Our projects
Optimizing electric vehicle charging
IVADO Labs, in close collaboration with a leading company in electric vehicle fleet management and charging, is addressing the complex challenges of transportation electrification. Leveraging our expertise in artificial intelligence, we offer innovative solutions for optimizing charging plans.
Manage the charging of an electric vehicle fleet to minimize costs and ensure availability for upcoming trips.
A smart platform based on real data that creates a customized charging plan for each vehicle.
Lower costs and longer battery life thanks to precise optimization of charging schedules, with reliable plans generated in seconds.
The electrification of transportation represents a significant evolution, driven by increased environmental awareness and supported by public policy incentives. This initiative is part of a broader effort to reduce greenhouse gas emissions and promote sustainability. For businesses, the adoption of electric vehicle fleets is not only a response to increasingly stringent regulatory requirements, but also an opportunity to reduce long-term operational costs.
This massive electrification of transportation brings about logistical and technical challenges. In particular, determining when and how to charge vehicles according to available infrastructure and usage, in order to minimize costs while ensuring the right level of charge and protecting against unforeseen events. Artificial intelligence technologies can offer novel solutions to these complex issues. IVADO Labs has developed an innovative system to optimize the management of electric vehicle charging. This smart tool is designed to help large companies optimize the economic and environmental benefits of electrification.
Optimized charging
We have designed a solution to optimize electric vehicle charging plans, to help our client support businesses in their transition to electric vehicle fleets. This solution aims to minimize electricity costs for the operator and maximize the robustness of the charging plan to ensure reliability, stability, and resistance to potential disruptions or unforeseen events. Integrating eco-design and artificial intelligence, this energy management system ensures intelligent fleet management.
Our solution reduces charging costs and improves the lithium-ion battery lifespan, ideally keeping the charge between 20% and 80%. More broadly, it helps avoid electrical grid overload and supports our client in remaining innovative in a context of energy transition.
Objective: Regulate maximum power demand to optimize costs
Businesses with high electricity consumption must consider power demand, meaning the amount of electricity consumed at any given moment. The objective is to distribute vehicle charging over time to reduce the Maximum Power Demand (MPD). The electricity supplier uses this to determine pricing. By preventing power peaks, we allow our collaborator to benefit from more advantageous pricing throughout the year.
Regionally, spreading out the charging over time helps avoid unnecessary overload of the electrical infrastructure. It is well known that a massive and spontaneous adoption of electric mobility without proper planning could prove problematic, potentially resulting in power outages.
By developing this solution, IVADO Labs is helping to facilitate society’s transition to carbon-neutral mobility.
Charging reliability and simplification of operations
IVADO Labs helped the client increase the efficiency of its platform for optimizing electric vehicle charging plans, enabling it to resolve real-world instances and eliminating the need for users to enter data manually.
The work began with a fleet of electric vehicles with relatively consistent routes. This initial phase was strategically chosen due to the regularity of trips and schedules, providing an ideal framework for testing and refining our solution in a controlled setting. The efficiency demonstrated in this specific context allows us to guarantee the robustness of our technology before deploying it on vehicles with more varied charging profiles.
A powerful AI-Driven approach
Our approach comprised two algorithms: one for predicting trips and another for optimizing charging.
The prediction model leverages telematics data, capturing the vehicle’s position, odometer, and charging status every three minutes. By analyzing this historical information, we can recognize behavioral patterns in vehicle use and predict trips, including abnormal situations such as early departures or late arrivals. By combining this data with weather forecasts, we can also accurately estimate consumption for future trips.
The optimization model leverages three variables provided by the prediction model to develop a charging plan:
- Departure and arrival times of trips over the next few days
- Energy consumption and distance associated with each trip
- Initial state of charge of the battery
The generated charging plans prioritize robustness and mitigate the operational risks associated with underestimated charging needs, while minimizing the maximum power demand.
Cutting-edge technology
We use advanced analytical techniques to examine and aggregate the historical trips of each electric vehicle, identifying similar temporal patterns. This facilitates more efficient and personalized planning.
Additionally, our optimizer employs sophisticated linear programming methods, structured around a hierarchical objective function, to further refine the energy efficiency of charging.
Concrete results
Our model demonstrated its ability to efficiently and accurately predict the majority of real-life trips, and in turn, anticipate the actual needs of vehicles. Although some projections did not correspond to actual trips, this rate of false positives was anticipated and remains well within the robustness margins targeted by our system.
This technology also enables precise management of schedules and energy consumption of vehicles, optimizing their use while minimizing inefficiencies.
Increasing organizational efficiency
The solution generates a weeklong charging plan for several hundred vehicles in just seconds. This plan provides fleet managers with granular information, reoptimized every 15 minutes, allowing them to adjust the power supplied to each vehicle.
Rather than fully charging each battery as rapidly as possible, the solution generates a charging plan that indicates how many kilowatts each station should provide to each vehicle at any given moment. It takes advantage of the predicted charging uptime and target charge level for each vehicle, reducing the maximum power demand while ensuring each vehicle is ready for its next day’s route. The AI application adapts to each vehicle’s specific needs. By updating the charging plan every 15 minutes, the application monitors capacity variations and withstands technical difficulties such as faulty connections, ensuring reliable and efficient energy management.
The algorithm developed by IVADO Labs ensures a calculation speed that allows a weekly charging plan to be generated for several hundred vehicles in just seconds.
The AI software was designed to ensure vehicles are always charged ahead of schedule. This approach allows for great flexibility and reliability of service in the face of unforeseen events. For fleet managers, this strategy increases operational efficiency, as one incident can quickly create a domino effect of delays. Overall, it reduces operational uncertainty. Granular charging management also ensures that the battery is charged to a level that guarantees trip completion, plus a buffer if an unforeseen event occurs, such as a significant drop in temperature.
The solution was rigorously tested in various configurations, including a high number of vehicles, to ensure optimal model performance.
Through the application of artificial intelligence, dispatchers are supported in managing vehicle charging and benefit from increased visibility on the future state of their fleet. This allows them to focus on their core business and ensure optimal management of their transport operations.
Main features of the tool
Key results
Future perspectives
Transport companies can benefit from this tool, which provides a complete solution from prediction to optimization to regulated Maximum Power Demand (MPD).
Through this project, we are strengthening user confidence in electric vehicles, facilitating a transition from combustion vehicle fleets in the transport sector.