Fibrous Web Tear Detection with Wet-Dry Anomaly Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Web breaks in fiber-containing webs during the production process lead to significant downtime and operational inefficiencies in industrial plants, particularly in the drying section of paper machines, often due to increased web tension caused by drive malfunctions or malfunctions in the wet section.
Innovation Solution
A method and device that utilize sensors and encoders to monitor and analyze parameters as time series, employing artificial intelligence to detect anomalies in both the wet and dry sections of the industrial plant, allowing for proactive adjustments to prevent web breaks by adjusting drive settings and process parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If web break monitoring is performed only after breaks occur, then response time is delayed, but system complexity is minimized
Solution Approach 1:
The system performs preliminary analysis of operational parameters (speed, torque, temperature, vibration) to detect anomalies that precede web breaks. By analyzing trends and patterns before actual breaks occur, the system enables proactive intervention and prevents downtime, transforming reactive monitoring into predictive maintenance.
Solution Approach 2:
The system continuously monitors operational parameters and provides feedback when anomalies are detected. This feedback mechanism allows real-time adjustments to be made to prevent web breaks, creating a closed-loop control system that improves reliability by responding to early warning signs rather than waiting for actual breaks.
2Reliability
If comprehensive parameter monitoring is implemented in both wet and dry sections, then web break prevention capability is improved, but device complexity increases
Solution Approach 1:
The monitoring system is divided into distinct segments corresponding to the wet section and dry section of the paper machine. Each segment monitors specific parameters relevant to its operational characteristics, allowing comprehensive coverage while maintaining manageable system complexity through modular architecture.
Solution Approach 2:
A unified anomaly detection algorithm is applied across multiple sections and parameter types. The same analytical framework processes diverse inputs (speed, torque, temperature, vibration) from both wet and dry sections, reducing overall system complexity by using a universal detection mechanism rather than section-specific algorithms.
3Measurement precision
If real-time parameter analysis is performed, then web break detection accuracy is improved, but computational resource consumption increases
Solution Approach 1:
The system applies full analytical rigor only when anomalies are detected, rather than continuously analyzing all parameters at maximum depth. During normal operation, monitoring is performed at a baseline level, and computational resources are intensified only when abnormal patterns emerge, balancing accuracy with energy consumption.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for monitoring an industrial plant (100), to a computer program product, and to an industrial plant (100). The industrial plant (100) comprises two parts (100a, 100b). In a first part (100a) of the industrial plant (100), in particular a paper machine, first parameters (x1,...,xn) are provided. The industrial plant (100), particularly designed as a paper machine, is used to produce and/or process a fibrous material web (1). Second parameters (y1,...,yn) are provided in a second part (100b). The parameters (x1,...,xn,y1,...,yn) are stored, preferentially as time series (16a,...,16n,17a,...,17n). In the case of a web break in the second part (100b), the second parameters (y1,...,yn) are first analyzed for a second anomaly. If no second anomaly can be detected, the first parameters (x1,...,xn) are analyzed for a first anomaly. During the analysis, the parameters (x1,...xn, y1,...,yn) which were stored in a time range before the web break (F) are preferably examined. If a first or second anomaly is detected, these, and optionally measures to avoid such web breaks (F), are displayed to the user. Optionally, the first parameters x1,...,xn and/or the second parameters y1,...,yn can be set so as to avoid future web breaks F.