Interrupts
In real life, studies would be conducted to determine the types of interrupts that happen on a line as well as their uptime and downtime distributions. For the example, we can be a little creative and imagine some interrupts to fit our scenario.
First, lets create a belt failure interrupt on both mixers. This will be a cumulative failure, because wear on the belt will increase whenever the line is running, and it will not reset whenever the line goes down. We will say that this interrupt has a normal uptime distribution with a mean of 120 minutes and a standard deviation of 5 minutes. Since it happens so frequently, the operators have gotten pretty good at fixing it. Therefore, the downtime distribution will also be normal with a mean of 5 minutes and a standard deviation of 1 minute.
Next, let's add a competing interrupt to the molding machines. Let's say that there is a very slight difference in speed for the actual molds and the conveyor which the cookies ride on. Given enough time, these can get out of sync enough to stop the line; however, the molds and conveyor get resynchronized each time the line goes down. As said, this will be a competing interrupt. For the uptime distribution, we will again use a normal distribution. This one is a bit more uncertain with a mean of 60 minutes and a standard deviation of 15 minutes. Usually, fixing it just involves restarting the line, which is nearly instantaneous. For this reason, we will assign the downtime a Weibull distribution with a lambda of 5 and a k of 1 (the default settings).
There is one more type of interrupt that we need to demonstrate. For an example of a wall clock interrupt, we will use the quality control step. The union that the members of the quality control team are a part of mandates that these employees must take a 15 minute break every 4 hours. To model this, we will set up a wall clock interrupt with fixed uptime and downtime distributions of 240 and 15 minutes, respectively.
For simplicity, we will delete all interrupts from the remainder of the steps using the "Delete All" button at the top of the screen.
Once you are sure your model has the correct interrupts, including duplicates for the two baking lines, we are ready to begin running experiments and viewing the results.
Last updated
