What Does Automation Cost?
Calculating total life-cycle costs of automated production equipment such as automotive component manufacturing systems isn't straightforward
By Ron Quaile
Vice President, Proposal & Estimating,
Division of MAG Industrial Automation Systems
Sterling Heights, MI
You know what your company paid for the equipment on your floor. But what does it cost to buy, operate, and maintain an automated manufacturing system that completely machines a casting and turns it into a component such as a finished engine block? The next question is a better one: what does it cost to make 4,000,000 of these engine blocks on the system over a 10-year period?
The answer: more than $110 million and less than $160 million. That $50 million spread has to do with not only what type of machines the automaker buys, but the extent of the modifications that must be made to the system during the 10 years to accommodate minor and major changes to the component (engine block) design. In addition, the type of equipment chosen changes the cost factors for consumable tooling, spare parts (such as spindles), utilities, and labor for operation and maintenance.
The stakes are high. That's why the calculation of a total life-cycle cost comparison using a comprehensive model is gaining popularity. The projected costs that are revealed can be very surprising. When the planned evolution of a given product design and the expected part volumes are factored in, the system that has more machines and higher upfront acquisition cost may turn out to be the least expensive over a 10–20 year period of use. But maybe not—if the production volumes are high enough.
Some automakers are using these models purely for general guidance, while others are basing specific purchasing decisions on them.
In the automotive industry, for decades, the only feasible, cost-effective way to make hundreds of thousands of large powertrain components per year was to build sequential-process transfer lines of dedicated, nonprogrammable machining stations, with part-handling automation that moved parts in "lock-step" from one station to the next. A raw casting went in one end and a completed engine block (or cylinder head), ready for assembly operations, came out the other.
Families of parts with small variations were dealt with via limited system flexibilities. Major design changes, however, almost always called for new stations to be built, with an investment of significant engineering hours, new iron, and inventory banking and downtime during changeover. Conversion costs to accommodate a frequently evolving product design were high.
Programmable CNC machines weren't originally created for high-volume production. They were meant to handle batches; low-to-medium volumes with multiple operations being performed in a single fixturing. This situation has changed. Today's automotive manufacturer has a choice between two significantly different types of systems, and a hybrid of the two. Agile systems that consist of robust CNC machines and intermachine part-handling automation have been developed for continuous production. The machines make parts using a parallel process; if one machine shuts down, it doesn't shut down an entire process. This type of system is also easily expandable.
Or, there is today's version of the traditional dedicated transfer line. These highly standardized systems have lower acquisition costs, and are favored for high-volume production when anticipated changes are mostly minor and/or infrequent. Conversion costs are typically higher than agile CNC systems, but these costs have become lower in recent years compared to prior generations of transfer lines.
A third option is a hybrid system that combines the two approaches, including both transfer lines and banks of CNC machines.
The key to deciding which one of these systems is right for the product manufacturing program is to take a comprehensive look at total life-cycle costs.
In the worksheets that we've developed at MAG Powertrain to create a total life-cycle cost model, product changes are categorized as minor, moderate, or major. A minor change might be the relocation of the engine's attachment holes so that it can be mounted in a new model. The spot-facing, drilling, reaming, and tapping operations may be altered. On a transfer line, several new multi-spindle heads and multipin probes are needed, plus the cutting tools, for this type of product change. An agile system merely requires a revision of the CNC part programs and the cutting tools. Changes for either production option will not break the bank.
An example of a moderate product change would be an engine block with new features that affect the surrounding features on the part, necessitating additional machining. In this case, the dedicated transfer line needs new multispindle heads and two or three new stations, as well as modifications to the part-holding fixtures. Installing the new or revised equipment requires an extended plant shutdown. Part banking is needed prior to the changeover.
The value of an agile system's flexibility increases when you look at moderate changes. One new CNC machine is probably called for, as well as revisions to part programs. But an agile system is easier to expand, and installation can be done during normal production hours, while production continues within the system on other CNC machines.
Thirdly, consider major changes to the base product. If you're using a transfer line, several completely new stations must be designed and built to replace existing stations. Multiple machines are involved. There are a significant number of new multiheads and fixtures. Plus, extensive updating to mechanical, electrical, fluids, PLC logic engineering, tool layouts, process sheets, and manuals is required. On an agile system, two new CNC machines may be necessary. Because the process is redistributed within the system, all CNC part programs must be revised.
During the 10-to-20-year life cycle of an engine or transmission, it's not uncommon to have two major changes, four or five moderate changes, and 10 to 20 minor changes. All of the above sounds like a clear case for choosing an agile system. But the fact remains, when the volumes are high, the initial acquisition cost of a transfer-line system is less. When you get up into the 600,000 parts-per-year range, cost of the transfer line can be 40% less than the cost of an agile system. Operating costs are also lower.
It's true, however, that the conversion cost of retooling later can offset the initial benefit of lower acquisition and operating cost in many applications. Minor product changes can cost several million dollars, moderate changes as much as $10 million, and major changes typically over $20 million.
The likelihood is that the frequency of product changes will go up in coming years. This stream of changes will affect cylinder heads more than blocks, causing some automakers to tend toward agile systems for heads.
Operating efficiency factors are included in the comparison of systems. Predicting the operational variables of manufacturing systems is an important part of determining total life-cycle costs. Based on the manufacturer's labor rates, our company calculates labor costs for the alternatives (for the system types and selected production volume levels). Keep in mind the fact that the number of operators and maintenance personnel needed to run a CNC-based agile system can be two to three times the number needed for a sequential transfer line. Utility costs (power and air), and the cost of coolant and oil are tallied for each alternative.
Scheduled and unscheduled maintenance, and spare parts, are evaluated and broken out on worksheets. Spindle replacement, including the cost of carrying spareparts inventory, can be a significant number. When used in a parallel process, CNC machines typically have more spindles operating than transfer machines. When you multiply the number of spindles in the system by the anticipated number of replacements required during the planned system life cycle, multiplied by spindle cost of $30,000–$50,000 each, this factor becomes substantial. Mean Time Between Failure (MTBF) and the varying labor-skill-level needed to fix faults and failures, are also calculated and entered as maintenance line items.
There is another disadvantage to the parallel process. Because the same operation or set of operations is simultaneously performed by several identical machines in a cell, these "streams of variation" mean it can be difficult to trace the source of problems when parts aren't meeting spec. The sequential process has a quality-assurance advantage, because there is only one stream. In reality, finding the source of a flaw is a minor problem, because of the high quality levels achievable with individual CNC machines, and the use of sophisticated part-tracking systems that capture the machining history of individual workpieces.
Cutting tool consumption is another variable. As speeds and feeds go up, tooling costs may also be higher. And there are some operational factors that positively affect life-cycle cost. How about eliminating most of the coolant traditionally required? Coolant is about 15% of operational cost. Minimum Quantity Lubrication (MQL) initiatives are getting more feasible every year due to new extraction systems. The trick is to effectively redesign the machine bases to let gravity help get rid of chips, as opposed to flushing them away. Dry machining may be the objective, but the fact is that once the special bases are used, it takes less coolant to flush the chips down in wet applications.
In many applications, a total life-cycle cost study points to a hybrid approach; a combination of CNC machining cells and transfer lines. Transfer lines can be used for roughing and finishing operations on features that don't change. Flexible CNC machines are employed to handle features subject to design changes or variations, such as a switch from four to six-cylinder blocks, different transmission-mounting faces, and engine accessories.
As another processing option, our company, (in partnership with DaimlerChrysler), has developed the XT525 transfer center. In a single unit, a fixtured workpiece is moved to multiple tooling spindles that are arranged in a U-configuration. The workpiece moves in the normal X, Y and Z axes, and it can rotate. Throughput is higher than that achievable by an HMC, because toolchange time is eliminated, and the part is cut with multiple spindles rather than a single spindle. Also, the transfer center is more reliable than a transfer machine, because there are fewer axes needed to accomplish the same amount of machining.
Life-cycle cost analyses we have completed to date indicate that agile, parallel systems almost always win the cost comparison if volumes are low; hybrid or agile systems are best for moderate volumes; and transfer lines rule the world of high-volume production only when minimum product evolution is forecast.
All three system approaches are valid, and can be deemed advantageous based on a life-cycle cost analysis. Agile systems cost more initially, but they provide an opportunity advantage in terms of allowing evolving products to be manufactured cost-effectively. Transfer lines cost less initially, and are viable if future opportunities are not a priority.
As system builders who understand the critical need to shift to long-term thinking to survive and thrive in the auto industry, we're glad that the industry has not only asked for total life-cycle cost studies, but that higher and higher levels of management are using them to decide direction for their manufacturing operations. The name of the game is to examine the whole situation, and get it right before the purchase order is issued.
This article was first published in the May 2007 edition of Manufacturing Engineering magazine.