English Japanese Spanish       

 

 <Figure #1>

This queue of the next scheduled tasks comprises the buffer (Figure #1). Some of the buffer will (or should) reside at the resource ready for processing at all times. It's called the buffer because it "buffers" the constraint against fluctuations in the system. As the name implies, fluctuations at upstream resources will cause the queue in front of the constraint to fluctuate.

 

<Figure #2>

As time passes, the constraint depletes the physical queue and the feeding resources replenish the buffer. It is a revolving stock of work for the constraint — the content is always changing (Figure #2). Since the components of the buffer represent available work for the constraint, it is referred to as a time buffer. The measurement of the buffer is always time.

This planned buffer content allows users to predict when jobs will arrive at the constraint. More importantly, it helps users determine when tasks are not at the constraint. These missing tasks, or holes in the buffer, report when and to what degree there are fluctuations in the system feeding the buffer.

Quantifying Variation

The buffer guards against the process variation preceding the buffer origin (in this case, the drum). Since variation is always present, the entire planned buffer physically present at the constraint resource is never seen.

Ideally, the actual buffer should contain 50% of the planned buffer (work ready to be processed at the drum). If there is more than 50%, the buffer is too large — the fluctuations in the system are not great enough to warrant the investment in inventory.

If there is less than 50%, the buffer is too small. The throughput of the system is in jeopardy. The risk of the constraint being idle increases as the physical portion of the buffer decreases.

Predicting Problem Areas & Responding

When the buffer falls below the ideal, it's a signal that there is greater than anticipated variation in the system. When this occurs, it is a trigger for action. Depending on how deep the penetration of the incursion into the buffer, the next action should be to check status, expedite, or reschedule. This “early warning” allows action BEFORE the constraint being idled and the subsequent loss of system throughput. In this regard, the buffer is an early warning system for the entire system, warning of potential problems in output.

In order to track disruptions, these buffer incursions are recorded in a database.

When there are penetrations into the buffer, the missing task (batch) will be found at the resource causing the penetration (Figure #3).

<Figure #3>

Resources that are on their way to becoming  bottlenecks will cause incursions in the buffer more frequently.  The Buffer Management and Diagnostics Module tracks these penetrations and their source over time to identify patterns and trends. (Figure#4)

<Figure #4>

Tracking these buffer incursions gives management the information necessary to identify the most troublesome resources and take the appropriate action.  This then allows reduction of the buffer size and subsequently the inventory and lead times.