Predictive Maintenance

Predictive Maintenance

Introduction

Maintainer: chet_metcalf@bmc.com

In the consumer goods sector, maintaining the reliability and efficiency of production machinery and supply chain systems is crucial. Unexpected equipment failures can disrupt operations and impact product availability. Predictive maintenance is key to anticipating these failures and resolving potential issues before they occur. However, managing the data lifecycle—from retrieval to consumption—poses various challenges.

Our predictive maintenance solution leverages Control-M to seamlessly integrate and automate the entire data pipeline, ensuring compliance with audit requirements, adherence to SLAs, and comprehensive reporting, ultimately enhancing operational efficiency and minimizing downtime.

Use Case Overview

The predictive maintenance workflow in consumer goods is designed to improve asset reliability by predicting equipment failures before they occur. The orchestration of the various processes is handled through Control-M, integrating data from multiple sources and using machine learning to generate maintenance insights.

The Consensus Recording for this use case can be found here: Watch this video!

Use Case Technical Explanation

Workflow Components:

1. Data Sources and Consumption:

   • The system captures data from customer records, parts data, warranty information, and field services. This data is processed from several data storage systems such as Hadoop Spark, AWS S3, and Azure Data Factory.
   • Key Components:
     • Customer Data (Hadoop Spark): Captures customer data for further processing.
     • Parts Data (AWS S3 & Azure Data Factory): Data related to spare parts is transferred and stored using AWS S3 and Azure.
     • Field Services Data (Informatica CS): This captures data from on-site service activities.
     • Warranty Data (Snowflake): Warranty data is handled through Snowflake’s integration to capture and store records.

2. Data Processing:

   • The captured data undergoes transformation and cleansing using tools like Talend, Hadoop, and Power BI. This prepares the data for machine learning model development.
   • Key Components:
     • Extract and Transform Data (Power BI): The raw data is transformed into structured formats and prepared for further use.
     • Metrics to Device Records (Talend): Talend processes raw data into device records, which can then be used for analytics and model training.
     • Process IoT Data (Hadoop MapReduce): IoT data is processed, and machine learning models are trained to predict maintenance needs.

3. Audit and Compliance:

   • Detailed logs of all job executions and data transformations are maintained using Qlik Cloud for audit and compliance purposes. These logs ensure that data integrity and job execution meet internal and regulatory standards.
   • Key Components:
     • Service Failure Analysis (Qlik Cloud): Maintains a log of all job executions and the transformations applied to the data.
     • Summary Reports (Tableau): Generates real-time dashboards to provide insights into the predictive models’ performance and maintenance workflow.

4. Maintenance Process SLA:

   • The service level agreement (SLA) is managed to ensure maintenance processes meet deadlines. Control-M monitors and evaluates the performance of the maintenance workflows, ensuring deviations from the standard are handled within tolerance limits.
   • Key Components:
     • SLA Management: Monitors maintenance workflows with deviation tolerance and generates alerts if the SLA is breached.

Orchestration by Control-M:

Control-M coordinates all data ingestion, processing, and reporting tasks to ensure seamless workflow execution. The platform manages dependencies between various tasks and ensures that the data pipeline is processed in the correct order by triggering events when one job completes and another is ready to start. Control-M ensures all jobs run as per the schedule and alert systems are in place to detect anomalies or failures.

Summary: Control-M orchestrates a sophisticated predictive maintenance solution for consumer goods by leveraging multiple data sources, transforming them using ETL tools, processing them with machine learning models, and providing insights via real-time dashboards. The system ensures compliance with auditing and SLA standards, making the maintenance process more efficient and predictive in nature.

This overview provides insight into how Control-M supports the end-to-end orchestration of predictive maintenance in the consumer goods industry, ensuring reliability and automation.

To view the demo flow code-base, and all artifacts, please navigate to the Predictive Maintenance Git Repository

Job Types Included

• Control-M Hadoop
• Control-M for Informatica CS
• Control-M Databases
• Control-M for Azure Data Factory
• Control-M for Snowflake
• Control-M Managed File Transfer
• Control-M for Talend Data Management
• Control-M for Microsoft PowerBI
• Control-M for Qlik Cloud
• Control-M for Tableau
• Control-M SLA Management

Demo Environment Information

Environment	Status	Folder
Helix Production	Available	zzz-predictive-maintenance
Helix Pre-Production	Available	zzz-predictive-maintenance
VSE CTM PROD	Available	zzz-predictive-maintenance
VSE CTM QA	Available	zzz-predictive-maintenance