Edge Pump Anomaly Detection for Predictive Maintenance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional pump systems, particularly in wastewater treatment, suffer from blockages and component damage due to solids, leading to costly and time-consuming system failures, with existing SCADA and cloud-based systems lacking the accuracy and scalability to effectively detect anomalies.
Innovation Solution
An asset health system utilizing edge computing and machine-learned models processes high-granularity pump data with contextual information to detect anomalies, validate them, and provide a health index, enabling proactive maintenance and preventing system failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional SCADA and cloud-based systems are used for pump monitoring, then system complexity is reduced, but anomaly detection precision and reliability are insufficient
Solution Approach 1:
The system segments anomaly detection into multiple layers: edge computing devices perform local real-time analysis of pump data, while cloud-based systems handle broader pattern recognition and model training. This segmentation allows high-precision detection at the edge without requiring complete system redesign, resolving the contradiction between detection precision and system complexity.
Solution Approach 2:
Edge computing devices serve as intermediaries between conventional SCADA systems and cloud-based analytics platforms. These intermediaries process and filter pump data locally, extracting anomalies before transmission to the cloud, thereby enhancing detection precision while maintaining compatibility with existing simple SCADA infrastructures.
2Measurement precision
If high-granularity pump data is processed continuously, then anomaly detection accuracy improves, but data processing time and computational resources increase
Solution Approach 1:
The system performs preliminary actions by pre-processing pump data at the edge computing level, extracting relevant features and filtering out normal operational variations before data reaches central processing systems. This preliminary filtering reduces the volume of data requiring intensive processing while maintaining high detection accuracy, thus reducing overall processing time.
Solution Approach 2:
Different processing granularities are applied locally: edge computing devices handle high-granularity real-time data for immediate anomaly detection, while cloud systems process aggregated lower-granularity data for trend analysis. This local differentiation of processing quality allows accurate anomaly detection without requiring all systems to process full-resolution data continuously.
3Reliability
If real-time anomaly detection is implemented, then system reliability improves, but implementation cost and infrastructure requirements increase
Solution Approach 1:
The system implements self-service capabilities through automated anomaly detection and alerting at the edge, where local computing devices independently analyze pump data and generate maintenance alerts without requiring constant human intervention or complex centralized infrastructure. This autonomy improves reliability while minimizing infrastructure complexity.
Solution Approach 2:
Maintenance activities are performed in advance based on predicted anomalies detected by the system. By identifying potential failures before they occur, the system improves reliability through proactive maintenance scheduling, reducing the need for complex emergency response infrastructure and specialized rapid-response systems.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An asset health system includes an edge-based computing device that receives high-granularity asset data for an asset from an asset data source or asset database and receives contextual data from a contextual database. The edge-based computing device processes the asset data and the contextual data with a machine-learned model. The machine-learned model includes processor-executable instructions for extracting run cycle data from the asset data to determine a run cycle of the asset, creating a feature vector based on the asset data that characterizes the run cycle, comparing the feature vector to a nominal feature vector to detect one or more anomalies, validating the one or more detected anomalies with the contextual data, determining a health index for the asset based on the one more anomalies, determining a mitigating action for mitigating the anomalies, and notifying a system operator regarding the one or more detected anomalies, health index, and mitigating actions such as a maintenance schedule.