Sensor Data Mapper for Power Plant Onboarding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The on-boarding of new power generation sites is complicated due to variations in sensor data tagging standards among manufacturers, leading to difficulties in identifying sensor data types and sources, especially when data is unlabeled or tagged in unfamiliar languages.
Innovation Solution
A sensor data mapper system that employs both rule-based classifiers and machine-learning models to normalize sensor data by assigning standardized data labels, regardless of manufacturer or language, by processing tagged and untagged data sets to provide consistent and understandable labels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If traditional manual methods are used to identify and standardize sensor data tags from different manufacturers, then data accuracy and understanding can be maintained, but the on-boarding time and complexity increase significantly
Solution Approach 1:
The patent introduces an intermediary system (the classification system with ML models and rule-based classifiers) that mediates between diverse sensor data tags from different manufacturers and the standardized internal data structure. This intermediary automatically translates and normalizes tags without requiring manual intervention, thereby reducing on-boarding time while managing complexity through automation.
Solution Approach 2:
The system enables self-service by allowing sensor data to be automatically classified and tagged without human intervention. The ML models and rule-based classifiers autonomously process incoming sensor data, identify patterns, and assign appropriate standardized tags, eliminating the need for manual data standardization efforts.
2Ease of operation
If sensor data from multiple manufacturers with different tagging standards is processed manually, then data understanding can be maintained, but the complexity and time required for site on-boarding increase
Solution Approach 1:
The system applies parameter changes by transforming diverse sensor data tags into a standardized format. The ML models learn the mapping between various manufacturer-specific tag parameters and the standardized internal parameters, automatically adjusting and normalizing the data representation to improve ease of operation while reducing on-boarding time.
3Measurement precision
If a comprehensive manual review process is implemented to understand all sensor data tags, then data accuracy is improved, but the on-boarding complexity and resource requirements increase
Solution Approach 1:
The classification system is segmented into multiple specialized components: ML models for learning patterns from training data, rule-based classifiers for applying explicit classification rules, and a hybrid approach that combines both methods. This segmentation allows each component to focus on specific aspects of classification, improving overall accuracy while managing complexity through modular design.
Data Source
AI summary
In some examples, sensor data comprising a first set and second set of sensor data can be received. The first set of sensor data can be provided to a rule-based classifier to identify a first normalized data tag. The second set of sensor data can be provided to a trained classifier to identify a second normalized data tag. The trained classifier can include a machine-learning model that has been trained based on tag sensor training data for classifying the second set of sensor data into a respective class associated with the second normalized data tag. The first and second sets of sensor data can be updated with respective first and second normalized data tags to normalize the first and second sets of sensor data to provide a standardized data label for each of the first and second sets of sensor data.