Dynamic changes to the labor market, shifts in order volumes and an unpredictable logistics environment can lead to unexpected plant downtime and inefficiencies. Agility in decision-making is essential to adapt to unanticipated evolutions in the business environment. Even without unforeseen changes, decision-makers must constantly strive to optimize KPIs to stay competitive.
Whole-plant simulation has become an industry standard for new machine and plant design. Simulation ensures that designs will operate as intended across mechanical, electrical, and automation functions, saving time and money during startup. After commissioning, these simulations are often not used in operations because the models are rigid at best or inaccurate at worst. Data snapshots are a great way to design a machine or plant, but obsolete data can’t help you make decisions about tomorrow’s operations.
See also: Machine learning helps IT, OT teams anticipate equipment breakdowns long before they happen
Simulation and modeling offer substantial cost savings for design and startup, but their potential to provide predictive results in operations is of immense additional value. The reality is that simulation can help identify critical production points to achieve aggressive operational targets, but real-world decisions necessitate a simulation calibrated with real-time data. That hasn’t always been feasible for every plant.
What is a ‘calibrated real-time simulation?
A comprehensive, calibrated simulation of a plant can provide daily predictive insights that simulate a whole plant or process. By operating with real-time data, simulation can augment production planning and process optimization. What is a calibrated simulation, and why isn’t it yet a “standard” like simulation in the design process?
Simulations are only as good as the starting data. Plants can’t make reliable predictions with outdated data, and historically, substantial custom system design and code have been required to feed live data from disparate machines and systems into a simulation. In addition, the siloed nature of some machine data often led to “data delays,” where most simulations could look back at data from last month or quarter to analyze past performance but could not provide visibility into operations this minute. The inability to run simulations with live data led many manufacturers to only use simulations as a design tool before operations were up and running, where a data snapshot of a moment in time is sufficient.
See also: Extending factory equipment life: ‘Legacy doesn’t necessarily mean ancient’
For simulation accuracy, constantly feeding real machine data from the physical world into the virtual model ensures the simulation is valid. The resulting simulation is “calibrated.” This calibration is essential to trust a simulation’s predictions and recommendations, such as on associate assignment, machine utilization or material handling bottlenecks. A modern simulation and automation platform directly fed by data from edge devices can enable direct visibility into operations and allow the simulation to predict outcomes.
Edge computing platforms with open architectures even allow users to bring their own FMUs to edge devices and feed real-time data into simulations to generate more valuable insights, which can be transferred to external systems for further analysis.
The adoption of Industry 4.0 has made open architecture automation platforms, simulation platforms and edge computing devices to process and “push” data streams from machines commonplace components. The key that unlocks the ability to use simulation as a predictive tool in automation operations is accurate, real-time data. Reliable predictions can’t be made with outdated data.
A calibrated simulation eliminates human guesswork in planning procedural changes and resource reallocation to increase efficiency and reduce downtime risk. Increased efficiency translates to streamlined operations and higher profit margins, empowering plants to compete more in demanding markets. In a shifting business environment where unexpected changes require flexibility, calibrated simulations ensure that strategies can be accurately tested before making decisions.
Best practices for implementing real-time simulation
It’s critical to effectively plan any use of simulation or whole-plant analytics for decision-making. Understanding what the goals are, what data is stored, and where, how often and what risks are easily avoidable needs to be the role of a broad team of plant stakeholders.
When implementing simulation and analysis:
- Build a stakeholder team to evaluate the adoption of real-time simulation. Include all levels and departments impacted, from potential users to managers and decision-makers, IT and OT teams, etc.
- Evaluate current in-place processes and systems, explicitly noting challenges faced regarding data and analytics. How is production data delivered currently? What data is stored where? When is it available?
- Identify fundamental needs for real-time analytics or predictive analytics and potential areas where agility in decision-making can increase efficiency. Ensure the notes reflect any potential difficulties the team raises and include them in the planning phases.
- Identify opportunities for implementation where the most value is added and create detailed implementation plans for edge hardware and your simulation platform.
- Ensure the solution delivers flexibility and open customization to provide insight and optimizations for unique applications.
- Choose a simulation solution with a user interface allowing real-time predictions by operators and managers, not only by analysts. Accessibility enables the simulation to benefit every level of design, process, production and optimization –every day, not just when an expert is on hand.
- Create an implementation timeline
- As you implement edge hardware and simulation software solutions, maintain the stakeholder team throughout the implementation phases and launch to identify potential improvements and ensure that new systems and processes are fully implemented as designed.
Predictive simulations are the foundation of future competitive advantage. As with any significant shift, Industry 4.0 technology standardization means manufacturers must keep abreast with digitization. There are many simulation approaches and solutions that offer some level of modeling and analysis. To stay competitive, one-off simulations are inefficient, while dynamic calibrated simulations that use real-time data provide ongoing value and a powerful tool for plants to reduce costs and improve efficiency.
Chris Liu is product manager for Siemens Industrial IoT technology Industrial Edge for the U.S.
