---

In this paper a neural network with a feed forward topology and a back propagation algorithm was used to investigate the effect of chemical composition on hardness and impact energy in API X65 microalloyed steel. Experimental data was obtained by cutting 100 specimens from pipes manufactured in industrial scale (with 1219 mm diameter, 14.3 mm wall thickness, with similar heats and manufacturing processes). The chemical analysis, Vickers hardness and Charpy impact tests were conducted then according to requirements specified by API 5L standard. The weight percent of C, Si, Mn, P, S, Ni, Cr, Mo, Al, Cu, V, Ti, Nb and Ca were considered as input parameters of the network; while Vickers hardness and Charpy impact energy were considered as output. Scatter diagrams and two statistical criteria: correlation coefficient and mean squared relative error were used to evaluate the prediction performance of developed ANN model. With regard to the exact performance of the developed neural network, it was used then to investigate the effect of chrome and vanadium on Vickers hardness and Charpy impact energy of tested steel.

---

The transformation behavior and microstructural characteristics of API X65 pipeline steel were investigated by dilatometry and microstructural observation. Microhardness measurements were used to verify the observed microstructures. The test steel is imported from abroad and is used extensively in Iran natural gas transmission projects. The continuous cooling transformation curves of the test steel were constructed. The results showed that with increasing the cooling rate from 0.5 to 40°C/s, the microstructure changes from polygonal ferrite, quasi-polygonal ferrite-pearlite to acicular ferrite. The microstructure was dominated by acicular ferrite in cooling rates higher than 5°C/s. The results can be used to design the optimum thermo-mechanical control process (through the selection of proper cooling rate) in domestic manufacturing process of the test steel.