The SAP Predictive Maintenance & Service (PdMS) mission at Customer Innovation & Strategic Projects is two-fold: co-innovation and productization. We are actively engaged in customer co-innovation projects and planning to release a new set of standard business applications that operationalize machine learning and enable maintenance operators, engineers, and data scientists to receive issue notifications, explore root causes, visualize sensor data, track issues, and take action. We have also developed the PdMS Foundation and the native HANA time series. Together, these make it easier and faster to meet the business needs and use cases of our customers and deploy production ready systems.
In a typical customer co-innovation project, the team works with customers to prioritize challenges, define use cases, and build the technical foundation that fuses and mines sensor and business data. Once the foundation is in place, along with the ability to get insights, the business benefits can begin to be realized. For example, the system generates a derived signal that a product on the production line will likely fail the quality standard. Action can be taken in real-time to analyze root cause and fix the issue. Additionally, informed decisions can be made to create work activities and optimize maintenance schedules, order spare parts, improve product quality, or decide whether it’s best to repair, replace, or refurbish parts. The benefits can be impactful, such as less downtime, quality and productivity improvements, engineering improvements, getting the right parts to the right place at the right time, and more.
The tools used for predictive analytics and real-time results include: R, Python and SAP Predictive Analytics 2.2 including SAP Expert Analytics, SAP Automated Analytics, and SAP Lumira.
The technology used is the SAP HANA Platform including: PdMS foundation, SAP HANA R Integration, HANA spatial, Predictive Analysis Library, and SAP HANA Text Analysis.
Data types and sources are typically a combination of transaction, unstructured, machine, HADOOP, real-time, location, and other app data; e.g., 4M work orders, 3M notifications, 6B sensor readings, and 1.2M machine components.
To learn more about Predictive Maintenance and Service solutions and co-innovation, please contact email@example.com. If you are interested in a Webcast episode I recently recorded specifically about this topic, sign up here in order to listen in.
Please add yourself to the Influence Council to provide input on the new standard applications being developed.