LDPE production developments require processing equipment with increased capacities and existing applications uprating. They include large reciprocating compressors with several cylinders operating at very high pressures. Selection of piping size, as these special components have very long delivery time, including reactors and intercoolers, requires a careful analysis to be performed at an early stage of a project to properly estimate the costs. A preliminary evaluation should be made during the bid stage using compressor data sheet and P&I diagram. A specific pre-study linked to compressor sizing program can be used for preliminary cost estimation. Prior to piping purchase order issue, based on the first issue of real layout around compressor, the above evaluation has to be confirmed by a preliminary pulsation study. Once the plant arrangement will be finalized, the final acoustic study will have to be performed. Usually such study will only result in minor adjustments to the design not involving the piping size.

Hyper-compressors are subjected to very high operating pressure, maximum and fluctuating, strongly impacting design and life of all cylinder components. Components subjected to fatigue (both high cycle fatigue and low cycle fatigue) can experience a failure in a very short time that may cause large production losses. Packing cups proved to be one of the most critical cylinder components and the fatigue behavior of the packing cups subjected to alternating loads, due to the suction and discharge conditions, is their key design factor. A thorough study has been performed in order to increase the safety factors in all critical regions of the packing cups analyzing a new manufacturing cycle that allows optimizing the compressive pre-stresses on the lube oil parts due to the autofrettage procedure. The approach included a dedicated sensitivity FEM analysis of the autofrettage pressure to determine the correct residual stresses at the lube oil parts in order to keep them under compressive stress in operating conditions and Fracture Mechanics Calculations based on BSI7910 aimed at achieving a design able to survive with a threshold defect larger than the minimum detectable defect.

Today market requires machinery with ever larger performance increasing compression system vibrations risks. Knowledge of dampers pulsation-forces phenomena and best practices allows their minimization. Loads acting on foundation are early defined allowing a proper design, while “Cylinder-Gas-loads” depending upon compressor data-sheet cannot be adjusted unless to change requirements. Considering their high amplitudes and frequencies spectrum, some exciting frequencies often coincides with system mechanical natural frequencies. This involves expensive efforts in preliminary Cylinder Manifold response studies to guide compressor General Arrangement design. Specific software that includes Compressor standard elements selection and that allow building dampers by parametric inputs is cost effective in model creation. Together with the cost benefits it facilitates the designer to simulate multiple configurations rebuilding the model in a short time and exploring several solutions to optimize the system vibration control. A FEM specialist is not required for the Model build-up, the Software allow automatically applying cylinder gas loads, run analysis and compare results vs allowable ones. In case of allowable limits exceeding or design changes, G&A Designer can easily change input and iterate the loop till satisfactory results are achieved in a timely and quality manner, optimizing dampers and supports.

The behavior of the valves of Hypercompressors on LDPE plants is challenging to predict because it depends on many factors and often the expected and macroscopic gas parameters, such as pressure, temperature and gas composition are not sufficient to properly evaluate the valve behavior in the field. In fact valve operation is highly dependent on local phenomena such as localized pressure losses and presence of vortexes which are in turn influenced by the geometry of the valve and by its behavior. To better understand all these phenomena it is needed to characterize these valves through experimental tests aimed at defining, with a good accuracy, the valve dimensionless parameters Cd (drag coefficient) and Ks (flow coefficient) as a function of the geometry of the valve itself. If the coefficients Cd and Ks are not accurate, the expected behavior of the valve may be completely different from the evidence of the field and could not properly explain certain types of failure modes. With a more accurate evaluation of Cd and Ks, some types of damage which in first hypothesis would seem caused by factors external to the valve, in reality are proven to be intrinsically related to valve design and often dependent on valve malfunctioning. As a final step, through to a deep understanding of the valve behavior in the field an improvement of valve reliability and efficiency can be achieved through optimization of the design for various operating conditions.

No abstract