New Research Advances Multi Functional Air Conditioning Modeling and Control for Battery Electric Vehicles

A comprehensive research program has advanced the modeling and control of multi functional air conditioning systems for battery electric vehicles, targeting a critical balance between cabin comfort, intelligent user experience, and energy efficiency. The work examines an internal heat exchanger heat pump architecture, develops a fuzzy PID control strategy, designs a vehicle-grade controller, and validates the system through simulation, electrical testing, electromagnetic compatibility testing, and real vehicle trials. The findings are positioned to support the next generation of electric vehicle thermal management, where air conditioning is no longer a stand-alone comfort device but an integrated, connected, and energy-aware subsystem.

The research comes as the electric vehicle market continues to expand at a rapid pace. In 2025, China’s automobile production and sales reached 34.531 million and 34.40 million units, up 10.4 percent and 9.4 percent year on year, respectively. New energy vehicle production and sales reached 16.626 million and 16.49 million units, up 29 percent and 28.2 percent, respectively, maintaining the world’s leading position for 11 consecutive years. The market share of new energy vehicles reached 47.9 percent of total new vehicle sales, 7 percentage points higher than the same period a year earlier. This growth has placed greater attention on the energy consumption of electric vehicle auxiliary systems, especially the air conditioning system, which directly influences driving range and passenger comfort.

The research notes that a traditional internal combustion engine vehicle can use engine waste heat for cabin heating and can drive the compressor directly from the engine, while a battery electric vehicle relies on the power battery for both propulsion and thermal management. As a result, the electric vehicle air conditioning system has a more independent operating character, and its control strategy has a direct impact on energy use. The work therefore focuses on an electric vehicle heat pump air conditioning system with an internal heat exchanger, a fuzzy PID control method, and an AUTOSAR-based controller architecture designed for multi functional cabin services.

Key technology themes: electric vehicle thermal management, electric vehicle heat pump modeling, electric vehicle cabin comfort, electric vehicle energy efficiency, electric vehicle intelligent control, electric vehicle remote connectivity.

1. Why Electric Vehicle Cabin Thermal Management Matters

Air conditioning is one of the most important auxiliary loads in an electric vehicle. It regulates cabin temperature and humidity, maintains air quality, supports windshield defrosting and defogging, and contributes to driver alertness and passenger comfort. In an electric vehicle, the air conditioning system is also closely linked to battery thermal management, because both the cabin and the battery require appropriate temperature ranges for safe and efficient operation. The research therefore treats the electric vehicle air conditioning system as part of a broader thermal management architecture rather than as an isolated comfort feature.

The study highlights that human thermal comfort is generally associated with a temperature range of approximately 21 °C to 26 °C, which overlaps with the preferred operating temperature ranges of key electric vehicle components. A well-designed electric vehicle thermal management system must coordinate the air conditioning state, cabin temperature, and battery temperature. This coordination is important not only for comfort but also for battery safety, charging performance, and overall electric vehicle range.

At the same time, user expectations for electric vehicle cabins have changed. Drivers and passengers now expect more than basic cooling and heating. They expect remote pre-conditioning, multi-zone temperature control, air quality monitoring, fatigue relief, call noise reduction, intelligent air circulation, negative ion purification, and rapid cabin ventilation. These functions require more sensor inputs, more actuator control, more communication interfaces, and more advanced software architecture. The research addresses this shift by combining thermal system modeling with a multi functional controller design.

2. Internal Heat Exchanger Heat Pump Architecture for Electric Vehicle Applications

The research examines an internal heat exchanger, or IHX, heat pump air conditioning system for electric vehicles. The IHX is a coaxial heat exchange device in which high-pressure liquid refrigerant flows in the outer tube and low-pressure gaseous refrigerant flows in the inner tube. The two refrigerant streams flow in opposite directions, allowing heat to transfer from the high-pressure liquid line to the low-pressure gas line. This internal exchange does not require an additional external energy input, because it uses the temperature difference already present within the refrigerant circuit.

On the high-pressure side, liquid refrigerant leaving the condenser enters the outer tube of the IHX and releases heat, becoming a subcooled liquid before it reaches the expansion valve. On the low-pressure side, gaseous refrigerant leaving the evaporator enters the inner tube and absorbs heat, becoming a superheated gas before it returns to the compressor. The increased subcooling improves the refrigerant’s heat absorption capacity in the evaporator, while the increased superheat reduces the risk of liquid slugging and supports compressor reliability. The research states that this internal heat exchange can improve the coefficient of performance of the electric vehicle heat pump system and contribute to more stable refrigerant circulation.

The working fluid considered in the research is R134a. In the ideal refrigeration cycle, the refrigerant undergoes four main processes: isentropic compression in the compressor, isobaric condensation in the condenser, adiabatic expansion through the expansion valve, and isobaric evaporation in the evaporator. With the IHX, the cycle gains an additional internal exchange stage. The condenser outlet becomes more subcooled, and the evaporator outlet becomes more superheated. The research links these changes to improved refrigeration capacity, improved cycle efficiency, and reduced compressor risk.

The electric vehicle air conditioning system described in the research includes an HVAC control section, a refrigerant circulation section, and a coolant circulation section. The HVAC control section receives user requests from the vehicle infotainment system, the air conditioning control panel, or a mobile phone remote control. It also receives signals from sensors such as solar radiation, cabin temperature, cabin humidity, outlet air temperature, and ambient temperature. The control section then calculates target outlet air temperature, air volume, operating mode, and fresh or recirculated air mode, and sends commands to mode motors, blend motors, circulation motors, and blowers. It also sends coordinated commands over the vehicle network to the refrigerant circulation section and the coolant circulation section.

The refrigerant circulation section receives target outlet air temperature, air volume, operating mode, and air circulation mode from the HVAC control section, while collecting ambient temperature, refrigerant temperature, and refrigerant pressure as feedback. It calculates refrigerant flow demand and adjusts the electronic expansion valve opening, electric compressor speed, and solenoid valve states. In a battery fast-charging high-temperature scenario, the system can prioritize compressor speed and request greater heat dissipation from the coolant circulation section, thereby supporting both cabin cooling and battery cooling in the electric vehicle.

The coolant circulation section reads ambient air temperature, coolant temperature, battery coolant temperature, battery temperature, vehicle speed, and state of charge from the vehicle network. Its controller decides whether heating or heat dissipation is required. During low-temperature startup, it can drive a high-voltage water heater and water pump, direct the three-way valve to the heater core circuit, and quickly raise cabin and battery temperature. Under high-load or fast-charging conditions, it can switch the three-way valve to the radiator circuit and adjust the coolant pump and cooling fan through PWM control to regulate coolant flow and heat dissipation intensity.

Electric vehicle operating mode Primary thermal target Key system action
Cabin-only cooling Reduce cabin temperature The compressor operates, the outdoor heat exchanger rejects heat, and the evaporator cools cabin air.
Battery-only cooling Maintain battery temperature within a safe range The chiller exchanges heat between refrigerant and battery coolant, removing battery heat.
Combined cabin and battery cooling Support cabin comfort and battery safety at the same time The compressor speed and refrigerant flow are coordinated, and the coolant loop increases heat rejection.
Cabin-only heating Raise cabin temperature The heat pump operates, the outdoor heat exchanger absorbs heat, and the water-cooled condenser delivers heat to the coolant loop.
Battery-only heating Improve battery activity in low temperature The three-way valve is switched so that cabin and battery coolant exchange heat through the chiller.
Combined cabin and battery heating Balance passenger comfort and battery performance The heat pump, water heater, and coolant circuit are coordinated to serve both thermal targets.

3. Model Development and Validation for the Electric Vehicle Heat Pump System

The research built a dynamic simulation model of the electric vehicle multi functional air conditioning system on the AMESim platform. The model includes the electric compressor, outdoor heat exchanger, evaporator, chiller, water-cooled condenser, IHX, expansion valve, water heater, water pump, and connecting pipes. The compressor model considers volumetric efficiency, isentropic efficiency, and mechanical efficiency. The heat exchanger models consider refrigerant-side and air-side heat transfer, pressure drop, wall temperature, and fin-and-tube or micro-channel structures. The IHX model considers the coaxial tube geometry, inner and outer flow areas, wetted perimeters, equivalent diameters, Reynolds number, Nusselt number, and convective heat transfer coefficients for the gas and liquid sides.

The research validated the compressor model using inlet pressure, inlet temperature, outlet pressure, and compressor speed as inputs. Under the same operating conditions, the simulated compressor power and outlet temperature were compared with experimental values. During the startup phase, the compressor power showed an experimental peak of 1.72 kW and a simulated peak of 1.68 kW, followed by oscillation and decay to steady state. The steady-state simulated power was 0.92 kW, while the experimental average was 0.95 kW, with an error of less than 3.2 percent. The compressor outlet temperature first rose quickly and then stabilized. After 200 seconds, the simulated value remained at 48.2 ± 0.3 °C, while the experimental value was 48.5 ± 0.4 °C, with a relative steady-state error of less than 5 percent. These results indicate that the electric compressor model has good engineering credibility for electric vehicle system simulation.

The heat exchanger model was calibrated using a heat exchanger calibration tool in AMESim. Boundary conditions were applied to both sides of the heat exchanger, and dedicated models were used to iterate heat transfer and refrigerant pressure drop until the simulation matched the experimental data. Five typical operating conditions were used for validation. The maximum error in heat transfer was 4.4605 percent, and the average error was 2.9093 percent. The maximum error in refrigerant pressure drop was 4.3445 percent, and the average error was 2.0010 percent. Because the outdoor heat exchanger, evaporator, and chiller are all heat exchanger types, the same calibration approach was applied to them. The validation results show that the heat exchanger models are sufficiently accurate for system-level simulation and optimization of the electric vehicle air conditioning system.

Validation item Experimental or reference value Simulation value Error or assessment
Compressor startup power peak 1.72 kW 1.68 kW Close transient match
Compressor steady power 0.95 kW 0.92 kW Error less than 3.2 percent
Compressor outlet temperature after stabilization 48.5 ± 0.4 °C 48.2 ± 0.3 °C Relative error less than 5 percent
Heat exchanger heat transfer maximum error Experimental data set Calibrated model 4.4605 percent
Heat exchanger heat transfer average error Experimental data set Calibrated model 2.9093 percent
Heat exchanger refrigerant pressure drop maximum error Experimental data set Calibrated model 4.3445 percent
Heat exchanger refrigerant pressure drop average error Experimental data set Calibrated model 2.0010 percent

4. Performance Comparison Between IHX and Conventional Electric Vehicle Air Conditioning

To evaluate the effect of the internal heat exchanger, the research compared an IHX electric vehicle heat pump air conditioning system with a conventional system under the same operating conditions. The simulation was set at an ambient temperature of 30 °C, a target cabin temperature of 22 °C, and an initial cabin temperature of 48 °C, which is 1.6 times the ambient temperature to account for solar heat load. The simulation time was 500 seconds. The comparison covered superheat, subcooling, compressor energy consumption ratio, and cabin temperature.

In terms of superheat, the IHX system showed a brief fluctuation of 6.1 °C during startup and then stabilized at 10 °C. The conventional system showed multiple larger fluctuations, with a maximum of 10.3 °C, and stabilized at 4 °C. This indicates that the IHX effectively suppressed superheat fluctuation and reduced the risk of liquid slugging in the electric vehicle compressor. In terms of subcooling, the IHX system maintained subcooling at approximately −4.8 °C, while the conventional system remained near 0 °C. The improvement of about 4.8 °C indicates that the refrigerant’s unit-mass heat absorption capacity in the evaporator was enhanced by the internal heat exchange.

In terms of compressor energy consumption ratio, the gap between the two systems gradually reached 0.6 percentage points during startup. The conventional system reached a maximum of 8.8 percent, while the IHX system reached a maximum of 8.5 percent. During steady operation, the gap remained at 0.2 percentage points. The research interprets this as evidence that the IHX improves cycle efficiency in the electric vehicle heat pump system. In terms of cabin temperature, the IHX system oscillated between 21.5 °C and 24.3 °C, with a range of 2.8 °C, showed clear oscillation from 75 to 116 seconds, and converged to the target temperature at 171 seconds. The conventional system oscillated between 20.8 °C and 24.6 °C, with a range of 3.8 °C, showed clear oscillation from 65 to 111 seconds, and converged at 184 seconds. The comparison shows that the IHX improved temperature control accuracy and dynamic quality while supporting thermal comfort in the electric vehicle cabin.

Performance indicator Conventional electric vehicle air conditioning IHX electric vehicle air conditioning
Superheat behavior Multiple fluctuations, maximum 10.3 °C, stable at 4 °C Brief 6.1 °C startup fluctuation, stable at 10 °C
Subcooling Approximately 0 °C Approximately −4.8 °C
Compressor energy consumption ratio during startup Maximum 8.8 percent Maximum 8.5 percent
Compressor energy consumption ratio during steady operation Baseline About 0.2 percentage points lower
Cabin temperature oscillation range 20.8 °C to 24.6 °C, range 3.8 °C 21.5 °C to 24.3 °C, range 2.8 °C
Convergence to target temperature 184 seconds 171 seconds

5. Fuzzy PID Control Strategy for Electric Vehicle Air Conditioning

The research notes that an electric vehicle heat pump air conditioning system is nonlinear, time-varying, and subject to multiple disturbances. A fixed-parameter PID controller can provide fast response and steady-state accuracy, but it may not maintain optimal performance across all operating conditions. Fuzzy control can handle nonlinearity and uncertainty through linguistic rules, but it may have limitations in dynamic response precision. The research therefore combines the two approaches and designs a fuzzy PID control method for the electric vehicle air conditioning system.

The fuzzy PID controller uses the error between cabin temperature and target temperature, together with the error change rate, as fuzzy inputs. The PID parameters are the outputs. The input and output domains are set to [−3, 3] and [−5, 5], respectively, and each variable is divided into seven fuzzy subsets: NB, NM, NS, ZO, PS, PM, and PB. Gaussian membership functions are used for the input variables to improve noise suppression, while triangular membership functions are used for the output variables to reduce defuzzification complexity and improve engineering portability.

The fuzzy rules follow a continuous adjustment logic based on temperature state, parameter adjustment, and steady-state accuracy. When the temperature error is large, the controller increases the proportional gain to improve response speed, limits the integral gain to prevent integral windup and overshoot, and introduces derivative action to suppress possible oscillation. When the error is moderate but the error change rate is large, the controller strengthens the derivative gain to provide sufficient damping and avoid temperature overshoot. When both the error and the error change rate are small, the controller increases the integral gain to eliminate steady-state deviation and reduces the proportional and derivative gains to avoid overreaction to small disturbances. This logic allows the electric vehicle air conditioning system to converge quickly without overshoot or oscillation.

The research built an AMESim-Simulink co-simulation platform to validate the fuzzy PID controller. AMESim provided the high-precision physical model of the compressor, heat exchangers, IHX, expansion valve, water heater, and coolant loop. Simulink provided the control system design environment and the fuzzy inference engine. The interface exchanged cabin temperature and target temperature from AMESim to Simulink, and compressor motor torque from Simulink back to AMESim. The fuzzy controller performed real-time inference, defuzzification, and PID parameter correction, forming a closed-loop control path for the electric vehicle heat pump system.

Control element Design choice Purpose in the electric vehicle air conditioning system
Fuzzy inputs Cabin temperature error and error change rate Capture both the current deviation and the trend of cabin temperature.
Fuzzy output PID parameter corrections Adjust proportional, integral, and derivative gains online.
Input membership function Gaussian membership function Improve noise suppression in the electric vehicle cabin sensor signals.
Output membership function Triangular membership function Reduce defuzzification complexity for controller implementation.
Rule logic Large error for fast correction, fast change for damping, small error for stable integral action Balance response speed, overshoot suppression, and steady-state accuracy.

6. Co-Simulation Results Across Ambient Temperatures and WLTC Conditions

The research evaluated the fuzzy PID method under different ambient temperatures and drive cycles. In one set of tests, the electric vehicle was simulated at 30 °C, 35 °C, and 40 °C while traveling at a constant speed of 100 km/h on a level road. The simulation time was 2000 seconds, and the print step was 1 second. The initial cabin temperature was set to 1.6 times the ambient temperature, namely 48 °C, 56 °C, and 64 °C. The initial battery temperature was set to the ambient temperature. The target cabin temperature was 22 °C, and the target battery temperature was 28 °C.

At 30 °C, the PID-controlled cabin temperature oscillated between 18.4 °C and 24.4 °C, showed clear oscillation from 180 to 200 seconds, and converged to the target at 400 seconds. The fuzzy PID-controlled cabin temperature oscillated between 18.6 °C and 21.8 °C, showed slight oscillation, and then gradually approached the target with significantly reduced overshoot. For battery temperature, the PID control exceeded the target at 1000 seconds and showed clear overshoot, while the fuzzy PID control gradually approached the target and never exceeded 28 °C, demonstrating better temperature-following behavior.

At 35 °C, the PID-controlled cabin temperature began to oscillate at 690 seconds, ranged from 21.3 °C to 25.0 °C, and stabilized near the target at 840 seconds. The fuzzy PID control began to oscillate at 688 seconds, ranged from 21.5 °C to 23.1 °C, and stabilized near the target at 743 seconds. The oscillation amplitude was smaller and convergence was faster. For battery temperature, the two methods performed similarly, but the fuzzy PID control stabilized more precisely near the target after overshoot, with a smaller steady-state deviation.

At 40 °C, the PID-controlled cabin temperature oscillated around the target at 1205 seconds, ranged from 20.7 °C to 22.4 °C, and settled into oscillation between 22.4 °C and 21.7 °C. The fuzzy PID control began to oscillate at 1183 seconds, ranged from 20.8 °C to 22.3 °C, and stabilized near the target at 1270 seconds. The oscillation amplitude was slightly smaller and convergence was better. For battery temperature, both methods performed similarly overall, but the fuzzy PID control approached the target more stably after overshoot and achieved better steady-state accuracy.

Ambient temperature PID cabin temperature behavior Fuzzy PID cabin temperature behavior Battery temperature observation
30 °C 18.4 °C to 24.4 °C, oscillation at 180 to 200 seconds, convergence at 400 seconds 18.6 °C to 21.8 °C, reduced overshoot, gradual convergence PID exceeded 28 °C at 1000 seconds; fuzzy PID never exceeded 28 °C
35 °C 21.3 °C to 25.0 °C, oscillation from 690 seconds, stabilization at 840 seconds 21.5 °C to 23.1 °C, oscillation from 688 seconds, stabilization at 743 seconds Fuzzy PID showed smaller steady-state deviation after overshoot
40 °C 20.7 °C to 22.4 °C, oscillation at 1205 seconds, settled between 22.4 °C and 21.7 °C 20.8 °C to 22.3 °C, oscillation at 1183 seconds, stabilization at 1270 seconds Fuzzy PID approached the target more stably after overshoot

The research also analyzed the electric vehicle air conditioning system under the WLTC driving cycle at ambient temperatures of 30 °C, 35 °C, and 40 °C. The simulation time was 1800 seconds, and the initial cabin temperature was again set to 1.6 times the ambient temperature. The target cabin temperature was 22 °C, and the target battery temperature was 28 °C. Under the WLTC cycle, the fuzzy PID control responded effectively to dynamic speed changes and maintained precise regulation of cabin and battery temperature.

For cabin temperature, the system reached the target at 210 seconds, 660 seconds, and 796 seconds at 30 °C, 35 °C, and 40 °C, respectively. The cooling rate decreased as ambient temperature increased. The overshoot values were 10.5 percent, 13.6 percent, and 22.7 percent, respectively. The steady-state adjustment time increased with ambient temperature, but the system eventually stabilized near 22 °C, with steady-state fluctuation controlled within ±0.5 °C. At 35 °C and 40 °C, the system experienced temperature rebound but responded quickly and converged again, showing strong robustness. For battery temperature, the system reached 28 °C at 200 seconds, 570 seconds, and 940 seconds at 30 °C, 35 °C, and 40 °C, respectively. The steady-state errors were −0.5 °C, −0.3 °C, and −0.1 °C. No obvious overshoot occurred, and the battery temperature-following behavior was strong. These results indicate that the fuzzy PID control method has good adaptability and convergence for electric vehicle thermal management under dynamic driving conditions.

7. Vehicle-Grade Controller Hardware for the Electric Vehicle Air Conditioning System

The research designed a multi functional air conditioning controller following a vehicle electronics V-model development process. The controller is the core hardware of the electric vehicle air conditioning system. It collects passenger commands and sensor data in real time, applies the control strategy, and sends commands over CAN and LIN buses to sub-controllers and actuators, including the compressor, positive temperature coefficient heater, blower, motors, and electronic expansion valves. The input and output signals were defined according to the control requirements, including temperature, pressure, and supply voltage signals, and a system electrical architecture was developed.

The main control chip selected for the controller is a Renesas RH850 series microcontroller. The design requirements included a 32-bit MCU with a main frequency above 80 MHz, Flash memory of 1 MB to 1.5 MB with pin-to-pin compatibility, built-in DataFlash with at least 100,000 erase and write cycles for EEPROM emulation, at least two CAN interfaces, LIN, SPI, IIC, and UART interfaces, at least 25 analog input channels, low-power mode with current below 40 µA, at least six PWM outputs with independent frequency configuration and ADC trigger capability, support for AUTOSAR 4.0.3 and above, and an automotive-grade operating temperature range of −40 °C to +105 °C.

For power management and communication, the research selected the MFS2323BMBA5EP system basis chip. This device integrates CAN and LIN transceivers with system and safety functions for automotive electronic control units. It provides a configurable power tree for the microcontroller and sensors, with scalable supply options from 100 mA to 400 mA. Voltage and power-up and power-down sequences are configurable. The chip includes run, low-power on, and low-power off modes to manage quiescent current and support fast wake-up. It also supports flexible data rate CAN and LIN transceivers up to 5 Mbit/s. The controller uses one high-speed CAN channel to communicate with the vehicle comfort CAN bus and three LIN channels to communicate with sensors and actuators inside the electric vehicle air conditioning system.

The CAN and LIN interface circuits include common-mode chokes, capacitors, termination resistors, filtering capacitors, transient voltage suppression diodes, current-limiting resistors, and Schottky diodes to suppress electromagnetic interference, reduce signal reflection and attenuation, and protect sensitive components from electrostatic discharge, surge, and overvoltage events. The LIN2 and LIN3 channels use the TLIN1021ADRQ1 automotive LIN physical layer transceiver, which is AEC-Q100 Grade 1 qualified, operates from 4.5 V to 36 V, supports LIN transmit data rates up to 20 kbit/s and receive data rates up to 100 kbit/s, and is suitable for the real-time requirements of the electric vehicle air conditioning controller.

Sensor acquisition circuits were designed for temperature and pressure sensors. The temperature sensor uses an NTC thermistor and the beta formula to calculate temperature from resistance change. The pressure sensor uses voltage linear fitting and compensation based on the pressure-voltage relationship, with correction for nonlinearity and temperature drift. Both circuits include pull-up resistors to maintain a high level when no signal is present, filter capacitors to remove high-frequency noise, and divider resistors to adjust the output signal to the ADC input range of the microcontroller. The controller also includes high-side drivers,低-side drivers, and stepper motor drivers for different load types.

Hardware requirement Design target for the electric vehicle air conditioning controller
Main processor 32-bit automotive-grade MCU with main frequency above 80 MHz
Program memory Flash size of 1 MB to 1.5 MB with pin-to-pin compatibility
Nonvolatile data storage Built-in DataFlash with at least 100,000 erase and write cycles
Communication interfaces At least two CAN channels, plus LIN, SPI, IIC, and UART
Analog input At least 25 analog input channels
Low-power current Below 40 µA in low-power mode
PWM output At least six channels with independent frequency configuration and ADC trigger capability
Software platform Support for AUTOSAR 4.0.3 and above
Operating temperature −40 °C to +105 °C

For actuator drive circuits, the research selected the VND7050 high-side driver for loads such as the blower and water pump. The device provides stable power supply, high voltage withstand capability, low on-resistance, and low conduction loss, and the circuit includes input filtering, current limiting, and motor current detection. The NCV8402 low-side driver is used for solenoid valves and the negative ion generator, providing high current drive capability and fast switching speed, with input capacitance, pull-up resistance, and output current monitoring. The DRV8889 stepper motor driver is used for the electronic expansion valve, mode motor, circulation motor, and blend motor, providing high integration, wide operating voltage range, micro-stepping control, stall detection, and low-power sleep mode. These three driver types together support the electric vehicle air conditioning system with efficient and reliable actuator control.

8. AUTOSAR-Based Software Architecture and Multi Functional Services

The research follows the AUTOSAR layered architecture to improve software reusability, maintainability, and hardware abstraction. The architecture includes the application layer, the runtime environment, the basic software layer, and the microcontroller layer. The application layer contains software components for temperature regulation, air volume control, air outlet mode switching, and fresh or recirculated air switching. The runtime environment provides communication services between software components and the basic software layer, decoupling software from hardware and improving portability. The basic software layer provides standardized functional interfaces and is divided into the service layer, ECU abstraction layer, microcontroller abstraction layer, and complex driver layer.

In the system service module, the research designed task scheduling, clock management, power management, and watchdog monitoring. The task scheduling mechanism combines interrupt-driven operation with a main loop polling structure. After system initialization, the main loop checks global flags to determine whether a new task needs to be executed. If a flag is set, the controller clears the task state, resets the flag, and executes the task. If no new task is pending, the system waits for the operating system timer interrupt. The interrupt service routine checks the interrupt every millisecond and increments a counter. The counter is used to track interrupt counts and determine task timing. This mechanism supports timely task response and stable system operation in the electric vehicle air conditioning controller.

Clock management provides a unified time base for task scheduling, event recording, and system coordination. It includes system clock initialization, calibration, and maintenance, and provides time stamp and timer services to upper layers. Power management dynamically switches power modes to optimize performance and energy consumption. It monitors the power state, configures wake-up conditions, writes configuration data to the power chip through the serial peripheral interface, calculates a cyclic redundancy check to ensure configuration accuracy, de-initializes power-chip-related general-purpose input and output pins to reduce power consumption, and enters sleep mode. During sleep, the system monitors wake-up conditions. When a wake-up condition is met, the power management module reconfigures the power chip, verifies the cyclic redundancy check again, and wakes the system. The watchdog monitoring module periodically resets the watchdog timer. If the timer is not reset within the preset time, a timeout interrupt is triggered and the system is reset to recover normal operation. The watchdog is fed every 100 milliseconds in the main loop to prevent timeout and improve reliability.

The storage module manages application data access, including read and write operations. The NvmDrv driver provides low-level access to Data Flash memory, including read, write, and erase functions. The bootloader is固化 in a protected storage area and is not erased, so the application can be reprogrammed if it is damaged. The flashing process uses unified diagnostic services and requires preconditions such as vehicle speed of 0 km/h, park gear, electronic parking brake applied, and not charging. The bootloader must pass security access authentication, and the application uses an A/B backup mechanism. At startup, the controller calls the latest software first; if it is invalid, the backup is used. During flashing, the client sends a functional addressing diagnostic online service every 4 seconds to maintain bus communication, and the programmed controller ensures that this message does not affect the flashing process.

The communication module supports CAN, LIN, and SPI. CAN communication uses a decentralized multi-master architecture and message-based broadcast to exchange data among vehicle electronic control units. The CAN driver configures low-level registers, handles message transmission, network management, and error processing, and uses built-in error detection and fault tolerance to maintain stable bus operation in complex electromagnetic environments. LIN communication uses a master-slave structure and is suitable for temperature sensors, air door actuators, and other nodes in the electric vehicle air conditioning system. The LIN driver implements message scheduling, fault detection, and signal encoding and decoding. SPI communication supports high-speed board-level serial communication between the microcontroller and peripheral chips such as sensors and memory devices.

The research also applies AUTOSAR network management and unified diagnostic services. Network management uses a multi-master cooperative strategy and broadcast network management messages to coordinate network state. The operating modes include bus sleep mode, prepare bus sleep mode, and network mode. Network mode is further divided into repeat message state, normal operation state, and ready sleep state. A node transitions between states based on its own network request state and received network management messages. When the network needs to be awakened, the node actively sends network management messages and enters the fast repeat message state, then switches to the normal repeat message state. If a request persists, the node enters the normal operation state and maintains periodic message transmission. When the request is released and no valid network management message is received, the node passes through the ready sleep state and prepare bus sleep state, and finally enters sleep mode after the bus sleep timer expires. This mechanism supports coordinated wake-up and sleep management among multiple nodes and optimizes bus load and system power consumption.

Unified diagnostic services provide standardized fault diagnosis and maintenance functions, including fault code reading, parameter configuration, data monitoring, and software update. The electronic control unit continuously performs fault self-detection during initialization and operation. When an abnormality is detected, it stores a diagnostic trouble code and records fault status, snapshot information, and extended data. If a diagnostic request message is invalid or cannot be executed, a negative response is sent with the corresponding return code. For sensitive operations such as program download and memory read and write, the security access service uses a seed-key authentication mechanism. The diagnostic tool first sends a request seed service, and the controller responds with a seed value. The tool calculates a key based on the seed and sends a send key service. The controller compares the received key with its internal calculation. If they match, access is granted and a positive response is sent. If they do not match, access is denied and a new seed is requested. This mechanism helps prevent unauthorized access and supports vehicle functional safety and emissions performance.

9. Multi Functional Application Layer for Electric Vehicle User Experience

The application layer software is designed around user experience, driving safety, and energy optimization. It uses vehicle connectivity and cabin perception technologies to deliver multiple intelligent functions. These functions are intended to transform the electric vehicle air conditioning system from a manually operated comfort device into a context-aware service platform.

  • Mobile phone remote control: The function uses a cloud-pipe-device architecture. A user sends commands through a mobile application, and the command is delivered through the cloud to the telematics box. After authentication with the body control module and confirmation that the vehicle is in remote mode, the user can send air conditioning on or off commands, one-touch cooling, or one-touch heating. The on command sets the electric vehicle air conditioning system to AUTO mode at 22 °C, one-touch cooling sets AUTO mode at 18 °C, and one-touch heating sets AUTO mode at 28 °C. The system reports status back to the application.
  • Intelligent air conditioning zoning: The function supports single-zone, dual-zone, and three-zone modes. In single-zone mode, the main driver, front passenger, and rear row blend motors use the same signal and are controlled by the main driver side. In dual-zone mode, the front row uses synchronized control while the rear row is independently controlled. In three-zone mode, the main driver, front passenger, and rear row are independently controlled. The controller reports the zoning state to the central display to keep the interface consistent with the system state.
  • Fatigue relief assistant: When the vehicle is powered on and driving, an intelligent camera monitors driver state. If fatigue is detected, the vehicle provides a warning, and the electric vehicle air conditioning system responds to a fatigue relief request from the central display by turning off automatic mode and adjusting temperature, air volume, and air outlet mode to help the driver stay alert.
  • Call noise reduction: When a Bluetooth call is detected, the electric vehicle air conditioning system automatically reduces air volume to a preset quiet threshold of level 3 while maintaining basic thermal comfort. Manual operation takes priority, and after the call ends, the air volume returns to its previous state. This function reduces driver distraction and improves call quality.
  • PM2.5 and intelligent air circulation: The system monitors PM2.5 and air quality signals and automatically switches between fresh air and recirculation modes. It responds to front defrost mode, windshield washing, poor outside air quality, high cabin PM2.5 concentration, reverse gear, and power-down events. Manual operation takes priority, and the system restores the previous circulation state when the triggering condition ends.
  • Negative ion generator: The negative ion generator can be turned on manually or automatically. When the cabin PM2.5 value is above 150 micrograms per cubic meter for 3 seconds and the automatic negative ion setting is enabled, the system activates the generator. If the front air conditioning is off, the blower operates at minimum duty, the air outlet mode changes to face or face and foot, and the air conditioning status remains off. If the front air conditioning is on, the state remains unchanged. The function exits when the vehicle powers down, when cabin PM2.5 stays at or below 65 micrograms per cubic meter for 30 seconds, or when the switch is pressed again.
  • Cabin ventilation and cooling: When the vehicle is powered down, cabin temperature is at least 5 °C higher than ambient temperature, ambient temperature is above 25 °C, all four doors are closed, state of charge is above 60 percent, and the ventilation function is enabled, a long press of the unlock key for more than 3 seconds triggers automatic air conditioning. The system enters manual mode, sets air volume to level 7, selects face and foot outlets, and switches to fresh air mode to quickly exchange cabin air. The compressor and positive temperature coefficient heater remain off, and the temperature setting remains at the previous value. The function exits when the driver door opens, after 5 minutes, when the vehicle powers up, or when state of charge falls below 60 percent.
Application function Trigger or condition Electric vehicle cabin benefit
Mobile remote control Authenticated remote command through the telematics box and body control module Pre-cooling or pre-heating before the driver enters the electric vehicle
Intelligent zoning Single-zone, dual-zone, or three-zone request from the central display Personalized comfort for different seats in the electric vehicle
Fatigue relief assistant Camera-based fatigue detection Adjusts cabin environment to help maintain driver alertness
Call noise reduction Bluetooth call event Reduces blower noise and improves call quality
Intelligent air circulation Defrost, washing, air quality, PM2.5, reverse gear, or power-down signals Balances safety, air quality, and energy use
Negative ion generator Manual switch or automatic PM2.5 threshold Supports cabin air purification in the electric vehicle
Cabin ventilation and cooling Long press of unlock key under defined conditions Rapidly reduces cabin temperature before entry

10. Electrical Performance, EMC, and Real Vehicle Validation

The research completed the integration of the bootloader and application software on the target hardware. The bootloader code was obtained from a version-controlled repository, compiled with the Renesas CS+ for CC integrated development environment, and flashed to the internal Flash of the target microcontroller using the Renesas Flash Programmer and E1 emulator. The application software was compiled with CMake and Ninja, generating a binary file and release version. The application was flashed over the vehicle CAN bus using CANoe and a CAPL script. The script follows unified diagnostic services to communicate with the bootloader, perform erase, download, and verification steps, and restart the controller after successful flashing.

Electrical performance tests were conducted in a laboratory environment using a programmable power supply, signal generator, oscilloscope, multimeter, and load simulation box. The tests covered supply voltage adaptability, fault mode protection, electrical safety, and initialization and recovery. The supply voltage tests included normal supply voltage, slow voltage rise and fall, abnormal supply voltage, micro-interruption of power, and vehicle power network voltage fluctuation. The controller remained normal without reset, hang, or functional abnormality. Fault mode protection tests included quiescent current, grounding and connection to the positive terminal, long-term overload, single-wire open circuit, multi-wire open circuit, and signal circuit short-circuit protection. The controller entered protection states when necessary and recovered automatically after the fault was removed. Electrical safety tests included dielectric strength and insulation resistance, with no breakdown or flashover and with insulation resistance meeting standard requirements. Initialization and recovery tests confirmed that the controller started normally and restored its pre-power-down working state. All tested items met the performance level requirements.

Test area Test item Test state Result
Power supply adaptability Normal supply voltage performance Standby Pass
Power supply adaptability Slow voltage rise and fall performance Standby Pass
Power supply adaptability Abnormal supply voltage performance Standby Pass
Power supply adaptability Micro-interruption of power Sleep Pass
Power supply adaptability Vehicle power network voltage fluctuation Standby Pass
Fault mode protection Quiescent current Sleep Pass
Fault mode protection Grounding and positive terminal connection Standby Pass
Fault mode protection Long-term overload Standby Pass
Fault mode protection Single-wire open circuit Standby Pass
Fault mode protection Multi-wire open circuit Standby Pass
Fault mode protection Signal circuit short-circuit protection Standby Pass
Electrical safety Dielectric strength and insulation resistance Standby Pass
Initialization and recovery Re-initialization test Standby Pass

Electromagnetic compatibility tests were conducted in an electromagnetic compatibility laboratory. The test platform included a receiver, signal generator, artificial power network, antennas, and coupling devices. Emission tests covered conducted emission by voltage method, conducted emission by current method, radiated emission, and magnetic field emission. Immunity tests covered transient coupling immunity, transient conduction immunity, bulk current injection immunity, radiated immunity, portable transmitter immunity, and electrostatic discharge immunity. The controller’s conducted emission, radiated emission, and magnetic field emission amplitudes were below the standard limits. During transient coupling, bulk current injection, radiated immunity, portable transmitter immunity, and electrostatic discharge tests, the controller maintained normal operation or recovered automatically without data loss or logic disturbance. All tested items met the performance level requirements.

Test area Test item Test state Result
Emission Conducted emission by voltage method Working Pass
Emission Conducted emission by current method Working Pass
Emission Radiated emission Working Pass
Emission Magnetic field emission Working Pass
Immunity Transient coupling immunity Working Pass
Immunity Transient conduction immunity Working Pass
Immunity Bulk current injection immunity Working Pass
Immunity Radiated immunity Standby and working Pass
Immunity Portable transmitter immunity Standby and working Pass
Immunity Electrostatic discharge immunity Working and idle Pass

Real vehicle validation was conducted after the controller was installed in a test vehicle. CANoe was used to monitor vehicle status and record data over the CAN bus. The tests focused on software and hardware cooperation, application layer and basic layer integration, and the actual performance of the intelligent functions in typical user scenarios. For mobile phone remote control, the tests verified air conditioning on and off, one-touch cooling, and one-touch heating. When the air conditioning was off, a remote on request caused the electric vehicle air conditioning system to start immediately in AUTO mode at 22 °C and report status every 100 milliseconds. A remote off request caused immediate shutdown and status reporting. One-touch cooling started the system in AUTO mode at 18 °C and fed back the one-touch cooling state. One-touch heating started the system in AUTO mode at 28 °C and fed back the one-touch heating state.

For intelligent zoning, the tests confirmed single-zone, dual-zone, and three-zone transitions. In single-zone manual mode, a zoning control signal from the central display did not change the main driver and front passenger temperature settings, and the system remained in single-zone mode. Operating the front passenger temperature entered three-zone mode, while operating the rear temperature entered front-rear dual-zone mode with synchronized front left and right control and independent rear control. In dual-zone mode, different front and rear temperature settings caused a return to single-zone mode when a zoning control signal was received. In three-zone mode, different main driver, front passenger, or rear settings caused a transition to dual-zone mode, and a single-zone control signal caused a transition to single-zone mode with the front passenger and rear settings forced equal to the main driver setting.

For the fatigue relief assistant, when the vehicle power mode was local on, the electric vehicle air conditioning system received a fatigue relief request from the central display. The request included forced automatic mode off and settings for temperature, air volume, and air outlet mode. The system turned off automatic mode, lowered the temperature, set air volume to level 7, and switched the air outlet to face mode to help the driver stay alert. For call noise reduction, when the air outlet mode was not windshield mode and a Bluetooth call event was received, the controller checked the current air volume level. If the level was above 3, it reduced the level to 3; if the level was 3 or below, it kept the level unchanged. Manual operation during the call took priority and exited the noise reduction function. After the call ended, the air volume returned to its previous state.

For intelligent air circulation, the tests covered front defrost automatic fresh air, windshield washing automatic recirculation, poor outside air quality automatic recirculation, high cabin PM2.5 automatic recirculation, reverse gear automatic recirculation, and power-down automatic fresh air. When the air conditioning received a front defrost mode signal, it switched to fresh air mode and reported the external circulation state. When the front defrost mode ended, the circulation mode returned to its previous state. When a windshield washing signal was received, the system switched to recirculation mode while the display and reported circulation state remained unchanged. When the washing signal stopped for 10 seconds, the system exited the function and returned to the previous circulation mode. For poor outside air quality, the system monitored the air quality signal and switched to recirculation when the threshold and duration conditions were met, then returned to the previous mode when the air quality recovered. For high cabin PM2.5, when the cabin PM2.5 value remained above 150 micrograms per cubic meter for 3 seconds, the system switched to recirculation and reported the internal circulation state. If the cabin PM2.5 value remained at or below 65 micrograms per cubic meter for 30 seconds, the system returned to the previous mode. For reverse gear, when the gear was R and vehicle speed was above 1 km/h, the system switched to recirculation. After the gear left R for 5 minutes, the system restored the previous mode. If R was engaged again within 5 minutes, the timer restarted. For power-down, when the vehicle power mode was off and the air conditioning was in recirculation, the system automatically switched to fresh air mode. After the vehicle powered up again, the circulation mode returned to the previous mode.

For the negative ion generator, the tests covered manual switch operation and automatic PM2.5-triggered operation. When the vehicle was powered on and the central display negative ion switch was off, pressing the switch turned the negative ion generator on. If the front air conditioning was off, the blower operated at level 1 in face or face and foot mode. If the front air conditioning was on, other settings remained unchanged. The system memorized the state before power-down and did not respond to the sent value within 1 second after the next power-up, sending the memorized value instead. The function exited when the vehicle powered down or the switch was pressed again. In automatic mode, when the vehicle was powered on, the automatic negative ion setting was on, and cabin PM2.5 was above 150 micrograms per cubic meter for 3 seconds, the system activated the negative ion generator. If the front air conditioning was off, the blower operated at minimum voltage or duty, the air outlet mode included face or face and foot, and the air conditioning switch status remained off. If the front air conditioning was on, the state remained unchanged. The function exited when the vehicle powered down, when cabin PM2.5 stayed at or below 65 micrograms per cubic meter for 30 seconds, or when the switch was pressed again while the generator was on.

For cabin ventilation and cooling, the tests confirmed that when the vehicle was powered down, cabin temperature was at least 5 °C higher than ambient temperature, the ventilation cooling feedback setting was on, ambient temperature was above 25 °C, all four doors were closed, and battery state of charge was above 60 percent, a long press of the remote key unlock button for more than 3 seconds triggered automatic air conditioning. The system entered manual mode, set air volume to level 7, selected face and foot outlets, and switched to fresh air mode. The compressor and positive temperature coefficient heater were off, and the main driver and front passenger temperature settings and zoning state remained as before power-down. The function exited when the driver door opened, after 5 minutes, when the vehicle power mode was on, or when battery state of charge fell below 60 percent. The air conditioning circulation mode, air volume level, and air outlet mode returned to their previous states, and the air conditioning turned off.

11. Outlook for Electric Vehicle Thermal Management and Intelligent Cabin Control

The research concludes that the IHX heat pump architecture, fuzzy PID control method, and AUTOSAR-based multi functional controller provide a practical path for improving electric vehicle cabin comfort and energy efficiency. The simulation results show that the IHX improves superheat stability, increases subcooling, reduces compressor energy consumption ratio, and improves cabin temperature convergence. The fuzzy PID control method shows faster response, smaller overshoot, and better steady-state accuracy than conventional PID control across different ambient temperatures and WLTC conditions. The controller hardware and software design supports remote control, intelligent zoning, fatigue relief, call noise reduction, intelligent air circulation, negative ion generation, and cabin ventilation cooling, and the electrical, electromagnetic compatibility, and real vehicle tests confirm stable operation and reliable function execution.

The research also identifies several directions for future work. First, data-driven intelligent control algorithms could be introduced. Current control rules rely partly on expert experience and prior knowledge. With the accumulation of vehicle operating data, machine learning methods such as neural networks and reinforcement learning could be used for offline training and online learning, enabling adaptive control parameter adjustment and self-optimization for the electric vehicle air conditioning system. Second, the architecture could evolve toward central computing platforms. As vehicle electronic and electrical architectures move toward domain-centralized and zone-controlled designs, traditional distributed controllers may be gradually replaced by domain controllers or central computing platforms. Future work could study integrating air conditioning control into a cockpit domain controller or body domain controller, including software and hardware partitioning, standardized communication interfaces, and resource scheduling optimization. Third, multi-modal cabin perception and fusion control could be further developed. Current intelligent functions are triggered mainly by single signal sources such as Bluetooth events, camera-based fatigue detection, and PM2.5 sensors. Future systems could fuse driver facial features, voice commands, gesture recognition, ambient light sensing, and navigation information to build a more comprehensive understanding of the cabin scenario and deliver more natural and proactive cabin environment adjustment for the electric vehicle.

The overall contribution of the research is a complete chain from electric vehicle heat pump system modeling and performance analysis to control method design, controller development, and real vehicle validation. The work supports the transition of the electric vehicle air conditioning system from a basic comfort load to an intelligent, connected, and energy-aware thermal management node. As electric vehicle adoption continues to grow, such integrated modeling and control research is expected to play an important role in balancing cabin comfort, driving safety, and energy efficiency.

Scroll to Top