Lean Manufacturing 306

Process manufacturers are generally classified as predominantly continuous flow or batch. These two sides of the process world are more similar in terms of products than in terms of processes. Both continuous flow and batch manufacturing produce "stuff," i.e., a nondiscrete substance that may be subject to yield variations and loss of potent strength over time.

However, batch processes are usually time sensitive, and are paced by the speed of a batch reaction which may occur in a vat, vessel, or thermally jacketed reactor. Batch processes typically include operations like aging, fermentation, cooking, dyeing,

compounding, and baking. Important batch industries include dairy, pharmaceutical, rubber, cosmetics, beverages, and tobacco. By nature, there is more information systems processing overhead associated with tracking production batches, so batch manufacturers have been more willing in the past to adopt MRP II technology.

From an operations standpoint, one of the biggest differences between continuous flow and batch processing lies in the produc­tion scheduling function. To schedule a continuous flow facility, one is primarily concerned with sequencing products into a fixed process flow. Run lengths, changeovers, and sequences are all variables, but product proceeds linearly down a fixed path. In scheduling batch production, the process routing also becomes a variable. For example, in a dairy, a batch may be mixed in any one of twelve tanks in a mixing room. Different products may be produced concurrently. Capacity may be added by working a third shift, or contracting some of the work out to an external resource.

One of the popular practices used in batch processing today is least cost formulation. This is the ability of computerized optimi­zation packages to balance a number of variables—e.g., cost of raw material lots, cost of labor, capacity availability—and to recommend least cost formulations to production at any given point in time. Although this is desirable in batch processing, it is usually not practical in continuous flow processing where there is much less discrete visibility of individual lot costs, where routing options are limited, and where homogeneous product is pumped out 24 hours a day.

Continuous flow manufacturing is frequently confused with and mistaken for repetitive manufacturing, which is more closely associated with discrete production than with process manufac­turing. Continuous flow and repetitive are similar in terms of processes, but not in terms of products. Both methods use fixed production lines with predetermined routings and dedicated equip­ment. In repetitive factories, product is often described as "flow­ing" down the line as it is assembled.

However, repetitively assembled products are not commodities, but are physically discrete units that can be disassembled back into their component parts. In fact, product design—design for manufacturability—plays a key role in repetitive line arrangement and balancing. Such products are often ordered by customers from a catalog. A final assembly schedule may be used to assemble the end items out of semifinished options. The Kanban principle applies here, with the objective of producing items in lot sizes of one. In a mixed-model scheduled line, you might see different products alternating every one or two positions on the line. This is clearly beyond the physical capability of a continuous flow processing plant.

Types of operations commonly seen in repetitive plants include stamping, punching, wiring, plastic injection molding, assembly, painting, and packaging. Important repetitive industries include automotive, consumer electronics, shoes, and household applian­ces.

One final distinction can be made in terms of key operating concerns. In continuous flow manufacturing as well as batch processing, the primary focus is on product yield. There is no guarantee that ingredient lots will yield a predictable quantity or type of end product. The history of process control technology has been geared to improving yield and minimizing yield loss. In repetitive manufacturing, component items and processes are more reliable and consistent, so product yield isn't nearly as much of a concern. More attention has been focused on labor yield— namely productivity and efficiency measurement.