Applying Business and Artificial Intelligence (BI/AI) concepts in optimized, risk-based inspection and maintenance of power and process plants components 45 th MPA Seminar October 1-2, 2019 A. Jovanovic 1 , S. Chakravarty 1 , T. Rosen 1 1 Steinbeis Advanced Risk Technologies, Germany
25
Embed
Applying Business and Artificial Intelligence (BI/AI ...€¦ · Applying Business and Artificial Intelligence (BI/AI) concepts in optimized, risk-based inspection and maintenance
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Applying Business and Artificial Intelligence (BI/AI) concepts in optimized, risk-based
inspection and maintenance of power and process plants components
45th MPA SeminarOctober 1-2, 2019
A. Jovanovic 1 , S. Chakravarty 1, T. Rosen 1
1 Steinbeis Advanced Risk Technologies, Germany
STEINBEIS ADVANCED RISK TECHNOLOGIES
The Industry 4.0 (“digitalization”): in the area of risk-based maintenance and inspection of power and process plants components.
Enormous increase in digital, but not the people dealing with them.
Difficulty in extracting the most relevant outputs and results out of the data.
How to “learn from past experience (cases)”
Potential application of: Business Intelligence (BI), Artificial Intelligence (AI) assisted tools for the optimization of future inspection and maintenance plans.
Intro
STEINBEIS ADVANCED RISK TECHNOLOGIES
What has changed 1996 to 2019? RBI 1998:
4 basic risk outcomes - flammable event, toxic releases, major environmental damage, and business interruption losses.
Risk: Probability x Consequence
Both a qualitative and quantitative process for understanding and reducing the risks associated with operating pressure equipment
RBI 2018:the same & more > safety, business, environment, reputation, … WIDE ACCEPTANCE & NEW UNDERSTANDING OF RISK!
Context #1 RBI: More than 20+ years “around”
STEINBEIS ADVANCED RISK TECHNOLOGIES
Context #2 RBI: AI, KI, ML, DL, NN, … DBA
STEINBEIS ADVANCED RISK TECHNOLOGIES
Context #2 RBI: AI, KI, ML, DL, NN, … DBA
STEINBEIS ADVANCED RISK TECHNOLOGIES
Despite the age of big data, 71% of organizations still rely on a single data source to analyze asset performance and risk management.
39% of asset managers already implement advanced digital methods of maintenance and risk management, of which 26% are using risk-based inspection (RBI) and 13% are using reliability centred maintenance (RCM).
Advanced digital methods: Big data, AI and machine learning are the top digital technologies
For a long time, inspections were seen as a (burdensome) cost factor in companies. But new technologies and developments enable companies to secure real competitive advantages through a properly designed inspection strategy.
Accessible link (with a PowerBI account)https://app.powerbi.com/groups/me/reports/0f5c00b1-3aa2-4509-ba7e-cb88a170df13/ReportSectionef9d48a09750d2d13199?ctid=de8151da-a3a9-4734-9a3e-80153f6411fe&openReportSource=ReportInvitation
Number of data points: 2,600 data points over the 6 variables.
Algorithm: CARET library Platform(IDE): R Studio Processing time: 1+ hour using Intel i7 7500U dual core
Case Study: Use of ML algorithm on RBI data
STEINBEIS ADVANCED RISK TECHNOLOGIES
The CARET (Classification And REgression Training) is a set of functions that attempt to streamline the process for creating predictive models.
The package tools contain: Unified interface for modelling and prediction Data splitting Pre-processing Feature selection Model tuning using resampling Variable importance estimation Parallel processing: computational efficiency
CARET package on R
STEINBEIS ADVANCED RISK TECHNOLOGIES
2,600 data points over 6 variables (used for training model)
Results (1/3)
STEINBEIS ADVANCED RISK TECHNOLOGIES
46 data points used for prediction
Results (2/3)
STEINBEIS ADVANCED RISK TECHNOLOGIES
Result (3/3)
The algorithm was able to correctly predict the correct DM mechanisms 89% of the time.