Virtual Sensor Generation via Inferential Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-end enterprise computing systems face challenges with increasing complexity and cost due to numerous physical sensors, which push against bandwidth limitations, necessitating a reduction in sensor numbers while maintaining diagnostic accuracy and system reliability.
Innovation Solution
A self-optimizing inferential-sensing technique that uses a training data set to replace physical sensors with virtual sensors, computed via cross-correlations, optimizing sensor deployment and reducing computational overhead by iteratively dropping least valuable signals and observations while ensuring accuracy criteria are met.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of physical sensors is increased to improve diagnostic accuracy, then measurement precision is improved, but device complexity and system cost increase
Solution Approach 1:
The patent creates virtual sensor signals that are computational copies of physical sensor readings. These virtual signals are generated through inferential modeling that replicates the function of physical sensors without requiring additional hardware, thereby maintaining measurement precision while avoiding increased device complexity
Solution Approach 2:
The patent replaces the mechanical/physical sensor system with a computational/inferential system. Instead of adding more physical sensors to the hardware layer, the system uses software-based inferential modeling to generate virtual sensor readings, substituting physical measurement mechanisms with computational ones
2Measurement precision
If the number of physical sensors is increased to improve diagnostic accuracy, then measurement precision is improved, but bandwidth requirements increase
Solution Approach 1:
The patent generates virtual sensor signals as computational copies that derive information from existing sensor data through inferential relationships. This approach provides additional diagnostic measurements without requiring proportional increases in bandwidth, as the virtual signals are computed rather than physically transmitted
Solution Approach 2:
The inferential modeling system serves multiple functions simultaneously: it processes existing sensor data, generates virtual sensor readings, identifies incipient failures, and optimizes observation rates. This multi-functionality allows the system to maintain comprehensive diagnostic capabilities without proportionally increasing bandwidth requirements for each individual function
3Quantity of substance
If the sampling rate of physical sensors is reduced to decrease bandwidth usage, then bandwidth requirements are reduced, but reliability deteriorates
Solution Approach 1:
The patent employs inferential modeling that continuously processes sensor data and provides feedback about system state and potential failures. The system uses feedback from multiple sensor signals to generate virtual readings that maintain reliability even when individual physical sensor sampling rates are reduced, as the inferential system compensates for reduced individual measurements through multi-signal analysis
Solution Approach 2:
The patent combines multiple physical sensor signals through inferential modeling to create virtual sensor readings. By merging information from multiple sensors, the system achieves reliable failure detection at lower individual sampling rates, as the combined inferential analysis compensates for the reduced frequency of individual measurements
4Device complexity
If virtual sensors are used to reduce the number of physical sensors, then device complexity is reduced, but computational overhead increases
Solution Approach 1:
The patent applies partial action by selectively generating virtual sensor signals only for the most critical parameters and using iterative optimization to drop less valuable signals. The system performs computational inferential modeling only when necessary to maintain accuracy criteria, avoiding excessive computation by focusing resources on the most important diagnostic functions
Solution Approach 2:
The patent implements dynamic optimization of the virtual sensor system through iterative signal and observation optimization. The system dynamically adjusts which signals are used for virtual sensor generation and optimizes observation rates based on actual system needs and accuracy requirements, allowing computational overhead to adapt to changing conditions rather than operating at fixed maximum levels
Data Source
AI summary
We disclose a system that optimizes deployment of sensors in a computer system. During operation, the system generates a training data set by gathering a set of n signals from n sensors in the computer system during operation of the computer system. Next, the system uses an inferential model to replace one or more signals in the set of n signals with corresponding virtual signals, wherein the virtual signals are computed based on cross-correlations with unreplaced remaining signals in the set of n signals. Finally, the system generates a design for an optimized version of the computer system, which includes sensors for the remaining signals, but does not include sensors for the replaced signals. During operation, the optimized version of the computer system: computes the virtual signals from the remaining signals; and uses the virtual signals and the remaining signals while performing prognostic pattern-recognition operations to detect incipient anomalies that arise during execution.