Electrochemical Device Optimization via Machine Learning Data Management

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current systems for managing and optimizing electrochemical devices, such as fuel cells, lack efficient data management and optimization techniques, leading to suboptimal operational reliability, efficiency, and resource utilization throughout their life cycles.

Innovation Solution

A system comprising a database for storing extensive data sets, a detection unit for acquiring data, a management unit for data storage and management, a trigger unit for optimization requests, a readout unit for data retrieval, and a computing unit using machine learning to generate optimization results, which can automatically determine recurring optimization results to improve resource efficiency and operational safety.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional data management systems are used for electrochemical devices, then device complexity is reduced, but operational reliability and optimization capability deteriorate

Engineering Contradiction:
Improveoperational reliabilityVSAvoiddata management system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The data management system is segmented into distinct functional modules: detection unit for data acquisition, management unit for data processing and storage, trigger unit for optimization requests, readout unit for data retrieval, and computing unit for machine learning operations. This modular segmentation improves operational reliability by isolating functions while managing complexity through structured organization of each module's specific responsibilities.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The management unit acts as an intermediary between the detection unit (data acquisition) and the computing unit (optimization processing). It standardizes and processes raw data before transmission to the computing unit, and coordinates with the trigger and readout units. This intermediary role enhances operational reliability by ensuring data quality and system coordination while managing overall system complexity.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Productivity

If machine learning algorithms are implemented for optimization, then resource efficiency and operational safety improve, but computational requirements and system complexity increase

Engineering Contradiction:
Improveresource efficiencyVSAvoidcomputational system complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The management unit performs preliminary data processing, cleaning, and organization before data reaches the computing unit for machine learning analysis. This preliminary action prepares data in advance, reducing the computational burden during optimization operations and enabling resource-efficient machine learning while managing system complexity through staged processing.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The computing unit automatically executes machine learning algorithms to generate optimization results without requiring manual intervention for each optimization cycle. The system self-services by continuously processing data, triggering optimizations, and implementing results, thereby improving resource efficiency while the automated nature manages computational complexity through consistent algorithmic execution.

Inventive Principle:
Principle #25Self-service

3Measurement precision

If comprehensive data sets are collected and stored, then optimization accuracy improves, but data storage requirements and processing time increase

Engineering Contradiction:
Improveoptimization accuracyVSAvoiddata processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The management unit continuously pre-processes and organizes incoming data from the detection unit, maintaining data in optimized formats and structures. This preliminary organization of comprehensive data sets enables the computing unit to perform machine learning operations more efficiently, achieving high optimization accuracy while reducing processing time through pre-prepared data readiness.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system maintains continuous data collection, processing, and optimization operations without interruption. The management unit continuously manages data flows between detection, storage, and computing units, ensuring comprehensive data sets are always available for optimization while maintaining steady-state processing that reduces overall time loss compared to batch processing approaches.

Inventive Principle:
Principle #20Continuity of useful action

Data Source

PatentEP4020644A1System and method for optimizing at least one electrochemical device
Publication Date: 2022.06.29 ROBERT BOSCH GMBH
  • EP4020644A1 patent drawingFigure 1
  • EP4020644A1 patent drawingFigure 2
  • EP4020644A1 patent drawingFigure 3

AI summary

A system (10) for optimizing at least one electrochemical device (12), in particular a fuel cell device, is proposed, comprising: a database (14) for storing data which includes at least one data record relating to the electrochemical device (12), a data acquisition unit (16) for acquiring the data, a management unit (18) for storing the data in the database (14), a trigger unit (20) for providing at least one optimization request relating to the electrochemical device (12), a readout unit (22) which, in an operating state, reads and provides data from the database (14), and a computing unit (24) which, based on the optimization request and the provided data, determines at least one optimization result by applying machine learning.