A typical eMobility depot such as an eBus depot can have hundreds of vehicles all entering and leaving the depot at different times. These vehicles can also have differing energy requirements due to different mission requirements. In addition, the electricity for the depot may have a different price according to the time of day it is used, the depot may have power constraints and different types of chargers. In the midst of all this complexity there is still an overriding need that the fleet of vehicles is charged at the right time, with the right energy at the optimal power for the lowest costs. To solve this problem, we need to harness the power of AI and IOT.
First Connect the chargers in the Depot to the cloud
To manage the chargers in the depot it is first necessary to connect them into a monitoring and control system. The mist efficient manner to do this is to connect the chargers to a cloud system e.g. www.siemens.com/evdepot-digital. With such a cloud system, it is now possible to monitor the chargers and to control them, the first step to optimizing the charging in the depot. Ideally, the connectivity is through an open protocol such as the Open Charge Point Protocol (OCPP) to ensure interoperability and the ability to operate different chargers within the depot if needed.
Collect all the necessary information
To optimize the charging of an EV fleet in a depot we need more information than just the charger information. The vehicle characteristics are required, in particular the battery size and the vehicle charging power. Next is the fleet schedule, what time is a vehicle expected to arrive and depart from the depot, how much energy will each vehicle needs, what are the expected state of charge of the vehicles at mission start and end. Then comes the details of the electrical infrastructure in the depot, are there any power constraints, what is the level of grid connection. Finally, the specifics of the electricity costs need to be addressed. Typically, the electricity will be governed by a Time of Use (TOU) tariff, which means that electricity will be expensive at certain times of day (typically around tea time) and cheaper time periods (typically in the middle of the night). Also, power may be more expensive at certain time periods and therefore incur demand charges. This information has to be collected by the cloud management system either through user input or automatically through machine to machine communication from other IT systems.
Choose an objective for the optimisation
An optimization problem needs three things. The first is a number of variables which “describe” the problem. In our depot this is the information that we have collected as described in the previous section. The second requirement is that we need an objective for our optimization, what is it we are trying to achieve through the optimization. For our depot this typically is formulated in either minimizing the power requirements of the depot or reducing the electricity costs for the depot. In a typical depot vehicles return to base in “waves” and as they return they are plugged in. If there is no control in place, then the charging starts as soon as the vehicles are plugged in and this results in large peaks in power. These peaks in power create two issues for the depot, firstly power is directly related to capital costs for the depot in terms of grid connection, transformers, switch gear etc. More power equals more cost, less power less cost. Secondly the power peaks may be at times of the day when electricity is expensive. So for the case where we wish to reduce the capital costs then we choose an objective function to minimize the maximum power. Where we wish to reduce the electricity costs then we choose an objective function to minimize the total electricity costs.
Apply AI optimization techniques to find the optimal charging solution
The third thing an optimization problem needs is an optimization agent or engine. Typically, the problem can’t be solved by a human directly but rather needs some clever mathematics and computing power to solve the optimization problem. There are many types of AI algorithm and these techniques are used for a bewildering array of applications. Some interesting information on AI in general can be seen here https://new.siemens.com/global/en/company/stories/home/artificial-intelligence.html. In our application we use these techniques to find the optimal charging profile for each charger in the depot. These optimal profiles when combined deliver the lowest power or energy costs for the depot as a whole. Some examples of the output of such an optimization is seen below.
Enact the optimized schedule.
After the AI optimization has produced an optimized schedule for each charger in the depot this has to be enacted in the real world. Here the monitoring and control capabilities of the cloud management system can be leveraged. Once the individual schedules are available the management system can distribute them through the cloud connectivity to each charger. Once this takes place the chargers will implement the schedules.
Continuous Monitoring and Dynamic Optimization
As every fleet operator and depot manager will know things never go according to plan. That’s why its essential not to follow a “fire and forget” strategy when optimizing charging and energy management within a depot. It is pretty much guaranteed that there will always be changes within a depot, Vehicles will be late returning to base, they return to base with less energy than expected, they may be parked in the wrong location, a charger may be unavailable. To keep the depot operations on track and to keep the charging management optimal it is essential for the charging operations to be monitored and compared with the base plan. Any deviations between planned and actual can then be highlighted to the fleet operator. More importantly though is to continuously manage the charging, to dynamically re-optimsie the charging schedules to account for late vehicles etc. This is done using the same underpinning AI technology as the initial optimization. This ensures that each vehicle is charged at the right time with the optimal energy at the lowest cost.
Want to know more?
Siemens has a range of digital products to support fleets and depots. If you want to know more about charging management for fleets and depots then please come and talk to us. The easiest way to do this is to visit our website www.siemens.com/evdepot-digital. Here you can contact us to ask for more info or to book a slot for a demo or to talk to our experts.