This white paper explains the meaning of Lead Times in instantMRP/MRPlitePlanner, and the relationship between Lead Time and Capacity Analysis. Keep in mind that instantMRP is included in MRPlitePlanner(and is simply called MRP in MRPlitePlanner). MRPlitePlanner adds Capacity Planning to instantMRP.
Lead Time is a parameter held against every part number.
Its purpose is to predict when an order to make or purchase that part will be delivered. You must think of your manufacturing process as a sausage machine. If I throw a manufacturing order into the sausage machine (i.e. release it), it probably won’t start right away due to previous orders. Even after it starts, it will spend time travelling between operations, then queuing up at each operation with other orders. Eventually the last operation will be completed, and it will emerge from the sausage machine.
Each part goes through a particular sausage machine. Fabricated parts go through fabrication operations. Assembled parts are made on assembly lines. Therefore Lead Time is a characteristic of the sausage machine the part must go through.
Lead Time must be long enough that you can be pretty sure that it will emerge from the sausage machine within the lead-time. Typically this is based on experience.
Lead Time is NOT the bare time to make the part. This is sometime called the ‘touch time’, or ‘value add time’. Often, the bare manufacturing time is only a matter of minutes, but the time for a manufacturing order for a lot or batch of that part to go through the sausage machine is counted in days. Lead Time is NOT the few minutes of bare time; it is an assessment of the number of days needed for a typical order to clear the sausage machine.
The hidden assumption is in those last few words above: “the number of days needed for a typical order to clear the sausage machine”, the key word being “typical”. The assumption is that a given manufacturing order will be thrown into its sausage machine with a good number of others, and this ‘load’ on the facility will statistically be fairly stable, so that a prediction of when that manufacturing order will emerge is itself fairly stable.
If the manufacturing order has a large load on the facility, and tends to monopolise the capacity of the facility, then the prediction of when it will complete is objectively a matter of when it can be scheduled. The concept of a ‘typical’ lead-time may not be relevant. In that case, MRP, with its dependence on the sausage machine model, may not be the best tool. A Finite Capacity Scheduling tool may be more relevant.
MRPlitePlanner’s Capacity Analysis is a means of indicating potential overloads released into the sausage machine, in aggregate. It should not be confused with a scheduling system, which attempts to calculate exactly when each operation will start, and its duration. We will try to clarify this distinction.
Capacity Analysis takes over once MRP has applied lead times. What has MRP done? To make a product, normally it is done in stages – fabrication, sub-assembly, final assembly. In process industries, this can be – process ingredients, mix, ‘cook’, bottle or package. MRP uses lead times to space out the stages. The calculation ends up with given manufacturing orders to be started at given dates.
A bunch of manufacturing orders starting off on a certain date is injecting a load into the sausage machine. Let’s say we are working in weekly periods and we have capacity of 500 hours of Machine Type A each week. If we release orders with a load of 1100 hours, then Capacity Analysis will show this overload for the current time period.
This can only be interpreted as a Warning, not an Actuality. It is possible that last period’s load released into the sausage machine was light – say 200 hours, and next period’s may also be light. Therefore the load spread over a number of periods could average out to be within capacity. This may be OK for some factories, but may not be acceptable for others. This is why we say that a Capacity Analysis overload is a Warning, which may or may not be a problem in the context of the bigger picture.
This point is best illustrated by an example. We have to release a manufacturing order OrdX for part A. Part A is made in three successive operations P, Q and R. Lead time is three weeks. When Capacity Analysis calculates the load arising from OrdX, how does it apportion the load over the 3 weeks? The answer is: it doesn’t know exactly in the 3 weeks when operations P, Q and R take place, so it apportions the load from P, Q and R all in week 1 of the 3.
Therefore an overload must be interpreted as a load released into the sausage machine, not as predicting precisely when the facility will be affected.
It’s a bit like saying a heavy snowfall has hit the hills, and we can expect the river to be swollen soon, but we don’t know exactly where, or when, or will it come in a wave or swell then retreat over an extended time.
A parameter called Lead Time Offset can be applied to an operation to move the timing of its load within the lead time. In the above example, it could be used to time |P’s load in week 1, Q’s load in week 2 and R’s load in week 3.
Scheduling is when you work out the precise timing of each operation – when exactly will OrdX be made on Machine A. It might be thought that if you use scheduling, you can predict the exact timing of overloads. Capacity Analysis doesn’t employ scheduling because it requires more data and adds complexity to the task, when the current approach is sufficient for the purpose. Scheduling is appropriate when a large, expensive resource is a bottleneck and specific orders will occupy the capacity of the resource over an extended period, but that kind of scheduling is not in MRPlitePlanner’s scope.