Dynamic Reconfigurable Battery Pack balanced discharge management using LLM guidance. Use when: (1) Planning battery cell connections for dynamic reconfigurable battery systems (DRBP), (2) Optimizing battery discharge to achieve balanced SOC across all cells, (3) Adjusting battery topology based on navigation information and battery state, (4) Implementing LLM-guided battery management strategies. Trigger keywords: "dynamic reconfigurable battery pack", "DRBP", "balanced discharge", "battery connection planning", "cell SOC均衡", "电池拓扑重构", "导航指导电池连接".
This skill enables intelligent management of Dynamic Reconfigurable Battery Packs (DRBP) using Large Language Model (LLM) guidance. The core objective is to achieve balanced discharge across all battery cells, ensuring each cell's State of Charge (SOC) approaches zero simultaneously at the end of discharge cycles, while extending battery pack lifetime.
Key architecture:
Primary goal: When battery pack discharge completes, every cell's SOC must be close to 0 (balanced discharge).
graph TD
A[Input: Navigation Info + Battery State] --> B[Calculate Power Requirements]
B --> C[LLM Strategy Selection]
C --> D[Select Cells with Bucket Effect Protection]
D --> E[Determine Series/Parallel Topology]
E --> F[Output]
Extract battery state summary parameters from user prompt
Extract navigation parameters from user prompt
references/vehicle_model.mdi_req,the discharge current of each cell must not exceed the continuous discharge current of 37.7A),then calculate v_req based on the current and powerOutput: v_req (V) and i_req (A) (The discharge current of each cell must not exceed the continuous discharge current of 37.7A), based on:
Call LLM API with a structured prompt containing:
references/strategies.md)Expected LLM output format:
{
"strategy": "",
"reason": "", //reason for the choice
"parameters": {
"cells_per_module": 4,
"priority_weights": {"soc": 0.9, "soh": 0.1}
}
}
Available strategies:
Execute scripts/cell_selector.py with LLM strategy parameters:
python scripts/cell_selector.py --battery input_file.json --strategy equilibrium --cells_per_module 4'
Bucket effect protection: The algorithm ensures the weakest selected cell (lowest SOC, highest resistance) can safely discharge for 10 minutes without dropping below minimum SOC threshold (default: 0.05). Critical Rule: When none of the options can be satisfied, directly return status as False and end the task.
In this step, there are no tools; the decision is directly handed over to the large model. Constraints:
Algorithm:
n_series = ceil(v_req / (20 * cell_nominal_voltage))n_parallel = ceil(i_req / (cell_max_current * efficiency))n_series × n_parallel ≤ selected_cells_per_moduleOnce the final result is generated, you can finish; just return the result directly.Do not write it into the file. Final output format:
{
"status": boolean, //Whether it meets the task requirements
"v_req": float, //Required voltage, calculated based on current and power
"i_req": float, //Required current, first match the current. The discharge current of each cell must not exceed the continuous discharge current of 37.7 A.
"selected_cells": [
{
"mod_id": int, //module id
"cells": [[...], [...]] // A two-dimensional list, where each sublist represents a parallel branch. For example, this is n in series and 2 in parallel
},
// ... 19 more modules
],
"reason": "" // reason for the choice
}
references/strategies.mdreferences/vehicle_model.mdreferences/safety_constraints.mdRemember the primary objective: When battery pack discharge ends, every cell's SOC should be close to 0. All decisions must prioritize this balanced discharge goal.