CRM Engineering » FEA

Design validation through FEA

Charles McCreary — Mon, 17 Dec 2007 17:25:13 +0000

This particular client often uses our services to validate design modifications to plant equipment. Their equipment is used to transport large sheets of glass from the production line to the shipping bays. Failure of the equipment during a lift operation can result in severe injury or death due to glass shards or the payload.

We first calculate the appropriate loads and safety factors using appropriate codes (ASME in this case). We then evaluate the design as far as possible using tradiational structural analysis methods. Once the member sizes are appropriate, we perform a final validation of the structural design using finite element analysis (FEA) methods.

Random response analysis of avionics tray

Charles McCreary — Fri, 14 Dec 2007 16:40:14 +0000

A client approached us with a difficult electronics packaging problem. Their customer required a commercial, off-the-shelf (COTS) board to be included in a device that must pass the MilSpec random response envelope for a particular aircraft. Since the COTS board could not survive the vibration unless the enclosure provided shock and vibration isolation, the objective of the analysis was to specify the required properties for the enclosure supports as well as to ensure that the enclosure itself had sufficient fatigue life to survive the vibration spectrum.

Machine Design – Fatigue

Charles McCreary — Fri, 14 Dec 2007 03:09:16 +0000

A client was experiencing failure in a cantilevered shaft which supported a heavy spool that was subject to frequent rapid starts and stops. The initial design incorporated a a 3″ diameter shaft that was supported by two bearings with a drive gear halfway between the bearings. A larger, 5″ diameter shaft was fillet welded onto the 3″ shaft just outboard of the outer bearing.

The massive spool was accelerated from rest to 60 rpm in a fraction of a second and braked several times an hour, 24/7. Thus the number of stress reversals seen in a year required a design in which the peak stress is well below the fatigue limit. The shaft would break at the fillet weld within a few weeks of service at the fillet weld (no surprise).

Examining the problem from first principles (beam theory and stress concentration factors) showed that the expected fatigue life was measured in hours. A detailed finite element analysis of the system corroborated the diagnosis.

Our recommendations:

Machine the shaft in one piece with a generous shoulder fillet. This eliminates the fillet weld and greatly reduces the stress concentration at the shaft diameter transition

Specify a generous root fillet in the gear keyway to reduce the peak stress

Use a material with a fatigue limit much higher than the mild steel used in the original design

Hanger Flanges

Charles McCreary — Thu, 13 Dec 2007 02:47:38 +0000

A client recently engaged our services to design hanger flanges to satisfy a prescribed load combination for each size. A hanger flange is a joint in a pipeline whose primary purpose is to react the axial load below the hanger flange into some structure (hence the term hanger). The requirements were that the design satisfy ASME BPV Section VIII – Div. 2 (primarily that the linearized stresses are within the allowables).

InjectSafe Safety Valve

Charles McCreary — Wed, 09 May 2007 22:23:17 +0000

Assume you wanted to deliver something via capillary tubing from topside to the producing zone(s) and there is (will be) a tubing retrievable sub-surface safety valve. You need a mechanism to bypass the safety valve without compromising it’s function. BJ-Services/Dyna Coil division has developed a technology that allows an operator to inject chemicals past the safety valve into the production zone without compromising the safety valve’s function.

Assisting BJ-Services in the design and qualification of a new size, CRM Engineering Services provided design and analysis support. As with any safety critical device, considerable thought must go into designing load cases for every conceivable operational event. Finite Element Models were built and analyzed for a number of load cases. As usual, there is a trade-off between the time required to build the model and the time required to solve the model. Fortunately, we have the computational infrastructure to handle very large, non-linear models and since computer time is relatively cheap, we usually choose to build very fine, detailed finite element models. Furthermore, we perform convergence studies on the models to verify that the discretization is sufficiently fine that we can have confidence in the predicted results. Thus the emphasis is on turn-around time, the time from when we are initially engaged to the delivery of the final report.