Systems Analysis Integration

Purpose & Goal

The purpose of the SAI Project is to reduce redundant and inconsistent information about systems and their analysis across the multiple engineering disciplines involved. Redundant and inconsistent information exists between:

1. Systems and analysis models: discipline-specific analysis describes some system components and their interactions in detail, but not all, while overall system models describe all components and interactions in less detail.
2. Analysis models on multiple tools: each kind of discipline-specific analysis model is supported by multiple, incompatible tools, which encode the same analysis information, usually inconsistently.

The project is pursuing analysis integration with centralized systems models in three areas, building on past work:

1D cyberphysical simulation (time-based physics + signals). Many design engineers use 1D (aka lumped parameter) simulation tools, such as Modelica and MathWorks Simulink/Simscape. The SAI project was critical to creation of a standard for translating between SysML models and these tools (SysPhS).  It is implemented by engineering software vendors, including Dassault in the most-widely used SysML tool.

The SAI project will develop SysPhS libraries to cover more physical processes, including translational, rotational, and thermal, with example manufacturing applications of each.  This will significantly increase the range of systems analysis problems it can address.

Logistics analysis. Logistics is an industrial activity in which things are moved but not modified, as happens in factories (moving resources and products between machines), supply chains (transporting products between locations), and warehouses (accepting, storing, and providing products). The SAI Project developed reference SysML models to support logistics analysis of system models, including multi-commodity flow optimization (to determine lowest cost movement paths), queuing analysis (to find bottlenecks in movement), and discrete event simulation (for more detailed prediction of movement).

The SAI project will refine its logistics reference models to better integrate the three kinds of analysis, developing an extension of SysML for those analyses, as well as building an automatic translator from the models to analysis tools. This is in preparation for submitting them for standardization.

Behavior model verification. This work developed techniques to automatically verify executability of system behavior models by treating them as logical constraints on execution order. This enables logical solvers to determine whether models are executable by trying to find executions that meet those constraints. The approach relies on earlier NIST work unifying behavior modeling in the Systems Modeling Language (SysML) under a logical framework.

In FY23 the SAI project will develop an open source translator from SysML behavior models to a second logical solver and contribute critical elements to the logical behavior models in SysML 2, a major overhaul of SysML in preparation.

Each of these areas aims to integrate various mathematically-based engineering analysis information with each other and with other aspects of systems design.

Technical Approach

The technical idea is to use overall system models to coordinate discipline-specific engineering analysis by identifying and eliminating inconsistencies between systems models and analysis models. The project will organize analysis information around overall system models to reduce redundancy and inconsistency, by developing standard methods and protocols that prevent redundancies and inconsistencies between 1) systems and analysis models, enabling manufacturers to coordinate multiple analyses of the same system and 2) analysis models on multiple tools, enabling manufacturers to use analysis tools most suited to their problems.

 The high-level plan for each area of engineering analysis is:

1. Identify a collection of similar and widely used engineering analysis methods and tools in manufacturing design and operations.
2. Identify redundancies and inconsistencies between systems models and the selected analysis methods and tools.
3. Develop methods and protocols that prevent redundancies and inconsistencies between systems and analysis information by consolidating analysis information in overall systems models:
   • Identify and abstract commonalities in analysis information between the selected methods and tools.
   • Consolidate the above in extensions of overall systems models.

Status and next steps in applying the high-level plan to each of the research areas are:

1D cyberphysical simulation (time-based physics + signals). Past work completed a first pass through the high-level plan, resulting in commercial support for methods and protocols that integrate multiple 1D analysis tools around standard extensions of SysML models (SysPhS).

The SAI Project will begin a second pass through the high-level plan, starting with step 2, identifying differences in the first pass analysis tools that were not addressed previously. This will focus on variation in engineering-specific libraries among the analysis tools. The SAI Project will develop new SysPhS libraries for these, aligned with existing ones, as well as example manufacturing applications of each. They will be proposed as additions to the SysPhS standard.

Logistics analysis. Past work developed reference SysML models to support application of the high-level plan to logistics analysis, including multi-commodity flow optimization (to determine lowest cost movement paths), queuing analysis (to find bottlenecks in movement), and discrete event simulation (for more detailed prediction of movement).

The SAI project will complete the first two steps in the high-level plan and begin the third. It will begin refining the reference models to better integrate the three kinds of analysis, developing an extension of SysML for those analyses, as well as building an automatic translator from the models to analysis tools.

Behavior model verification. Past work developed techniques to automatically verify executability of system behavior models by treating them as logical constraints on execution order. It developed these for Satisfiability Modulo Theory (SMT) solvers and Alloy Analyzer. The SMT method was validated by developing an automated translator to a standard SMT file format, applying it to many example SysML behavior models, and verifying them on the Z3 solver. The Alloy Analyzer method was validated on the same examples, with documented patterns of translation to Alloy.

The SAI Project will contribute critical elements to the logical behavior models in SysML 2, a major overhaul of SysML onto a logical foundation, under preparation by the Object Management Group.

Tasks