Overview

The No Production Period Detector Processor identifies the periods of time in which the machine is not producing due to an unplanned stop and allows to compute the Overall Equipment Effectiveness parameter. It analyzes patterns that indicate why the stop occurred.


THIS PROCESSOR IS HYPER CUSTOMIZED: It was specifically designed for one certain machine and probably has to be modified before it can work with other input data from a new project.


First, the periods of time when the machine in not producing (STOPs) are detected employing the values in the “output counter” column. Time domains in which the counter does not increase for longer than 10 seconds are considered as a STOP. Isolated time domains shorter than 10 seconds, both in the production mode or STOPs, are not considered. This step generates the list of the total number of STOPs, together with the related indexes of begin and end of each single STOP.


Then, the length of the production cycles is computed during the normal functional periods of the machine. The length of the production cycles is detected analyzing the cyclical variations of the values of the three power consumptions. This information allows to monitor the variations of the production rate (speed) of the machine.


Next, the pattern of the three power consumptions together with the trend of the production rate are analyzed to determine, whether each single detected STOP is a planned stop or an unplanned stop. Six different characteristics patterns in the measured quantities have been defined. The planned/unplanned stops can hence be identified and distinguished accordingly to specific combinations of these six patterns.


Input

A dataset containing the three power consumptions and the output counter of a packing machine are provided as a function of time at a sampling frequency of 4 Hz.


Configuration


  • P1: Select the first column with power consumption data
  • P2: Select the second column with power consumption data
  • P3: Select the third column with power consumption data
  • Index: Select the index column (Timestamp)
  • Production Rate: Select the column containing machine output data
  • No Production Period Detection - Sample Rate: Set the sample rate. Default is 4
  • Wavelength computation - Coefficient for difference Threshold: Define the coefficient for multiplication with the sigma during wavelength computation. This is for checking, whether the wavelength should be set to 0 at a certain point in time or not. Default is 0.66
  • Maximum wavelength used during wavelength measurements: Set the maximum wavelength to be used. Default is 12
  • Maximal Production Rate: Set the maximal production rate. Higher values will be set to 0. Default is 10
  • SlowDown - Minimal increasing trend: Specify the minimal inclination of the linear-fit (B) of the smoothed period to detect a significant increase approaching the beginning of the STOP phase. Default is 0.008 for 4Hz sampling
  • SlowDown - Smooth length: Specify the parameter smooth_len (A) that allows to tune the smoothing of the input period signal. This parameter is expressed in “pixel”. Default value is 11 pixel
  • SlowDown - Number of frames to look before and after stops: Specify the number of frames to look before and after stops for pattern detection. Default is 200
  • SlowDown - Minimal Wavelength before the stop: Specify the parameter wlendmin (C) that allows to set a lower limit on the smoothed period right before the beginning of the STOP phase. Default value is 5 for 4Hz sampling
  • 2Steps - Minimal decreasing trend: Specify the trend_min that defines the minimal inclination of the linear-fit (B) of the smoothed period to detect a significant decrease after the end of the STOP phase. Default is 0.01 for 4Hz sampling
  • 2Steps - Smooth length: Specify the parameter smooth_len (A) that allows to tune the smoothing of the input period signal. This parameter is expressed in “pixel”. Default is 11 pixel
  • 2Steps - Number of frames to look before and after stops: Specify the number of frames to look before and after stops for pattern detection. Default is 200
  • 2Steps - Initial minimal Wavelength: Specify the parameter wlinimin (C) that allows to set a lower limit on the smoothed period right after the end of the STOP phase. Default is 6 for 4Hz sampling
  • PowerSave - Difference threshold: Specify the parameter n_sig (C) that allows to define the level of significance of step-like changes in the data. Default is 6 and ensures that only strong drops are considered
  • PowerSave - Minimal length: Specify the parameter pslenmin (A) that defines the lower bound of the time window after the begin of the STOP in which the three signals simultaneously show a significant drop. Default is 28s∙4Hz = 112 pixel.
  • PowerSave - Maximal length: Specify the parameter pslenmax (B) that defines the upper bound of the time window after the begin of the STOP in which the three signals simultaneously show a significant drop. Default is 34s∙4Hz = 136 pixel
  • BOXall - Level of significance for step-like changes: Specify the parameter nsigstep (A) that allows to define the level of significance of step-like changes (both up-ward and down-ward) in the data. Default is 10 and ensures that only strong steps are considered.
  • BOXall - Maximal differential change: Specify the parameter nsigdiff_max (B) that allows to define the maximal differential change allowed across the three signals (how different they can be). Default is 8. Smaller values require stronger similarities across the three signals to detect this pattern
  • BOXall - Minimal number of steps: Specify the parameter nstepmin that allows to define the minimal number of steps (both up-ward and down-ward) that the three signals must show in order to detect this pattern. Default is 10. A single peak counts as 2 steps (one up-ward followed by one down-ward) as well as a box-shaped profile.
  • BOXdiff - Difference threshold: Specify the parameter n_sig (C) that allows to define the level of significance of step-like changes in the data. Default is 20 and ensures that only strong drops are considered
  • BOXdiff - Minimum length: Specify the parameter boxlenmin (A) that defines the range of the time length in which two of the three signals show an off-set to greater values. Default is 2.5s∙4Hz = 10 pixel
  • BOXdiff - Maximum length: Specify the parameter boxlenmax (B) that defines the range of the time length in which two of the three signals show an off-set to greater values. Default is 3.5s∙4Hz = 14 pixels
  • Church - Coefficient for the difference threshold: Specify the parameter n_sig (D) that allows to define the level of significance of step-like changes in the data. Default is 6 and ensures that only strong drops are considered
  • Church - Coefficient for the difference threshold of the two church sides: Specify the parameter step_min (C) that allows to define the minimal difference between the value at the feet of the tower (on the left side) and the value at the begin of the roof of the church (right side of the tower). Default is 3
  • Church - Maximum length: Specify the parameter towerlenmin (A) that allows to define the maximal time length of the feature recognisable as the tower of the church. Default is 3 pixel for 4Hz sampling
  • Church - Minimal length of the roof: Specify the parameter rooflenmin (B) that allows to define the minimal time length of the feature recognizable as the roof of the church. Default is 6 pixel for 4Hz sampling

Consult Additional Information section for more information about the patterns SlowDown, 2Steps, PowerSave, BOXall, BOXdiff and Church.


Output

The processor has two output ports. The left port delivers all input rows, the right port delivers a table with all detected stops.


Additional Information


F_SlowDown

This function analyses the trend of the production rate (period) of the machine right before the STOP phase starts. The pattern is recognized, if the period decreases gradually (the length of the production cycles becomes longer) approaching the begin of the STOP.


2Steps

This function analyses the trend of the production rate (period) of the machine right after the STOP phase ends. The pattern is recognized, if the period increases (the length of the production cycles becomes shorter) after the end of the STOP. The increase of the production rates take place in two defined steps.


F_PowerSave

This function analyses the trend of the signals of the three power consumptions during the STOP phase. The pattern is recognized, if all three signals simultaneously show a significant drop about 30 seconds after the begin of the STOP.


BOXall

This function analyses the trend of the signals of the three power consumptions during the STOP phase. The pattern is recognized if all three signals show simultaneous step-like changes, i.e. significant increase or decrease of the signals in short time.


BOXdiff

This function analyses the trend of the signals of the three power consumptions during the STOP phase. The pattern is recognized, if two signals show an off-set to greater values of about 3 seconds while the third signal is constant.


Church

This function analyses the trend of the signals of the three power consumptions during the STOP phase. The pattern is recognized, if all three signals show a pattern that visually remind a “Church with Tower” or a 180-degree rotated “Square root symbol”.