Machine Monitoring Warning Lines for Stable Sensor Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for establishing warning lines in Fault Detection and Classification (FDC) charts are ineffective when data is stable or lacks variation, making it difficult to accurately monitor machine parameters.
Innovation Solution
A method involving sensors to detect first data, calculate minimum change amounts, establish initial and final warning lines using shift values, and adjust ranges to account for data stability, thereby improving accuracy in monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If statistical methods (normal distribution curves and standard deviations) are used to establish warning lines, then the method works well for varying data, but it becomes ineffective when data is stable or has no variation
Solution Approach 1:
The patent changes the fundamental parameters used for warning line establishment from statistical distributions (mean, standard deviation) to actual observed data ranges (maximum and minimum values). This parameter transformation allows the method to work effectively for both varying and stable data by directly using the extreme values observed in the data itself rather than assuming a statistical distribution.
Solution Approach 2:
Instead of using statistical theory to predict data behavior (top-down approach), the patent inverts the approach by using actual observed data to define the warning lines (bottom-up approach). The warning lines are established based on the maximum and minimum values actually observed in the data, rather than deriving them from statistical assumptions about data distribution.
2Measurement precision
If traditional statistical warning lines are used, then the calculation is simple for varying data, but the accuracy deteriorates when data is stable
Solution Approach 1:
The patent transforms the calculation parameters from statistical metrics (mean ± standard deviation) to direct observational metrics (maximum and minimum values). This parameter change improves measurement precision for stable data while actually simplifying the calculation process, as finding max and min values is computationally simpler than calculating standard deviations and assuming normal distribution.
3Productivity
If warning lines are established without considering data resolution, then the process is faster, but false alarms increase due to stable data characteristics
Solution Approach 1:
The patent performs preliminary actions by collecting data over an extended period and identifying the maximum and minimum values before establishing the warning lines. This preliminary data collection and analysis phase ensures that the warning lines are based on comprehensive observed ranges, thereby reducing false alarms while maintaining efficient one-time calculation setup.
Solution Approach 2:
The patent cushions against false alarms by incorporating the resolution of the measuring device into the warning line calculation. By adding and subtracting the resolution value from the maximum and minimum observed values respectively, the method creates a buffer zone that accounts for measurement precision limitations, thereby preventing false alarms caused by minor fluctuations within the resolution range.
Data Source
AI summary
A method for monitoring a machine includes: manufacturing products by at least one machine, and the machine includes a sensor used to detect first data of a parameter of the products during manufacturing the products. The first data of the parameter are transmitted to a monitoring system by the machine. A final upper warning line and a final lower warning line are established by the monitoring system to determine whether the parameter is abnormal. The steps for establishing the final upper warning line and the final lower warning line include: using the monitoring system to calculate a plurality of change amounts in the first data between seconds, and selecting a minimum change amount among the change amounts corresponding to each of the products. The minimum change amounts form a minimum change amount set, and a minimum value is selected as a resolution among the minimum change amount set.