Dissertations / Theses on the topic 'Cold Rolling'

The research work presented in this thesis is concerned with analytical, numerical and experimental studies of the effect of cold rolling on the fatigue behaviour of threaded drillstring connections.

A comprehensive literature study is made of the various effects on the fatigue behaviour of residual stresses introduced by mechanical deformation of notched components. Some of the effects studied are cyclic hardening behaviour after prestraining, cyclic creep, fatigue initiation in prestrained materials, short cracks and crack growth models including crack closure.

Residual stresses were introduced in the surface of a smooth pipe by a rolling device to simulate a cold rolling process and verify the calculated residual stresses by measurements. Strain hardening and contact algorithm of the two bodies were incorporated in the FE analyses. Two significant errors were found in the commercial software package for residual stress evaluation, Restan v. 3.3.2a also called SINT, when using the Schajer method. The Schajer algorithm is the only hole-drilling algorithm without theoretical shortcomings, and is recommended when measuring large residual stress gradients in the depth directions. Using the Schajer method solved by in-house Matlab-routines good agreement between measured residual stress gradients and residual stress gradients from FE analyses was found.

Full scale fatigue tests were performed on pipes cut used drillstrings with notches of similar geometry as threads used in drillstring connections. The simulated threads consisted of four full depth helix notches with runouts at the surface. The pipe threads were cold rolled and fatigue tested in a full-scale four-point rotating bending fatigue testing rig. The test results showed that cold rolling had an effect on the crack initiating period. A major part of the fatigue life was with cracks observed at the notch root, but due to the increased fatigue crack propagation resistance the final fracture initiated at pits inside the pipe. Therefor, an optimisation of the roll geometry and rolling parameters was not possible. However, a significant fatigue life improvement was achieved. Based on experiments, a roller with similar profile as the thread root is recommended. A rolling force of maximum 20 KN is recommended to minimise the possibility of damaging the thread profile. Shallow cracks were observed typically when 5% of the fatigue life had expired. Re-rolling after 50% of expected improved fatigue life, when also short cracks were observed in the notch roots further increased the fatigue improvements.

Pretensioned small steel specimens with a notch were used to simulate cold rolled threats. The specimens were fatigue tested in tension with minimum load close to zero. Pretensioning increased the fatigue life form approximately 50 000 cycles to an infinite number of cycles. In these test non-propagating cracks of typically 0.4 mm length were found. The benefit from pretensioning gradually disappeared with increasing mean stress. FE analyses indicated that an almost instant relaxation of residual stresses to a level with no monotonic strain hardening from preloading would take place when cycled to moderate mean stress. Cycled at low mean stress, an instant relaxation of the surface layer was found in analysis. All observations from notched pretensioned fatigue specimens were in good agreement with the available literature. However, preloading was found to be strain rate dependent in tests where a pretension load held for 2 minutes gave a longer fatigue life than a sinusoidal loading-unloading cycle performed over a one minute interval.

Strain hardening was found not contributing to the fatigue life improvement, whereas the polishing effect from improved surface quality after cold rolling increased the fatigue initiation period. However, residual stress and subsequent early crack closure was the dominating effect at moderate cyclic mean loads.

The material data required to perform FE fatigue simulation studies of a full threaded cold rolled coupling incorporating make-up torque, include cyclic stress strain behaviour at various amplitudes and mean stress caused by various degrees of prestraining. Such data are not readily available today, and are only possible to obtain in carefully planned and executed experiments. Also, 3D FE model required for cold rolling analysis is extremely CPU time consuming. Consequently, cold rolling simulations could not be successfully implemented in this work.

One of the main conclusions from this work is that drillstring connections will respond differently to thread rolling at the pin or box. A significant improvement in the fatigue life of box threads from residual stresses is expected mainly from increased resistance to crack propagation. However, the compressive residual stress is sensitive to overloading in compression, and the improvement from residual stress depends strongly on the mean stress (or R-ratio). At values of R of approximately 0.6 or higher the beneficial of rolling therefore tends to disappear. At the critical locations of the pin, which are the last engaged thread or the stress relief groove, the effect of residual stresses introduced by rolling is therefore likely to be severely reduced by the high mean stress imposed during make-up of the connection. However, a beneficial effect of rolling is expected to remain due to improved surface condition and due to a possible effect of strain hardening. The net results of these factors on the fatigue performance of actual drillstrings can only be determined in full scale rotating bending tests.

Cold-rolling mills reduce the thickness of an incoming metal strip to produce longer, thinner strip with desired mechanical, dimensional, and metallurgical properties. A primary factor in manufacturing high-quality cold-rolled sheet is accurately predicting the required rolling force. The rolling force directly influences roll-stack deflections, which correlate to the resulting rolled sheet flatness quality. Increasingly high demand for thin and ultra-thin cold-rolled sheet metal gauges, along with the correspondingly larger sensitivity of flatness defects, makes it more important to accurately and rapidly predict the rolling force before the rolling operation begins. Accurate rolling force predictions enable assignment of appropriate pass schedules and flatness mechanism set-points early in the rolling process, thereby improving quality and reducing time and scrap. Cold rolling force predictions have traditionally employed two-dimensional analytical models such as those proposed by Roberts and Bland & Ford. These simplified methods are prone to inaccuracy, however, because of several uncertain, yet influential, model parameters that are difficult to establish deterministically for wide-ranging products. These parameters include, for example, the rolled strip average compressive yield strength, frictional characteristics relating to low and high mill speeds, and the yield strength strain rate dependency. Conventionally, these unknown parameters have been evaluated deterministically by comparing force predictions with industry force data and using a best-fit regression approach.

In this work, Bayesian updating using a probability function is applied to identify joint posterior probability distributions of the uncertain parameters in rolling force models. It is shown that the non-deterministic Bayesian updating approach is particularly useful as new rolling force data becomes available and the models can “learn” from this available production data. The goal is a model that can better predict necessary mill parameters based on accurate probability estimates of the actual rolling force. The rolling force data used in this work for applying Bayesian updating is actual production data of grades 301, 304L (low carbon), and 304 stainless steels, rolled on a 10-inch wide 4-high cold rolling mill. This force data was collected by observing and averaging load cell measurements at steady rolling speeds.

Recrystallisation annealing, a repeated heat treatment between different stages of cold rolling of martensitic chromium steel strip, is successful when neither high rolling forces nor wear of the working rolls occur during the subsequent cold rolling. Mechanical properties as tensile strength, yield, elongation or hardness have been, by tradition, the criteria that described the quality of the annealing process. In recent years, the development of the measurement equipment in the rolling mills and of the instruments for material investigations has accentuated more and more the role played by the microstructural properties in the evaluation of the heat treatment. Two microstructural characteristics of the degree of annealing are, firstly and most important, the recrystallisation degree, and, secondly, the secondary carbide density.

The sample manufacturing and heat treatment, modelling and microstructure investigations by light optical- (LOM) and scanning electron microscopy (SEM) described in this article were carried out at Sandvik Materials Technology’s R&D Department and Bell Furnace Line in Sandviken, Sweden, while microstructure investigations and evaluation by scanning electron microscopy with field emission gun (FEG-SEM) and electron back scatter diffraction (EBSD) were done at the Corrosion and Metals Research Institute (KIMAB) in Stockholm, Sweden.

The first part of this work shows that, in contrast to the traditional methods LOM and SEM, that use chemical etching for the preparation of the samples, EBSD can successfully characterise recrystallised structures in annealed martensitic chromium steels. Unlike conventional microscopy with LOM and SEM, EBSD is able to reveal the grain geometry, as well as to separate and identify the different phases in this kind of steels (ferrite, M23-, M6-carbides). Important parameters such as grain size, particle size and recrystallised fraction can be measured with high accuracy. This information can be used to understand, evaluate, control and even predict the recrystallisation annealing of martensitic chromium steel.

The second part of this work presents how the results from microstructure description by EBSD can be directly used in relatively simple empirical models for determination of recrystallisation degree as function of the annealing parameters and the deformation history. EBSD was applied to evaluate the degree of recrystallisation in a series of annealing tests, with the purpose to model recrystallisation temperature in two types of martensitic chromium steel strip, a traditional one and one alloyed with molybdenum, cold rolled with different amounts of reduction and annealed with different temperatures, soaking times and heating rates. The empirical quadratic models were built with Umetrics’ software for experimental design, MODDEÒ 8.0 and they defined the recrystallisation degree (limits for LAGB and HAGB were set to 1.5° and 7.5° for the first grade and 2.5° and 10° for second one) and the secondary carbides density as functions of annealing temperature, soaking time and cold reduction (the factor heating rate was removed as nonsignificant). To be observed that these empirical models were fit much better for the recrystallisation degree than for the secondary carbides density.

The modelling work described above, together with the implementation of online physical temperature models in the bell annealers may lead to an increased productivity in the production plant by shortening the annealing cycle and minimising scrap and thus to an economical gain of ca 1,5 MSEK per year at Sandvik Materials Technology.