A new variant of the nonlinear kinematic hardening model is proposed which accommodates both nonlinear and linear strain hardening during initial tensile loading and reduced elastic modulus during initial load reversal. It also incorporates the Bauschinger effect, as a function of prior tensile plastic strain, during the nonlinear compressive loading phase. The model is shown to fit experimental data from a total of five candidate gun steels. The numerical fits will be employed in subsequent work to predict residual stresses and fatigue lifetimes for autofrettaged tubes manufactured from the candidate steels.

The Bauschinger effect has been evaluated in high-strength pressure vessels. A simple initial test suggested that a biaxial Bauschinger effect was present and that it was different from previously published uniaxial Bauschinger results. The difference was believed to be significant, so further investigation was undertaken. Several full-size A723 steel gun sections were heavily overstrained and subjected to slit tests in order to measure opening angles and displacements. These geometries were then modeled with finite element (FE) analysis using both ideal autofrettage stresses and Bauschinger modified stresses, which were based on previously published uniaxial Bauschinger test results. Because techniques available for predicting reverse yielding for overstrained pressure vessels were limited, a simple methodology for predicting the yield surface upon reverse yielding from a series of uniaxial Bauschinger test data was developed and is presented. This methodology, when used in the FE predictions, compares favorably with analytical predictions made previously. Comparisons of slit-opening results measured from pressure vessel sections with FE calculations using uniaxial Bauschinger data are made. The opening displacements comparison between the uniaxial predictions and those measured from the heavily overstrained sections with biaxial stresses are so subtle (<1mm) that the tests appear to be inconclusive.

Fast fracture in cannons can be well described using elastic-plastic fracture toughness, in combination with comparisons of cannon section size relative to the size required to maintain plane strain fracture. Fatigue fracture of cannon tubes is modeled from results of full-size fatigue tests that simulate cannon firing. These tests are also the basis of fatigue-intensity-factor modeling of fatigue life, which incorporates material strength, initial crack size and Bauschinger-modified autofrettage residual stress into life predictions. Environment-assisted fracture in the thermally damaged near-bore region of fired cannons is shown to be controlled by hydrogen. High strength cannon steels are susceptible to hydrogen; cannon propellant gases provide the hydrogen; and the source of sustained tensile stress is the near-bore thermal damage and compressive yielding. A thermo-mechanical model predicts tensile residual stress of similar depth to that of observed hydrogen cracks. Coating fracture in the thermal-damage region of fired cannons is characterized and modeled. The Evans/Hutchinson slip zone concept is extended to calculate in-situ coating fracture strength from observed crack spacing and hardness in the damaged region.

Calculations of Bauschinger-modified residual stress in autofrettaged tubes using a simple elastic-plastic model based on the Hill method are compared with neutron diffraction measurements of hoop residual stresses in sections from a swage-autofrettage cannon tube. These residual stresses combined with pressurization stresses are used to describe tube yield pressure and fatigue lives, for comparison with laboratory measurements from cannon tubes from recent publications and new results. Gun steels in the yield strength range from 1000 to 1400 MPa are considered. Results from neutron diffraction measurements show a significantly shallower depth of overstrain in a swage autofrettaged tube compared with that predicted from earlier work with hydraulic pressure autofrettaged tubes, for the same amount of total bore hoop strain. This difference between depth of overstrain for swage and hydraulic autofrettage is incorporated into calculations of yield pressure corresponding to a small amount of permanent strain on the tube outer diameter. Comparison is made with measurements of yield pressure at the same amount of permanent strain. Comparison between calculated and measured yield pressure is good, suggesting a simple design method. Calculations of fatigue life using the classic Paris method compare well with measurements from cannon tests, once Bauschinger-modifications are made in near-bore hoop residual stresses and the difference in overstrain between swage and hydraulic autofrettage is incorporated in the model calculations.

The uniaxial Bauschinger effect has been evaluated in several high strength steels being considered for armament application. The steels investigated include ASTM A723 (1130 and 1330 MPa), PH 13-8 Mo stainless steel (1380 MPa), PH 13-8 Mo super tough stainless steel (1355 MPa), and HY 180 (1180MPa). Tests were conducted at plastic strains up to 3.5%. Results of testing show a progressive decrease in Bauschinger effect up to plastic strains of approximately 1% (for all materials investigated), after which there is little further decrease in the Bauschinger effect. Several key features were discovered during testing. First, all of the materials tested exhibited a changing modulus, where the elastic modulus on unloading after tensile plastic straining is consistently lower than that observed in the original loading of the specimens. The amount of modulus reduction is dependent upon the material tested, and larger reductions are observed with increasing amounts of tensile plastic strain. Prior work by Milligan reported Bauschinger effect factor β for a modified 4340 steel (old vintage A723 steel), which compares well with the present work. However, his results failed to mention any observations about a modulus reduction. The second observation was the expected strength reduction where a reduced compressive strength is observed as a result of prior tensile plastic straining. Numerical curve fits used to calculate residual stresses, which take into account both the modulus reduction and strength reduction are presented for all materials. Fatigue life calculations, utilizing the numerical curve fits, show good agreement with full size A723 laboratory fatigue test results.