Building & Understanding Predictive Maintenance Solutions for the Aerospace Industry
This post is by Rahee Ghosh, a Program Manager in the Data Group at Microsoft.
We are pleased to announce the availability of new resources on GitHub to help businesses in the aerospace industry better understand their opportunities to benefit from advanced analytics solutions for predictive maintenance.
These predictive maintenance solutions, creating using the Cortana Intelligence Suite, allow businesses to maximize the utilization and performance of their assets, avoid unscheduled downtime and gain a competitive edge. Our previous work in the predictive maintenance space is discussed in this blog post. The new material builds on our earlier work and provides both business and technical audiences with what they need to start gaining real business improvements by using their data.
This package contains a solution overview targeted at business audiences that explains the current opportunity for improvement as well as key outcomes to expect from a fully operationalized predictive maintenance solution. Business owners can use this overview to understand the benefits of putting their data to work and see how Cortana Intelligence enables the next set of business and process improvements.
We have put together a detailed deployment guide that walks through the steps of building an end-to-end predictive maintenance solution. The solution described by the deployment guide is analogous to the one presented in our deployable solution template for predictive maintenance. By following the deployment guide, implementers will gain a deeper understanding of the components of the solution, how they work together, and how to do basic validation and debugging. Understanding these concepts will help you customize the solution to meet the specific needs of your organization, working with your existing data and infrastructure.
Data scientists looking for guidance on building models for predictive maintenance can visit the Predictive Maintenance Modelling Guide, which covers the steps needed to implement a predictive maintenance model, including feature engineering, label creation, training and evaluation. This resource includes modelling tips specific to the predictive maintenance space.
This tutorial walks through the steps to create an on-premises version of this solution using SQL Server R Services.
Do try out these new resources and share your thoughts and comments with us below.