Production time at workstation can be divided into following categories:
- Transfer time
- Queue time
- Setup time
- Process time
- Wait-to-bath time
- Wait-in-batch time
- Wait-to-match time
Tranfer time is the time job spends moving from the previous station. Queue time is the time job spends waiting to be processed or moved. Setup time is the time job spends waiting for the workstation to be set up. After this comes the actual process time. Then there is wait-to-batch time that means the time job spends waiting for other parts in that batch to be finished. Wait-in-batch is the time job spends in process batch waiting its turn to be processed. Wait-to-match is the time job spends waiting its buddies in order to be able to start assembly.
Only the process time is the time that something is done for the part. Besides that the transfer time is the only necessary evil that will occur even though workstations are close each other. Other times are waste (muda) in the JIT point of view and they are often combined into one waiting time. For the better examination they are divided here as specific as possible.
Previous post already discussed some of the lot size related things so we will now focus on other factors.
Almost all manufacturing includes assembly. Parts are inserted into circuit board or bigger machines are assembled. Assembly can not happen if we don’t have all the parts. Waiting for parts to match adds waiting in product flow.
Error in one production line can affect other lines too if assembly can not be done and WIP storages are exceeded in assembly. We can’t produce more parts so we have to stop other lines too.
When purchasing parts from supplier it’s challenging to coordinate them efficiently. Simplest way would be to purchase all the parts from the same supplier when they would be delivered at the same time. But usually we have to deal with different suppliers when their different reliability and variability means that we won’t have parts at the same time.
That’s why we have to schedule safety margin into assembly. Example product is made from two parts which lead times are 4 and 6 days and assembly takes another 4 days. All times with the service level of 95 %. We could say that lead time to the customer is 10 days. But in reality our 10 day lead time is not with 95 % service level because we did not consider wait-to-match time.
Reducing Cycle Time
In most systems the actual process and transfer time is only 5-10 % of the cycle time. So there is possibility to reduce cycle time. Here is the most common places to start at.
Queue time is caused by utilization and variability so it can be reduced by decreasing utilization. Specially by increasing the speed of the bottleneck station. Another way is to reduce variability in process times or incoming speed. Specially the incoming rate of the bottleneck station.
Wait-to-batch and wait-in-batch times are dependent on the batch size. Batch size can be reduced example by decreasing the setup times.
Wait-to-match time is caused by bad synchronization in product manufacturing or purchasing. This can be improved in similar ways than reducing the waiting time. Another way is to control the time that different products production is released.
Station overlap time is the time that certain batch spends similarly in different work stations. Example from the batch of 100 parts first 50 parts are produced in machine 1 and then taken into the machine 2 at the same time that the rest 50 parts are processed in the machine 1. Unlike the other times this is meant to be increased! This can be done by using lot splitting anytime it is possible. Example using smaller transfer batches than production batches.
Make-to-order manufacturer had the practice to take orders in 2 week periods and lead time was 10 week from the end of the period. This caused that most of the orders were made at the end of the period which caused overtime for sales people and difficulties for production scheduling.
Problem was that competitors offered products with 5 weeks lead time from the order day. In addition service level of the example company was only 70 %. Certain production functions were made in big batches. Biggest problem was long cycle time that was caused in the name of factory physic laws from big variability and utilization.
First of all the demand is unnecessarily variable because of the practice used. By moving into “live” orders we can smooth over the demand. Secondly the squared coefficient of variation (c2e) for one work phase was 7. This was caused by utilization and it was tried to reduce by decreasing setup times.
Major factor for variability was lot sizes that were caused by handling and production. Workstations were far from each other so the transfer batches were big, typically full containers. Because of the long setup times the production batches were big. Target was to organize the production so that it could be possible to use smaller transfer batches and reduce setup times.
Furthermore in case of most of the common products it could be reasonable to stock products and make them with MTS model in stead of MTO.
Generally speaking production batches for metal parts are usually big so the make-to-stock is often good practice. Inventory can be reduced with design that takes use of as much common parts as possible. This way we can make common parts to stock and then assembly-to-order (ATO) them from common and product specific parts.