What is a bottleneck?
A 'bottleneck' is a stage of the production chain (a sub-process), that constraints the throughput of the whole process. Its capacity is smaller than the capacity of the up-stream and down-stream sub-processes.
When a process has no variability (as in bottling lines) there are no bottlenecks. However, in the presence of variability bottlenecks constitute an inevitable fact. It is the nature of the beast!
Bottlenecks appear every time there is a process composed by Dependent Stochastic Events . . . as in mining!
There is a number of reasons why there are bottlenecks in a production chain:
- Inadequate design/planning
- Unplanned events that constraint the production
- The inherent variability of the processes
Are bottlenecks a bad thing?
We use to think that bottlenecks are a bad thing. Often the word 'bottleneck' has a negative connotation in our minds. However, in production systems the 'bottleneck' plays an invaluable dual role.
- it sets the production pace, and
- it allows the manager to program the operations by subordinating everything to the bottleneck.
Should we eliminate the bottlenecks?
Mining operations are not bottling production processes. Inherent variability in each sub-process (e.g. loading, unloading, transportation) is present all the time.
When trying to achieve the goal of 'no bottlenecks', what happens is that 'floating bottlenecks' appear. Managing moving/floating bottlenecks in mining is a nightmare!
A simple example will help us to understand this: let's simulate a process line composed by four sub-processes. Processing times in each sub-process is simulated using dice. Since the variability in this case is big (1 to 6), a bottleneck will appear in the line. Since capacity and variability are equal for all the 4 stations (1 die), you will see that the bottleneck moves from one sub-process to other over time: 'floating bottlenecks'.
This proves that trying to 'balance the capacity' in en each of the sub-processes to 'eliminate the bottlenecks' is futile. The reason: variability.
The good news is that mining operations are easier to manage when having a bottleneck. Indeed, the manager can synchronize the operations by planning for the 'bottleneck' and subordinating the rest of the processes to it.
In summary, the bottleneck CONTROLS the system's flow.
Where to locate the bottleneck?
I think that the location of the bottleneck is a strategic decision.
For example, if the decision is to have the truck fleet as the bottleneck, it will become very difficult to synchronize the operations to achieve a smooth production.
Some argue that the bottleneck should be located at the operational point where the capital cost to rectify that bottleneck becomes uneconomic. Most mines will got through a series of operational modifications to increase the bottleneck capacity until such a point is reached.
An expert told me once: "The only commanding process at the mine should be the processing plant capacity. Everything else should be driven by that capacity" . . . and I totally agree!
I often suggest managers should locate the bottleneck in the crusher at the entrance of the processing plant, and program the rest of the sub-processes in a way that the crusher is fed as smoothly, uniformly, and continuously, as possible.
The reason is simple: since there are always other operations inside the mine which are different from the ore feed (e.g. site preparation, stock), you will always have extra resources when needed (e.g. trucks).
Should all processes in the mine be busy all the time?
Having all resources busy at all time should not be the goal. The goal should be having the bottleneck busy all time!
Since by definition all other resources/sub-processes in the mine have more capacity than the bottleneck, if you keep them busy all the time, no synchronicity will be achieve and lines/overstock will appear all over the place.
How to administrate the bottleneck?
Managers should put emphasis in augmenting the capacity of the bottleneck, and reduce its variability. To reduce the variability Six-sigma & TOC tools work at its best!
Remember what Dr. Eli Goldratt said: "an hour lost on a bottleneck is an hour lost to the whole plant".
On the other hand, if you increase the capacity of a resource up-stream from the bottleneck, you will get over-stock.
Finally, if you increase the capacity of a sub-process downstream from the bottleneck, nothing will happen!
A VP of Operations complained that "we doubled the transport capacity, but we are loading dirt onto the trucks". Well, in his case the transport capacity was not the bottleneck . . .
In summary:
- Decide where to locate the 'bottleneck'
- Program the operations in such way that the 'bottleneck' is working smoothly at maximum capacity
- Take measures (using continuous improvement techniques) to reduce variability and increase capacity in the 'bottleneck'
Planning the system's flow is simpler if you 'plan the flow at the bottleneck' and make the whole system to be synchronized with it!
Eli Goldratt describes this process as 'Drum-Buffer-Rope'. I can summarize it as follows: Let the bottleneck set the pace of the production and let the rest of the sub-processes work full capacity ONLY when it is required by the bottleneck. The rest of the time they should be ready to produce but idle in order not to waste resources and generate overstock. This is know as the road-runner rule.
Can Computer Simulation aid in planning the operations?
Totally! SmartSimulation has developed a computer simulation program to help managers in planning mine operations and test different scenarios. Computer simulation becomes a powerful tool to analyse different production scenarios, and to decide where to locate the bottleneck.
More information can be found HERE.
0 comments:
Post a Comment