5th and final course in the Algorithms for Battery Management Systems Specialization

Instructor: Gregory Plett, Ph.D., Professor

In this course, you will learn how to design balancing systems and to compute remaining energy and available power for a battery pack.

Prior knowledge needed: ECEA 5730, ECEA 5731, ECEA 5732, ECEA 5733, a Bachelor’s degree in Electrical, Computer, or Mechanical Engineering, or a B.S. degree with undergraduate-level competency in the following areas: Math: Differential and integral calculus, operations with vectors and matrices (mechanics of linear algebra), and basic differential equations, Engineering: Linear circuits (modeling resistors, capacitors, and sources), Programming: MATLAB, Octave, or similar scientific program environment

### Learning Outcomes

• Evaluate different design choices for cell balancing and articulate their relative merits.
• Design component values for a simple passive balancing circuit.
• Use provided Octave/MATLAB simulation tools to evaluate how quickly a battery pack must be balanced.
• Compute remaining energy and available power using a simple cell model.
• Use provided Octave/MATLAB script to compute available power using a comprehensive equivalent-circuit cell model.

### Syllabus

Duration: 3 hours

In previous courses, you learned how to write algorithms to satisfy the estimation requirements of a battery management system. Now, you will learn how to write algorithms for two primary control tasks: balancing and power-limits computations. This week, you will learn why battery packs naturally become unbalanced, some balancing strategies, and how passive circuits can be used to balance battery packs.

Duration: 3 hours

Passive balancing can be effective, but wastes energy. Active balancing methods attempt to conserve energy and have other advantages as well. This week, you will learn about active-balancing circuitry and methods, and will learn how to write Octave code to determine how quickly a battery pack can become out of balance. This is useful for determining the dominant factors leading to imbalance, and for estimating how quickly the pack must be balanced to maintain it in proper operational condition.

Duration: 2 hours

This week, we begin by reviewing the HPPC power-limit method from course 1. Then, you will learn how to extend the method to satisfy limits on SOC, load power, and electronics current. You will learn how to implement the power-limits computation methods in Octave code, and will see results for a representative scenario.

Duration: 4 hours

The HPPC method, even as extended last week, makes some simplifying assumptions that are not met in practice. This week, we explore a more accurate method that uses full state information from an xKF as its input, along with a full ESC cell model to find power limits. You will learn how to implement this method in Octave code and will compare its computations to those from the HPPC method you learned about last week.

Duration: 5 hours

Present-day BMS algorithms primarily use equivalent-circuit models as a basis for estimating state-of-charge, state-of-health, power limits, and so forth. These models are not able to describe directly the physical processes internal to the cell. But, it is exactly these processes that are precursors to cell degradation and failure. This week quickly introduces some concepts that might motivate future BMS algorithms that use physics-based models instead.

Duration: 5 hours

This capstone project explores the design of resistor value for a switched-resistor passive balancing system as well as enhancing a power-limits method based on the HPPC approach.

Duration: 2 hours

Final exam for the course.

Assignment

Q​uiz for week 1

7.5%

Q​uiz for week 2

7.5%

Q​uiz for week 3

7.5%

Q​uiz for week 4

7.5%

Q​uiz for lesson 5.5.1

1.5%

Q​uiz for lesson 5.5.2

1.5%

Q​uiz for lesson 5.5.3

1.5%

Q​uiz for lesson 5.5.4

1.5%

Quiz for lesson 5.5.5 1.5%
Quiz for 5.5.6 1.5%
P​rogramming project "Designing a switched-resistor passive-balancing system" 5.5%
P​rogramming project "Improved HPPC power-limits estimator" 5.5%
F​inal exam 50%

A

93.3%

A-

90.0%

B+

86.6%

B

83.3.0

B-

80.0%

C+

76.6%

C

73.3%

C-

70.0%

D+

66.6%

D

60.0%

F

0%