Subscribe and listen to AW’s podcast!
Subscribe and listen to the Automation World Gets Your Questions Answered podcast!
Listen Here

Simulation for Intelligent Cyber-physical Systems

As products continue to get smarter, so does the complexity and integration requirements. Learn how designers of these cyber-physical systems are using advanced simulation platforms as a guide in this increasingly complex manufacturing space.

Dick Slansky, senior analyst, ARC Advisory Group
Dick Slansky, senior analyst, ARC Advisory Group

The level of intelligence now embedded in our cars, homes, communication devices, consumer electronics, and other devices increases every day. In the very near future, not only will humans interact with a rapidly growing array of smart products, but many of these products will interact autonomously with each other and other systems.

Moreover, factory production lines, process plants for energy and utilities, and smart cities will depend on cyber-physical systems (CPS) to self-monitor; optimize; and even autonomously run infrastructure, transportation, and buildings. In the future, cyber-physical systems will rely less on human control and more on the intelligence embedded in the artificial intelligence (AI)-enabled core processors.

While manufacturers across all industrial sectors are ramping up to meet demand for this growing “smart product” market, they face major challenges developing and manufacturing these new and increasingly more complex products and systems. These cyber-physical systems require tight coordination and integration between the computational (virtual) and the physical (continuous) worlds. To meet these complexity and integration requirements, designers of cyber-physical intelligent systems are using advanced simulation platforms that cover model-based mechatronic systems engineering, embedded system design integration, and simulation models that validate product and system design in the physical world.

Cyber-physical systems will run business and industry

A CPS is an engineered system or mechanism that is controlled or monitored by computer-based algorithms and tightly integrated with both the internet and its users. In CPS, physical and software components are deeply intertwined, each operating on different spatial and temporal scales, exhibiting multiple and distinct behavioral modalities, and interacting with each other in a lot of ways that change with context. Examples of CPS include smart grid, driver-assist and autonomous automobile systems, transportation systems, health and biomedical monitoring, manufacturing and process control systems, smart cities, robotics systems, intelligent edge devices, and new agricultural technologies.

CPS can operate in the presence of uncertainty, often due to external circumstances not under system control. For CPS, operating in the physical world, unplanned natural events could include weather, natural disasters, and of course, unplanned human error, or intentionally malicious human actions. System failures such as faulty sensors and actuators and inaccurate or interrupted data streams could also create uncertainty. The research community is constantly exploring new approaches for simulation and modeling to deal with uncertainty.

Today, AI and machine learning (ML) are being applied to the problem of uncertainty. Probabilistic algorithms can deal with predictive and prescriptive analytical models. Intelligent CPS get much of their intelligence from both the use of ML, which introduces approximation and requires probabilistic and statistical training algorithms; and from inferencing engines embedded in intelligent edge devices. Use of current advanced simulation applications to model CPS will be critical to the development and implementation of the next generation of Internet of Things (IoT) ecosystems.

Simulation technology enables advanced cyber-physical systems

The sheer complexity of many CPS use cases requires state-of-the-art simulation applications. These simulation application platforms typically include the traditional CAE testing applications like FEA, CFD, multi-physics, electro-magnetic, stress analysis, and other product design testing applications. Today, these platforms also include applications for modeling and simulating multi-discipline systems engineering. These allow engineers to apply a model-based, systems engineering approach to mechatronic product and process development.

As mechatronic systems take advantage of more powerful microprocessors, and the software that runs on them, the interaction between hardware and software becomes more complex. Managing this complexity can prove challenging for hardware and software engineering teams that develop requirements, describe functionality, and test and implement the concepts in a variety of ways. Most of these systems include closed-loop control methodologies that compensate for electromechanical interactions and other variables, adding further to the complexity.

Mechatronic design methods used today typically initialize the design process with mechanical modeling, followed by electrical design. Traditionally, when developing new software, engineers addressed software validation at a late stage in the development process, only testing the software through emulation on hardware prototypes. Just as electrical design-imposed constraints on mechanical systems, software typically imposes significant constraints on the overall electromechanical system design. Compensating for constraints and errors found in hardware or software at this late stage creates costly delays in the development process, since it can be time consuming to trace problems back to their root cause. Errors related to incomplete, incorrect, or conflicting requirements may even require a fundamental redesign.

In contrast, embedded systems developers typically use a concurrent approach in which the electromechanical device or equipment is targeted based on the overall function of the system and the software developed to meet the functional system requirements. Without a system engineering approach for software development, late stage integration often requires costly and time-consuming engineering design change.

Simulation platforms offer concurrent design and validate capability

Today’s multi-functional simulation platforms offered by product lifecycle management suppliers allow system simulation engineers to virtually assess and optimize the performance of CPS and mechatronic systems. The idea is to enable engineers to simulate, test, and validate their systems design in a concurrent lifecycle process from early development stages until the final performance validation and controls calibration.

Technology and science of simulation modeling

Creating a cyber-physical system depends upon key stages across a sensing/simulation modeling/analysis/action process. These include:

  • Sensing (sensors) that provides accurate system data;
  • Simulation modeling to simulate multiple conditions (adverse and failures) the system is designed to predict and prevent;
  • Condition monitoring that reliably monitors anomalies from expected system behavior;
  • Diagnostics that contain inferencing engines that isolate and infer root causes of faults and failures within sub-systems;
  • Prognostics (predictive analytics) that use system simulation models and data to predict a system’s useful life span probabilistically; and
  • Actionable recommendations (prescriptive analytics) based on decision-theoretic algorithms (ML) for accurate planning.

The development process is predicated on well-established, physics-based simulation modeling techniques and the science of inferencing and probabilistic outcomes. Today’s CPS are being designed to use currently developed machine learning algorithms and state-of-the-art AI processors.

>>Dick Slansky is a senior analyst at ARC Advisory Group. His responsibilities at ARC include directing the research and consulting in the areas of PLM (CAD/CAM/CAE), engineering design tools for both discrete and process industries, Industrial IoT, Advanced Analytics for Production Systems, Digital Twin, Virtual Simulation for Product and Production.

Test Your Machine Learning Smarts
Take Automation World's machine learning quiz to prove your knowledge!
Take Quiz
Test Your Machine Learning Smarts
Field-tested advice on machine vision & AI
View the latest trends in using machine vision, smart instruments and artificial intelligence.
Get the report
Field-tested advice on machine vision & AI