Preliminary battery pack design methodology based on interval constraint satisfaction problem resolution and feasible domain analysis

Battery pack design is an interdisciplinary field in which electrical performances are closely related to safety, longevity, environmental impact, and cost considerations. Designers strive to optimize battery design variables while meeting these requirements. However, the combinatorial nature of the system makes exhaustive testing too time-consuming and therefore intractable. To reduce the problem's complexity, methodologies focus on the optimization of subsystems (thermal management, safety devices). Which requires prior specification of key design variables (number and type of cells, thermal management system). This initial choice, known as the architecture, may lead to suboptimal solutions at the battery system scale or even a system not complying with all requirements. This article presents a two-step preliminary design methodology to help designers select this initial architecture, considering the various aspects of battery design. The first step formulates the battery design problem as an analytical constraint satisfaction problem (CSP) and solves it using an interval-based solver. The second step, based on previously computed solutions, provides designers with visualization analysis tools to explore the obtained feasible domains. The novel approach enables a rapid assessment of feasible architectures while integrating a broad range of battery design aspects upstream. By reducing the risk of failing to meet system specifications, it also encourages designers to innovate by exploring new architectures. A case study is presented to illustrate this methodology, exploring the selection of the cell and cooling system.