Cloud Pak for Data Group

Intelligent Maintenance 

Thu August 29, 2019 05:45 AM

Reduce your costs by scheduling maintenance at just the right time.

Unplanned maintenance is a key issue in a number of asset intensive industries such as telecommunications, manufacturing, oil and gas, and transportation. It can be extremely costly due to extended production downtimes. Not only do they result in high urgency repairs and investment in new machinery and parts, but unexpected production loss may affect supplier obligation and force a company to purchase products from its competitors, resulting in significant unanticipated costs.

Minimizing maintenance cost for one machine

Predictive maintenance uses predictive analytics and machine learning (ML) to determine the lifespan of a machine and the likelihood of it failing on a given day. An ML model can take into account all available information, such as sensor data, historical tasks and orders, past failure history, and maintenance history. As a result, it can help determine the optimal maintenance day for each machine based on its specific characteristics. The idea is to use this information to schedule maintenance before equipment fails but not before it is actually needed, avoiding unnecessary costs associated with repair and production loss.

Minimize Expected Maintenance Cost for One Machine

Minimizing maintenance cost for many machines

This methodology works great when we consider one machine at a time and there are not conflicting priorities. However, once we start introducing some complexity, dependencies, and limited resources, determining the optimal schedule becomes more challenging. For example, how do we take into account limited resources? If our ML model predicts that Day 3 is the optimal maintenance day for three machines but only two maintenance crews are available, when is the optimal maintenance day for each machine?

Minimize Expected Maintenance Cost for Multiple Machines

Intelligent Maintenance: Machine Learning and Decision Optimization Working Together

While machine learning can take into account all available data and past history to predict the likelihood of failure for a given machine, decision optimization (DO) can take it a step further and generate a schedule that is optimal for a set of machines, subject to limited resources (e.g. maintenance crew availability), other constraints and dependencies (production plan and repair costs), and optimization metrics (minimizing total cost, maximizing customer satisfaction/meeting planned production, minimizing late maintenance). Not only does it offer us valuable insights, but it also generates an actionable schedule or plan.

Intelligent Maintenance: Machine Learning and Decision Optimization Work Best Together
Minimize Expected Maintenance Cost for One Machine

For a detailed description of the Intelligent Maintenance accelerator please refer to the following blog

What's included?

  • A structured business glossary of business terms.
  • Sample data science assets

How does it work?

The glossary provides the information architecture that you need to govern the data about your equipment. And your data scientists can use the sample notebooks, predictive models and decision optimization model to accelerate data preparation, machine learning modeling and finding an optimal solution. The Machine Learning and Decision Optimization models are deployed for use in production.

The accelerator uses predictive analytics and machine learning to determine the optimal day to performance maintenance on each machine. The accelerator helps you schedule maintenance so that you:

  • Don't waste time and money fixing equipment that is working well.
  • Don't let equipment failure leave you idle and accruing unnecessary costs.

Optimize your maintenance schedule based on the costs that you will incur.

The accelerator includes a Sample application which shows an example of how the deployed models can be embedded into a Machine operators tooling to plan for scheduled maintenance.

A sample of the Intelligent Maintenance accelerator deployments integrated in a Maintenance Planning application

Prerequisites

Required services: To use the industry accelerators, you must install one or more of the following services on IBM® Cloud Pak for Data

Service Required for
Watson Knowledge Catalog Importing data governance artifacts, such as business terms and categories. See Installing Watson Knowledge Catalog.
Watson Studio Importing data science assets to an analytics project. See Installing Watson Studio.
Watson Machine Learning Deploying analytical models. See Installing Watson Machine Learning.
Decision Optimization add-on Optimizing the maintenance schedule
(Optional) SPSS Modeler Flow add-on DataBuilding machine learning model

Importing the accelerator

To use this accelerator on Cloud Pak for Data v2.5.0.0, complete the following steps:

  1. Download the intelligent-maintenance-industry-accelerator.tar.gz file, which is available on the https://github.com/IBM/Industry-Accelerators repository.
  2. Extract the contents of the package.
  3. Follow the instructions in the README.pdf.

To use this accelerator on Cloud Pak for Data v2.1.0.0, complete the following steps:

  1. Download the intelligent-maintenance-industry-accelerator.tar.gz file, which is available on the https://github.com/IBM/Industry-Accelerators repository.
  2. Determine how you want to import the accelerator:
    • If you want to pick which components you import, complete the following steps:
      1. Extract the contents of the package.
      2. For the business glossary, use the Cloud Pak for Data web client to import the XML file. For details, see Importing a data dictionary in the Cloud Pak for Data documentation.
      3. For the data science assets, use the Cloud Pak for Data web client to import the ZIP file. For details, see Creating a project in the Cloud Pak for Data documentation.
    • If you want to install all of the components, complete the following steps:
      1. From a command prompt, run the following command to authenticate to Cloud Pak for Data:

        curl -s --output /dev/null -w '%{http_code}' -k -X POST https://{HOSTNAME}:{PORT_NUMBER}/v1/preauth/signin -F username={USERNAME} -F password={PASSWORD} -c auth_cookie.txt

      2. Run the following command to import the TAR.GZ file to Cloud Pak for Data:

        curl -k -X POST https://{HOSTNAME}:{PORT_NUMBER}/zen-watchdog/v1/accelerator -F project_file=@{FILEPATH} -b auth_cookie.txt

        For the {FILEPATH}, use the fully qualified path of the TAR.GZ file that you downloaded.

      3. Run the following command to verify that the import completed successfully:

        curl -k -b auth_cookie.txt -X GET https://{HOSTNAME}:{PORT_NUMBER}/zen-watchdog/v1/accelerator/status/{REQUEST_ID}

        For the {REQUEST_ID}, use the ID that was returned by the preceding command.

Release Notes

This accelerator has been verified on:

  • Cloud Pak for Data v2.1.0.2
  • Cloud Pak for Data v2.5.0.0

About the developer:

IBM

Licensing

This project contains Sample Materials, provided under license.
Licensed Materials - Property of IBM.
© Copyright IBM Corp. 2019. All Rights Reserved.
US Government Users Restricted Rights - Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.

Terms and conditions

Before you use the Industry Accelerators that are being linked to from this IBM Cloud Pak for Data Community page, you must agree on the terms and conditions that are set forth here:
This information contains sample modules, exercises, and code samples (the code may be provided in source code form) (collectively “Sample Materials”). You may copy, modify, and distribute these Sample Materials for your internal use only provided such use is within the limits of the license rights of the IBM agreement under which you are licensing IBM Cloud Pak for Data. Notwithstanding anything to the contrary, IBM PROVIDES THE SAMPLE MATERIALS ON AN "AS IS" BASIS AND IBM DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, ANY IMPLIED WARRANTIES OR CONDITIONS OF MERCHANTABILITY, SATISFACTORY QUALITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE, AND ANY WARRANTY OR CONDITION OF NON-INFRINGEMENT. IBM SHALL NOT BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY OR ECONOMIC CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR OPERATION OF THE SAMPLE MATERIALS. IBM SHALL NOT BE LIABLE FOR LOSS OF, OR DAMAGE TO, DATA, OR FOR LOST PROFITS, BUSINESS REVENUE, GOODWILL, OR ANTICIPATED SAVINGS. IBM HAS NO OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS OR MODIFICATIONS TO THE SAMPLE MATERIALS.

The Industry Accelerators might include applicable third-party licenses. Review the third-party licenses before you use any of the Industry Accelerators. You can find the third-party licenses that apply to each sample in the notices.txt file that is included with each sample.



#Featured-area-2
#datascience
#governance
#IBM
#cross-industry
#Jupyter
#IndustryAccelerator
#CrossIndustry

Statistics

0 Favorited
129 Views
0 Files
0 Shares
0 Downloads