Anomaly Detection in Multivariate Asset Data via PCA Transformation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing asset data platforms face challenges in accurately detecting anomalies in multivariate data from assets, leading to unnecessary alerts and inaccurate predictions due to issues like malfunctioning sensors and correlated variables, which complicate anomaly detection.
Innovation Solution
The asset data platform employs a process involving data transformation using Principal Component Analysis (PCA) to reduce dimensions, standardization, comparison with threshold values defined from training data, and inverse transformation to identify anomalies, thereby enhancing anomaly detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional anomaly detection methods are used on multivariate data, then the detection process is simple, but the accuracy is low leading to unnecessary alerts and inaccurate predictions
Solution Approach 1:
The patent transforms the multivariate data from the original coordinate space to a new coordinate space using Principal Component Analysis (PCA). This dimensionality transformation converts correlated variables into uncorrelated principal components, creating a new dimensional representation that separates signal from noise and enables more accurate anomaly detection while managing data complexity.
Solution Approach 2:
The patent changes the parameters of the data by standardizing the transformed data and applying modifications in the new coordinate space. By transforming the data representation and applying statistical modifications, the system improves detection accuracy without being constrained by the original complex multivariate structure.
2Reliability
If multivariate data with correlated variables is analyzed directly, then all variables are considered, but anomaly detection becomes difficult due to variable correlations
Solution Approach 1:
The patent addresses variable correlations by transforming the data into a new coordinate space where principal components are uncorrelated. This dimensional transformation eliminates the interference of correlated variables, making anomaly detection more reliable and less difficult while still considering all original variables through their contribution to the principal components.
3Measurement precision
If data from malfunctioning sensors is processed, then all sensor data is utilized, but false anomalies are detected leading to unnecessary alerts
Solution Approach 1:
The patent modifies the standardized transformed data by comparing it with threshold values and applying modifications to reduce the impact of malfunctioning sensors. This parameter modification in the transformed space allows the system to process all sensor data while filtering out false anomalies, maintaining data volume utilization while improving measurement precision.
Data Source
AI summary
Disclosed herein are systems, devices, and methods for detecting anomalies in multivariate data received from an asset-related data source, such as signal data and/or other data from an asset. According to an example, a platform may receive multivariate data from an asset in an original coordinate space and transform the data in the original coordinate space to a transformed coordinate space having a relatively fewer number of dimensions. Additionally, the platform may standardize the data in the transformed coordinate space and modify the standardized data based on a comparison between the standardized data and a set of threshold values previously defined via training data reflective of normal asset operation. Thereafter, the platform may inversely transform the modified data back to the original coordinate space and perform an analysis to detect anomalies.