VARIMOS: Automated Optimisation
VARIMOS is our unique optimisation solution for plastic injection moulding. It uses artificial intelligence to save you time and provide better results, by automating the search for optimal solutions.
How does VARIMOS work?
To get going, VARIMOS will ask you three things:
1) What goals / quality measures do you want to optimise?
2) Which variables / parameters may VARIMOS change, to achieve the goals?
3) What restrictions must not be violated?
Based on your answers, VARIMOS will create a Design Of Experiments for you, automatically run these experiments as simulations, create a meta model from the results, and visualise the results for you, simply and interactively.
Benefits
- Automatically search the solution space using artificial intelligence, instead of tedious and repetitive manual trial and error.
- Test significantly more parameter combinations than can be tested manually.
- Identify optimal and creative solutions that could not have been found by manual simulation.
- Work faster and more efficiently.
- Unmatched speed and efficiency, thanks to parallelisation of simulations and the highly efficient CADMOULD simulation algorithm.
- More substantiated and constructive discussions with customers and colleagues. Have engineers focus on interpretation and decision making, rather than tedious repetitive work.
- As a result, get a higher first-time-right rate, fewer correction loops, greater time savings and lower overall costs.
Sensitivity
Sensitivity answers the question: How do the variables influence the goals or quality measures? You can interactively vary the variables with simple sliders and immediately see what changes in the results (see Figure 3). This allows you to see at first glance which variables are most important and which have only a minor influence on your results. This is useful for several reasons:
- Check alternatives quickly – get more than a “point estimate” Instead of getting only a “point estimate” of the set variables / parameters, as in a simple simulation, you can immediately see how the results would change, if the variable values were changed. The awesome thing about it is: you can actually manipulate the variables dynamically, and the results update dynamically, immediately. This makes it easy to see cause and effect relationships.
- Recognize goal conflicts. Because you can immediately see the impact on quality characteristics when you change the variables, you can quickly check which goals are well matched and which are in conflict.
- Better communicate and discuss with customers and colleagues Because the presentation of the results is so simple and intuitive, you can discuss them with colleagues and customers quickly and easily, based on facts. Discuss which goals can actually be achieved at the same time and how you would like to prioritize when different goals conflict with each other.
- Where can I save? Where can I not? For example, variables that have only a minor impact can be designed to be particularly cost-effective. Variables that have a large influence should be examined more closely and, if necessary, should be examined in more detail during sampling.
Optimum
VARIMOS mathematically determines an optimal suggestion for improving component, tool and process parameters in order to optimally achieve your goals. The way this is designed has several advantages for you:
- Unmatched speed. For an optimization, an experimental design is simulated through – this requires many simulations to thoroughly search the parameter space. This is where the unique speed advantage of CADMOULD as a simulation engine comes into play. Only CADMOULD offers the cutting-edge 3D-F algorithm, which is specially optimized for speed and precision in plastic injection molding. It combines speed with accuracy in a unique way. In addition, each individual simulation runs on a separate CPU core, so that the experimental design can be parallelized wonderfully during optimization. That way, VARIMOS can run several simulations simultaneously, each on its own processor core. On multicore computers, this multiplies the speed advantage that CADMOULD already has at the individual simulation level. The result: Where conventional calculation methods may take days or weeks for a complex experimental design, CADMOULD and VARIMOS are often faster by a factor of 4-10 (depending on the component and computer configuration). CADMOULD’s decisive speed advantage is what makes simulative optimization tractable in practice.
- Systematic, automated analysis, instead of human trial and error instead of laboriously simulating various parameter combinations manually by trial and error, let VARIMOS’ artificial intelligence do the repetitive work. This saves you valuable human engineering time and allows your team to concentrate on the really important points: interpreting the results and making decisions. You save yourself the repetitive, manual trial and error.
- True optimum – finding unconventional solutions. With VARIMOS you can be confident that all important parameter combinations have been checked. Instead of simply stopping when you have a satisfactory simulation result, you can be sure that VARIMOS has systematically checked the entire solution space, and so that the result is not merely satisfactory, but truly optimal. This means that even unconventional solutions that you might not have tried by hand are investigated. Because no feasible option was left out, the solution will indeed be optimal. This certainty enables you to have fact-based discussions with colleagues and customers, and really zoom in on the trade-offs that warrant discussion.
Flexibility with more than one goal
Of course, the optimum depends on how you weight different goals, relative to each other. You can set this weighting freely and dynamically, and also change it interactively, later on. The best thing is that you don’t need to run any more simulations if ou’ve changed the weighting. Instead, VARIMOS will update on the fly and give you an adjusted optimum, usually within seconds.
Use Cases
Our customers use VARIMOS along their entire value chain:-
A. Analyze feasibility, to check whether customer specifications can really be fulfilled at the same time. This allows you to give your customer fast, well-founded feedback and to make better calibrated offers.
B. Efficiently simulate more variants. Don’t just test your first hypothesis. Let the computer do the work and simulate variants around this solution. Your head remains free for more important things because you have automated repetitive work. Focus your energy on the important things: Interpret results, discuss options and make decisions.
C. Easy and fast comparison of variants. See not only a “point result” but also variants. Check the sensitivity. What is really important? Discuss the advantages and disadvantages of different solution options with your customers.
D. Achieve better results, be it in shrinkage and warpage, energy consumption, or other goals. You define the goals and VARIMOS will find solutions for you.
E. Optimize process parameters. This is possible even if the mold is already built. At this stage, it is also possible to use the VARIMOS REAL product variant, which designs a sampling plan to be run physically, on the real machine, instead of using simulation to test variants (more on this in the next section).
Two flavours of VARIMOS
There are two variants of VARIMOS. Firstly, VARIMOS can be used in the VIRTUAL variant. Here, the design of experiments is executed in simulation, as described above.
This variant is best suited if the mould has not yet been built so that the part and mold can be optimally designed from the start. Secondly, VARIMOS can also be used in the REAL variant. In this case the experimental design is not simulated, but physically carried out on the machine. This is suitable for optimising the process parameters for existing tools or parts that are already in production, or to optimise the set-up process for new moulds.
