URBANIZED D4.4: Optimised self-adaptive, multi-layer EMS design and virtual validation fleet management algorithm

Urbanized develops and demonstrates the next generation of modular vehicle architectures for urban commercial e-vehicles. Such architectures are based on optimal design principles and vehicle right-sizing for a particular mission. To help on accomplishing this goal, a multi-layer Energy Management Systems (EMSs) is to be developed for the next generation modular vehicle. Such EMSs are developed in several stages throughout the project scope, as Figure 1 shows. (Figure in Report) Figure 1: Work package structure of Urbanized in the context of EMS-related developments. The design of the multi-layer energy management system is done in several steps. First, a clear desired functionality overview, requirements specification and analysis, etc. is agreed upon partners. Such a functionality guarantees that the resulting algorithm is within feasible boundaries of what is achievable by the vehicle and realistic within the end users operation. Second, a clear interfacing (i.e., signals and outputs) is agreed between the multiple software components. This is meant to verify that all the required information (i.e., inputs) is available upon implementation, and to allow for independence at the moment of designing the actual software. Third is the actual design (i.e., software development) of the algorithm and simulations are carried out to verify the desired functionality. The fifth is testing in real-life vehicle. Steps one and two were developed in work packages (WPs) 2 and 3, respectively; while the last step is to be implemented in WP 6. The work presented in this report corresponds to the third step of the development of the multi-layer EMS (actual software development) and validation in virtual validation environment, which is carried out in work package 4, task 4.2. The goal of this task is to develop the actual algorithms (i.e., software) and test the algorithm capabilities via a virtual validation environment. The multi-layer EMS is composed of several Eco-functions, which are running on board of the vehicle or in the cloud. The four ECO-functions developed in this report as:
• Two Eco-functions are deployed at fleet layer: Eco-routing for optimising the mission profile / route of the vehicle, and Eco-charging for smart charging of the fleet at both, depot and on mission operation. The aim is to minimise the charging and operational cost at fleet level. Both of these algorithms correspond to fleet management tools. Eco-routing is developed based on a typical Vehicle Routing Problem (VRP) including the constraints of the case study and the vehicle characteristics (i.e., electric). Eco-charging is developed to provide a charging schedule that complies with the waiting times in the VRP and minimizes vehicle related metrics such as electricity cost and battery degradation.
• Two Eco-functions are deployed at vehicle layer: Eco-driving for optimising the vehicle speed for given mission profiles using traffic flow information and Eco-comfort for the vehicle thermal management of the cabin / cargo / battery. Both algorithms, use predictions of on road conditions (e.g., speed limits, weather forecast) to achieve energy savings at vehicle level.
To quantify the performance of the multi-layer EMS, a virtual validation environment is developed. Such an environment compiles together the mission profiles defined in WP2, the vehicle models developed in WP3 and the multi-layer EMS. Therefore, the virtual validation environment is then used to assess the performance of the Eco-functions, while running the vehicle prototype under realistic operational conditions. Further, an initial evaluation of EMS-related KPIs is performed. This initial evaluation is to be used as a comparison point of the real vehicle performance, which is to be tested in latter work packages.
Simulation results indicate that the four Eco-functions have a positive effect in the project KPIs related to energy consumption. This is valid not only they the Eco-functions are used individually (each Eco-function separately) but also when they are combined (all Eco-functions at the same time). Simulation results also show some interactions between multiple Eco-functions, and how they can be further optimized.