3D Storage Matrix for Time-Independent Battery Usage Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for storing battery usage data in electrified vehicles face challenges due to limited on-board non-volatile memory, inability to efficiently store time-dependent data, and lack of mechanisms for future retrieval and analysis of battery aging and customer behavior.
Innovation Solution
A method utilizing a three-dimensional storage matrix in non-volatile memory to efficiently store battery usage data by identifying and incrementing counts for charge-discharge full and half cycles, with calculations for average state of charge, depth of discharge, and current flow rate, allowing for memory-efficient storage and future retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If time-dependent battery usage data is stored in traditional memory structures, then complete battery usage history can be retained, but the amount of memory required exceeds available on-board non-volatile memory capacity
Solution Approach 1:
The patent transforms the traditional one-dimensional time-series storage into a multi-dimensional histogram structure with bins representing different state of charge ranges, current flow directions, and temperature ranges. This dimensional transformation allows aggregation of similar operating conditions, dramatically reducing memory requirements while preserving essential battery usage patterns for aging analysis.
Solution Approach 2:
The patent changes the storage parameters from continuous time-dependent values to discrete binned categories (state of charge ranges, current direction bins, temperature ranges). By discretizing continuous parameters into finite bins and counting occurrences in each bin, the system compresses large volumes of time-series data into compact histograms that fit within limited on-board memory.
2Adaptability or versatility
If on-board non-volatile memory is allocated for storing time series data, then battery usage data can be stored, but the memory allocation requires knowledge of time series length which is not practically available
Solution Approach 1:
The patent implements a dynamic histogram approach where bins are created based on predefined operational ranges (minimum and maximum state of charge, current thresholds, temperature ranges) rather than fixed time periods. This dynamic structure automatically adapts to varying data collection frequencies and time series lengths, eliminating the need for advance memory allocation planning while ensuring sufficient storage capacity.
3Measurement precision
If detailed time-dependent battery data is stored, then accurate battery aging analysis can be performed, but the data storage requirement becomes impractically large for on-board memory
Solution Approach 1:
The patent extracts only the essential features needed for battery aging analysis (state of charge ranges, current direction and magnitude, temperature ranges) and stores them in histogram bins, discarding redundant temporal details. This selective extraction preserves the critical information required for aging models while eliminating excessive data volume, achieving a balance between analysis accuracy and storage feasibility.
Data Source
AI summary
Time domain battery usage data of a battery of an electrified powertrain of a vehicle in terms of average SOC, DOD and a set including average current flow rate and average battery temperature for each charge-discharge full cycle and half cycle of the battery during a use period are used to identify a location in a 3-D storage matrix (of fixed size and predetermined discretization levels for each dimension) in memory of an electronic control unit and a count in that location incremented. In an aspect, the charge-discharge full cycles and half cycles are identified using four-point rainflow cycle counting.