How data analytics and AI are critical to EV charging anxiety

Electric vehicle adoption is on the rise, but there is mounting apprehension regarding malfunctioning chargers, which poses challenges for both consumers and infrastructure operators.

The global transition towards electric vehicles (EVs) has seen remarkable growth in recent years, with governments worldwide incentivising EV purchases and charging infrastructure. However, the rapid expansion has been marred by a lack of charging infrastructure, and frustratingly unreliable charging networks. Ford Motor Company CEO Jim Farley aptly captured the sentiment when he remarked “We’re going into the mass consumers who have a lot of charging anxiety. They don’t have range anxiety; they have charging anxiety.”

Cost of unreliability
The frustration of encountering malfunctioning or out-of-service chargers is a common experience for many EV owners. Shockingly, more than one in five charging attempts fail, with a staggering 72 per cent of these failures attributed to charger issues. This not only results in inconvenience for users but also represents a significant financial burden. It's estimated that $20 billion of the $100B global investment in EV charging infrastructure is currently wasted due to non-functioning chargers.
To address this issue, significant investments are being made to repair or replace existing EV charging ports. In early January, United States departments of transportation and energy awarded nearly $150 million for projects to repair or replace nearly 4,500 existing EV charging ports, highlighting the urgency of the matter. With each repair/replacement averaging $33,000 per EV charger, the financial impact extends beyond just repair costs, encompassing lost revenue from idle charging stations, and dissatisfied customers.

Navigating complexities
The reliability of EV chargers is influenced by a myriad of factors, ranging from faulty installations to substandard maintenance practices. A key challenge lies in ensuring seamless communication between vehicles, smartphones, chargers, and cloud-based management platforms. Any disruption in this intricate network can lead to charging failures, frustrating users, and undermining confidence in EV technology.
For example, connectivity issues, often attributed to multiple stakeholders including charging network operators, equipment vendors, cellular network providers, and utilities, pose a significant challenge. Addressing these challenges necessitates a multi-faceted approach, involving timely diagnosis, proactive problem prevention, and resilient solutions.

Solutions through data analytics, AI
Data analytics and artificial intelligence (AI) have emerged as indispensable tools in addressing the reliability of EV chargers. Proactive maintenance strategies, powered by data analytics and AI, enable operators to predict and prevent charger failures before they occur. By analysing historical data and real-time sensor readings, machine learning algorithms can identify patterns and anomalies, facilitating targeted maintenance interventions.
Some of the applications of data analytics and AI in charger reliability are:

Establishing Causality: Unlike correlation-based techniques, Causal AI techniques enable the identification of system-level failures based on component-level issues. By modelling complex relationships between various factors, neural networks provide determinative approach that may elude human observers, facilitating automatic anomaly detection and root-cause analysis of the underlying cause of an event, and its precise relationship to the outcome.

Tailored Predictive Maintenance: Not all chargers are the same, and neither are their maintenance needs. As more drivers utilise a set of stations, and their frequency of use increases, the risk of potential damage also rises. AI algorithms analyse usage patterns, environmental conditions, and historical performance data to create customised maintenance schedules for each charger. This approach minimises unnecessary expenditures by scheduling maintenance only when needed, and ensures that chargers receive timely attention.

Anomaly Detection: The condition of chargers is significantly impacted by their location and prevailing weather patterns. Moisture from humidity, rain, severe cold, and high temperatures can be particularly detrimental. Inaccuracies within the data set or unforeseen issues can compromise predictive maintenance approaches, leading to incorrect recommendations and predictions. By leveraging historical data, AI models can automatically identify abnormal meter readings or records, and alert technicians in real-time for verification and corrective action.

Inventory Management: Effective inventory management is one of the most crucial elements of any maintenance strategy. AI-driven inventory management optimises the supply chain, predicting when spare parts will be needed and managing inventory accordingly. This minimises carrying costs while ensuring that necessary parts are always available.

Conclusion

As the electrification of transportation gathers momentum, addressing the challenge of unreliable EV chargers is imperative for fostering widespread EV adoption. By harnessing the power of data analytics and AI, stakeholders can proactively tackle charger reliability issues, enhance user experience, and accelerate the transition to a sustainable mobility ecosystem. Through investments in predictive maintenance and infrastructure optimisation, we can ensure the long-term viability and success of electric mobility.

(Mohanakrishnan P is Chief Growth Officer, and Akshay Sasikumar is CEO, 82Volt Technologies.)

Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH.