James I. Finkel
Edited by Leland Teschler
Today’s new engineers are faced with an unprecedented suite of problems. There is little question that significant numbers of engineers get frustrated with the work and leave the profession. The irony is that the chief problems that many engineers find exasperating have little to do with technical issues. They arise more from the fact that corporations today have decimated their engineering departments, leaving little in the way of support for those entering the profession. This has negative implications for companies trying to ramp up with innovative new products and recover sales lost in the latest economic crunch. Increasingly, the intellectual horsepower to get significant products to market just isn’t there.
Of course, converting green engineers into useful contributors has never been easy or quick. Engineering at the best companies has traditionally encompassed apprenticeship programs. The senior staff would take in the newbies and, after a few years, turns them into productive personnel. During this growth process, the engineers would learn not only the firm’s products, but also the relationships between departments and how to use their firm’s CAD and analysis tools.
These sorts of relationships have, in many firms, become impossible dreams in the hollowed-out engineering departments that now characterize so many manufacturers. Moreover, the engineering environment is increasingly characterized by multiple constraints from multiple departments. With the engineering bosses gone, the engineer must fend for himself or herself to meet such demands. Among the typical day-to-day requirements: Marketing wants geometry for ads, manufacturing wants tooling, external compliance wants adherence to the appropriate ANSI standards. And the boss wants it Tuesday. But when it comes to advice, the newbie engineer is typically all alone.
“Alone” is a scary place to be, especially because CAD has done a great job of ensuring that all engineering departments function the same way, independent of scale, whether or not this is an appropriate way to do business. To understand the problems confronting the newbie, consider some typical engineering scenarios with the example of a steam-turbine manufacturer. As with many firms of this kind, the turbine maker has been making turbines for a number of years. In a typical scenario, the new engineer might be asked to incorporate the latest set of metals in blades to let the turbine run hotter and more efficiently. But there are a few problems. First, the casing was created before the engineer was born, so there are no CAD drawings. Any sketches might exist on vellum, but they are likely stored in an inaccessible archive. So the first job is to recreate the casing (messy seals included) in the current flavor of CAD.
This cannot be a halfway job. The “new” casing must be able to seat any old parts that may have fit all the prior designs. Next, the engineer must recreate the rotor. As with the casing, drawings are a wishful fantasy. But because this is now a “new” design, the engineer must run a full vibration analysis on the new shaft. Of course, he or she must now also recreate all the other rotor stages — in the new CAD system to get accurate response numbers.
And because the boss believed the ease-of-use claims of the CAD salesperson, the engineer may now be expected to run a full fluid-dynamic analysis on each stage to verify peak efficiency. Let us ignore the fact that the newbie engineer may only have done the analysis on a single blade sometime as an undergraduate. He or she is now expected to run a full thermodynamic and aerodynamic analysis. Did I mention the one-week project deadline to recreate 50 years of design work?
Yet another significant issue is the lack of meta data associated with the old part drawings. The drawing may exist in the last iteration, but all the history is gone. Drawing notes are likely to be sparse and may not fully explain the “why” of the most recent change, let alone the half century of prior changes. The fit issues — where the guys on the floor used stainless-steel shims to make the stator-blade sections fit better in seals — are not part of drawing records.
As if the engineer has not yet seen enough problems, the new crop of materials to be used in the part may have different rates of expansion. This may become an issue when creating the root section of the turbine-blade disks. Adding another twist to the mix, the original tooling was likely based on high-speed steel. Can the old geometry be shaped with the new set of carbide tooling? This is not a simple question. The original machinist is probably long gone. The drawings on vellum may not be sufficient to enable the new tooling to be interpolated. And the toolpath might have to be created on the latest generation of CAD for the CNC profiles.
None of these steps is easy. The tooling expertise may have been pushed overseas and the deadline is looming. And the current crop of engineering managers likely has no advice to give on any aspect of this suite of issues.
Henry Ford was wrong?
Henry Ford’s ideal was to automate the production line to a point where each person did a single task, in the minimum time allotted, and repeated the task ad infinitum. Modern engineering turns that notion on its head. Now, the engineer is responsible for everything. The engineer creates the geometry, selects the tooling and the material, designs the production, and establishes the QA procedures to verify that the part was made correctly. Unlike the worker who grows bored, the engineer gets too much excitement.
That brings us to communication difficulties the new engineer is likely to encounter when working with his or her somewhat nontechnical managers. To illustrate, consider the accompanying image. Which is the best element? Pick 1, 2, or 3. Your mission is to explain to your boss that the best element is clearly number 1, as this is computationally efficient, takes little time to set up and will give the best results in the shortest time. Your goal will be to explain this to your boss in less time than it took you to run the analysis. Points will be taken away if you use the words “boundary conditions.” Bonus points will be added if your boss understands that simpler is better. If you can complete this test in five minutes then you are on your way to a career in sales or management.
Engineers understand that the Finite-Element Method is a mathematical modeling technique based on an abstraction of geometry. In this case, the movement of the beam is based on the well-established “classical” equations. The trick is to show your boss the framework and have your manager understand that the simplified geometry represents the “real” structure. Also, the beam model will produce more repeatable results. Remembering the loading on a fancy model is a huge memory feat.
The more-attractive geometry that looks realistic is actually a far worse model. The boundary conditions (I’m losing points here) are more difficult to set and the compute times become infeasible. For most analytical models, complexity reduces the model’s usefulness. Bigger models take more time to create, have more problems when setting boundary conditions and give less reliable results. The fillets and holes necessary to make the part detract from the ability to get good, fast answers.
This is another reason so little time is spent in doing “real” engineering. “The” engineer in the hollowed-out organization must constantly explain the results to a technologically naive audience. The better the engineer is at explaining the results, the more time available to do real work.
As with any engineering argument, we must consider the other side: What is real work? Real work is when your simplistic model uncovers an unsuspected problem. Unsuspected is a nice way of saying that the results clearly indicate you need to bring out the “big guns” of analysis, nonlinearities.
When you consider the geometric and material nonlinearities, you now must spend a lot more time creating an appropriate model. One of the continuing ironies is that while the model behavior becomes more complex, the geometry on which the more-stringent analysis takes place often becomes less detailed. Instead, the engineer will focus on the loading and boundary conditions, a better material definition, and a more-realistic response. The fancy carbon-fiber bicycles on which racers depend are first built with FEA to make sure they can stand up to the stresses of the road. Figuring out the exact stiffness of the multiple layers at varying angles is far from child’s play. The key in engineering is knowing when to use the better analysis, and this is not a skill learned in a few days.
Finally, we must acknowledge that the new engineering structure is no longer strictly hierarchical. It used to be that the engineer reported to a boss who reported up the managerial food chain. There were clear lines of responsibility, and the engineering boss acted both as a buffer and as a final judge of the “worth” of the engineer. All this changed as the organization hollowed out. The engineer is now faced with a slew of external constraints, few of which are negotiable.
In dealing with nontechnical outsiders, new engineers also typically have a great deal of trouble trying to explain why a very real design constraint cannot just be arbitrarily changed. Engineers have been trained to make the choices, but few have been trained to tactfully say, “No.”
There is another aspect to dealing with nonengineers in the organization: With so much time spent on noncore engineering, the time left for other required tasks is increasingly spread thin. Inevitably, some tasks will not be done well. So the engineer is thrust into an increasingly political role. Consider this question, “How do you best get the product out the door while meeting the required ANSI standards, internal cost constraints, and get a well-deserved financial reward?”
The last part is tricky. The engineer is a target for groups that provide a constraint. Because a good engineer will balance the conflicting goals, some constituencies will be annoyed. And this annoyance can translate into a withheld raise. The product may make millions of dollars. But because a single department is disappointed, the raise may not happen.
All in all, the lure of engineering is the ability to make something. Today the punishment for choosing the profession is that engineers are constantly criticized even when they do well; in many ways, engineering in the hollowed-out company can be the classic thankless job. There is no easy way to fix this situation and it is unlikely to go away on its own. Still, understanding the pressures from the viewpoint of the newbie is the first step towards a cure.
Engineering product development is not quite as painful as a “Kill Kurt” drill but it does entail aggressive people streaming in at engineering all at once. Instead of arm pads, they are armed with product requirements and their own agendas. The engineer must absorb all these inputs and deliver an end product that meets enough of these goals.
New engineers will find that, as with pass rushes, it becomes easier with practice to take the pressure and deliver. The difficulty is that in a hollowed-out company, there are few old hands around to remind newbies that the reward for handling the pressure is a killer product.