Techniques for EV Battery State Monitoring

A hybrid model for predicting the remaining useful life (RUL) of batteries in vehicles. The model combines physics-based battery models with machine learning models to improve prediction accuracy. The physics-based models generate battery properties from sensor data, simulations, manufacturer data, and lab experiments. These properties are then used by machine learning models to predict cell RUL.

A method for safely charging a traction battery in electric vehicles to avoid overcharging and potential safety issues. The method involves monitoring battery parameters like state of charge (SOC) during charging, and discharging or stopping charging when the parameter changes by a certain threshold. This prevents continuous charging if the battery is already full, preventing overheating, lithium plating, and other problems.

Estimating the degradation state of a battery non-destructively and rapidly without disassembling it. The method involves charging the battery in short intervals with pauses between until it reaches full charge. The voltage and charge capacity are recorded at the end of each pause. Differences between voltage and capacity changes are analyzed to identify peaks which indicate degradation.

Battery management system for accurately estimating the state of a battery. The system uses a battery state estimation algorithm that involves estimating a "surface SOC" representing the electrode surface potential based on current measurements. This surface SOC value is then combined with the overall SOC to estimate the battery's terminal voltage. By accounting for the dynamic effects of current flow on the surface potential, the algorithm provides more accurate state estimation during charging/discharging compared to just using overall SOC.

Optimizing power supply to multiple batteries in a power supply device. The device has a power generator, batteries, and loads connected by power lines. Diodes allow power flow only from generator to batteries. To supply power, a control system sums each battery's power deficit and adds load requirements. The generator output is then adjusted to meet this total. This compensates for power losses to the batteries. By accurately controlling generated and supplied power, it better manages multiple batteries compared to complex systems analyzing each battery's state.

Diagnosing battery abnormalities in electric vehicles to prevent failures and fires. The method analyzes voltage variation rates in battery cells over driving distances. A threshold of voltage change per distance is set based on vehicle state. Battery cell voltage changes exceeding the threshold indicate problems. By tracking voltage changes over distance, battery issues can be detected and diagnosed early to prevent accidents.

Device for accurately measuring battery impedance using separate current and voltage sensors. The device has two components - a current sensor that measures battery current and detects current feature quantities, and a voltage sensor that measures battery voltage and detects voltage feature quantities. The current sensor transmits its feature quantities to the voltage sensor. The voltage sensor calculates battery impedance using both sets of feature quantities, but only if the current feature quantities were received within a predetermined time period of the voltage feature quantities. This ensures the current and voltage measurements are synchronized for accurate impedance calculation.

A battery diagnosis method for electric vehicles that can detect faulty cells likely to catch fire. The method involves collecting vehicle and battery state information from a fleet of vehicles, tracking maximum voltage change rates per mileage for each vehicle, and diagnosing battery abnormalities if change rates exceed thresholds.

A high-voltage battery used in electric vehicles that can detect damage to the battery housing. The battery contains pressure sensors inside the housing and an evaluation device. It identifies damage like deformation or leakage by comparing internal pressure changes to an external pressure reference. If severe deformation is detected, the battery can be shut off to prevent further damage or safety risks. The battery can also warn the driver of minor damage to get the battery checked.

Estimating the remaining useful life (RUL) and cumulative wear of an electric vehicle battery to optimize battery life, performance and cost. The method involves periodically monitoring battery parameters like state of charge (SOC) and health, using that data to estimate degradation parameters like capacity loss and internal resistance, and then using those estimates to predict RUL and cumulative wear. This allows proactive optimization of battery usage to extend its lifetime.

Method to accurately estimate the temperature of individual battery cells in a battery pack using a single temperature sensor. This reduces cost and complexity compared to using a sensor on every cell. The method uses measured voltages, currents, and resistances from one cell along with the temperature sensor to compute the temperatures of other cells. It does this by creating resistance relationships between cells based on the reference resistance and measured temperature.

Rapidly detecting a thermal runaway of a battery in an electric vehicle and notifying the vehicle occupant to prevent dangerous situations like battery fires. The detection method involves comparing the voltage of each battery cell with the voltage of the battery module. If the battery module voltage is higher than the summed cell voltages, it indicates a sensor problem. But if the battery module voltage is lower, it indicates a cell in thermal runaway.

Monitoring battery cell health and electrolyte properties using external conductive portions on the battery enclosure. The conductive portions are attached to the cell's outer surface and capacitance measurements are taken between them and internal conducting structures like cell terminals. These capacitance values are used to assess battery health, electrolyte level, remaining life, and identify thermal runaway.

Secondary battery system that detects and suppresses high-rate deterioration caused by uneven salt concentration in the electrolyte. The system measures the battery impedance during charge/discharge. High-rate deterioration is detected as an increase in DC resistance component compared to initial. When detected, the system reduces the charge/discharge current to mitigate the deterioration.

Detecting thermal events in battery packs of electric vehicles using temperature sensors. The method involves flagging potentially faulty temperature readings from sensors as unreliable and excluding them from thermal event detection. This allows accurate identification of true thermal events even when some sensors malfunction.

Predicting the remaining useful life of a battery by using a hybrid model combining physics-based and machine learning models. The physics-based model generates battery properties from sensor data while the machine learning model predicts battery life from those properties. By combining the strengths of both models, accurate remaining life predictions can be made to optimize battery management.

Method for diagnosing health of a battery using in-vehicle impedance analysis. The method involves applying a current profile using an onboard charging module to cells in the battery and analyzing the voltage response to calculate impedance. Changes in the impedance components compared to thresholds indicate battery health problems.

Reducing battery usage and increasing efficiency of wireless communication between master and slave battery management systems in vehicle battery packs. The system uses an on/off schedule to periodically power the slave wireless communication module, instead of leaving it on continuously. This reduces standby power consumption while still allowing wireless data transfer between the master and slave BMS. The master BMS can control the schedule of when the slave BMS wireless module powers on/off to receive commands.

Monitoring secondary battery status to prevent over-discharge and extend battery life. The method uses sensors to detect voltage, current and temperature, then calculates the polarization overvoltage by subtracting the voltage drop due to resistance from the difference between open circuit and closed circuit voltage. This allows accurate tracking of the battery voltage drop due to polarization effects during discharge.

Improving the accuracy of state-of-health (SOH) estimation for silicon-dominant lithium-ion batteries using enhanced models that consider factors unique to silicon anodes. The models utilize data beyond the battery itself, like manufacturing and operating data, to calculate SOH. This allows accurate prediction of remaining life even with complex degradation relationships involving hysteresis, temperature, current, voltage, etc. The enhanced SOH models can be used in battery management systems to optimize battery operation and prolong useful life.