AH2020.PPS2 Frameworks

FAMTS

Frameworks for Adaptive Maintenance Time Scheduling

  • Project Frameworks:

  • FAMTS_sys:

  • Status Phase: In development phase

  • Common Data Format: FAMTS_sys (cdf)

  • Description:

Framework for Product-UseCase services management

  • FAMTS_bDinG-BTT01:

Read the [BTT] file format for Smart Presses and inject it to the internal analytics database.

  • FAMTS_bDinG-OLI01:

Read the [OLI] file format for Injection Mold Machine and inject it to the internal analytics database.

  • FAMTS_edaG:

  • Status Phase: In development phase

  • Common Data Format: FAMTS_edaG (cdf)

  • Description:

Exploratory data analysis reporting.

  • FAMTS_fGeneralizedFaultTreesG:

From sensors data, this framework finds the best Generelized Fault Trees (GFTs) structures that models several classes of machine failures (see [P08] [P07]).

  • FAMTS_maGeneralizedFaultTreesG:

In real-time and using Generelized Fault Trees (GFTs) structures, this framework makes predictions about failure events of components in a machine that may be repaired or replaced, answering questions as:

  • What is the expected time of failure (e.g. break) of the current component?

  • What is the probability of failure of the current component?

FBIP

Frameworks for Bottleneck Identification and Prediction

  • Project Frameworks:

  • FBIP_sys:

  • Status Phase: In development phase

  • Common Data Format: FBIP_sys (cdf)

  • Description:

Framework for Product-UseCase services management

  • FBIP_bDinG-BTT01:

Read the [BTT] file format and inject it to the internal analytics database.

  • FBIP_fBottleneckIdentificationG:

Calculate several bottleneck metrics. This framework has the following capabilities [level>=A]:

  • CA01: Automatic determination of the locationId sequences;

  • CA02: Calculation of the transition probability matrix from location i into j;

  • CA03: Identification of reprocessed partnumberID at the same locationID;

  • CA04: Identification of partnumberID belonging to the previous shift;

  • CA05: Identification of partnumberID not concluded in the current shift;

and the capabilities [level>=B]:

  • CA06: Determination of the Ideal Cycle Time (ICT) for each locationID;

  • CA07: Calculation of several bottleneck metrics as sole bottleneck, shifting bottleneck, queue bottleneck, beside others.

  • FBIP_fBottleneckPredictionG:

Train several machine learning classifiers to determine the best model that forecast the bottleneck machine in the next N minutes. See details in [P04].

  • FBIP_maBottleneckPredictionG:

Deploy the best model, trained with FBIP_fBottleneckPredictionG (cdf), to forecast the bottleneck machine in the next N minutes.

FEFF

Frameworks for Equipment Failure Forecasting

  • Project Frameworks:

  • FEFF_sys:

  • Status Phase: In development phase

  • Common Data Format: FEFF_sys (cdf)

  • Description:

Framework for Product-UseCase services management

  • FEFF_bDinG-BTT01:

Read the [BTT] file format for Smart Presses and inject it to the internal analytics database.

  • FEFF_bDinG-OLI01:

Read the [OLI] file format for Injection Mold Machine and inject it to the internal analytics database.

  • FEFF_edaG:

  • Status Phase: In development phase

  • Common Data Format: FEFF_edaG (cdf)

  • Description:

Exploratory data analysis reporting.

  • FEFF_fFailurePredictionG:

From sensors data and event data, this framework do feature engineering, dimension reduction, anomaly detection, and trains several machine learning classifiers. See the details in [P05].

  • FEFF_maFailurePredictionG:

For each failure event class, this framework deploys the best model in order to forecast failures in the next T minutes.

FNPRC

Frameworks for NOK Prioritization and Root Cause

  • Project Frameworks:

  • FNPRC_sys:

  • Status Phase: In development phase

  • Common Data Format: FNPRC_sys (cdf)

  • Description:

Framework for Product-UseCase services management

  • FNPRC_bDinG-BBT01:

Read the [BTT] csv files and inject then to the internal analytics database.

  • FNPRC_fStepsAnalysisG:

Produce statistics about the steps modes OK vs NOK in quality tests, namelly, finding the most frequent steps that fail and the pairs of steps the fail simultaneously, beside other metrics.

  • FNPRC_fStepsPredictionG:

From step information, this framework do feature enginerring and trains several machine learning classifiers to find a good model for step NOK forecast.

  • FNPRC_maStepsPredictionG:

Uses the (best) model trained in FNPRC_fStepsPredictionG (cdf).

  • Supplementary Frameworks:

  • FNPRC_bDinG-BTT01:

Read the [BTT] file format and inject it to the internal analytics database.

  • FNPRC_fStepsOptimalOrderG:

Analysis the order of the steps in a set of test and proposes a new order that minimized the global time of testing.

FPRVF

Frameworks for Partnumber Rejection Validation and Forecasting

  • Project Frameworks:

  • FPRVF_sys:

  • Status Phase: In development phase

  • Common Data Format: FPRVF_sys (cdf)

  • Description:

Framework for Product-UseCase services management

  • FPRVF_bDinG-BTT01:

Read the [BTT] file format and inject it to the internal analytics database.

  • FPRVF_fRejectionPredictionG:

From sensors data and event data, this framework do feature engineering, dimension reduction, anomaly detection, and trains several machine learning classifiers. See the details in [P09].

  • FPRVF_maRejectionPredictionG:

Use the (best) model trained in FPRVF_fRejectionPredictionG (cdf).