No abstract

The growing accessibility of CAE environments has made FEM more egalitarian, calling into question the reliability of some analyses. This paper aims to investigate the influence of some popularly taken assumptions on results of FEM thermal analyses on the example of a LNG tank support saddle. The temperature distribution of that structure is of interest since the classification societies require that the temperature of construction materials should not exceed allowed service conditions. THE GEOMETRYGeometric model of the tank support saddle is shown at Figure 1. In order to build a cost effective structure, the tank saddle includes a polymer layer (Grudziński 2014) used as an insulation. The order and thickness of layers is shown at detail view B (Fig. 1). The overall dimensions of the tank support saddle, not including the deck, are 0.4 m × 1.4 m × 4.6 m. PRESUMING A TWO DIMENSIONAL TEMPERATURE FIELDTwo-dimensional models are often used in order to decrease the calculation time. Such a simplification is allowable if one of the overall dimensions of the concerned body is an order of magnitude smaller than the other dimensions, or if two opposite border surfaces are insulated. In other cases, a distortion in analysis results occur. A difference between 3D and 2D, evaluation is shown at Figures 2 and 3 on the example of a LNG tank support saddle.

The paper presents a compensation system of thermal deformation for conventional feed axes applied in CNC machine tools allowing for an effective reduction in the impact of heat generated during its operation on the positioning accuracy of the axis. The result has been achieved by equipping feed screws with thermistor temperature sensors. Wiring sensors was led out through an axial bore in the screw and through a rotating electrical connector to an acquisition device coupled with the control system of the CNC machine. An algorithm based on neural networks was implemented in the machine control system, which allows for the online calculation and compensation of heat deformation of feed screws. The algorithm takes into account a variation of thermal deformation values as a function of the table position and the current distribution of the temperature field of the screw and machine. The paper presents a user-friendly method for implementing algorithms containing neural networks in the machine control system. The proposed compensation method has been verified by measuring the linear accuracy of the feed axis positioning. The obtained results confirm the effectiveness of the proposed method in reducing the impact of thermal deformation errors on the positioning accuracy of the axis in CNC machine tools.

In numerous papers it is proposed to use IR measurements of feed axis ball screw temperature distribution in order to compensate CNC machine tool thermal errors. The paper aims to validate reliability of the IR measurements in application to the feed axes ball screws. The identification of key factors influencing the accuracy of the IR measurements of ball screw temperature distribution has been conducted. A test-bench utilizing a ball screw assembly with built-in temperature sensors was introduced and the experimental data are presented along with conclusions.

No abstract