Implementing MIMOSA Standard

Abstract

Listen

Translate article

A common challenge to Prognostic Health Management (PHM) systems is the management of data across different organizations based on a standardized format and meaning. The Open System Architecture for Condition-based Maintenance (OSA-CBM) and the Open System Architecture for Enterprise Application Integration (OSA-EAI) are complementary reference architectures for domain-independent asset and condition data management. In previous papers, we reported on our experiences with implementing a data integration layer based on these two architectures. In this paper, we report on our experience implementing code generators for binary OSA-CBM and OSA-EAI Tech-CDE (Compound Document Exchange), and the utilization of the resulting components within the OMAHA project. OMAHA aims towards an overall management architecture for health analysis, incorporating manufacturers, operators and maintainers of fleets of aircraft.The OSA-CBM standard specifies a message structure but leaves the assembly and disassembly of OSA-CBM data up to the implementor. Our solution is a builder/reader Application Programming Interface (API) for a binary OSA-CBM message codec which we have implemented under the constraints of a real-time computing environment. The required C code is automatically generated from the provided technical documentation for OSA-CBM. We discuss the properties of the resulting codec and point out future improvements for the OSA-CBM binary protocol to improve consistency and to add the capability of streaming. Using the same generative approach we have implemented a code generator for a Tech-CDE-compliant middleware system, consisting ofclient libraries (currently C++ and Java), a network layer, a server portion, and a database backend. Analogously to OSA-CBM, the code generator processes the documentation provided for Tech-CDE, creating both productive and testing code. We discuss the properties of the resulting system, report specific limitations of the Tech-CDE protocol and suggest mitigations. The paper concludes with an experience report from utilizing our work in the OMAHA project. While Tech-CDE was generally found sufficient, we identified areas of improvement, including protocol properties and entity coverage. We were able to make customizations using our generative coding approach and present these as suggestions for future standard extensions.