White Paper: Using Machine Learning and AI to Augment Industrial Plant Asset Diagnostics

#News(IoTStack) [ via IoTGroup ]

Headings…
White Paper: Using Machine Learning and AI to Augment Industrial Plant Asse
Overview
Introduction
Problem Statement
Line of Business Definitions
Operations
Maintenance
Monitoring and Diagnostics
Breakdowns in the Operations, Maintenance, and M&D Processes

Auto extracted Text……

This White Paper details the Operations, Maintenance, and Monitoring and Diagnostics challenges faced by owners and operators of Industrial Process plants, and outlines and defines methods of information management, data visualizations, analytic pipelines, and Machine Learning and Artificial Intelligence to improve the performance and reliability outcomes of each of these disciplines.
The remainder of this paper describes a set of five use cases, or “microservices,” that can be used to compose a set of data services, visualizations, analytic pipelines, and ML and AI techniques to improve the performance of the Operations, Maintenance, and Monitoring and Diagnostics lines-of-business, for process industry plants.
M1 – Semantic Model: a graph-structured organization of data, or a data services layer on top of the source databases, that allows technical personnel and/or data scientists to navigate and access Asset data, across a number of disparate systems-of-record.
For example, an M&D engineer could query for instrumentation or asset management data from “all plants that have turbine-driven boiler feed pumps;” a tree structure or graph query allows intuitive access to the data.
As mentioned in the previous discussion on “Breakdowns in the Operations, Maintenance, and M&D Processes,” it is typical for failure event data to be recorded via unstructured text; to be able to perform robust reliability analyses, the events must be classified according to Problem Code, and Failure Code.
Use Case M2 (described later) will discuss how unstructured technical maintenance data can be converted into structured data; the semantic model can be used to create and maintain a taxonomy of problem and failure codes, and the codes can use the “inheritance” feature of the model, i.e. a boiler feed pump is derived from a process pump, therefore the feed pump inherits the standard pump’s Problem and Failure Codes, as well as adding its own

Read More..
AutoTextExtraction by Working BoT using SmartNews 1.0299999999 Build 26 Aug 2019