This post is part of a five part series on the frameworks you need to maximize production and cut costs through a Simulation Digital Twin. In this installment, we take a closer look at how to build a digital twin and walk you through outcomes.
With digital transformation now driving the competitive landscape across manufacturers, now is the time for businesses to replace outdated and disparate tools with using manufacturing simulation software that takes full advantage of Industry 4.0 value-creating systems.
In previous content, we’ve covered what is a Simulation Digital Twin, the right types of data to include, and the different types of digital twins that will help you meet your business objectives.
Before we roll up our sleeves, let’s look at the top questions we hear from your peers in the industry about simulation digital twins.
Question #1: Which production planning processes should we replicate?
This may be asking yourself whether you want to model and refine a single factory (either as a pilot program or to concentrate on a key specific area), or the entire production ecosystem.
For instance, one of our global automotive clients needed manufacturing simulation software that showed where they could boost the production of specific engines to meet an increasing demand with the same capacity, requiring a digital twin across multiple product production processes, but with impact across the entire organization. This resulted in a 10% boost in operational efficiency, in just 20 minutes to first test results, across 3100 decision variables.
Because a simulation digital twin has the power to unlock value across your entire production process, this is your opportunity to work across teams and departments to create a fully-optimized, holistic production plan. See part 3 in the series to read more on selecting the right digital twin type for your organization.
Question #2: Do we need historical data and does it all need to be in one database?
Like a lot of manufacturing simulation software on the market, one of the biggest barriers to entry to some digital twin types comes down to data. Some digital twins require historical data to operate, and ideally this data needs to be already cleaned and stored in the same source. Our simulation digital twin does not. Meaning you can get started more quickly and with fewer resources.
Historical data can be difficult to extract from your current systems, or impossible, if the right sensors/infrastructure are not currently in place. By allowing enterprises to build a digital twin from existing systems, this requires far less time investment over solutions that require this type of data.
Head back to part two in the series to learn more about what type of data you do need for your digital twin.
Question #3: What are the steps needed to build a simulation digital twin?
This is what we will examine in this guide. Building a digital twin can generate next quarter value when all steps are followed correctly.
By the time you have your first simulation digital twin built, you’ll be able to:
The digital twin represents your factories, production steps, products, machines, and transport flows. With a complex system of systems, mapping real-world assets and processes to digitized data is also an intricate activity.
This involves many associated tasks, including:
There is also room for more in-depth data, including customer feedback, vendor models, weather, ERP/ business systems, and product data – all mentioned as key areas that companies are focusing on in driving their digital twins further, according to a study on digital twins from LNS Research.
We talked previously about the unification of data that comes along with creating a digital twin.
It not only helps simulate current operations to see where optimizations lie and where we can anticipate roadblocks, but also allows different stakeholders (e.g. customers, investors, the sales team, operations) to all have an up-to-date view of the data relevant to their interests. Having a unified data model of your systems available on demand means decision-making and reporting for specific user groups can be done at speed.
Our product comes with Cosmo Tech Studio, manufacturing simulation software designed specifically to help create a simulation digital twin via modeling. The software has a graphics-based interface and uses minimal coding, making it easier to learn for systems modelers.
This powers efficient mapping of real-world business and operational systems – including all of their interconnections – as they really are and as they evolve over time.
By creating a full system model, you have the ability to view all your assets, processes, and interconnections in a top-down, organized manner. This digital and graphical representation not only drives the simulation, it makes the system comprehensible to the human eye.
The further advantage of a simulation digital twin is it leverages the real-time extraction of data, giving your team the most up-to-date manufacturing simulations. This includes data flows from third-party applications, as well as building out new extraction methods from internal systems where none currently exist. This may involve monitoring from IoT devices, RFID tagging of components, and building out APIs.
The beauty is in the technology’s potential – a simulation digital twin is flexible and powerful enough to scale as you grow. Teams can start with a pilot program, such as modeling and simulating a singular factory or another smaller system within your wider ecosystem. This is a good opportunity to not only make a quick impact but also identify any gaps in your data that will need to be filled as you build for more comprehensive use cases.
Starting with a pilot quickly pinpoints where you need to begin capturing all relevant data, such as quality reports or customer feedback. Enterprises should work hard to ensure digital twins match their physical representations as accurately as possible to verify the validity of projections and ‘what-if’ simulated scenarios.
Once you have in place a system model along with the real-time data flowing into the system, you are able to visualize the current state of the system, then run simulations with this data. Your first simulation will identify current inefficiencies, or projected roadblocks. From here, you will be able to alter variables within the model and then run further simulations to project new system states and attain better outcomes.
In the Cosmo Tech suite, you are able to get a complete, real-time overview of your systems as they currently stand, with the ability to toggle on and off certain components (like transport operations) for enhanced comprehension at a higher level.
Once you run a simulation on your current data, you are able to see the entire overview of what is happening with your systems. From here, we can identify bottlenecks, do forecasting, and improve performance.
Running manufacturing simulations allow us to see where inefficiencies lie in our current systems. With simply a click of a button, the digital twin can run a simulation and we can uncover a range of current inefficiencies, from stock levels, to transportation timelines, and quality shortcomings.
With the right intelligence suite, we can drill down to see the root cause of system inefficiencies – and thus update system parameters to address the underlying issues.
For instance, when running initial system data of a digital twin through the Cosmo Tech Business Intelligence module, we can see that demand for a particular product will not be met with the current expected stock projections.
From here, we can go into production analysis to see that machines are not utilized at their full operating time, affecting stock upstream, causing missed production in Factory A.
We then examine further to determine that there is a bottleneck at Heat Treatment TT1 in Factory A.
Gaining these insights means we can investigate further to determine what’s happening at the machine level. If the machinery is operating as expected, we can run new scenarios to determine the optimal solution to the problem.
When there is an inefficiency in our system, we can alter variables within the digital twin and then run each newly developed scenario to uncover the optimal solution.
Within our digital twin in Cosmo Tech, once you’ve run your first optimization, you can easily identify the bottleneck. After which, you can create multiple what-if scenarios and then re-run an optimization for each scenario. From there, the team can analyze each result and better determine a path forward.
In evaluating the simulated outcomes, both running three 8 hour shifts or selecting a different contractor could solve the bottleneck issue and increase service levels. From there, we can choose between the two optimizations to best hit KPIs.
In 2020, we have come to expect the unexpected. COVID-19 has required businesses to redefine strategy, for best recovery. As EY has noted, you must ‘adapt operations and increase resilience’ to identify robust solutions and have confidence in decision making under uncertainty.
By running scenarios with ‘way out’ parameters, such as a complete halt in demand in a particular product, we are able to see the effects on systems and make plans for if that scenario should happen in the future. We can here use analytics to determine the best way to adapt to minimize the impact on the organization.
The “what-if” scenarios a digital twin can simulate are almost boundless, so long as you have the right data driving them. Having contingency plans already drawn up for various situations makes the company more resilient and robust.
It is important to note that a digital twin allows data to flow into the digital representation of the system from the physical system, and data can flow out of the digital system. This means that if we notice an inefficiency, we actively change the physical system parameters by updating our digital system.
This data exchange means that times to implement system changes can be rapidly decreased, shifting forward ROI on optimizations.
Just as your systems change, or new data is discovered, so too will your digital twin. Like Agile software development, you should iterate over the process, bringing in new information or making changes due to feedback. Refining your digital twin is essential to ensuring it’s the most accurate representation of your physical systems.
Choosing Cosmo Tech for your digital twin ensures that your digital twin product owners have all the skills necessary to update, delete, or build on the digital twin as necessary.
