On occasion of the SAP Conference for Utilities (Europe, Middle East and Africa) which took place in Mannheim last April ( http://uk.tacook.com/utilitiespresentations ) we had the opportunity to discover and discuss on several interesting subjects being one of them which may be conceptually a unified HANA platform for Preventive Maintenance and transform a SAP EAM system into a continuous improvement engine. This exercise was done together with my colleagues Tony Bellini and Sandro Marcotto.
As can be read at Wikipedia the main value of Predictive Maintenance is to allow convenient scheduling of corrective maintenance, and to prevent unexpected equipment failures. The key is ‘the right information in the right time’ ” so we are discussing on an instance complementary to a platform where routine or time-based preventive maintenance is done, we identify this with our EAM application located in the ERP, in order to determine the condition of in-service equipments and therefore predict when maintenance should be performed.
This single HANA platform may have three parts: Information repository, Rules and Tools and Analysis layer all tied by a unified design enviroment summarized in the following concept schema:
The information repository part may contain a communication and data integration layer. In this area is necessary to define how data is gathered from IT and OT sources, including M2M communications. Elements like SAP Mobile Platform, SAP Gateway and Process Orchestration together with Event Stream Processor will play a key role to secure the right data at the right time. An important role of the HANA platform is the Geoenablement of the information as, with the exception of the power generation business, the main focus of entery transmission, transportation and distribution require the geoespational attributes of the data as the assets are allocated along territories. Also a datamodel set up is necessary in order to describe the assets relationships and dependencies (topology), from our point of view based on IEC standards (International Electrotechnical Commission: www.iec.ch ). On this area, SAP Master Data Governance would facilitate the required modelling. The huge volume of historical data cummulated (both transactional and operational) will be processed optimally by HANA in-memmory technology complemented with IQ database for Big Data volumes.
The rules and tools part of the platform is mainly designed to support the ProblemIdentification processes and rule based decision support, enabling experienced product or equipment experts the identification of potential future problems based on analytic algorithms from Infinte Insight component, the Predictive Analysis Library (PAL) or the “R” Integration for SAP HANA. The HANA Rules Engine will also enable to introduce conditions and design the monitoring of rules and patterns configuring a new generation of Condition Based Maintenance Platform.
Hana Rules Framework screenshot for Data modeling, mapping tool and
The analysis layer contains all the main functions to explote the informaiton gathered by the Information repository. First achievable level is to set up a Monitoring and basic FMEA analysis layer (failure mode and effective analysis) by implementing business objects technology and work on the data mainly provided by the SAP EAM system: FMEA codes used during Breakdown Reporting and Scheduled plans. However, the key part of this instance is the Asset Health Monitor and criticality analysis, which should be a customizable overview of assets, showing a health indicator for each of them. The calculation of the health indicator is defined in the rule engine where the indicator is calculated. Due to the diversity of calculation formulas and methodologies is relevant to highlight the easy configuration capabilities of the HANA Design Enviroment which may enable to have both a Design and an Runtime Enviroments.
Asset Health monitor concept where can be displayed together with the indicator, a detailed 360° information view about this asset
Additonal scenarios / layers to provide are Maintenance and service strategies optimization, this is the prediction of results to be used for optimizing maintenance strategies. Instead of using time or performance based rules the definition of the next maintenance date is based on the prediction of a potential future problem with a clear impact in the cost and effort by extending the maintenance cycles while reducing efforts; Equipment Benchmarking and Retrospective analysis by comparing, for example, the asset health and the technical condition of assets from the same type to identify the impact of environmental conditions and asset usage on the asset health and condition (weather correlation) or siply to identify errors like transformers spike investigations ; maintenance/service and spare parts demand forecast based on the prediction of potential future asset problems can be used to predict future service and maintenance demands resulting from works that will need to be done to prevent or to resolve these problems.
Implementing such this kind of platform is not a one time project effort but part of a continuous Improvement Process. Most of the techniques are known for decades but still hard to implement so limiting the scope where possible and focus, for example, on realistic business impact of failures instead of every (theoretical) risk is key for success. Offerings like Space Time Insight and Rolta One View match these requirements.