CDEDS Methodology

The Engineering Manager can be in three different situations relating to this requirement and he would react differently for each situation:

Situation 1: The specification is an extension of an existing design within the factory: then the kinematic lengths of boom, dipper and bucket are almost known and the problem reduces to iterating the four-bar mechanism at the bucket end and arriving at the cylinder sizes.

Situation 2: The specification is of a competing machine not within the factory:  then the manager would conduct a competitive bench-mark exercise and arrive at starting lengths from the study.

Situation 3: Specification is not a known configuration anywhere in the market: If you address this last situation, then other situations are automatically addressed.

Current engineering practices use elaborate and iterative methods to synthesize the geometry. The marketing specification are the design targets to reach and there aren’t many software, commercially available, that can synthesize the workgroup kinematics to satisfy all parts of specification at one go. You can work out preliminary kinematics in any 3D or 2D CAD software to satisfy the reach dimensions but then vertical wall depth may not be met. You could get the vertical wall and swing radius too if you are lucky, but then the forces become a consequence of that particular geometry. Worse still, the force calculations have to be accomplished on another software, and often, these two software cannot be dynamically linked. You then either live with what you get if it is “close enough”, or spend more hours iterating the kinematic design between multiple software.

Additionally, there are parts to workgroup design that are unstated in the specification that an engineer has to recognize upfront, to avoid problems later in the designs. For instance, an engineer has to recognize that heaviest pieces of the machinery are better closer to the machine to avoid problems in lift performance during working. Weight and forces automatically relate to stresses, stresses to section sizes and so on. Similarly there will be space claim issues relating to hydraulic cylinders, transport dimensions in folded condition, clearances to other machine parts like undercarriage, cab etc. Therefore “All” stated specification and unstated design criteria have to be recognized and, like in Concept Phrasing, have to be translated to boundary conditions or design constraints. Once the constraints are identified, solution to the problem can be found by applying the following methodology, described broadly below:

Each specification parameter, say maximum reach, is a configuration arising from two things – the kinematic lengths (of boom, dipper arm and bucket) and a particular orientation of the parts, in this case all stretched out. Therefore, if the kinematic lengths are defined as variables, then a configuration, say max reach, is a just a mathematical function of these variables.
Since different configurations arise from just changes to orientation of the parts, we recognize that lengths between pivot points should remain same between the different configurations. We can model other specification parameters, max depth, cutting height etc., using the same variables for length between pivots.
Other design criteria, can now be brought in as either more variables or more mathematical functions. For instance the SAE standard defines the digging forces to be applied in a particular orientation of dipper and bucket and at certain operating pressures conditions. Once we have equations for forces, section stresses can be described and consequently weight of parts can be estimated.
This methodology now suggest that the design synthesis problem can be solved numerically by applying multi-criteria optimization methods to the specification targets now described as Objective functions. Simpler problems like those of door flap or Situations 1, 2 described above, can be solved using simpler numerical optimization methods.
The seed/start/guess values for variables of lengths between pivots can always be generated from regression analysis of dimensions over weight of machine for different sizes. This is known to work.
Without getting into deeper discussion on Optimization, problems like the full-blown excavator work group design do not have a unique solution and so we seek a Pareto optimality which is a trade-off between different objective functions. There is a better chance of convergence if more variables are given bounding values reflecting design engineers preference or experience.