Multidisciplinary Model Transformation through Simplified Intermediate Representations Bjorn Cole Kevin Dinkel Jet Propulsion Laboratory Jet Propulsion Laboratory California Institute of Technology California Institute of Technology 4800 Oak Grove Dr. 4800 Oak Grove Dr. Pasadena, CA 91109 Pasadena, CA 91109 [email protected] [email protected] Abstract— There has long been a challenge of making 5. APPLICATION .......................................................... 8 engineering tools from multiple disciplines interoperate. This BSERVATIONS AND UTURE ORK problem extends to system modeling practices. This challenge 6. O F W .................... 9 has been confronted with a wide variety of techniques. These 7. SUMMARY .............................................................. 10 techniques include attempting to interface tools together into ACKNOWLEDGEMENTS ............................................ 10 combined suites, attempting to find underlying commonalities REFERENCES ............................................................. 10 in mathematics, supporting connections through semantic encoding, various graph mappings and transformations, and BIOGRAPHY ............................................................... 11 code wrappers. All of these approaches have strengths and weaknesses. These are measured in multiple areas: relative freedom of action of individual domain engineers in developing 1. INTRODUCTION their own tools, speed of execution, ease of creation, traceability, fidelity of information transfer, and degree of alignment The Europa project is designing a proposed large-scale between the concepts of different domains. science mission that would study one of the largest moons of Jupiter. The mission concept involves orbiting Jupiter and This paper presents an approach to this interoperation problem flying by Europa a large number of times in order to currently being used in the World-Wide Web. The approach is investigate the moon’s potential for harboring conditions to develop easy-to-parse formats that allow flexibility to both the suitable for life. Deep-space probes have been steadily file author and file interpreter. Many of the formats that are increasing in complexity, especially in the area of software currently deployed sacrifice runtime performance for the ability of third parties to easily understand what to do with the data. and scientific instrumentation. In addition, the environment XML became popular earlier as a de-facto standard format for of Europa is in the midst of an intense radiation field many web applications, but is now being replaced by JSON to generated by Jupiter accelerating electrons trapped in its enhance human readability and provide a simpler data model. magnetic field. Because there would be only a limited This is the basis for work in this paper. number of Europa flybys, there would be a premium on assuring that they go well. Our approach, which provides the key to interoperation, is a simplified “shrapnel” intermediate collection of objects and Like any other large space project, there are many relationships that is the result of a breakdown of the system engineering teams working together. Typically, they have model into minimal pieces. It is then reassembled on the collaborated via documents and other communications. destination side, forming a two-step transformation. Previous efforts with single-step transformations have proven too However, the information exchange provided by documents difficult to create efficiently. In contrast, the use of this leads to a great deal of hand re-entry of information across approach leads to an almost automatic procedure for multiple systems, simulators, and analysis tools. The natural transformation development. rate of human error can be greatly improved upon by relying on common data sources that are electronically translated. The Europa project is a large engineering project that must coordinate the efforts of many different teams with different The first set of tools to be integrated where those that specialties. The traditional form of exchanging engineering developed a detailed picture of the generation, storage, and information has been documentation. The vision of model-based use of power as well as science data generation and downlink. systems engineering is to make this information exchange much For this paper, it will be called the Europa Project System more digital. This paper presents the application of our simplified format to connecting two different engineering tools Resources Toolchain. This toolchain is more fully described to the system model, with a focus on a dynamic mission in a companion paper [1]. simulation encoded in Modelica. 2. KEY NEEDS TABLE OF CONTENTS The motivating problem for this work is that engineering analysis is still a very fragmented enterprise from an 1. INTRODUCTION ....................................................... 1 interoperability standpoint. In running a system-level 2. KEY NEEDS ............................................................. 1 analysis, it is often necessary to take the output of one 3. RELEVANT BACKGROUND ..................................... 2 analysis tool and use it as the input to another analysis tool. 4. APPROACH .............................................................. 3 This is because tools are typically developed and used by 978-1-4673-7676-1/16/$31.00 ©2016 IEEE (Copyright not yet transferred. © California Institute of Technology. Government sponsorship acknowledged. 1 subsystem experts. The system perspective is gained by requires either side to negotiate every detail of an input chaining together domain specific tools, gathering a system specification adds friction to the interchange. analysis perspective. But there are a large array of engineering software tools built upon many different A previously suggested solution to the engineering mathematical and logical structures and theories, tailored to integration problem has been to utilize the techniques of their domain of interest. There are islands of integration that Model-Driven Architecture vision promoted by the Object have been built up between tools used by frequently Modeling Group [2], [3], [4]. The issues that have arisen in collaborating engineers, such as between mechanical design this approach is that a fully declarative, two-way tools and stress analysis tools. Many connections in large transformation in real engineering models can be difficult to engineering organizations such as JPL are implemented as ad formulate. The languages and tools for the transformations hoc software developed for specific projects or efficiency are effective, but finding patterns for using them more improvement tasks. When a series of point-to-point efficiently is needed. That is described in the Approach part integrations is attempted in scripts and codes, the result is a of this paper. nest of hard-to-maintain connections. The objective for this work can be stately thusly: Develop a There are frameworks that work to tame the chaos. Examples style of interchange between engineering tool file formats include Dassault’s iSIGHT, Glenn Research Center’s that allows the owners of each tool to retain control of their OpenMDAO, and Phoenix Integration’s ModelCenter. These own format, is straightforward to develop, and does not tools abstract the details of specific tool Application require complete knowledge of both formats before starting. Programming Interfaces (API’s) into a general series of This interchange approach does not in principle preclude or software components with inputs and outputs. This seems prescribe the use of any particular tool for implementation. like a small abstraction at first, but it does encapsulate a fair amount of execution detail into individual components. The 3. RELEVANT BACKGROUND weakness in these frameworks is that they were designed to The need to transform data objects is not new to the object- expose collections of parameters rather than structural model oriented world, in either the programming or modeling sides information. In many tools, model structure can be just as of it. On the modeling side, the Open Modeling Group is one important as parameter values. Thus, parts of the structure of of the advocates of “model-driven engineering,” which is model A may need to be transformed into the form for model oriented toward creating useful source code from Unified B in order to truly transfer all of the information. There are Modeling Language (UML) models. There is a parallel emerging frameworks that get deeper into this problem such collection of tools and languages that aims to create models as Vanderbuilt University’s CyPhyML, Comet Solutions’ in one domain from models in another. On the programming CoMET tooling, and InterCAX’s Syndeia tool. However, side, there are working tools such as Object-Relational these tools do not provide full user control of the model Mappers that transform class specifications into relational transformation, which may be needed to connect together table schemas. tools that were built in-house. This paper describes how to make these transformations much easier to develop. The complexity and compromises inherent in an ORM provides a reason to have caution about a general method for During the early development of the Europa Project System object-to-object transformation. However, there are a number Resources Toolchain, it was realized that different of sources to this complexity. For one, the ORM must support engineering