Virtual Sensor Data Reconstruction via Topology-Based Candidate Selection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In complex systems with multiple sensors, reconstructing faulty sensor data reliably and efficiently is challenging due to the high computational requirements and human intervention needed in existing approaches, which complicates the detection of critical operating states and system control.

Innovation Solution

A monitoring device with a receiving and evaluation unit that identifies sensor candidates based on structural and functional relationships, calculates similarity values, and generates virtual sensor data using the data from the most similar functioning sensors, reducing computational load and reliance on human knowledge.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a statistics-based approach is used to identify relationships between sensor data, then the reliability of sensor data reconstruction is improved, but the computing power requirement and hardware complexity increase significantly

Engineering Contradiction:
Improvesensor data reconstruction reliabilityVSAvoidhardware complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the sensor set into groups based on spatial proximity and functional relationships. Instead of calculating statistical relationships for all sensor pairs (which would be O(n²) complexity), the system only evaluates sensors within localized groups, dramatically reducing computational requirements while maintaining reconstruction reliability through structurally-informed candidate selection.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary structural analysis to identify candidate sensors before fault occurrence. By pre-establishing spatial and functional relationships between sensors during normal operation, the system creates a ready-made candidate list that can be quickly consulted when faults occur, avoiding the need for complex real-time statistical calculations.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If a knowledge-based approach is used to identify related sensor data, then the reliability of sensor data reconstruction is improved, but the complexity of system analysis and human intervention increase

Engineering Contradiction:
Improvesensor data reconstruction reliabilityVSAvoidsystem analysis complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent enables the monitoring system to automatically identify candidate sensors through structural analysis of the monitored object's geometry and sensor arrangements. The system self-configures by analyzing spatial coordinates and functional relationships without requiring external expert knowledge or manual knowledge base creation, thereby maintaining high reliability through objective structural criteria while eliminating human intervention complexity.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the manual knowledge-based approach (requiring expert analysis and knowledge base creation) with an automated structural analysis mechanism. By using mathematical models of spatial relationships and functional dependencies, the system objectively identifies candidate sensors without human intervention, substituting expert judgment with algorithmic structural analysis.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If all sensor pairs are evaluated for statistical relationships, then the completeness of candidate identification is improved, but the computing time and processing power increase

Engineering Contradiction:
Improvecandidate identification completenessVSAvoidprocessing speed
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent applies local quality by evaluating statistical relationships only for sensor pairs that are spatially proximate or functionally related, rather than all possible pairs. This localized evaluation focuses computational resources on the most relevant sensor combinations, maintaining identification completeness for fault reconstruction while dramatically improving processing speed by eliminating evaluations of unrelated sensor pairs.

Inventive Principle:
Principle #3Local quality

4Measurement precision

If detailed system analysis is performed to create a knowledge base, then the accuracy of sensor candidate identification is improved, but the time and resources required for system setup increase

Engineering Contradiction:
Improvesensor candidate identification accuracyVSAvoidsystem setup time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary structural analysis during system initialization to establish spatial and functional relationships between sensors. By pre-computing candidate sensor lists based on the monitored object's geometry and sensor arrangements before operation begins, the system achieves high identification accuracy without requiring detailed manual system analysis during setup, thereby reducing initial configuration time and resources.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3631593B1Monitoring device and method for monitoring a system
Publication Date: 2021.04.14 SIEMENS AG
  • EP3631593B1 patent drawingFigure 1~3

AI summary

The invention relates to the monitoring of a technical system by means of sensor data. Upon failure of a sensor, virtual sensor data is created for the failed sensor on the basis of the remaining functional sensors. The sensors for calculation of the virtual sensor data are selected in two stages. In a first step, possible candidates of sensors are determined on the basis of a scientific approach and the topology of the system. In a second step, a mathematical relationship between the sensor data of a faulty sensor and the possible candidates of sensors is calculated for the calculation of the virtual sensor data. In this way, those sensors forming a suitable basis for the calculation of the virtual sensor data can be identified.