Histogram-Based Input Error Detection for Manufacturing Control Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing input error detection methods for numerical controllers in manufacturing machines are inadequate in accurately determining input errors, particularly when different input trends exist for the same item and fail to ensure correct functioning due to mixed machine coordinate values and varying tool offset values.
Innovation Solution
An input error detection device that creates a histogram of past input values and compares it with a histogram based on the control program to detect errors, enabling advanced error detection based on trends and simple deviations, such as wrong decimal point inputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard deviation and average based error detection is used, then some input errors can be detected, but the accuracy of incorrectness determination is reduced when different input trends exist for the same item
Solution Approach 1:
The patent segments the input values into multiple groups based on their distribution characteristics. Instead of treating all input values as a single homogeneous set, the system divides them into distinct groups (e.g., workpiece vertex values and workpiece center values as separate groups). This segmentation allows the error detection to accurately handle mixed input trends by analyzing each group's characteristics independently, thereby resolving the contradiction between detection accuracy and adaptability to different input patterns.
2Reliability
If formal error detection (decimal point check, range check) is used, then some obvious errors can be detected, but the detection capability is limited for subtle input errors in control programs
Solution Approach 1:
The patent performs preliminary analysis of the control program structure and input value distributions before executing the numerical controller. By creating histograms of command values and offset values in advance and comparing them against expected patterns, the system can detect subtle errors that formal checks would miss. This preliminary action enhances both the reliability of obvious error detection and the capability to detect subtle errors, as the system is prepared with reference data before actual processing begins.
3Ease of operation
If mixed machine coordinate values and tool offset values are used without separate settings, then the system is simpler to operate, but the correctness determination of input values cannot be ensured
Solution Approach 1:
The patent applies local quality by analyzing input values according to their specific context and type. Instead of using a uniform detection method for all inputs, the system creates separate histograms and applies different detection criteria for command values, offset values, and different tool types. This localized approach maintains ease of operation (users don't need to manually configure different settings) while ensuring high correctness determination accuracy by adapting the detection method to each specific input context.
Data Source
AI summary
An input error detection device can detect an input error based on input values input in the past and the trend of values from the view point of the whole of a control program. In order to detect an error in inputting a value used to control a manufacturing machine, the input error detection device is provided with a histogram creation unit configured to create a histogram, based on a set of values used to control the manufacturing machine, and a determination unit configured, based on the histogram, to determine input errors of the values used to control the manufacturing machine.