All-in-one electric vehicle monitoring with virtual sensors

Monitor 12V battery, high voltage battery, electronics, tires, and brakes

Monitor 12V battery, high voltage battery,
electronics, tires, and brakes

Extensive condition monitoring of electric vehicle beyond GPS tracking

Telematics systems for tracking fleets of vehicles are widespread around the globe. Fleet managers all over the world make increasing use of such monitoring solutions to get a centralized view of their vehicles and to optimize fleet operations. Historically, these telematics solutions have developed around GPS tracking. The main pain point of fleet managers was at that time to know “where are my trucks?”. Current telematics solutions for logistics use satellite navigation not only to prevent vehicles from being stolen, but also for optimized route calculation. The next step in fleet optimization resides in failure prediction and condition-based maintenance. Software-based solutions can help to reduce time-consuming and error-prone manual inspections, and avoid vehicle failures in the field which have a negative impact on costs, delivery time and customer satisfaction, not to speak of organizational issues.

Predictive maintenance and failure prediction provide also cost reduction and increased efficiency to shared car fleets, whether the drivers are company staff or individual consumers.

Our CAN-bus based electric vehicle monitoring solution provides deep insights into mechanical and electrical vehicle components, from B like battery until T like tires.

12V battery monitoring and failure prediction

Did you know that not only conventional cars with an internal combustion engine, and hybrid electric vehicles need a 12V starter battery, but also fully electric vehicles (EVs) ? It allows opening the doors, powers up the on-board infotainment system, and it activates the high voltage battery system. When the 12V battery is “flat”, an electric vehicle cannot start.

In a conventional vehicle, with a mere battery voltage signal over time, our algorithms are able to detect aging and malfunction in a 12V battery in a very reliable manner before the driver or the onboard diagnostics system (OBD) would notice any alteration.
In 2022, we will be able to provide a similar solution for electrified vehicles (PHEV/HEV/BEVs). Stay tuned!

Read more about 12/24V battery monitoring

High voltage battery system monitoring

High voltage (HV) batteries of EVs undergo two types of normal aging processes: calendar aging means that the battery performance decreases over time, even when you drive softly or when you do not drive at all. Cyclic aging means that the battery performance decreases by small incremental steps every time you actually charge or discharge it. Cyclic aging is higher in case of fast charging and in case of extreme charging (state of charge too close to 100%) and extreme discharging (too close to the lower limit).

Besides these aging, different other anomalies can happen in an HV-battery system. Our algorithms are able to detect thermal runaway and battery cell malfunction, among other issues.

High voltage battery monitoring requires continuous data regarding voltage, current and temperature together with basic battery specifications.

Monitoring of anomalies and isolation in the E/E system

Another frequent cause of defects in electric vehicles is linked to the cables. Animals like martens can bite the cables and damage electrical insulation. Demanding operating profiles may cause overheating of cables. We are currently developing algorithms for these types of anomalies with pilot customers

Tire wear monitoring

One of our algorithms calculates tire tread depth from vehicle data like wheel speed and GPS position. It provides a convenient centralized status indicator of all tires in a fleet without the need for manual inspection.

Brake wear and brake anomaly monitoring

We have developed two dedicated algorithms for the brake system of a vehicle. One that determines brake pad wear based on signals like motor speed, torque and brake pressure. The second algorithm detects anomalies in the braking system before they are detected by the on-board diagnostics.

Contact us for more detailed information about required input data.

All-in-one electric vehicle monitoring with virtual sensors | COMPREDICT

We have developed two dedicated algorithms for the brake system of a vehicle. One that determines brake pad wear based on signals like motor speed, torque and brake pressure. The second algorithm detects anomalies in the braking system before they are detected by the on-board diagnostics.

All-in-one electric vehicle monitoring with virtual sensors | COMPREDICT

Contact us for more detailed information about required input data.

System architecture for real-time EV (electric vehicle) fleet monitoring

The overall architecture of our solution consists of three main parts: data ingestion, algorithm hub and data visualization. The standard architecture is cloud-based: the input data (IoT data) from the vehicles are processed in a data lake, ingested and prepared for the further analysis. Our solution is quite flexible, since COMPREDICT can provide the data storage, or we can directly connect to customer’s databases through a pull or push approach.

The heart of our solution consists of our algorithm hub AI-CORE. The calculations, predictions and forecasting are executed in this cloud-based algorithm hub, which we call the virtual sensors’ hub. In parallel, we are working with selected customers on decentralized algorithm execution (edge implementation) on board of their vehicles or vehicle subsystems.

The results are delivered either to a customer’s own fleet management platform through standard interfaces like REST API, or customers can use our convenient web-based platform AI-VIEW to get a quick overview of the whole fleet, to look at the trip history and to check out each individual vehicle and to deep dive into it on component level. Our deep vehicle knowledge can bridge the gap between simple navigation-based vehicle tracking and an extensive electric vehicle monitoring system.

The solution is suitable for electric fleets, mixed and conventional fleets (with internal combustion engines).

Further reading about the functioning of our CAN-bus based virtual sensors for vehicles.

Get in touch with us to discuss about your Use case!

Contact

+49 (0)6151 3844614

contact@compredict.ai

COMPREDICT GmbH

Rheinstra├če 40-42

64283 Darmstadt, Germany