Traditional Design Teaching – Chief Concerns by Kotur Raghavan

 Design Approaches

Broadly there are two classes of approaches to Engineering Design.

The first is the traditional approach. In this approach the designs are arrived at using the information available in Codes, Standards and Handbooks. Industrial equipment like boilers and civil engineering structures like bridges, transmission towers and buildings are designed using this approach. As an example, we can quote the ASME’s Boiler and Pressure Vessel Code, Section VIII, Division One. Whereas the codes and standards yield satisfactory designs, the main objection will be that they are often too conservative and the designs tend to be unnecessarily bulky and heavy. For simpler designs the traditional method involves the use of formulae from Applied Mechanics and Strength of Materials. Most of the Machine Design books use this as basis. The approach is useful for simple mechanical components like shafts, pulleys and springs. There are limitations and shortcomings in this approach which will be discussed in the next section.

The second major category is Design by Analysis (DBA). Usually this will not be a single step procedure. Iterations will be involved. In today’s context, the only suitable method is the finite element analysis. For boilers and pressure vessels, ASME prescribe the guidelines available in in its B & PV Code, Section VIII, Division Two. Similar guidelines are available for many industrial equipment also. Design of engineering components using FEA calls for in-depth engineering knowledge and a sound knowledge of strength of materials and elasticity theory.  If properly used, the use of FEA will result in satisfactory optimum designs.

Approach based on Strength of Materials – Main Concerns

In the strength of materials based approach, which is to be found in all machine design books, a common feature is simplifications, often gross, are introduced in the geometry, loads and load path. Frequently, a few assumptions are made with regard to the structural behaviour of the component. These assumptions are needed so that simple closed form expressions of strength of materials can be invoked. Quite often, such assumptions result in statically indeterminate structure being solved as statically determinate one. In effect, all the design problems covered in the books can be considered to be “closed ended”. This means to say that solution is obtained for N unknown parameters using N number of equations.  The simplifications may result in conservative or non-conservative designs.

With regard to the treatment of stresses, the main drawback is that only one component of stress, like direct, bending or shear comes into reckoning. Consequently, equivalent stress, say von Mises, or principal stress do not form the design basis. The effect of neglecting the combined effect of several stress components is likely to result in non-conservative designs. Consequently, the factor f safety  will be effectively lower than what is actually intended.

Another shortcoming is that the stresses are considered at their face value only. The damage potential of stresses is not accounted for. The recognition of damage potential is one of the major aspects of ASME’s DBA rules. Stresses are assigned tags such as PRIMARY, SECONDARY, MEMBRANE, BENDING and PEAK. The categories are dependent on the spatial distribution of the stress and the type of load causing the stress. The allowable stresses strongly depend on the category to which it belongs. For example, the allowable bending stress is 1.5 times the allowable membrane stress for the same material. By treating stresses of different categories on equal footing, the designs tend to be conservative. If a bit too conservative, the designs will turn out to be non-optimal.

Design Evaluation Case Studies

A series of case studies will be published in order to demonstrate the foregoing aspects. The approach will be analyse the component using a high fidelity models of the components concerned. The models are going to be rigorous with recourse to any simplifying assumptions. The design assumptions will be assessed based on the results of analysis.

Fig. 1

The contents of this article and a few of the forthcoming case studies are based on a conference presentation shown above.