This paper examines the buckling behaviour of unstiffened mild steel cone-cylinder assembly under (i) axial compression and (ii) combined loading (i.e., axial compression and external pressure). Experimental results on ten (10) laboratory scaled axially compressed cone-cylinder models and their accompanying numerical predictions of collapse load are provided. The tested models are assumed to have the following range of geometric parameter: rcone/rcyl = 0.565–0.721, rcyl/t = 72.17–72.73, β = 8.536° – 16.69° and a constant wall thickness, t = 1.0 mm. Result confirms the repeatability of the experimental data. Besides that, there is a good agreement between experimental and numerically predicted collapse load with discrepancy calculated to be within 10%. Furthermore, numerical analysis of stability domains for cone-cylinder assembly subjected to simultaneous load actions of axial compression and external pressure was calculated for a range of geometrical parameters of: (i) 50 < rcyl/t < 400 and (ii) 10° < β < 30°. For comparison purpose, the equivalent cylinder approach was also deployed to complete calculation based on ASME code case 2286-2. For the case of dimensionless radius-to-thickness ratio, rcyl/t, the result confirms that the ASME code case 2286-2 is unsafe in designing the cone-cylinder shell with rcyl/t > 400. Whereas, for the case of different cone angle, β, it can be said that the ASME code case 2286-2 may perhaps be safe to use in designing the cone-cylinder shell with β < 20°. In addition, the current analysis confirms that increasing the (a) dimensionless radius-to-thickness ratio, rcyl/t and (b) increasing the cone angle, β, may further shrink the combined stability plot. Finally, imperfect cone-cylinder shell under combined loading confirms that (i) the shell is remarkably sensitive at the pressure dominant region compared to the force dominant region and (ii) the location of dimple/dent imperfection at cylinder mid-section proves to be the worst-case scenario.
Read full abstract