Recursive Variance-Based Steady State Detection for Building Equipment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current building management systems (BMS) face challenges in reliably detecting steady states for equipment operations involving multiple variables, which is essential for accurate performance analysis and fault detection, especially in low-cost and minimally supervised industries like HVAC, where existing methods are not robust and require tuning of multiple parameters.
Innovation Solution
A BMS with a predictive diagnostics system that includes a steady state detector capable of recursively updating mean and variance, identifying steady or transient states based on combined slope and second derivative of variance, and adjusting equipment operations accordingly, using models constructed from steady state data to predict performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If typical steady state detection methods are applied to multiple variables, then detection capability is improved, but system complexity and parameter tuning requirements increase significantly
Solution Approach 1:
The patent combines multiple variable detections into a unified steady state detection framework. Instead of treating each variable independently with separate parameter sets, the system integrates them through a common variance-based metric that simultaneously evaluates multiple variables, reducing overall system complexity while maintaining detection reliability.
Solution Approach 2:
The invention creates a universal steady state detection method that works across different variables and equipment types without requiring variable-specific parameter tuning. The variance-based approach with combined slope and second derivative criteria provides a multi-functional solution applicable to HVAC equipment, industrial processes, and other systems, eliminating the need for separate detection mechanisms for each variable.
2Measurement precision
If typical steady state detection methods are used, then detection accuracy is improved, but the cost and supervision requirements become unaffordable for low-cost industries
Solution Approach 1:
The system performs self-adjustment through recursive updating of mean and variance calculations. The detection algorithm automatically adapts to changing operating conditions by continuously recalculating statistical parameters from incoming data, eliminating the need for external supervision or manual parameter adjustment while maintaining high detection accuracy.
Solution Approach 2:
The invention dynamically changes detection parameters through recursive updates of mean and variance. Instead of using fixed thresholds, the system adapts parameters based on real-time data characteristics, allowing accurate detection across varying operating conditions without requiring expensive calibration equipment or expert supervision.
3Reliability
If steady state detection is implemented for multiple variables independently, then detection coverage is improved, but false detection and prediction errors increase
Solution Approach 1:
The patent merges multiple variable detections into a unified decision framework. By combining the slopes and variances of multiple variables into a single steady state assessment, the system achieves comprehensive detection coverage while reducing false positives that would occur when variables are evaluated independently. The combined approach ensures all variables agree on the steady state condition before triggering detection.
Data Source
AI summary
A building management system includes connected equipment and a predictive diagnostics system. The connected equipment is configured to measure a plurality of monitored variables. The predictive diagnostics system includes a communications interface, a steady state detector, a controller. The communications interface is configured to receive samples of the monitored variables from the connected equipment. The steady state detector is configured to recursively update a mean and a variance of the samples each time a new sample is received, identify whether each of the samples reflects a steady state or a transient state of operation of the connected equipment using the mean and the variance, and associate each of the samples to the steady state or the transient state as identified. The controller is configured to adjust an operation of the connected equipment based on the steady state or the transient state as identified.