Battery Fast Charging Optimization (Project 256)

[robertogl]

Contribute to the discussion by asking and/or answering questions, commenting, or sharing your ideas for solutions to project 256

---

🤖 An AI bot may reply to your comments with suggestions and guidance. It will not solve the project for you. Your comments may be processed by a third-party LLM. To direct a question to a specific person, tag them with @username.

[Shantanu2852]

Hello sir
I am interested in doing this project but I don't understand where to start.Can you please guide me?

[robertogl]

Hi @Shantanu2852,

As suggested in the project description, get familiar with battery modeling, in particular the Single Particle Model (SPM), and battery fast charging. If you are not familiar with this topic, it might be useful to leverage related videos and webinars and Simscape Battery onramp.

[Shantanu2852]

Thank you for your guidance, sir.

I will start by studying the Single Particle Model (SPM) and understanding the fundamentals of battery fast charging. I will also go through the recommended videos, webinars, and the Simscape Battery Onramp. If I have further doubts during the process, I hope it's okay to reach out again.

Thanks once again for your support.

[obayreaps]

[Shantanu2852]

Hello, Could you please share more about your background and how you think you could contribute to this project? It would be helpful to know your skills or any specific areas you're interested in working on.

[rds510]

Hello, my group is having a problem modeling the charging and discharging portion of the battery (Single Particle), we are running into an issue where the State of Charge (SOC) in regards to the Constant Current - Constant Voltage (CC-CV) Algorithm is not working properly and we can not get the battery to meet the threshold to discharge. To add on, we found that when the SOC is at 0 (0% capacity), the voltage is at 3V. When the SOC is at 1 (100% capacity) the voltage is just under 3.5V. Our Voltage readings through out are really wacky and not the ideal voltages. Eventually we want to add temperature but we want to make sure our CC-CV works.

Please anything helps, at least some direction, our group has been stuck on this for a while. If you need any more information feel free to let me know. Thank you.

Battery Builder: We Integrated the Single Particle Battery into our battery, we are using a pouch battery.

Here are our reference values:
Battery capacity of 3,349 mAh or 3.349 Ah
Energy rating of 12.98 Wh.
The battery voltage 3.869V (nominal) (our voltage lies a bit below this)
Maximum of about 4.48V during charging (our voltage never gets to this point but I believe it is possible if configured correctly)

Here is our Model:

Here is a snapshot of our graphs:

Here is a list of the blocks and their parameters from the images:
ModuleType1 (iphone15_s)
Main
Extrapolation method for all tables: Nearest
Initial Targets
Cell state of charge Priority: High Value: 0.3 , 1
Convective Heat Transfer Parameters
Convection type: Constant Area: 0.0054 m^2
Heat transfer coefficient: 10 W/(m^2K)
SOC Estimator (Coulomb Counting)
Parameters Specify Current input as cell current(s): Checked Cell capacity, AH (AHr): 3.349
Sample time (-1 for inherited): -1
Algorithm data type: Inherit: auto
Battery CC-CV4 Parameters
Specify CellVoltage input as [Min,Max]: Unchecked
Cell voltage threshold (V): 4.2
Controller proportional gain: 100
Controller integral gain: 10
Controller anti-windup gain: 1
Gain of signal-tracking feedback loop: 1
Sample time (-1 for inherited): -1
Temperature Source Parameters
Temperature: 298.15 K
Unit Delay Main Initial condition: 1
Input processing: Elements as channels (sample based)
Sample time (-1 for inherited): -1
Relay Main
Switch on point: 0.9
Switch off point: 0.3
Output when on: 0
Output when off: 1
Input processing: Elements as channels (sample based)
Enable zero-crossing detection: Checked

[xzhang96485]

Hi there,

If the goal is to model CC-CV charge, instead of designing controller (PID tuning), we can use this block provided by Simscape Battery: https://www.mathworks.com/help/simscape-battery/ref/charger.html.

Regarding your point about "wacky" voltage, in R2024b, the breakpoints for the open circuit potential (OCP) of both electrodes in the Battery Single Particle block are meant to be normalized stoichiometry instead of absolute stoichiometry. The normalized stoichiometry normalizes absolute stoichiometry by taking into consideration of the min and max stoichiometry defined cycling. Because of this, we need to assign the breakpoints and OCP tables properly in order to capture the right voltage. However, if we are using R2025a, the absolute stoichiometry breakpoint option is available for defining electrode OCP.

[rohankelkar23-boop]

Following are the issues i am facing in the simulation of the simscape block network we have built for project titled " Battery Fast Charging And Optimisation" -

1.) Plotted One charging profile for voltage vs time , current vs time for the spm battery pack in simscape the objective is to plot multiple such profiles and optimise on other battery parameters like the temperature substantial increase in which can comprise on the lifetime and health of the battery pack...we are suppose to optimise the network in such a manner that the impact of these parameters of soh will be as minimum as possible, in this while plotting multiple charging profiles what are the parameters we are suppose to fill in to various blocks in our network as nothing as such has been mentioned on the problem statement page , What other steps we have to follow in order to get those charging profiles im attaching our output here for reference -

2.) Temperature sensing network was to be integrated in the main system to monitor the temperature rise and variation thus to work on reducing the impacts of this temperature rise on overall SOH Of the battery pack( In this case the Single particle model battery pack) we were facing issue in how to build this particular circuit if not circuit what block should we use to get the readings because the blocks from thermal domain like the temperature sensor is able to transfer the output even with the help of converters ( ps-simulink and simulink-ps) data transfer is suppose to happen between the thermal and electrical domain systems for the entire system to behave as a complete unit taking inputs and and delivering outputs.So how to build this particular thermal network what blocks are to be used, if not network in itself what seperate block should we use also instead of thermal system, is it feasible to use a math function block which is coded with the same transcript as that of some RTD Or any particular block feasible in this application if yes what equation we are suppose to keep in mind while coding the block.

3.) While attaching the thermal system ,The thermal system didn't work but now that we have to remove the thermal system and check outputs of the electrical system (initially functioning) is now giving errors ,we are not able to decode

4.) The Electrical system is showing either error or the entire simulation is getting stuck at time stamp particularly "T=299.265"

This model is running with the solver configuration type of ode15s earlier outputs were with ode23s.

Thank You.

[xzhang96485]

Hi there,

While thermal management is critical during battery fast charging, this project was created to focus on the aspects of how to optimize battery fast charging current so that potential lithium plating is minimized for prolonging battery life.

To answer some of your questions, the Battery Single Particle block does a lumped thermal model internally, and it would require a proper heat transfer network to describe how heat is dissipated to the environment. To access the battery temperature from the lumped thermal model, we can use a Probe block (https://www.mathworks.com/help/simscape/ref/probe.html).