Binary Signal Unsteadiness Detection Using Expectation Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting unsteady states in factory automation systems using multivalued sensor data are ineffective when binary digital signals are employed, as they fail to accurately identify operation states in facilities utilizing binary digital signals.
Innovation Solution
An unsteadiness detection device is developed, comprising a model generation unit to create a normal model from binary digital operation data, an expectation value calculation unit to calculate expected operation data, and an unsteadiness detection unit to compare measured data with expected values to determine if an operation state is unsteady.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a monitoring method using deviation matrix with multivalued sensor data is applied, then the system can detect operation states in systems with multivalued signals, but it cannot effectively detect operation states in facilities utilizing binary digital signals
Solution Approach 1:
The patent transforms the detection approach by changing the parameter representation from direct binary values to probability-based expectation values. The model generation unit creates transition probability matrices that capture the statistical behavior of binary signals, and the detection unit compares actual binary signals against these probabilistic expectations, enabling effective detection for binary digital signals while maintaining adaptability to different signal types
Solution Approach 2:
The patent introduces an intermediary layer (the normal model and expectation value calculation) between the raw binary sensor data and the unsteadiness detection process. This intermediary transforms discrete binary signals into continuous probability distributions, allowing the system to apply sophisticated detection methods originally designed for multivalued signals to binary signal environments
2Measurement precision
If comprehensive cause identification operations or custom program development is performed to identify unsteady state causes, then detection accuracy improves, but the time and man-hours required increase significantly
Solution Approach 1:
The patent performs preliminary action by automatically generating the normal model and transition probability matrices during the setup phase, capturing the steady-state behavior patterns of the facility. This pre-computed knowledge base enables rapid cause identification during unsteady states without requiring time-consuming manual analysis or custom program development, as the detection unit can directly compare current signals against the pre-established model
Solution Approach 2:
The system performs self-service by automatically learning and storing the normal operation patterns of the facility through the model generation unit, which processes historical sensor data to create the transition probability matrices. This automated knowledge acquisition eliminates the need for manual programming or expert configuration, reducing both setup time and ongoing maintenance efforts while maintaining high detection accuracy
Data Source
AI summary
An unsteadiness detection device (30) is provided which is capable of detecting the operation state of facilities using binary digital signals, the unsteadiness detection device including: a model generation unit (313) to generate a normal model for determining operation states of a plurality of facilities (11) on the basis of operation data which are binary digital signals obtained from the facilities (11) in their steady operation states; an expectation value calculation unit (315) to calculate an expectation value of operation data by applying the normal model to past operation data of the facilities (11); and an unsteadiness detection unit (316) to detect whether or not an operation state of one of the facilities (11) is unsteady by comparing the expectation value of the operation data and a measured value of the operation data.