Field Device Production Verification for Series Error Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting series errors in field devices in automation technology often result in false-positive predictions, leading to unnecessary administrative efforts and delayed detection of actual issues, which increases damage and costs.
Innovation Solution
A method that involves accessing a service platform to collect and statistically evaluate data from field devices, using a machine learning or prognosis system to detect anomalies and predict series errors, thereby reducing false-positive predictions by verifying the cause of detected anomalies and providing a probability index for predicted errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If statistical evaluation of repair cases is used to detect series errors, then detection capability is improved, but false-positive predictions increase leading to unnecessary administrative effort
Solution Approach 1:
A machine learning model is introduced as an intermediary between statistical anomaly detection and final series error confirmation. The model evaluates multiple data dimensions (operating conditions, environmental factors, device configurations) to distinguish true series errors from false positives caused by operational variations, thereby reducing unnecessary verification efforts while maintaining high detection accuracy
Solution Approach 2:
The system transitions from single-parameter statistical evaluation to multi-parameter analysis by incorporating operating conditions, environmental data, and device configurations. This comprehensive parameter assessment enables the machine learning model to accurately differentiate between genuine series errors and anomalies resulting from varying operational contexts
2Reliability
If statistical evaluation is performed to identify series errors, then error detection is improved, but detection time delay increases reducing response speed
Solution Approach 1:
The machine learning model is pre-trained on historical field device data including repair cases, operating conditions, and environmental factors. This preliminary training enables the model to rapidly evaluate new data and provide immediate predictions about series errors, eliminating the time delay associated with traditional statistical evaluation while maintaining high detection accuracy
Solution Approach 2:
The manual or systematic statistical evaluation process is replaced with an automated machine learning system that processes multiple data parameters simultaneously. This substitution transforms the detection mechanism from a time-consuming sequential analysis to a rapid parallel computation, significantly reducing detection time while improving accuracy
Data Source
AI summary
The present disclosure relates to a method for verifying the production process of field devices, including a step of accessing a service platform on which data from field devices, including identification data, the respective type of field device, configuration data, containing application-specific data, environment information of the field devices or parameter data, data relating to the production date of a respective field device and repair or troubleshooting cases of the field devices are stored. The method also includes steps of detecting anomalies by statistically evaluating the repair or troubleshooting cases stored on service platform and creating a notification in the event of a detected anomaly, supplying the data of the field devices and the notifications to a machine learning or prognosis system, and evaluating the data of the field devices and the notifications by means of the machine learning or prognosis system for forecasting series errors of the field devices.