March 5, 2001
Merely staying busy is not a determinant of success.
The most popular book among steel fabricators—other than the Manual of Steel Construction—is probably The Goal by Dr. Eli Goldratt.
This groundbreaking work, cast in the form of an easy-reading novel, uses everyday events to introduce the concepts of the author's theory of constraints (TOC). TOC is an innovative management philosophy that encourages a shift in focus from local efficiencies to the critical chain of events required to complete a project. Merely staying busy is not a determinant of success, according to Goldratt. To truly improve your throughput, you must concentrate your resources on the bottlenecks and keep your noncritical functions ready when needed.
The following illustrates how fabricators can benefit from the application of TOC through an actual structural steel plant example. The facts, though, have been changed a little to protect the company's anonymity and clarify the point.
This shop recently had finished a large contract and had a bit of a lull before the next major project began. Naturally, they were anxious to get production started quickly—as soon as the first drawings were approved. Because of problems in the past, the project manager ordered the detailers to do the repetitive piece parts before releasing the drawings of the assemblies. Without the detail parts, obviously, the fitters could not put the assemblies together.
Once the detail part drawings hit the shop floor, they were immediately started through processing. Because the shop had been low on work for some time, the prep shop foreman could choose whatever parts to produce on each machine. For efficiency's sake, he chose the largest batches of required parts and started them first. A large quantity of identical baseplates were started on the plate duplicator, and the angle line began cranking out clip angles.
Meanwhile, the production manager was anxious to get the fitters and welders working, so he released a batch of column drawings to the saw. Soon, a batch of wide flanges were running through the beam line. Because the prep shop was busy cutting baseplates, however, none of the cap plates had been cut. Thus, none of the columns could actually be assembled.
Next, the production manager released a batch of beams drawings to the saw. There were plenty of clip angles ready to go, but because the plate duplicator was still cutting baseplates, none of the stiffener plates had been cut. Thus, the beams could not be assembled either.
At this point, the saw and the fitters were standing idle. The production manager released another batch of drawings, just to get something into production. By the time this author saw it, the beams were piled head high between the beam line and the fitting stations. There also were piles of baseplates and clip angles waiting for fitting, but no other detail parts.
It was easy to identify the bottleneck in this shop by applying the theory of constraints. Everyone was waiting for the plate duplicator to produce some parts. Yet, was this really the constraint? The plate duplicator was one of the fastest machines in the plant; it simply was working on the wrong parts. In an effort to gain maximum efficiency, the prep foreman had the duplicator operator make all of the baseplates in one run, even though some of those plates would not be required for several weeks. Meanwhile, the entire shop was waiting for that machine to process other parts before the assemblies could be finished.
For better work flow, the production manager should have reduced the batch size of the baseplates to produce just enough for the first few days. Within hours, the plate duplicator would have been free to produce the cap plates and stiffeners needed for the first batch of assemblies, which could have been fitted immediately.
So, where was the actual constraint? As the job proceeded through the shop, it became obvious that the real bottleneck was in fitting. Even when the prep shop produced adequate quantities of the right parts in the right order so the fitters never had to wait for them, the backlog of beams did not diminish. The fitters simply could not put the assemblies together as quickly as the CNC beam line and the other prep shop machines could produce the parts. In fact, the huge pile of beams hampered the fitters' ability to assemble finished parts because they had to move the beams around to find the ones to work on next.
Goldratt uses a concept he calls "drum, buffer, rope" to help control this process. The drum is the speed at which the slowest process can operate. In this case, that constraint is the fitters. The plant can produce only as fast as the fitters can produce; therefore, everyone else should proceed no faster than this pace. Having the beam line or the prep shop produce parts faster than the fitters can work on them offers no advantages, and can be a disadvantage.
The buffer is the backlog in front of the constraint—here, the stack of parts in front of the fitters. An adequate buffer is important, because the constraint never should have to wait on another process before it can produce.
The rope is the control valve at the front end of the production process. Work should be released into the process only as quickly as the constraint can process it, with an adequate buffer. If an oversupply of materials are pushed through nonessential machines, the production process will bog down with backlogged parts.
What this means is that, contrary to conventional wisdom, the sawyer should not necessarily cut beams as fast as he can. He should cut more parts only when the fitters signal that they are ready for more—that is, when they pull on the rope. Instead of a huge pile of beams in front of the fitters, the buffer should be just big enough to ensure that the fitters do not run out of work. Since the sawyer no longer is busy all the time, he can use his free time to help organize the beams for the fitters to save them time.
The goal of production is not to maximize the speed of each individual machine. It is to produce finished parts as quickly as possible at the end of the line. Because a plant can produce no faster than its slowest process, the other processes should and must be idle at times. These processes would be inefficient anyway, stockpiling undemanded parts. The trick is to use their slack time to improve the performance of the constraining process.
The theory of constraints offers practical, measurable methods for increasing throughput while reducing work-in-process. Its widespread application in the steel fabrication industry has the potential to vastly improve shop floor performance.