June 2007 Vol. 138 No. 6 Continuous Processes Can Be Lean http://www.sme.org/cgi-bin/findarticles.pl?&ME07ART42&ME&20070610&&SME&#article Although developed for use in producing discrete parts, lean manufacturing and six sigma can also be applied to continuous-process manufacturing SiewMun Ha Consultant Implementation Services LLC Highlands Ranch, CO E-mail: siewmunha@yahoo.com To order a hardcopy reproduction of this article, click here. To purchase digital reprints or reproduction licenses, please contact the resource center at service@sme.org or call (800) 733-4763. Process manufacturing is fundamentally different from discrete manufacturing: material flows in a continuous stream in the former, while parts move in discrete batches in the latter. The breadth of work on process improvement in discrete manufacturing is vast. And it seems very logical to apply popular existing methodologies tailored to discrete manufacturing, unaltered, to the process-manufacturing industries. This approach, however, is akin to fitting square pegs into round holes. A better approach is to adapt lean techniques developed for use in discrete production within a new processimprovement framework for process manufacturing. The new framework identifies the various forms of waste in the process-manufacturing value stream, and manages the wastes with the appropriate concepts and tools. The two most popular process-improvement methodologies in use today—lean manufacturing and six sigma—originated at Toyota and Motorola, respectively. These pioneering companies were discrete manufacturers, and the subsequent evolution and development of the two methodologies has focused mostly on improvements in discrete manufacturing. Much less development work has been carried out on applications of lean and six sigma to process manufacturing. How is process manufacturing critically different from discrete manufacturing, and why do the differences matter to the processimprovement approach? The key characteristic of process manufacturing that distinguishes it from discrete manufacturing is material movement. In this type of manufacturing, material is processed continuously through a series of machines from raw material to final product. It flows in a continuous stream from one machine to the next, without periods of stopping and waiting in between. Some examples would be iron-ore processing, petrochemical refining, paper manufacturing, and sugar refining. In contrast, the fundamental characteristic of discrete manufacturing is that material and parts move in discrete batches between machines, with the very real possibility of work-in-process (WIP) buildup between processing steps. Lean manufacturing and six sigma each has a central focus that has been the basis for its structure and tools. For lean manufacturing, it's the delivery of value to the customer through the elimination of waste, where waste is defined as anything that is non-valueadded from the customer's perspective. In the case of six sigma, the central focus is the elimination of defects, where a defect is defined as a part or service that does not conform to the customer's specifications. How do these concepts apply to process manufacturing, when you consider its differences with discrete manufacturing? Lean manufacturing defines seven types of waste that make a production system "unlean" and inefficient at delivering value to the customer. These are: Over-production: Producing too many parts, too soon. Inventory: Extra parts required to buffer process variability. Transportation: Movement of parts without adding value. Waiting: Increasing lead time without adding value. Movement: Movement of operators without adding value. Defects: Parts that do not conform to customer specifications. Over-processing: Processing a part more than is necessary to meet customer specifications. The first four types of wastes relate to a lack of material flow in the lean sense, a major problem at many discrete manufacturers. Here, parts and material are processed, transported, and wait in large batches, instead of moving smoothly in a one-piece flow from raw material to finished product. In process manufacturing, however, the very nature of the process dictates that the material already flows (literally!) from one machine to the next. Hence, the lean ideal of flow occurs by default. Also, there is typically little or no WIP between machines in process manufacturing. What WIP that exists tends to be of the order of hours of production rather than days or weeks of production, as is usually the case with discrete manufacturing. As a result, over-production, inventory, and waiting are either nonissues or only minor issues in process manufacturing. If the plant is suboptimally laid out, transportation is potentially an issue in process manufacturing, as material must be moved over greater distances than absolutely necessary. Suboptimal transportation increases the extent and complexity of the conveyance system (e.g. pipes, conveyors), which then requires increased investment and maintenance—waste. In reality, however, the nature of the machinery in process plants is such that any movement of machines to optimize material flow is usually difficult and expensive, and is consequently a nonstarter in most process-improvement efforts. Movement waste is also less relevant to process manufacturing. Operators in these plants typically monitor automated machinery, and even if they have to move from one piece of equipment to the next, that movement usually does not have an adverse impact on the ability of the equipment to continue processing the material. In discrete manufacturing, on the other hand, operator involvement is required to process a part on a machine. The last two types of waste—defects and over-processing—exist in process manufacturing just as they do in discrete manufacturing. Defects in process manufacturing result in the production of material that does not meet the specifications of the downstream internal/external customer. In iron-ore processing, for example, the ironcontaining rock must be crushed to a certain size before further processing can begin to separate the iron from impurities. If the rock is not crushed to the specified size, then the downstream separation process suffers from degraded performance. Over-processing occurs when the material is processed to a greater extent than is required by the downstream customer. In the iron-ore processing example, over-processing occurs if the ore is crushed more finely than required, with no commensurate benefit to the downstream separation process. This over-processing unnecessarily consumes crushing capacity and energy, and thus constitutes waste. Six sigma's focus on defect reduction/elimination aligns well with certain types of waste that exist in process manufacturing. It is, however, not the complete solution. While the six-sigma tool set is very powerful and works well to optimize process performance with respect to quality, throughput, and efficiency, the methodology fails to address other root causes of waste in process manufacturing. The fully successful process-improvement effort thus takes advantage of a complete armament of techniques, and applies the ones best-suited to achieving the desired results. What might some of these root causes be? To answer that question, it's instructive to analyze the process-improvement opportunities in process-manufacturing from first principles, rather than trying to force-fit the opportunities into the structure of lean manufacturing and/or six sigma. We do so by visualizing a perfect process-manufacturing value stream, and then identify improvement opportunities by mapping the gaps between this perfect value stream and its real-world counterpart. What would such a perfect value stream look like? The perfect process value stream is one where all the machines are: Capable: Able to produce product within the quality specifications of the downstream internal/ external customer. Available for production with no unplanned downtime. Efficient: Consumes no more energy and raw material than absolutely necessary. Adequate: Possess sufficient capacity to meet demand. With these characteristics, the perfect value stream produces good-quality product reliably, efficiently, and in sufficient quantity, at the individual machine level as well as for the value stream as a whole. A suboptimal condition with respect to each of these four ideal characteristics constitutes a type of waste. Because the perfect value stream is an idealization, all real-world process-manufacturing value streams will contain one or more of these wastes, representing targets of opportunity for any process-improvement effort. The root causes of these wastes may be identified by means of fish-bone (Ishikawa) diagrams. A putative first-level fish bone for machine capability is shown in the illustration on page 108. Ishikawa diagrams are not meant to be a complete and exhaustive identification of all causes of waste, but rather a starting point for further investigation. You can use such diagrams as "straw models" to identify the root causes of wastes in your specific value stream. As befits a root-cause analysis, each of the firstlevel causes under Man, Machine, Environment, Method, and Material may be deconstructed, as appropriate, into successively higher-level causes, until the root cause has been found. At that point, the appropriate methodology/tool can be applied to address it. While the causes of process-manufacturing waste are many and varied, a few occur with regularity across all or most of the four categories of waste tracked in Ishikawa diagrams. First and foremost is mechanical condition. This refers to machines that are at a suboptimal state of maintenance, a very common condition at many process plants. Machines in poor mechanical condition have poor availability, produce poor-quality product in inadequate quantities, and operate inefficiently. In short, they operate wastefully! A maintenance kaizen event is the appropriate process-improvement tool to return the machine to an optimum mechanical condition. To sustain the improvement, a long-term maintenance program such as Total Productive Maintenance (TPM) must be installed. A second major cause of waste is suboptimal operation. Process manufacturing typically involves a combination of physical parameters such as temperature, pressure, density, flow rate, moisture level, and chemical concentration that are set at the machine to process the material. If the settings of these physical parameters are suboptimal, then the machine also operates suboptimally in terms of throughput, quality, and efficiency. To fix this problem, we need to identify the optimum settings at which to run the machine. Optimization problems of this nature are ideally solved using the six-sigma methodology and tool set, employing designed experimentation to characterize the behavior of the machine with respect to its settings. Design and technology are another two major causes of waste. It's not uncommon to find machines at brownfield process plants that have been in use for many years, if not decades, that are obsolete with regard to both design and technology. Such obsolete machines operate wastefully in terms of availability, quality, throughput, and efficiency, much like a machine that is in poor mechanical condition. Improving machine design/technology is an engineering problem requiring technical analysis and designed experimentation. In addition to the characteristic of material flow, another distinction between discrete and process manufacturing is that changeovers are often not an issue in the latter, as most production lines only produce one type of product. Availability can still be adversely affected, however, by long setup times after a machine has been taken down for maintenance. In this case, quick changeover techniques such as Single Minute Exchange of Die (SMED) may be applied to reduce setup times and improve availability. If the overall management objective is to increase throughput, then the Theory of Constraints (TOC) can be a useful device to focus and guide the process-improvement effort. Briefly put, TOC posits that in any production system, at any one time, there usually exists only one constraint that limits the overall system throughput. We increase overall throughput by improving the performance of the constraint. Improvements elsewhere make no difference. As the existing constraint is lifted, a new constraint appears at another location, and we repeat the process to achieve additional throughput improvements. TOC enables us to focus our scarce resources on the critical part of the value stream that will make a difference, instead of diluting our efforts where they will not. With regard to process improvement, process manufacturing differs from discrete manufacturing in critical ways. Applying discrete-manufacturing process-improvement methodologies without modification results in disconnects and gaps in the improvement effort. Instead of blindly applying methodologies in ways that are more suited to discrete manufacturing, a more rewarding approach is to realign them to a framework that accounts for the unique nature of process manufacturing, identifies the specific wastes, and applies the appropriate tools to tackle them. To order a hardcopy reproduction of this article, click here. To purchase digital reprints or reproduction licenses, please contact the resource center at service@sme.org or call (800) 733-4763.