5 ways to improve High-Mix, Low-Volume processes with simulation
Those of us working with High-Mix, Low-Volume (HMLV) processes can face some challenges along the way.
High-variation, low-volume production tends to have all of the complexities associated with a bigger product range, but without the benefits of the economies of scale and flexibility that comes with higher volumes.
As an industrial engineer, operational researcher or process improvement team, it can become even more of a challenge to achieve buy-in on HMLV projects as stakeholders demand more justification on process improvement initiatives, validation of the potential benefit versus risk, as well as confidence on ROI.
But there’s hope! If you face these problems and you’re looking for a better approach, simulation software could be the solution. Take a look at five ways simulation can be used to improve High-Mix, Low-Volume processes and win stakeholder buy-in for projects.
1. To batch or not to batch, that is the question
This is always a hot topic for SIMUL8 users with HMLV processes. Yes, we all recognize the prescribed wisdom of single or one piece flow. That is the ideal. However, with low volume runs there are often machines or processes (typically ovens and testing chambers) that present a practical case for batching.
With simulation we can prove the best approach as we can watch each work item progress as a single piece flow (pull process) then compare that to a run with work items batched (push process). A full set of results from each run, along with the ability to repeat the test hundreds or thousands of times with variable conditions, enables you to confidently assess and find the best long-term solution for your process.
We can then take this to the next level. If we do determine that batching is necessary then lots of other questions become important:
- What is the best batch size?
- How big should the oven or autoclave be?
- Should we have one or two test chambers?
- How should we prioritize jobs at the bottleneck?
- Do we need more storage space to Kanban work items before the bottleneck?
But don’t let these hold-up your implementation! Again, simulation provides a risk-free, flexible environment where you can quickly assess any of these questions, backed by visuals and in-depth results that will really engage your stakeholders in the process.
An example of how simulation has been used
The throughput of a semi-autonomous palletizer unit was being restricted because operators needed to wait for quality checks on one set of parts before being able to process the next group of parts.
However, an increase in demand extended the size of these batches to a point where this approach was unsuitable for supplying downstream operations e.g. they would get 20 A parts immediately but have to wait three weeks for the 20 matching G parts.
To solve this, the team were able to use simulation to replicate the palletizer’s process behavior and Monte Carlo analysis to calculate the best batch size and sequence of parts to use.
Using the simulation, they discovered that substantial time could be saved by loading a much larger batch of parts into the machine but to then also use this loading time to run an initial quality check on each part type.
2. Dealing with variation: use labels to control process times and route decision-making
Labels are a favorite feature of SIMUL8 users and will enable you to quickly build simulations of high-mix processes. A label is a smart way that SIMUL8 can apply an attribute to an individual work item in your process, allowing you to distinguish between different work item types and apply process behaviors accordingly. For example, if you wanted two different product types to have different cycle times for the same activity, or if you wanted to use the size, color or re-work status of an item to make pathway selections.
As work items can be given multiple labels, they also support decision-making for high-variation norms like customized product options, routing by sequence and express jobs.
Labels are a smart tool because they can be updated or edited as items progress through the process. This means we can track events in the simulation as they occur; how many items required re-work, how many cycles of a looping operation have been completed, or even the total count of waiting time that a customer has experienced.
This insight will help you to get HMLV process right first time. When dealing with low volumes it is critical to be able to trace and understand the movement of all items through the system. The visual and animated aspect of simulation enables you to easily track this flow and see any blockages or non-value add time in the process.
An example of how simulation has been used
In laboratory-based processes the smart label features within SIMUL8 are particularly useful for replicating the real-world rules and restrictions involved with dealing with samples and cases.
Let’s take an example of simulating a busy DNA evidence laboratory; we needed to split cases into sets of samples, where the number of samples would vary based on the number of evidence items collected.
Here, the team used batching out by label routing to direct samples to the correct laboratory processes. For example, samples requiring chemical treatment receiving this step but others bypassing it. Labels were also used to control the allocation of resources to cases; a rule in the system dictated that the staff member who processed the first part of a case could not be be involved in reviewing that case at the end of the process. By linking label selection to pooled resources in the simulation we could replicate this behavior to make an accurate assessment on how this restriction would influence laboratory lead times and throughput.
Read our ‘4 ways to improve laboratory workflows with simulation‘ blog post to see more examples of how SIMUL8 has been utilized for laboratories.
3. Consider a more radical approach: automation, outsourcing or streamlining
A lot of HMLV process have evolved to their current state through a series of incremental changes. For example, parts of the process that have been added on over the years to cope with a new product line or increased demand. When this approach starts to become stretched it can be difficult to convince stakeholders that just ‘adding another machine’ will no longer be as effective.
Improving the process then presents two challenges; how can you safely and accurately test out options that are radically different to the current state and once you have found a better approach, how do you satisfy and excite stakeholders to get them on board?
The three most popular strategies that we see from our users demonstrating a new approach with simulations are automation (AGVs, palletizers, robots, online forms), outsourcing (typically on Kanban items or for specialized operations) or streamlining (setting up a dedicated working cell for tackling the most common items types in the system or a particular bottlenecking activity).
In all these cases simulation is an effective way to walk stakeholders through the proposed changes to the process and visualize how to move from the current state to the future state.
An example of how simulation been used
As part of the development of the Ares 1, a crew launch vehicle that was being developed by NASA as part of the Constellation program, SIMUL8 was used to optimize the manufacturing of ‘Tuna Cans’ – the cylindrical sections that stack together to form the launch vehicle.
By simulating the current state process and trialling out the adoption of a number of possible options key improvements were found. The simulation demonstrated that streamlining the welding processes through mechanization offered significant time saving over the existing process.
Read the full case study to learn more about the project and the changes that were adopted by the team as a result of the simulation.
4. Test schedules against your exposure to bottlenecks and rework rates
Demand is not the only factor that will determine if a process becomes high-volume or not. For example, within aerospace there are many processes where demand is high but the processes remain low-volume due to the intricacy and time-consuming nature of the work, as well as rigid testing regimes and high failure rates.
In some cases you might need to start building 10 items to have a higher than 80% chance that one of these will make it through testing in time for scheduled delivery date. The million dollar question (often literally!) comes when there is a rise in demand for this item and suddenly everyone wants assurances that you can deliver X amount for a specific date – most typically, a date promised by the sales team!
Simulation can offer many benefits in these situations. Firstly, it will help you to quickly identify the bottlenecking or critical factor that is delaying the process flow. Secondly, it allows you to test out multiple options for solving this:
- Should we bring in more staff / teams?
- Can we improve the pass-rate and what would that mean for throughput?
- Can we rework failed parts and test those again?
- Should we prioritize reworked parts over new parts or will that extend lead times?
Thirdly, and in my view the most valuable benefit, if the results indicate that you will not be able to achieve target throughput in the allotted time (even with changes to the strategy) then the simulation serves as evidence to formulate another plan of action. Having the ability to spot this and feed back long before implementation is a distinct advantage of simulation. Although this might not always be the most welcomed news for stakeholders to hear, it places you in a far better position than to have reached the same conclusion days before the scheduled delivery date!
An example of how simulation has been used
The manufacturing and approval of a highly specialized aerospace product involved a number of atmospheric tests that were known to result in a high fail rate for this type of device.
Taking this fail rate and other process parameters (cycle times, staff levels, re-work times), the team produced a simulation which showed that with the existing setup they would be unable to hit schedule. Using this simulation to explore and cost alternative options, the team discovered that two measures, a small increase in staff and additional robustness to a specific test, offered significantly improved odds on achieving schedule. Presenting these options using the simulation and contrasting to the current state process helped to win decision-maker approval for additional resources to be implemented.
5. Provide ongoing process communication, management and documentation
The rate of process change tends to be higher for HMLV systems than with other production strategies. From bringing new product streams online, to introducing better tech replacing legacy machines, as well as ongoing Lean Six Sigma activity – these all represent changes to the process that need to be understood, documented and communicated to staff.
From shop-floor to boardroom, simulation is a highly effective tool for visualizing these changes to the process and engaging staff to draw out their input and questions well before implementation.
In my own projects I have seen first-hand the real benefits of this type of engagement. Using the simulation to walk through the current process, identify any issues and assess the proposed improvements helps everyone to fully understand the current process and the reasons for making changes. It is this shared understanding that really motivates staff and helps to frame the project as an opportunity for collaboration, rather than a top-down management driven exercise where operational staff might feel that they have had no involvement.
Simulations even offer a number of advantages at the end of a HMLV project. For process documentation, it is a concise record of the decision-making and process rules that manage the system. It can even be used for training new staff or for testing the quick what-if questions that tend to crop up day-to-day, for example, managing higher than normal volumes or developing coping strategies if key activities are suspended.
You can also share the simulation with other parties to improve other parts of the process. For example, passing it on to a supplier to allow them to better coordinate deliveries to your process timings, or sending the simulation to another plant within your organization to help replicate the same or a similar process. SIMUL8 Studio is a great new way of doing this online so you can quickly and easily share your simulation with anyone.
Finally, when that next project comes around, it’s always a great boost to already be starting with an accurate simulation of the process. You will find that keeping an ongoing simulation of your process to use for risk-free experimentation helps promotes a continuous improvement culture and keeps the discussion on new initiates as an active focus for the team.
An example of how simulation has been used
Upgrading an existing automotive line with overhead conveyors to a more flexible system using Automated Guided Vehicles (AGVs) proved to be a major undertaking for a team of industrial engineers.
They needed to identify the cost and sizing of the investment required (how many AGVs were needed, downtime required to install, if additional floor space would be needed) and provide evidence of the increase in throughput that would be possible with the new system.
Using SIMUL8 simulation software for the first time, the team were able to comprehensively meet both of these objectives, identifying the best implementation of the new system to achieve increased throughout, without overspending. As the team had developed the simulation using a flexible setup (using labels and internal spreadsheets in SIMUL8) it also was possible to use this as a baseline simulation where parameters could be easily adjusted to predict line behavior going forward as new car models came on-stream.
Learn more about SIMUL8
If you’d like to learn more about how you could use simulation to improve High-Mix, Low-Volume processes, get in touch with the SIMUL8 team today.