Battery Management and Thermal Management

Introduction

An electric or hybrid vehicle is only as safe, durable, efficient, and usable as its battery and thermal-management systems. The battery management system (BMS) supervises electrochemical energy storage, while thermal management maintains batteries, power electronics, electric machines, charging equipment, and the passenger cabin within acceptable temperature ranges. These functions are strongly coupled: temperature affects battery resistance and available power; current produces heat; cooling power consumes electrical energy; and a fault in sensing or contactor control can create a serious safety hazard.

A functional BMS is therefore more than a voltage monitor. It is a distributed measurement, estimation, control, protection, communication, and diagnostics system. Its principal responsibilities are to measure cell and pack variables, estimate state variables, balance cells, control high-voltage contactors, detect faults, communicate limits, and coordinate thermal actuators.

Learning Objectives

After completing this chapter, the reader should be able to:

1. Explain the architecture and operating functions of a BMS.
2. Design basic sensing, balancing, estimation, and contactor-control strategies.
3. Calculate state of charge (SOC), coulombic energy, and thermal generation using appropriate assumptions.
4. Distinguish SOC, state of health (SOH), state of power (SOP), and state of energy (SOE).
5. Analyse battery cooling, heating, thermal propagation, and thermal-runaway mitigation.
6. Coordinate thermal control of batteries, motors, inverters, chargers, and cabins.
7. Diagnose common faults while managing safety, cybersecurity, and regulatory obligations.

1. Battery-Pack and BMS Architecture

Cells are connected in series to increase voltage and in parallel to increase capacity and current capability. A module contains a group of cells and mechanical, electrical, sensing, and sometimes cooling components. A pack normally includes modules, busbars, fuses, service disconnects, contactors, current sensors, insulation monitoring, cooling hardware, and a BMS.

The BMS commonly has three logical levels:

• Cell-monitoring units (CMUs): measure individual cell voltages and module temperatures; many also control balancing.
• Battery-control unit (BCU): estimates states, controls contactors and pumps, performs diagnostics, and communicates limits.
• Vehicle-control interface: exchanges torque, charging, thermal, and fault information with the vehicle control unit, charger, inverter, and supervisory controller.

A series-connected pack is constrained by its weakest cell. During charging, the cell with the highest voltage reaches its limit first; during discharge, the cell with the lowest voltage reaches its minimum first. Consequently, pack voltage alone cannot guarantee safe operation.

Important measurements include cell voltage, pack current, cell and coolant temperatures, isolation resistance, contactor auxiliary feedback, coolant pressure or flow, and sometimes gas, pressure, or humidity signals. Sensors require calibration, plausible-range checks, rate-of-change checks, redundancy where the consequence of failure is high, and a defined response to communication loss.

2. Electrical Sensing and Cell Balancing

Cell-voltage and current sensing

Cell-voltage measurements must tolerate common-mode voltage, switching transients, electromagnetic interference, and open-wire faults. The measurement chain should include input protection, filtering, diagnostic excitation or plausibility checks, and galvanic isolation where required by the architecture.

Current may be measured using a shunt resistor or a Hall-effect or fluxgate sensor. A shunt provides good linearity but dissipates power and requires isolation at high voltage. Magnetic sensors provide galvanic isolation and low insertion loss but may have offset, temperature, and bandwidth errors. Current-sensor offset is particularly important because even a small bias accumulates during SOC integration.

Passive balancing

In passive balancing, a resistor is switched across a high-voltage cell to convert excess energy into heat. If a cell voltage difference is ΔV and the balancing resistor is R_b, the approximate balancing current is:

The resistor power is:

These equations assume a constant cell-voltage difference and neglect switch and wiring losses. For example, with ΔV = 0.10 V and R_b = 20 Ω, I_b = 5 mA and P_b = 0.0005 W. In practice, the useful balancing current may be specified directly by the circuit, and the cell’s electrochemical voltage does not remain constant.

Passive balancing is inexpensive and robust but slow and thermally inefficient. It is often used near the end of charge, when cell voltage is sensitive to SOC differences.

Active balancing

Active balancing transfers energy from higher-energy cells to lower-energy cells or to a common bus using inductors, capacitors, or isolated converters. It can reduce heat and improve balancing speed, but it increases cost, component count, electromagnetic compatibility challenges, and control complexity. Balancing should not conceal a defective cell: abnormal self-discharge, excessive temperature rise, or a persistent voltage divergence requires diagnosis rather than prolonged equalisation.

A practical strategy is to enable balancing only when measurement confidence is high, the pack is in a suitable SOC and temperature window, the charger current is controlled, and the estimated benefit exceeds balancing losses. The BMS should impose maximum balancing time, resistor-temperature limits, and a clear fault response.

3. State Estimation

State of charge

SOC is the estimated fraction of usable charge remaining. Coulomb counting is fundamental:

Here, Qₙ is the usable nominal capacity in ampere-hours, I is current in amperes, positive for discharge, η is charge efficiency or a signed efficiency convention, and T is temperature. The equation assumes that current measurement, initial SOC, capacity, and efficiency are known sufficiently well. A current offset produces an SOC error that grows with time.

For a constant discharge current, the special case is:

If a 50 Ah battery supplies 25 A for 0.5 h, the ideal SOC decrease is 0.25, or 25 percentage points. This is an estimate, not a guarantee of available energy, because capacity depends on temperature, current, ageing, and voltage limits.

Open-circuit-voltage (OCV) correlation can correct accumulated coulomb-counting error when the battery has rested sufficiently and the chemistry has a reliable OCV–SOC relationship. Under load, terminal voltage is better represented by an equivalent-circuit model:

where Vₜ is terminal voltage, R₀ is instantaneous resistance, and Vₚ is the polarisation voltage represented by one or more RC branches. Sign conventions must be defined consistently. The model is useful for observers such as extended Kalman filters, but parameters vary with SOC, temperature, ageing, and direction of current.

SOH, SOP, and SOE

SOH describes degradation relative to a reference condition. Capacity-based SOH is commonly expressed as:

Resistance-based health may instead compare present internal resistance with a reference value. These definitions are not interchangeable.

State of power (SOP) is the maximum allowable charge or discharge power for a specified duration while respecting voltage, current, temperature, and safety constraints. State of energy (SOE) estimates remaining usable energy and is more informative than SOC when voltage varies substantially.

A production estimator normally combines coulomb counting with voltage-model correction, temperature compensation, capacity adaptation, and diagnostic confidence. Estimation outputs should include uncertainty or validity flags. A precise-looking SOC value without confidence management can be unsafe.

4. Contactor Control and High-Voltage Safety

Contactors electrically connect or isolate the battery from the vehicle DC bus. Typical components include positive and negative main contactors, a precharge contactor, a precharge resistor, a service disconnect, and a fuse. The precharge path limits inrush current into inverter and charger capacitors.

A typical start sequence is:

1. Confirm valid low-voltage supply, no critical faults, acceptable temperature, and adequate isolation.
2. Verify that contactors are open and the bus voltage is near zero or within the expected condition.
3. Close the negative main contactor.
4. Close the precharge contactor and monitor DC-bus voltage.
5. When the bus reaches a defined fraction of pack voltage and the rise rate is plausible, close the positive main contactor.
6. Open the precharge contactor and verify the final topology through auxiliary contacts and voltage measurements.

If a pack voltage is V_pack and the precharge resistance is R_p, the initial ideal precharge current is:

For a capacitive DC bus C_bus, the ideal voltage response is:

This assumes constant pack voltage, a linear resistor, negligible wiring resistance, and an initially uncharged capacitor. The time constant is τ = R_p C_bus. A failed-open precharge circuit causes slow or absent bus rise; a failed-short precharge path can create damaging inrush current. The BMS should detect welded contacts by comparing commanded state, auxiliary feedback, pack voltage, and bus voltage after isolation.

High-voltage interlock loops, isolation monitoring, discharge resistors, orange-cable routing, service procedures, and warning labels support electrical safety. A contactor must not be opened under conditions that exceed its interrupting capability unless the system has been designed for that event. Crash signals, water ingress, overcurrent, severe isolation loss, and thermal-runaway indicators may require immediate isolation, controlled discharge, or both.

5. Fault Diagnosis and Functional Safety

Fault diagnosis combines thresholds, temporal filtering, model residuals, redundancy, and fault-history management. Typical faults include:

Thresholds should include hysteresis and persistence times to prevent nuisance trips. A diagnostic trouble code should record fault type, severity, occurrence conditions, freeze-frame data, and whether the fault is active or historical. Diagnostics should distinguish a sensor failure from the physical condition being measured.

Functional-safety design uses hazard analysis, safety goals, fault-tolerant concepts, independence where required, and verification of safety mechanisms. The precise safety integrity target depends on the hazard analysis and applicable vehicle standards; it should not be assumed merely from the presence of a BMS. Cybersecurity is also relevant: unauthorised commands, spoofed measurements, or denial of communication can compromise contactor and thermal control. Access control, authenticated communications, secure updates, logging, and incident response are engineering requirements, not optional software features.

6. Battery Thermal Behaviour

Heat is generated by irreversible electrochemical losses and, depending on chemistry and operating condition, reversible entropic effects. A useful first-order pack model is:

where Q̇_gen is heat-generation rate in watts, I is current in amperes, and R_int is effective internal resistance in ohms. This approximation is appropriate for preliminary design when entropic heat and reaction heat are small or included in an effective resistance. A more complete electrochemical model may include entropic heat, reaction overpotential, mixing, and tab or busbar losses.

Heat transfer through a coolant interface can be approximated by:

where h is convective heat-transfer coefficient in W·m⁻²·K⁻¹, A is heat-transfer area in m², T_s is surface temperature, and T_f is fluid temperature. It assumes a representative average coefficient and temperature difference.

For a coolant stream, the rate of sensible heat removal is:

where ṁ is mass-flow rate in kg·s⁻¹, c_p is specific heat capacity in J·kg⁻¹·K⁻¹, and temperatures are in °C or K differences. The numerical temperature difference is identical in °C and K. This relation neglects phase change, leakage, and heat losses to the surroundings.

Worked example. A coolant flows at 0.08 kg·s⁻¹ with c_p = 3,800 J·kg⁻¹·K⁻¹ and warms by 5 K. The removed heat rate is:

Q̇_cool = 0.08 × 3,800 × 5 = 1,520 W.

If the battery generates 1,200 W, the remaining 320 W must be rejected through other paths or stored temporarily, and the coolant temperature will not remain steady indefinitely. The calculation illustrates why flow rate, allowable temperature rise, heat-exchanger capacity, and control margins must be considered together.

Battery temperature affects reaction kinetics, resistance, power capability, charging acceptance, ageing, and safety. Cold batteries generally have increased resistance and reduced charge acceptance. Hot batteries may deliver power temporarily but age faster and can become unsafe if cooling is inadequate. Thermal design should control both mean temperature and cell-to-cell temperature variation.

7. Thermal Systems and Coordinated Control

A liquid-cooled battery typically includes cold plates or cooling channels, a pump, valves, a radiator or chiller, expansion volume, temperature sensors, and a coolant reservoir. Air cooling is simpler and lighter in some applications but usually provides lower heat-transfer capability and less uniform temperature control. Refrigerant direct cooling can achieve high cooling performance but adds pressure, sealing, service, and control complexity.

The vehicle thermal system often contains several loops:

• a low-temperature loop for battery and power electronics;
• a motor/inverter loop, sometimes at a higher allowable temperature;
• a refrigerant loop with a chiller or heat pump;
• a cabin heating and cooling loop.

A coordinated controller allocates limited cooling and heating capacity according to safety and performance priorities. Battery safety normally has priority over cabin comfort, while propulsion availability may be reduced rather than allowing a critical component to exceed its limit. Motor and inverter heat can sometimes be recovered for cabin heating or battery warming. A heat pump can provide efficient heating, but its output decreases at low ambient temperature and may require supplementary electric heating.

Control strategies include thermostat control, proportional-integral control, feed-forward control based on predicted current, and model-predictive control. Feed-forward cooling anticipates a known acceleration or fast-charge event; feedback corrects model error. Valve commands should avoid oscillation, pump operation should be checked against flow and pressure feedback, and actuator faults should lead to graduated derating rather than sudden loss of control whenever safe.

A simplified energy balance for a battery is:

Here, m_b and c_b are effective battery mass and specific heat capacity, T_b is representative battery temperature, Q̇_gen is internal heat generation, Q̇_removed is controlled heat removal, and Q̇_env represents heat exchange with the environment. The lumped model assumes a nearly uniform battery temperature; it cannot predict local hot spots and must not replace cell-level thermal analysis for safety-critical design.

8. Thermal Runaway and Its Mitigation

Thermal runaway is a self-accelerating process in which heat generation exceeds heat removal and triggers exothermic reactions. Initiation may result from internal short circuit, mechanical crush, overcharge, manufacturing defect, external heating, contamination, or severe electrical abuse. Warning signs can include rapid temperature rise, abnormal voltage behaviour, venting, gas generation, smoke, swelling, or unexpected pressure.

Protection is based on layers:

1. Prevention: controlled materials, manufacturing quality, current and voltage limits, mechanical protection, isolation, and robust charging control.
2. Early detection: distributed temperature sensing, rate-of-rise detection, voltage divergence, gas or pressure sensing where justified, and model-based anomaly detection.
3. Propagation limitation: spacing, barriers, ceramic or mica materials, directed vent paths, thermal insulation, and module-level containment.
4. System response: stop charging and discharge where safe, open contactors when appropriate, disable restart, warn occupants, and communicate hazard information to emergency responders.
5. Post-event management: assume delayed re-ignition may be possible; follow validated emergency and transport procedures.

Opening contactors does not remove stored chemical energy from a damaged cell. Conversely, immediate isolation may not always be the only action: the safe response depends on crash condition, fire exposure, isolation status, and the vehicle’s validated safety concept. Thermal barriers should be tested under realistic abuse conditions, not only normal operating temperatures.

9. Applied Case: Fast-Charging Fleet Vehicle

Background. An illustrative urban delivery vehicle uses a liquid-cooled lithium-ion pack. Fleet operators request repeated fast charging, while the vehicle must operate in both winter and summer conditions.

Problem. During summer fast charging, the highest cell temperature rises rapidly and the BMS repeatedly interrupts charging. In winter, charging is slow because the cells are too cold. A service investigation also finds occasional disagreement between a coolant outlet sensor and a module temperature sensor.

Analysis. The engineering team separates the problem into estimation, sensing, and thermal-capacity issues. It verifies temperature-sensor calibration, checks wiring and communication timing, compares sensor rate-of-rise values, and tests pump flow. It then uses current-based heat prediction to command cooling before the highest-power charging interval. The battery model estimates cell temperature rather than relying only on coolant temperature. In winter, a controlled preheating phase is enabled before high-current charging.

Approach. The revised strategy includes sensor plausibility diagnostics, a validated fallback when one sensor is unavailable, feed-forward pump and valve commands, maximum cell-temperature and temperature-gradient limits, and charging-power derating based on the hottest estimated cell. The charger receives an explicit allowable-current limit rather than simply an enable/disable signal. The cabin heat-pump controller may recover waste heat, but battery protection has priority.

Outcome. Charging becomes more consistent, nuisance interruptions decrease, and cold-weather charge acceptance improves. The vehicle still derates under extreme ambient conditions, which is an intended safety function rather than a control failure.

Learning points. Thermal control must be based on the limiting cell or component, not only average pack temperature. Sensor diagnostics are part of thermal safety. Predictive control improves energy efficiency, but it must be bounded by independent protection thresholds. Operational requirements such as fleet turnaround time should be translated into measurable limits for temperature, current, flow, and ageing.

10. Design Best Practices and Limitations

A robust design should:

• define signal ranges, accuracy, sampling rates, latency, and diagnostic coverage;
• use conservative voltage, current, and temperature limits with validated hysteresis;
• estimate uncertainty and maintain traceable calibration data;
• distinguish normal derating from emergency shutdown;
• test sensor opens, shorts, stuck values, communication loss, contactor welds, pump failure, blocked channels, and isolation faults;
• validate models over temperature, SOC, ageing, production variation, and real duty cycles;
• verify precharge time and contactor interruption capability;
• use hardware-independent protection for hazards that software alone cannot safely manage;
• document service isolation, residual energy, emergency response, and end-of-life procedures.

No single model is sufficient. Equivalent-circuit models are computationally efficient but may miss local electrochemical behaviour. High-fidelity electrochemical and computational-fluid-dynamics models improve insight but require parameters and computing resources. Machine-learning anomaly detection can identify patterns in fleet data, but it needs representative training data, explainability, protection against data drift, and a deterministic fallback. AI should support—not replace—validated safety mechanisms.

Regulatory compliance involves electrical safety, functional safety, electromagnetic compatibility, battery transport, crashworthiness, environmental protection, and recycling or producer-responsibility requirements. Exact requirements depend on jurisdiction, vehicle category, chemistry, and use case; designers must consult current applicable standards and regulations rather than treating a textbook limit as a legal specification.

Chapter Summary

The BMS integrates sensing, estimation, balancing, contactor control, protection, diagnostics, communication, and thermal coordination. Cell-level information is essential because the weakest or hottest cell limits pack performance. Coulomb counting provides the basis of SOC estimation but requires correction for sensor bias, capacity variation, temperature, and ageing. OCV and equivalent-circuit models improve estimation when their assumptions are satisfied.

Contactor sequencing and precharge protect the DC bus from inrush current, while auxiliary feedback, voltage comparison, and isolation monitoring detect dangerous conditions. Fault diagnosis should combine thresholds, persistence, plausibility, redundancy, and model-based evidence. Battery thermal management must control temperature, gradients, heat generation, and cooling capacity. Coordinated loops can share cooling, heating, and waste heat among batteries, motors, inverters, chargers, and cabins.

Thermal-runaway safety requires layered prevention, detection, propagation resistance, controlled response, and emergency planning. Fast-charging and fleet applications demonstrate the value of predictive control, explicit current limits, validated sensor fallback, and temperature-based derating. Finally, safe systems require verification across ageing, environmental conditions, manufacturing variation, cybersecurity threats, and realistic abuse scenarios.

Key Terms

Battery management system: Electronic system that supervises battery measurement, estimation, control, protection, and communication.

Cell-monitoring unit: Circuit that measures cell voltages and temperatures, often with balancing control.

State of charge: Estimated fraction of usable charge remaining in a battery.

State of health: Measure of battery condition relative to a defined reference, commonly capacity or resistance.

State of power: Maximum allowable charge or discharge power under specified conditions.

State of energy: Estimated usable energy remaining.

Coulomb counting: SOC estimation by integrating measured current over time.

Open-circuit voltage: Cell voltage measured after sufficient rest with negligible current.

Passive balancing: Equalisation by dissipating excess cell energy as heat.

Active balancing: Equalisation by transferring energy between cells or to a bus.

Contactor: Electrically controlled high-voltage switching device.

Precharge: Controlled charging of the DC-link capacitance before main connection.

Isolation monitoring: Measurement or supervision of resistance between high-voltage conductors and vehicle chassis.

Equivalent-circuit model: Electrical model using resistors, capacitors, and voltage sources to approximate battery behaviour.

Thermal runaway: Self-accelerating exothermic battery failure process.

Thermal propagation: Spread of damaging heat or failure from one cell or module to others.

Derating: Intentional reduction of current, power, or performance to remain within safe limits.

Thermal gradient: Temperature difference across cells, modules, or components.

Functional safety: Engineering discipline for controlling hazards caused by malfunctioning behaviour.

Review Questions

Short Answer Questions

1. Why is cell-level voltage measurement required in a series-connected battery pack?
2. Distinguish passive balancing from active balancing.
3. What is the purpose of a precharge circuit?
4. Define SOC, SOH, and SOP.
5. State two early indicators of possible thermal runaway.

Long Answer Questions

1. Explain the architecture and responsibilities of a modern BMS from cell monitoring to vehicle communication.
2. Describe a complete contactor-control sequence and the diagnostic checks required at each stage.
3. Explain how temperature affects battery power, charging acceptance, ageing, and safety.
4. Compare air, liquid, and refrigerant-based thermal-management approaches.
5. Discuss the layered strategy used to prevent and mitigate thermal runaway.

Analytical / Application Questions

1. A 60 Ah battery supplies 30 A for 40 minutes. Assuming ideal coulomb counting and constant capacity, calculate the percentage-point SOC decrease. State why the practical value may differ.
2. A precharge circuit uses R_p = 15 Ω and a DC-link capacitance of 2 mF. Calculate the initial ideal current for a 400 V pack and the RC time constant. Explain two reasons the measured values may differ.
3. A cooling loop carries 0.06 kg·s⁻¹ of coolant with c_p = 3,900 J·kg⁻¹·K⁻¹. Determine the heat-removal rate for a 4 K coolant temperature rise and assess whether it can remove 1.0 kW of steady heat.

Multiple Choice Questions with Answers

Question 1

Which quantity is most directly obtained by integrating battery current over time?

A. State of charge

B. State of health

C. Insulation resistance

D. Coolant flow rate

Correct Answer: A. State of charge

Question 2

A passive balancing circuit primarily removes excess cell energy by converting it into:

A. Mechanical work

B. Heat in a resistor

C. Chemical energy in the lowest cell

D. Magnetic energy in the motor

Correct Answer: B. Heat in a resistor

Question 3

What is the principal purpose of precharge?

A. Increase cell capacity

B. Limit inrush current into DC-link capacitors

C. Measure insulation resistance

D. Cool the battery before charging

Correct Answer: B. Limit inrush current into DC-link capacitors

Question 4

A 50 Ah battery supplies 25 A for 0.5 h. Under ideal coulomb counting, the SOC decrease is:

A. 12.5 percentage points

B. 20 percentage points

C. 25 percentage points

D. 50 percentage points

Correct Answer: C. 25 percentage points

Question 5

Which event most strongly indicates a possible welded main contactor after an open command?

A. Coolant temperature decreases

B. Bus voltage remains present despite the commanded isolation

C. Cell balancing current becomes zero

D. Cabin fan speed decreases

Correct Answer: B. Bus voltage remains present despite the commanded isolation

Question 6

For a coolant flow of 0.08 kg·s⁻¹, specific heat capacity of 3,800 J·kg⁻¹·K⁻¹, and temperature rise of 5 K, the heat-removal rate is:

A. 152 W

B. 760 W

C. 1,520 W

D. 19,000 W

Correct Answer: C. 1,520 W

Question 7

Which state variable specifically describes degradation relative to a reference condition?

A. SOC

B. SOH

C. SOP

D. OCV

Correct Answer: B. SOH

Question 8

Why should a BMS monitor temperature gradients as well as average temperature?

A. Averages always overestimate the hottest cell

B. Local hot spots can exist even when the average is acceptable

C. Gradients determine pack nominal voltage directly

D. Gradients eliminate the need for current sensing

Correct Answer: B. Local hot spots can exist even when the average is acceptable

Question 9

Which is an example of thermal-runaway propagation mitigation?

A. Increasing uncontrolled charging current

B. Removing all temperature sensors

C. Using thermal barriers and directed vent paths

D. Disabling contactor diagnostics

Correct Answer: C. Using thermal barriers and directed vent paths

Question 10

For a 400 V pack and a 20 Ω precharge resistor, the initial ideal precharge current is:

A. 5 A

B. 20 A

C. 80 A

D. 8,000 A

Correct Answer: C. 80 A