Distributed Rule-Based Fault Detection for Large Building Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing building management systems rely on individual automated fault detection diagnostics, which are limited in monitoring and detecting faults across multiple systems and equipment within large environments, such as buildings and campuses, and are often costly and not fault-tolerant when dealing with large data sets.
Innovation Solution
A distributed rule-based automated fault detection system that utilizes a data extractor engine, AFD engine, and fault generation engine executed on a distributed execution platform, using defined rules and user inputs to monitor and detect faults across various systems and equipment, executing on commodity hardware and managing large data sets from multiple locations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If individual automated fault detection diagnostics are used for each equipment type, then fault detection can be performed for specific equipment, but the system cannot effectively monitor and detect faults across multiple systems and equipment within large environments
Solution Approach 1:
The system segments fault detection into hierarchical levels: individual equipment-level diagnostics, system-level aggregation, and facility-wide coordination. Each level processes data independently but contributes to the overall fault detection capability, enabling comprehensive coverage without requiring a monolithic complex system
Solution Approach 2:
The automated fault detection system is designed with universal components that can handle multiple equipment types and systems. The same diagnostic engine and analysis methods are applied across HVAC, electrical, plumbing, and other building systems, allowing a single system to perform diverse fault detection functions
2Reliability
If traditional building management systems are used to handle large data sets from multiple locations, then comprehensive monitoring is achieved, but the systems are costly and not fault-tolerant
Solution Approach 1:
The system implements fault-tolerance through redundant data collection pathways and distributed processing nodes that can compensate for individual component failures. Before critical failures occur, the system maintains backup diagnostic capabilities and can switch to alternative processing routes, ensuring continuous operation without requiring expensive fault-tolerant hardware
Solution Approach 2:
The system uses software-based virtual copies of diagnostic engines and data processing logic that can be replicated across multiple nodes. Instead of relying on expensive hardware redundancy, virtual instances of the fault detection system are deployed across standard computing platforms, providing fault-tolerance at lower cost
3Quantity of substance
If individual automated fault detection diagnostics are used, then specific equipment faults can be detected, but the system lacks the capability to handle large data sets from multiple locations
Solution Approach 1:
The system adds a temporal dimension to data processing by implementing continuous data streaming and real-time analysis capabilities. Instead of batch processing large data sets periodically, the system processes data continuously as it arrives from multiple locations, transforming the productivity challenge into a manageable stream-based operation
Solution Approach 2:
The system introduces intermediary data aggregation layers that collect, pre-process, and filter data from multiple equipment sources before passing it to the central analysis engine. These intermediaries reduce the raw data volume and prepare standardized inputs, enabling efficient processing of large data sets without overwhelming the fault detection system
Data Source
AI summary
Devices, methods, and systems for distributed rule based automated fault detection are described herein. One system includes a data extractor engine configured to: extract configuration data relating to an environment based on a number of defined rules, and receive monitored data relating to the environment, an AFD engine configured to evaluate the monitored data in view of the configuration data to determine a state of the environment, and a fault generation engine to determine whether the state of the environment is outside a range defined by the number of defined rules.