2024-06-15 Meeting notes

 Date

Jun 15, 2024

 Participants

• @Nathan Lemma

• @Aneri Patel

• Marcelo Araujo Xavier: Bio

 Goals

• Get a review of the proposal

• Understand our first steps to completing an estimation

 Discussion topics

1. Review of the proposal. Click to see the SoC Research Proposal

2. Getting educated on battery modeling. Below are notes from the lecture

“What you should implement is the crawl, walk, and run method as you need to start simple to build up a complex battery system.”

- Marcelo Xavier

Lets start simple with a constant, zero-resistance, linear, iso-thermic, time-invariant, limitless voltage source, otherwise known as an ideal voltage source.

Open-circuit voltage

Open

To model such a source, you can use the open-circuit voltage method (OCV).

• Voltage is measured when the battery is not connected to a load and has been at rest for a sufficient period of time.

Real batteries exhibit internal resistance, which causes a voltage drop when current flows.

To model a battery more realistically and to calculate the state of charge (SOC), the next step is to consider methods used to determine SOC. One common method is the Coulomb Counting method.

Coulomb Counting Method

This method calculates SOC by integrating the current flow into or out of the battery over time. The SOC is updated as follows:

• Equation 1: This assumes the charge current is negative and the discharge current is positive.

Estimation Techniques

The point of these techniques is so you don’t have to fully parameterized every time and temperature interval. They usually take voltage measurement and compares against a voltage estimate to compute the error which is used to correct SOC.

• Estimators are model-based: you need a model of your system

  • ECM (equivalent circuit model): an equivalent electrical circuit that mimics the electrochemical behavior of the battery

  • Topology: voltage source as a function of SOC connected in series to a resistance and n-resistor-capacitor-pairs.

  • Open

    

    Example of a model

  • Behaviors:

    1. OCV

    2. Polarization voltages

    3. Ohmic resistance

• Terminal Equation

• System of equations: V1, V2, z

• Example: Kalman filter

  1. Time propagation (open-loop).

  2. Correction (compute the error, compute the Kalman gain, correct the states and the covariance.

• Example of a chemistry with a "flat" OCV-SOC (LFP)

  1. Voltage sensor has 10mV accuracy

  2. 3.3 = 50%, 3.305 = 50.5%, .. 3.38 = , 3.310 = 60%

  3. dOCV/dSOC is close to zero at the flat region

COMSOL: a cfd

 Action items

 Decisions

1. Start at the bare minimum before adding so many options