Sequence Variant and select the cases one by one as shown below.

CASE 1:

From Test Sequence Variant dialogue box select 1 and click apply as depicted below.

 

 

This selection continues as selection changes made in masked BMS_State request subsystem as,

                                                             

 

 

This 1st test case consists of following shown sequence of operations.

Here in the above test sequence editor figure it is set as,

• Driving (Discharging) for first 3000 seconds.
• After driving, the balancing operation exists for next 1000 seconds.
• Then Charging & Balancing operation happens for next 5000 seconds.
• After these 3 operations, post charge balancing operation continues till the end of simulation period.

 

Output Result:

After running simulation the results will be shown as,

Here it can be seen that for overall simulation, the lamp indications are shown in green which impies that there aren't any fault occuring during these sequences.

 

From this plot we can understand the following.

 

Pack Current & Cell Voltages:

For the initial 3000 seconds all 6 cells are in driving (Discharging) mode where discharging happens majorly.

During this period, the cell voltage varies between 3.1V(minimun) to 3.9V(maximum).

The cell balancing can be seen between 3000-4000 seconds wherein the module is at standby mode & the voltage is approximately constant at 3.7V.

The charging stars after 4000 seconds, where initially the voltage is increased to 4V from 3.7V. The constant current charging can be seen between 4V to 4.2V at nearly 30A current and after 4.2V the constant voltage charging can be seen where the current reduces to 3A from 30A upto 9000 seconds which the end of the charging period.

The post charge cell balancing starts after 9000 seconds where all cell voltages are balanced to a voltage with respect to cell 1 to 4.1V as depicted below. In this stage from 9000 seconds to upto 20000 seconds (end of the simulation period), the pack/BMS state is in standby mode. 

 

Cell Temperatures:

Initially during driving i,e for 3000 seconds, the cell temperature gradually increased from ambient temperature of 302K from 288K.

The temperature of cell 1 is lesser than compared to other cells & temperature of cell 6 is higher for all stages.

Due to cell balancing during standby mode(no current flow through the cells), the temperature of these cells reduced slightly during 3000 sec to 4000 seconds.

After 4000 seconds i,e during charging period the cell temperature again increases but in exponential manner upto constant current region and later temperature reduces linearly as current flow reduces in constant voltage & cell balancing region which is upto 9000 seconds.

The maximum temperature of cell 1 & cell 6 are raised to 302.2K & 316.9K respectively.

The temperature still reduces after 9000 seconds & remains stable at reaching ambient temperature.

 

BMS State:

Upto 3000 seconds the battery module is in driving mode (discharging)

From 3000 seconds to 4000 seconds, the battery module is in standby mode where cell balancing occurs.

The battery module is in charging & balancing operation from 4000 seconds to 9000 seconds.

Only post charge balancing operation occurs after 9000 seconds upto end of simulation.

 

SOC:

Initial SOC of the battery module is 80%. In the plot the coulomb counting method for SOC estimation is indicated by yellow plot.

Whereas the Unscented Kalman filter technique & Extended Kalman filter technique for SOC estimation is indicated by blue and red plot respectively. Among these three, the EKF & UKF technique estimates the SOC accurately than the coulomb counting method.

During driving mode which is upto 3000 seconds, the SOC reduces upto 45% of its total.

As cell balancing occurs during standby mode (3000 to 4000 secs) thge SOC remains constant at 45%.

Upto 9000 seconds from 4000 seconds the battery pack is in charging mode where 4000 to 7000 sec is CC mode & 7000 to 9000 sec is CV mode.

The SOC becomes constant approx. to 94% after 9000 seconds.

 

Cell Balancing:

At the starting upto 3000 seconds balancing wasn't carried out.

Cell balancing starts after 3000 seconds upto 4000 seconds on cell 2 & cell 6 as they have raised to 1 as depicted in above plot since its voltage is higher compared to all other cell voltages.

As the cell 1 (yellow) has lesser voltage compared to other cells it is in 0 state throughout the simulation.

All cells try to balance equally at the end of the simulation.

 

 

CASE 2:

Selecting case 2 in test sequence variant and click apply as shown in the dialogue box.

 

 

This selection continues as selection changes made in masked BMS_State request subsystem as,

                                     

 

 

The test case includes driving mode for 10000 seconds and followed by a charging mode upto end of the simulation.

 

Output Result:

After running simulation the results will be shown as,

Here it can be seen that for overall simulation, the lamp indications are shown in green which impies that there aren't any fault occuring during these two sequences.

 

 

Output plots:

The below figure gives the plot for only driving mode.

 

Cell Voltages & Pack Current:

The cell voltage varies between 3.9V(maximum) to 3.02V(minimum) as the cutoff voltage upto 9000 seconds. At 9000 seconds the entire capacity is drained.

The battery module current discharges upto max. 70 A & slight charging spikes can be seen during the entire drive.

We can see some of standby mode like operation in between the discharge mode.

 

 

Cell Temperature:

The cell 1 temperature varies to 302.4K from 288K which shows the lesser temperature variations.

The cell 6 temperature varies to 322.5K from 288K which shows the higher temperature variations.

In between the drive modes the temperature of cells slightly reduces as those durations comes under standby modes.

 

BMS State:

As seen from the plot the entire simulation is performed in discharge/drive mode.

Small duration of standby plots are not plotted.

But in actual practical case we can't find the standby modes as the other small auxiliaries of EVs drains the battery current.

 

 

SOC:

For this case the SOC reduced to 0% from 80% as the current discharged for a longer duration as shown in the plot.

The SOC remains constant whenever there occurs a standby mode.

 

 

Cell Balancing:

As there aren't any voltage differences between the cells, the cell balancing is not carried out.

 

 

CASE 3:

Selecting case 3 in test sequence variant and click apply as shown in the dialogue box.

 

 

This selection continues as selection changes made in masked BMS_State request subsystem as,

                                        

 

The third case consists of charging operation for entire simulation time.

 

Output Result:

After running simulation the results will be shown as,

Here it can be seen that for overall simulation, the lamp indications are shown in green which impies that there aren't any fault occuring during the simulation of charging mode.

 

Result Plot:

The below figure gives the plot for only charging mode.

 

Cell Voltages & Pack Current:

It is seen that te cell 2 voltage to maximum 4.19V from 3.9V(approx.) and stabilised to 4.15V.

Cell 1 voltage rises to maximum 4.12V from 3.85V and stabilised to 4.08V.

At 2 second, the battery pack current rised to maximum 30A from zero and gradually reduced to 25A at the end of full charging which is upto 2000 seconds.

The pack is fully charged to 100% at the end of 2120 seconds.

After full charge of the battery pack, current flow will be cutoff  as depicted in result plot.

 

 

Cell Temperature:

Upto battery pack gets full charged the cell temperature continued to increase & after that the temperature started reducing to ambient temperature.

It can be seen from the plot that after full charge(at 2120 seconds) the temperature starts to decrease.

 

BMS State:

As seen from the plot the entire simulation is performed in charging mode.

 

SOC:

The battery pack SOC gets fully charged at 2120 seconds and remains at full charge for rest of the simulation period.

The actual state of charge by using UKF & EKF technique is fond to be around 93%. 

 

Cell Balancing:

From beginnig itself the cell 6 & cell 2 try to balance with respect to cell 1 & all cells try to balance equally at around the end of the simulation.

 

 

CONCLSION:

Hence from above 3 test cases the performance of Battery Management System on various factors are discussed and analysed.

 

-------------------------------------------------------------------------------------------------------------------------------------------------

 

2.What is coulomb counting? Refer to the above model and explain how BMS implements coulomb counting for SOC estimation? 

ANS:

Coulomb counting is method to estimate battery state of charge(SOC). State is one of the most main parameters of battery and it is denoted as the ratio of current capacity to the nominal capacity of the battery. There are different methods to estimate State of Charge like – open circuit method, terminal voltage method, coulomb method, direct measurement, modified coulomb method etc.

Coulomb counting is an attarctive solution for specially Lithium-ion batteries, which consists high coulombic efficiency. The total energy delivered will be reduced by losses and available energy will be less than delivered, but coulomb counting works well. Coulomb counting technique takes the values of current(charge flow) and integrates them over the period of time to find SOC. The ameter will provide the current value.

SOC equation is-

`S O C ( t ) = S O C ( t − 1 ) − ∫ ( eta⋅ I ( t ) . d t) /(Qn )`

where,

Qn = battery capacity

`eta` = discharge efficiency

$SOC\left(t\right)$= previously estimated SOC

$I\left(t\right)$ = discharging current

 

Coulomb counting method has easy implementation & simple structure, however it does not consider cycle life, battery history and battery temperature during SOC calculation. Battery temperature influences the battery efficiency which causes inaccurate SOC calculation. For this method, it is very important to calibrate initial SOC value.

 

Implementation in BMS 

With a suitable algorithm the Coulomb counting method can be implanted in BMS. The below shown chart gives the general idea, how the SOC estimation is carried out. This algorithm which can be implemented with state machine in Simulink, which can be further uploaded in flesh memory of BMS hardware using code generation option.

 

Now look at inside of a Simulink model of a BMS(Battery Management System). In this model SOC is estimated with different methods – the coulomb counting & the Kalman filter. Coulomb counting method is taking the battery pack current values and divided by capacity values, which is given by a look up table to show variation of capacity with temperature. Further values are integrated by using a disrete integratorblock to estimate the state of charge. This estimated SOC is given to the multiplexer along with the other two SOCs estimated by UKF & EKF method & given as the overall SOC of the pack to the Bus management block which inturn connected to the plant block(battery pack module).