Industrial Machine State Determination with Statistical Estimation Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing state determination methods for industrial machines face challenges in accurately diagnosing anomalies when there are variations in equipment or materials, leading to poor work efficiency, higher costs, and reduced versatility, as they require multiple devices or models and adjustments for different conditions.
Innovation Solution
A state determination device that uses a learning model to estimate abnormal degrees based on time-series data, calculates statistical quantities before and after a state transition, and applies a correction function to determine the abnormal degree, allowing for accurate anomaly detection without needing multiple devices or models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple learning models or state determination devices are prepared for different equipment and materials, then determination accuracy is improved, but device complexity and cost increase
Solution Approach 1:
The patent creates a single learning model that can universally handle multiple types of incidental equipment (molds, temperature controllers, dryers) and production materials (resin types) by training the model on diverse data covering various equipment-material combinations. This universal model eliminates the need for multiple specialized models while maintaining determination accuracy across different scenarios.
Solution Approach 2:
The patent incorporates operation conditions (parameters such as injection speed, injection pressure, and program settings) as input features to the learning model. By dynamically adjusting these parameters based on the current equipment and material configuration, the single model can adapt to different scenarios without requiring retraining or multiple models, thus resolving the contradiction between accuracy and complexity.
2Measurement precision
If determination reference or method is changed when equipment or materials are replaced, then determination accuracy is maintained, but work efficiency decreases and cost increases
Solution Approach 1:
The learning model automatically adapts to equipment and material changes by using the operation conditions and sensor data as inputs. The system performs self-adjustment through the model's inherent ability to process varying input patterns, eliminating the need for manual intervention to change determination references or methods when equipment or materials are replaced, thus maintaining both accuracy and efficiency.
Solution Approach 2:
The system continuously monitors operation conditions and sensor data, providing feedback to the learning model which adjusts its predictions accordingly. This feedback mechanism enables the system to automatically adapt to changes in equipment or materials without manual intervention, maintaining determination accuracy while avoiding the inefficiency of manual reference changes.
3Measurement precision
If correction coefficients for different machine models and materials are prepared in advance, then determination accuracy is improved, but adjustment work and complexity increase
Solution Approach 1:
The learning model performs automatic correction by learning the relationships between operation conditions, sensor data, and actual states from training data. Instead of requiring manual preparation and adjustment of correction coefficients for different machine models and materials, the model self-corrects its predictions based on the input data patterns, eliminating adjustment work while maintaining accuracy.
Solution Approach 2:
The patent incorporates machine model and material type as input features to the learning model. The model dynamically adjusts its internal parameters and predictions based on these inputs, automatically handling corrections for different machine models and materials without requiring external correction coefficients or manual adjustment work.
Data Source
AI summary
A state determination device includes: a data acquirer configured to acquire data related to an industrial machine; an estimator configured to perform estimation using a learning model based on the acquired data; and a statistical data calculator configured to calculate a statistical quantity in accordance with a predetermined statistical condition and uses the calculated statistical quantity to calculate a statistical estimation value corrected from the estimation value estimated by the estimator and, accordingly, can adapt a state determination result calculated by the learning model to a change in the operation status or the like of the industrial machine.