Modelling

As part of your continuous improvement efforts, you will need to experiment, and that means making models. Models are essentially, at their heart, experiments to figure out how to do things on a larger scale. They are intended to get all the mistakes and faults in ideas out before you invest in a full-sized change.

There are several different kinds of models that are available to you. Bear in mind that this is the layman’s terminology specific to your kaizen activity. There are more generally accepted lists of models in the scientific community, but those are not relevant to your CI efforts.

• Mathematical Modelling: Using math to calculate out whether something will meet your goals.
• Scale Models: Building small replicas to learn the best way to configure the finished item
• Full-size Prototyping: This is the building of a life-size, working replica. The main difference between it and the finished product is the roughness of it, and the lack of durability.
• Layout Modelling: This is an effort to make sure that your new equipment will fit. It can be done simply by moving shapes around on a piece of software, or more complex by laying cardboard cutouts in a parking lot to see how the flow actually feels in real life.
• Pilot Programs: This probably pushes the envelope of what a model is, but it fits the bill in that it is experimental to find answers before implementing an idea fully.
• Other: Clearly, this list is not exhaustive. Any other type of model you can think of is fair game.

Why Do Modelling?

The biggest reason is cost. If you make a mistake with a model, you are out a bit of money and the time of the team. If you make a mistake on an actual implementation of something and mess up, the production line could shut down, or worse. The impact on customers and the resulting loss of revenue, or even the cost of fixing problems, can be massive.

So, models are risk management tools.

When to Do Modelling?

Modelling should be done whenever the risk of failure is high, when you don’t know for sure what to do, or if you aren’t positive that your solution will work.

On small, just do it projects, there is no need to model. I like to use the garbage can example. If you think the garbage can should be in a different location, move it, or buy another one. If you are wrong, there is very little lost. If you are right, things get better. The cost of modelling that would greatly exceed the gain.

Types of Modelling

Mathematical Models

There are some robust simulation software packages available that can do advanced simulations of flow. Excel also has a solver tool as an add-in that can crunch numbers for you and come up with solutions to optimization problems.

In Excel, you can also just us a randomizer function and manually refresh to see how things look.

When is this useful? Let’s say you have a problem with line stops and want to see where you can get the best bang for the buck. You could observe the cycle times of your 10 stations and come up with an average and a standard deviation for both.

Put that into a spreadsheet that uses random numbers to simulate each station, and then adjust the inputs. See if dropping the average cycle time in station 2 reduces the line stops more than cutting the variation in station 6. Or see what happens if you add a station and move some work around. You might find that the cost of one extra person would cut out a lot of delays, and the related overtime.

The point is that you don’t have to guess when you can math things out.

That basic stuff is possible with core Excel functions. Learn solver, and you open up more options. Get one person in the company up to speed on a good software simulation package, and soon the demand on him or her will be spiking. Or you can even invest in outsourcing expertise now and again. The point is that sometimes knowing an answer (nearly) for certain through math can save you a lot down the road.

Just for clarity, what can these sophisticated simulation models do? They can track your parts flow through a series of fabrication machines, and figure out the optimum routing and numbers, or can game out ‘what if’ scenarios about changeover time reductions and things like that. With a good model, you can play around with moving things with very low risk.

Scale Models

I’ve worked in the past with a ‘moonshine shop’ that experimented all the time and had piles of failed models and other toys in a boneyard that they could pick through for future experiments.

They also had several of their experiments make it to the production line.

Scale modelling requires materials, though, and expertise. A moonshine shop will be set up with both, but that doesn’t do much for your project teams.

It is a good idea to create a modelling kit for use by your kaizen teams. This should include a variety of materials and tools. You can provide clay, foam, cardboard, erector sets, popsicle sticks, and whatever else you can think of. Put in hot glue guns, tools sets, tape, Velcro, and a variety of nuts and bolts.

I’d also add in a manual with best practices on it. The point is to take that load off the inexperienced teams. Don’t make them chase down items just to get ready to do what you have assembled them for.

Full-Sized Prototyping

Life-sized, working prototypes provide proof of concept and make sure that the machine does what it is supposed to do on actual parts.

This doesn’t have to be complicated. On a project team, I worked on, there was a problem with using an automatic grease gun to pack some bearings. The grease would drip as the tool was moved around and would make a big mess. Plus, the hose and power cable were unwieldy.

During the project, I modelled up an L-shaped slot on a large piece of PVC pipe, screwed it to a 2×4, and clamped to the table near the hub fixture. The grease gun could be rotated out of the slot, slid 4 inches to where the hub was in the fixture, and returned with minimal mess and no cord management issues.

Despite the intention of being a prototype, the tool was still there years later in its original form. But the presumption was that it if didn’t work perfectly, or if it wasn’t durable, it would have…

Extended Content for this Section is available at academy.Velaction.com