How to Replicate Digital Twin Technology for EV Battery Management

Author: Tomáš Slaný, Green Delta

The Digital Twin’s primary job is to monitor internal battery conditions, like how charged it is –represented by the state of charge–, cell voltage and the temperature, similar to the existing Battery Management System. But crucially, the Digital Twin can also predict the battery’s future, such as its remaining useful life, allowing maintenance to be scheduled right when it is needed.

The Digital Twin (DT) technology is very useful for managing Electric Vehicle (EV) batteries. Think of a DT as a perfect, living virtual copy of a physical battery. It uses Internet of Things (IoT) sensors to gather real-time data and combines this with advanced models (that can be based on Artificial Intelligence (AI) and Machine Learning).

This virtual copy allows us to monitor, predict, and optimize the battery’s entire life. By seeing exactly how the battery is performing and predicting when it might degrade, the DT enables better decisions about repurposing and recycling used batteries. This helps us move towards a Circular Economy by extending the battery’s useful life and improving environmental outcomes.

The Most Important Requirements for Success

Key Components for Your Digital Twin

To build this system, you need three key components working together:

• IoT Sensors: These are the eyes and ears on the physical battery, collecting data on voltage, current, and temperature in real-time. These sensors allow the state of the battery to be monitored continuously and in real-time.
• Cloud Computing Platform: This is the central hub where the DT model is stored and operated. It stores large amounts of data collected from the sensors and runs the complex analysis, using the digital replicas.
• Digital Replica Models: They simulate the battery’s behavior and use various modelling approaches, such as machine learning to learn from historical data, constantly improving their accuracy in predicting future performance.

Your Digital Twin Replication Checklist

Follow these steps to create your virtual battery replica:

• Install the Sensors: Install the IoT sensors onto the EV batteries to collect crucial real-time data (voltage, temperature, etc.). Alternatively, ways of using existing sensors collecting data for the battery management system can be explored.
• Select the Communication Protocol: Use standard communication protocols (like MQTT or CAN) to ensure the data is sent securely and quickly from the sensors to the central system.
• Establish the Cloud: Set up a cloud platform to act as the central hub for storing, organizing, and processing all the data. Consider the different computing techniques such as edge or fog computing for optimizing the data flow speed
• Standardize the Data: Create a common way (a “schema”) to format the data so that different systems can easily understand and use it.
• Select and Build the Prediction Model: Choose the right digital model for the DT. You can use simpler electrical circuit models (ECMs) for speed, more complex chemical models for detail, or smart models based on AI or ML. A hybrid model (combining physics and AI) is often a good balance.
• Ensure the Data Quality: Always ensure you have high-quality data by cleaning and filtering it before it’s used. Methods like normalization, filtering, feature extraction can be done to pre-process the data.
• Test and Verify: Rigorously test your DT’s predictions against the real-world battery performance to make sure the model is accurate and reliable.

Technology

This practice requires specific technologies that need to be integrated in one system. IoT sensors (like current and temperature sensors) are needed to constantly monitor the battery. A cloud computing platform is required for storing the massive amounts of data and running the models. Using local processing (edge computing) can help keep the system running in real-time.

Additionally, technology such as AI or ML might be needed for the digital replica models to predict battery’s future state.

Finance

The implementation of the DT technology requires a considerable investment in advanced hardware and software technologies, as well as potentially a change of manufacturing processes. However, these upfront costs are to be compensated in the long run.

The major financial appeal is that DTs can reduce costs by up to 80% and improve development efficiency. Therefore, demonstrating this cost-effectiveness, that the money you save on maintenance, design, and failure prevention outweighs the setup cost, is crucial for attracting public or private financing.

Stakeholders

The collaboration is crucial for a successful deployment of the DT. The following groups must work together:

• Battery manufacturers: They hold crucial knowledge about the battery’s design and should integrate the sensing technology in the production of the batteries.
• Repurposers: They need the data to efficiently track and trace the battery’s life, which is necessary for smart repurposing.
• Technology providers: They supply and implement the IoT sensors, cloud platforms, and predictive models.

Society

The DT technology helps society by promoting a sustainable energy future. By extending the useful life of EV batteries and improving their performance, it supports cleaner energy storage and helps reduce the carbon footprint associated with battery manufacturing and disposal. It helps establish a circular economy for energy storage technologies.

Environment

The implementation of the DT is expected to have positive impact on the environment. It makes energy storage more sustainable by optimizing performance and extending the battery’s useful lifespan. Moreover, it facilitates repurposing old EV batteries for second-life applications (like stationary energy storage) instead of discarding them. In this way, it helps reduce waste and the carbon footprint linked to battery production and disposal.

Safety

The DT’s primary role is to improve safety. It uses its predictive power to achieve the early detection of anomalies and critical failures, allowing for intervention before a dangerous event occurs. The main safety risk is if the system fails due to data transmission delays or inaccurate models, which could compromise its ability to detect these critical issues in real-time.