Search

This article presents a novel state estimation structure developed for a skid-steered, six-wheeled, ARGO® all-terrain vehicle (ATV). The ARGO ATV is a teleoperated unmanned ground vehicle custom fitted with an inertial measurement unit, wheel encoders and a Geographic Positioning System (GPS). This paper offers the following contributions: (1) a hybrid extended Kalman filter/Kalman filter state estimation technique that estimates the motion, orientation and wheel slips for an ARGO and (2) a virtual GPS point generation algorithm that can be used to adjust position estimates using a GPS sensor. Our field experiments reveal that the proposed estimation structure is able to estimate the position, velocity, orientation and longitudinal slip of the ARGO with a reasonable amount of accuracy. In addition, the proposed estimation structure is well suited for online applications and can also incorporate offline virtual GPS data to further improve the accuracy of the position estimates. The proposed estimation stru...

Any meaningful initiative to improve the teaching and learning in introductory control systems courses needs a clear test of student conceptual understanding to determine the effectiveness of proposed methods and activities. The authors propose a control systems concept inventory. Development of the inventory was collaborative and iterative. The diagnostic test was administered to students before (pre) and after (post) student learning activities. Test responses were analyzed to provide instructors with feedback on their teaching. Using classical test theory and item response theory, aggregated results were analyzed to assess internal consistency and measurement error, respectively. Students demonstrated an improvement from pre- to post-test scores, showing gains of 23%-34% in understanding of new concepts learned. The internal consistency of the test has ranged from 0.61 to 0.68. It can be shown that the precision of the test is highest in the score range of 33%-66%, which is where most post-test scores occurred.

This paper proposes an approach for online learning of the dynamic model of a robot manipulator. The dynamic model is formulated as a weighted sum of locally linear models, and Locally Weighted Projection Regression (LWPR) is used to learn the models based on training data obtained during operation. The LWPR model can be initialized with partial knowledge of rigid body parameters to improve the initial performance. The resulting dynamic model is used to implement a model-based controller. Both feedforward and feedback configurations are investigated. The proposed approach is tested on an industrial robot, and shown to outperform independent joint and fixed model-based control.

This paper focuses on evaluating Locally Weighted Projection Regression (LWPR) as an alternative control method to traditional model-based control schemes. LWPR is used to estimate the inverse dynamics function of a 6 degree of freedom (DOF) manipulator. The performance of the resulting controller is compared to that of the resolved acceleration and the adaptive computed torque (ACT) controller. Simulations are carried out in order to evaluate the position and orientation tracking performance of each controller while varying trajectory velocities, end effector loading and errors in the known parameters. Both the adaptive controller and LWPR controller have comparable performance in the presence of parametric uncertainty including friction. The ACT controller outperforms LWPR when the dynamic structure is accurately known and the trajectory is persistently exciting.

A survey of mining accidents has revealed that over 30% of all truck loading accidents can be addressed by providing dipper positioning feedback to the shovel operator. In this paper, a novel approach is presented for estimating a mining shovel’s dipper pose to obtain its arm geometry in real-time utilizing a two-dimensional laser scanner. The low spatial resolution of laser scanners and the need for accurate initialization challenge the reliability and accuracy of most laser-scanner-based object tracking methods. This work addresses these issues by using the shovel dipper’s kinematics model and position history, in conjunction with the dipper geometrical model, to track the dipper in space. The proposed method uses a bootstrap particle filter with a distance transformation in order to perform a global search in the workspace. The particle filter’s result is then used as the initial pose for an Iterative Closest Point algorithm that increases the accuracy of the pose estimate. The proposed method can be applied to other laser scanner-based object tracking applications in outdoor environments. Experiments performed on a mining shovel demonstrate the reliability, accuracy, and computational efficiency of the proposed approach. Moreover, using a single proximal sensor can simplify mounting, reduce maintenance costs and machine down time, and enhance tracking reliability.

Recent automation research has proposed robotic machining as an alternative to traditional CNC machining. However, unlike traditional machine tools, robots are subject to relatively larger dynamic disturbances and operate closer to their torque limits. As a result of these conditions, combined with inaccurate models of the manipulators and the machining process, joint actuators can often saturate during machining operation, resulting in damage to the tool, or the part, or both. As off-line planned trajectories are executed by the controller, unmodeled disturbances arising due to inaccurate manipulator and blank models may necessitate on-line trajectory re-planning to provide alternate joint trajectories. Such joint trajectories must be constrained to maintain a desired tool path relative to a blank, to respect the manipulators’ torque and joint limits, and to avoid collisions. This paper presents a new method for selecting the null-space motion based on torque limits and conflicting performance criteria. The effectiveness of the null-space motion algorithm has been demonstrated via extensive simulations.

An advanced control solution, specifically Model-based predictive control (MPC), has been developed for a low-pressure die casting (LPDC) process used to manufacture aluminum alloy wheels. The control solution was developed offline using a high-fidelity 3D finite element model as a virtual process on which open and closed loop trials were performed. Open loop trials were used to assess the influence of process variables and to facilitate the development of a reduced-order state-space model to predict the approximate input–output behaviour of the casting process. An MPC controller employing this state-space model was used to regulate die temperatures in the virtual process during disturbance scenarios. Two disturbances typically found in the industrial process were simulated, that is, the temperature behaviour of the molten metal in the industrial process and the length of time the die remains open after the cast wheel is ejected from the dies. These disturbances cause the die temperatures to deviate from their optimal values, which can result in defective wheels. The developed control solution improved the process performance by rejecting the simulated disturbances and maintaining the die temperatures near their optimal values.

Recent research has considered robotic machining as an alternative to traditional CNC machining. However, one of the problems with serial link manipulators is that they have a low inherent stiffness due to the cantilever design of the links and low torsional stiffness of the actuators. This problem is addressed during the resolution of a Cartesian-space tool path, defined relative to the blank, into joint space using the pseudo-inverse method of redundancy resolution. The two-armed system is modeled as a single system using a relative Jacobian to relate the joint motions to the relative motion of the tool and blank. The null space of the relative Jacobian is used to optimize the joint trajectories by minimizing the system compliance factor. A min-max optimization and a null-space search are investigated in order to produce a global optimum. Simulation results show that stiff trajectories lower the level of torque consumption, which allows for an increase in productivity. The deflection of the tool relative to the blank is decreased as well.

PurposeRecent research has considered robotic machining as a dextrous alternative to traditional CNC machine tools for complex sculptured surfaces. One challenge in using robotic machining is that the stiffness is lower than traditional machine tools, due to the cantilever design of the links and low‐torsional stiffness of the actuators. This paper seeks to examine this limitation, using optimization algorithms to determine the best trajectories for the manipulators such that the stiffness is maximized.Design/methodology/approachThe issue of low stiffness is addressed with an integrated off‐line planner and real‐time re‐planner. The available manipulator stiffness is maximized during off‐line planning through a trajectory resolution method that exploits the nullspace of the robot machining system. In response to unmodeled disturbances, a real‐time trajectory re‐planner utilizes a time‐scaling method to reduce the tool speed, thereby reducing the demand on the actuator torques, increasing the robot's dynamic stiffness capabilities. During real‐time re‐planning, priorities are assigned to conflicting performance criteria such as stiffness, collision avoidance, and joint limits.FindingsThe algorithms developed were able to generate trajectories with stiffer configurations, which resulted in a reduction in the actuator torques. The real‐time re‐planner successfully allowed the process plan to continue when disturbances were encountered.Research limitations/implicationsSimulations are presented to demonstrate the effectiveness of the approach.Practical implicationsAddressing the limitation of stiffness in serial‐link manipulators will enable robots to become more suitable for machining tasks. The real‐time re‐planning approach will allow robots to become more autonomous during the execution of a given task.Originality/valueAn integrated off‐line and real‐time planning approach has been applied to robotic machining.

Hydrostatic transmissions can be used to drive mining machines when the size of the machine restricts the use of mechanical transmissions. However, hydrostatic transmissions lack a differential effect and when the machine navigates around corners there is an increase in the amount of vibration and discomfort to the operator. Thus, a method of introducing a differential effect is required. Three different approaches to generate a differential effect are presented: the application of cheater lines (hydraulic sharing between the inlets and outlets of the hydraulic pumps), power sharing (hydraulic sharing between the hydraulic displacement control), and a method of incorporating articulation angle corrections to provide the differential effect in the desired speed. The effectiveness of a proportional–integral (PI) controller is also considered for comparison. This work found that cheater lines performed better than power sharing and performed similar to methods where the wheel speeds are adjusted base...

Recent research has considered robotic machining as an alternative to traditional computer numerical control machining, particularly for prototyping applications. However, unlike traditional machine tools, robots are subject to relatively larger dynamic disturbances and operate closer to their torque limits. Combined with inaccurate models of the manipulators and the machining process, joint actuators can often saturate during operation. Once a joint is saturated, tool-path tracking may not be possible and the blank and/or tool may be damaged. This paper presents a real-time trajectory planner designed to mitigate the effect of unmodeled disturbances, thus avoiding controller saturation and potential tool/blank damage. The forces acting on the end effectors are monitored to identify the onset of a disturbance so that the system can be slowed down before saturation actually occurs. In response to disturbances, a time-scaling method reduces the tool speed, thereby reducing the demand on the joint torques and allowing the precomputed process plan to continue. When there is sufficient torque available, the tool speed is returned to its planned magnitude. The effectiveness of the proposed time-scaling algorithm has been demonstrated with simulations. © 2006 Wiley Periodicals, Inc.

Recent research has considered robotic machining as an alternative to traditional computer numerical control machining, particularly for prototyping applications. However, unlike traditional machine tools, robots are subject to relatively larger dynamic disturbances and operate closer to their torque limits. These factors, combined with inaccurate manipulator and machining process models, can cause joint actuator saturation during operation. This paper presents a trajectory planner that will reduce torques that are near saturation by generating trajectories with a weighted pseudoinverse. Using a relative Jacobian, the tool path is resolved into joint trajectories at the acceleration level. This paper presents a new method for selecting the weighting matrix based on the proximity of the joint torques to saturation limits. This weighting reduces the joint accelerations contributing the most to the torques near saturation, thereby reducing the joint torques. The accelerations of other joints increase to satisfy the increased demand. The effectiveness of the acceleration and torque redistribution algorithm has been demonstrated via extensive simulations.

The stick-slip friction phenomenon is observed near zero relative velocity, during the transition from static to dynamic friction, when static friction is greater than dynamic friction. This nonlinear change in friction force over a small change in velocity results in difficulties in achieving accurate and repeatable position control. In some cases, the actuator position controller reaches a limit cycle (hunting effect). Friction compensation at low speeds has traditionally been approached through various control techniques. This paper proposes an alternative solution, namely, friction avoidance. By rotating the piston and rod, the Stribeck region of the friction-velocity curve is avoided and the axial friction opposing the piston movement is approximately linearized. Simulation and experimental results are presented to validate this approach.

Strain aging of Hadfield steel is discussed in terms of the interstitial octahedron, local-order model, which defines order as the probability that a C atom in an octahedral cluster of metal atoms has n (an integer between 0 and 6) Mn nearest neighbors. Equilibrium order is assessed by a Monte Carlo procedure using pair exchange energies derived from an established thermodynamic database and a Boltzmann distribution function. The disorder produced by the passage of a slip dislocation, the resulting change in free energy and, consequently, the stress opposing dislocation motion are calculated both for a single isolated dislocation and for a sequence of dislocations moving on the same slip plane. The model is extended to analyze aging effects involving diffusion of carbon before or during deformation. It is assumed that, during aging, atoms on the metal sublattice are frozen on sites determined either by the high-temperature equilibrium anneal or by prior deformation. Only diffusion of carbon is allowed. The fully aged condition at selected aging temperatures is simulated using a Monte Carlo procedure to assess local order when the free energy of the system is minimum (para-equilibrium). It is shown that the increase in strength on aging is a direct result of the relatively small thermal energy at the aging temperature favoring an increase in the number of Mn–C atom pairs. The predictions of the model are supported by the results of static aging experiments and the model provides a complete phenomenological description of dynamic strain aging in Hadfield steel.

This paper presents a laser scanner based reverse engineering (RE) system that generates accurate 3D solid models of bottles. The proposed system is designed for ease of use and has an open architecture, making it an ideal solution for bottle conveying system manufacturers. Conveyor system manufactures need a versatile RE system that can create 3D models of bottles to develop conveyor systems for their clients. This paper also presents a simple, robust filtering technique that can effectively remove the noise from the scanned laser data, yet retaining the shape of the scanned bottle. An error analysis is performed to compare our technique against another common data filtering technique, and sample 3D models are generated to contrast the effects of the different filtering approaches.

Model-based control strategies for robot manipulators can present numerous performance advantages when an accurate model of the system dynamics is available. In practice, obtaining such a model is a challenging task which involves modeling such physical processes as friction, which may not be well understood and difficult to model. This paper proposes an approach for online learning of the inverse dynamics model using Gaussian Process Regression. The Sparse Online Gaussian Process (SOGP) algorithm is modified to allow for incremental updates of the model and hyperparameters. The influence of initialization on the performance of the learning algorithms, based on any a-priori knowledge available, is also investigated. The proposed approach is compared to existing learning and fixed control algorithms and shown to be capable of fast initialization and learning rate.

This paper presents a method to detect three dimensional (3D) objects using an occupancy grid based mapping technique. This paper adapts the 3D occupancy grid based mapping technique from the two dimensional (2D) occupancy grid based mapping technique commonly used in Simultaneous Localization and Mapping applications. The 3D occupancy mapping technique uses a 3D inverse measurement model and has been developed for a LIDAR based off-road ground rover vehicle that drives on 3D terrain. The technique is developed and simulated in MATLAB to demonstrate its 3D object detection capabilities. This technique was developed as part of off-road autonomous vehicle research being conducted at the University of Waterloo.

Recent approaches to model-based manipulator control involve data-driven learning of the inverse dynamics relationship of a manipulator, eliminating the need for any knowledge of the system model. Ideally, such algorithms should be able to process large amounts of data in an online and incremental manner, thus allowing the system to adapt to changes in its model structure or parameters. LocallyWeighted Projection Regression (LWPR) and other non-parametric regression techniques have been applied to learn manipulator inverse dynamics. However, a common issue amongst these learning algorithms is that the system is unable to generalize well outside of regions where it has been trained. Furthermore, learning commences entirely from 'scratch,' making no use of any a-priori knowledge which may be available. In this paper, an online, incremental learning algorithm incorporating prior knowledge is proposed. Prior knowledge is incorporated into the LWPR framework by initializing the local linear models with a first order approximation of the available prior information. It is shown that the proposed approach allows the system to operate well even without any initial training data, and further improves performance with additional online training.

A Monte Carlo simulation model which uses thermochemical experimental data to predict short-range atomic order in face-centered-cubic (f.c.c.), solid solution alloys of Fe containing both substitutional and interstitial solute elements is described by reference to the ordering of an FeNiCrN alloy. The interaction energies between pairs of atoms and between N and an interstitial vacancy are evaluated by matching the molar Gibbs free energy, predicted by a quasichemical model, to that deduced from thermochemical data. These are then used in a Monte Carlo simulation to determine the equilibrium short-range order, defined by the distribution of Ni, Cr and N atoms among the octahedral cells of the metal lattice. Applying the model to two Fe40Ni15Cr alloys—one with zero and the other with 0.25 wt.% N—at 1273 and 298 K revealed that, at the higher temperature, the distribution of the Ni and Cr atoms in the N-free alloy is nearly random but, in the alloy containing 0.25 wt.% N, there is appreciable short-range order involving both the metal and N atoms. Decreasing the temperature to 298 K increases the order in both alloys. The implication of the short-range order and N apportionment for the deformation mechanism in FeNiCrN f.c.c. alloys is discussed.

High performance control of robotic systems, including the new generation of humanoid, assistive and entertainment robots, requires adequate knowledge of the dynamics of the system. This can be problematic in the presence of modeling uncertainties as the performance of classical, modelbased controllers is highly dependant upon accurate knowledge of the system. In addition, future robotic systems such as humanoids are likely to be redundant, requiring a mechanism for redundancy resolution when performing lower degree-of-freedom tasks. In this paper, a learning approach to estimating the inverse dynamic equations is presented. Locally Weighted Projection Regression (LWPR) is used to learn the inverse dynamics of a manipulator in both joint and task space and the resulting controllers are used to drive a 3 and 4 DOF robot in simulation. The performance of the learning controllers is compared to a traditional model based control method and is also shown to be a viable control method for a redundant system.

Among the essential characteristics of the mechanical behavior of nitrogen strengthened austenitic Fe-Ni-Cr alloys are planar slip in all alloys containing more than a small concentration of N over wide ranges of temperature and strain rate and a linear dependence of the athermal flow stress on the mole fraction of nitrogen. These characteristics of the deformation mechanism are analyzed using a model which assumes that in the undeformed, equilibrium crystal, the metal atoms are randomly arranged on the f.c.c. lattice but that the N atoms are distributed among the interstitial sites so as to minimize the free-energy. The flow stress is identified by equating the increase in free energy resulting from the redistribution of the N atoms with the work done in moving a dislocation through the lattice. Decrease in the atomic order, caused by the passage of dislocations, reduces the friction on their glide planes promoting planar slip.

Studies of the interactions between the magnetic transition and the first-order displacive martensitic transformation in FeNi and ordered and disordered FePt alloys are reviewed in an effort to find answers to two questions. Is lattice softening a necessary pre-condition for transformation? How does pre-transformation softening, when it occurs, affect the nucleation and growth of a martensitic plate? The clear demonstration that the low temperature transformation which produces martensite with the well-known acicular morphology occurs only in the ferromagnetic f.c.c. phase suggests that softening of the parent phase is necessary. However, the fact that a transformation to lath martensite occurs, at temperatures above room temperature, in dilute alloys in which the parent phase is paramagnetic, casts doubt on the validity of this generalization. There are a number of indications that the lattice softening which accompanies the magnetic transition influences the kinetics of the displacive transformation but the exact mechanisms by which nucleation is affected is still unclear. Single-crystal measurements of the soft modes in disordered FePt clearly show that they are not coupled directly to the transformation lattice strains and, consequently, it is unlikely that there is a direct connection between the soft mode lattice displacements and the displacements which produce the Bain strain. It is probable, however, that there are important indirect effects of softening on nucleation in these alloys. The influence of softening on growth, on the contrary, is clearer. Pre-transformation lattice softening is a necessary but not sufficient condition for thermoelastic reversible growth and shape memory properties.

Manufacturing and vehicle-life cost comparisons are made for selected steel and aluminum automotive body-in-white structures using a computer-based technical cost model developed by IBIS Associates, with the objective of assessing the sensitivity of the costs to change in various input parameters. From a manufacturing-cost standpoint, neither an aluminum unibody nor a spaceframe design is competitive with the steel unibody at production volumes greater than about 40,000 vehicles per year. However, both aluminum structures have lower vehicle-life costs than the steel unibody at all production volumes.

Using a dislocation model of interfacial structure, kinetic theories of dislocation motion are adapted to predict the mobility of martensitic interfaces. Defining generalized driving forces and activation parameters, analytical models are developed which describe the kinetics of motion controlled by various types of obstacle interactions. The behaviors of martensitic interfaces and slip dislocations in identical microstructures are compared. For a lattice-invariant shear by slip, the martensitic interface behaves similarly to a collection of glide dislocations. The interface/obstacle interaction is much weaker if the martensite is internally twinned, giving a higher relative mobility.

The mobility of theβ 1-γ 1 ′ martensitic interface in thermoelasticβ-Cu-Al-Ni alloys has been investigated using stress-assisted single interface transformation over the temperature range 180 to 410 K. The imposed interfacial velocities varied between 10-6 and 10-2 ms-1. The kinetic behavior is found to be consistent with thermally-activated interfacial motion, although an anomalous temperature dependence is observed below 210 K. This low temperature anomaly is attributed to the effect of the experimentally verified elastic softening. At higher temperatures, measured activation energies of 1.5 × 10-20 to 3.5 × 10-20 J and activation volumes of 103 to 104 atomic volumes are rationalized in terms of the rate-controlling interaction of the moving interface with fine-scale lattice displacements, the tweed structure, observed by transmission electron microscopy. Coarse particles ofγ- and 2H-phase, detected by the TEM, are found to give rise to noncontact particle/interface interactions which are too long-range to be surmounted by thermal activation. Consequently, they produce athermal friction stresses opposing interface motion.

The effects of vanadium/nitrogen additions on dynamic and static recovery and recrystallization have been studied in a set of aluminum-killed HSLA steels containing 0.1 pct carbon, 0.01 to 0.02 pct nitrogen, and either vanadium (0.1 or 0.2 pct), niobium (Cb) (0.03 pct), or vanadium and niobium together. Most, but not all, of the tests were carried out at 1173 K (900°C), a temperature at which precipitation of VN might be expected under some conditions. The net effect of dynamic recovery, recrystallization, and precipitation was monitored by measuring the change in compressive flow stress with strain at a constant temperature. Static changes were followed by measuring the change in compressive flow stress on isothermally holding unloaded specimens after a hot precompression. These kinetic data were supplemented by metallographic and electron-microscopic examinations of quenched specimens and of carbon extraction replicas taken from them. Evidence is presented which indicates that, at a holding temperature of 1173 K (900°C), static recrystallization occurs in vanadium steels containing 0.1 pct vanadium before any precipitation is detected. The progress of this recrystallization is arrested by the precipitation of vanadium nitride. At a higher vanadium concentration, 0.2 pct, recrystallization does not start. The effects of V/N ratio, austenitizing temperature (between 1373 K (1100°C) and 1523 K (1250°C), and isothermal holding temperature (between 1173 K (900°C) and 1273 K (1000°C)) on the kinetics of static softening and hardening are compared in some vanadium steels and plain-carbon and niobium steels of similar base-composition.

The interaction of the twinnedβ -1-γ 1 ′ , martensitic interface with various experimentally observed obstacle particles is analyzed using a specific dislocation model for the interface. The strain interaction of particles with simple shears and tetragonal distortions and their modulus interactions are treated as functions of particle crystallographic orientation, particle position with respect to interfacial intersection, and particle size. Differences from previous predictions of a simple general interface model arise primarily from differences in the assumed interfacial trajectory relative to the particles. The finite-particle calculations indicate that the point-particle approximation is valid for a particle radius less than one-tenth the interfacial twin period. Overall agreement with the experimentally measured interfacial mobility behavior is greatly improved over the previous simple model prediction. The measured athermal component of the driving force for interfacial motion is consistent with the strain and modulus interaction with 2H-phase particles. The activation-energy /driving-force relations obtained from the thermally activated component are reasonably represented by the strain interaction with the fine-scale atomic displacements of the tweed structure.

Analyse du mode de deformation: plastique ou elastique de chacune des phases en fonction des contraintes appliquees, a partir du modele de Tomota

Static precipitation and recrystallization following hot compression of austenite and the interactions between the two processes have been studied in a set of aluminum-killed HSLA steels containing 0.1 pct carbon, [0.016 - 0.026] pct nitrogen and 0.1 or 0.2 pct vanadium. Two steels containing both vanadium (0.1 and 0.2 pct) and niobium (0.03 pct) were included for purposes of comparison. The compression and the static tests were all carried out isothermally at temperatures between 800 and 900 °C. The course of recrystallization was followed by measurements of the rate of softening and by optical metallography of specimens quenched from the test temperature after different times. Precipitation was studied by measurements of the rate of hardening, by transmission electron microscopy of thin foils, carbon and aluminum extraction replicas, and by X-ray dispersion and energy-loss spectroscopy from individual precipitates.

The mobility of s-γ[MATH] martensitic interfaces in thermoelastic s-CuAlNi alloys is investigated using stress-assisted single interface transformations over a temperature range of 180 to 410K and imposed interfacial velocities of 10-6 to 10-2 m/s. The behavior is consistent with thermally-activated interfacial motion, although an anomalous temperature dependence is observed below 210K which may be due to elastic softening. Measured activation energies of 0.3 to 0.4 eV (4.5 x 10-20 to 6.5 x 10-20 J) and activation volumes of 103 to 104 atomic volumes suggest that motion of the interface above 210K is controlled by interaction with discrete particles or clusters.

The effects of austenitizing conditions on the kinetics at the start of martensite formation in Fe-31Ni and Fe-31 Ni-0.28C alloys have been studied using electrical-resistance measurements during cooling of the specimens to follow the course of the transformation. The primary object of the study was to decide whether or not a change in austenitizing temperature, in the absence of a change in austenite grain size, has any effect on the Ms temperature or the burst characteristics of athermal martensite. It is concluded that it does not, suggesting that the potential nuclei (embryos) of martensite are mechanically stable crystal defects. Another interesting observation is that when the austenite grain size is small, the Mb temperature increases with increasing grain size and the burst is always small. When the austenite grains are coarse, the Mb temperature is independent of the grain size and the burst is large. It is suggested that this phenomenon is a result of the elastic shear stress concentration being related to the size of the first martensite plate and, in turn, to the size of the austenite grain.

A series of iron-platinum alloys containing 25 or 27 at. pct platinum and with ordering of the γ-phase varying from substantial disorder to nearly complete order have been thermally cycled between 25°C and - 196°C. The kinetics of the γ⇌α transformations, the hysteresis revealed by electrical resistanceJtemperature plots, the thermoelastic growth and the reappearance of an identical microstructure after thermal cycling (the microstructural memory effect) were studied as a function of the ordering of the γ-phase. Thermoelastic growth does not appear to be affected by changes in the degree of order of partially ordered specimens but the microstructural memory was imperfect in the most highly ordered specimen examined. In agreement with earlier observations by Dunne and Wayman,3 the difference between the As and Ms temperatures and the hysteresis decrease markedly as the order is increased. It is shown by transmission electron microscopy that in all but the most highly-ordered specimens the α- γ transformation produces plates of austenite with a high density of dislocations. These plates are separated from the surrounding untransformed parent austenite by arrays of dislocation loops lying in the interfaces between untransformed parent austenite and the original martensite plates. All the dislocations have a Burgers vector direction which is the same as that of the usual slip dislocation in austenite. Such dislocations lying in a habit plane must be sessile. In the well-ordered specimen dislocation pairs, typical of glide dislocations in a crystal with long-range order, were formed in the austenite formed by the reverse transformation. These dislocations were segregated into roughly plate-like clusters, but the number of clusters in unit volume was appreciably less than the number of original plates of martensite. In this case, no arrays of sessile loops of dislocations mark the locations of the original martensite-austenite interface. It is deduced from the microscopic and kinetic results that the inherited nuclei responsible for the microstructural memory effect are located in localized volumes of highly dislocated austenite formed by the α- γ transformation. No unique dislocation configurations which could be associated with specific nuclei were found. The effects of ordering on the various kinetic effects and the microstructural memory are discussed in terms of the concept of inherited nuclei, the change of the flow stress of the γ-phase with ordering and temperature and the variation of To and the transformation driving-force with ordering.

Laws passed and regulations approved by the Congress of the USA in recent years require new automobiles in the USA to conform to standards of NOx emission, impact resistance, and fuel consumption which will test to the limit the technological and manufacturing resourcefulness of the American automobile industry. Radical changes will have to be made in the materials used. Changes in engine design and a reduction in the size of the automobiles will be insufficient. Steel is the cheapest acceptable material for many components, but steel will retain its position in the automobile market only if it can be made much stronger (thus reducing the section thickness and weight) without impairing its ductility, formability, weldability, and low cost. New steels for automobiles have been developed from the laboratory to full-scale production with astonishing speed in the last three years. The most promising of the new steels – ‘dual-phase steels’ – have more than twice the strength of traditional automobile s...

Five different dual-phase microstructures have been produced in a steel containing 0.11 C, 1.60 Mn, 0.73 Si. By careful design of the heat treatment schedules, it was arranged that all the specimens had the same fine austenite grain size before intercritical annealing and that they all contained close to 30 vol pct martensite and a negligible fraction of retained austenite after annealing. The microstructures were characterized by measurement of the mean ferrite path and the density and distribution of the extrinsic transformation accommodation dislocations in the substructure of the ferrite. Specimens representative of each of the microstructures were tested in tension at room temperature. The strength, work-hardening, and fracture of three of them were examined in detail and correlated with the microstructural parameters. The microstructural features which most influence the tensile properties are identified. Two specimens, representative of the finest and coarsest microstructures, were the subject of a detailed exploration of the initiation of persistent slip and microcracks and of crack propagation in bending fatigue. The fatigue endurance limit was also measured. It is shown that the initiation of persistent slip is primarily influenced by the heterogeneity in the density of the dislocations in the ferrite. The microstructural features of greatest importance in controlling the formation and propagation of microcracks are the local dislocation density and the mean ferrite path.

The variations of the lattice parameter, the Curie temperature (determined from specific heat measurements), and the temperature variation of the elastic constants C, C′, and Cl in an applied magnetic field of 0.8 T have all been measured as a function of the long-range order in a single crystal of iron-platinum alloy with very nearly the stoichiometric composition Fe3Pt. It is found that the Invar effect in partially ordered Fe-25 Pt is larger than in other Fe-Pt alloys or in well-known Invar alloys such as Fe-35 Ni. The results for Fe-25 Pt are analyzed and interpreted in terms of a model of nearest-neighbor exchange interaction which takes account of the spatial variation of the average exchange integral. The best fit with the experimental data is obtained by assuming an average spin of unity. Using ordering as the controlled variable avoids the difficulties inherent in studies which have relied on a compositional variable. The model satisfactorily describes the magnetic contribution to the elastic constants and the variation of Curie temperature with changes in the long-range order. It is concluded that Fe-Pt coupling is ferromagnetic and Fe-Fe coupling antiferromagnetic.

The variation in the equilibrium compositions of the parent phase and the (V,Nb)(C,N) precipitate phase in Fe-C-N-V-Nb-Mn is considered at two temperatures at which the parent phase is austenite and over a range of composition of interest in relation to microalloyed steels. The f.c.c. austenite phase is described by the Kohler temperature dependent-subregular solution model. In an earlier discussion of this system, we assumed the carbonitride phase to be an ideal solution of VC and NbC. Here, we distinguish between C and N and describe the (V,Nb)(C,N) precipitate phase by the Hillert-Staffansson sublattice model for stoichiometric phases.

The low-temperature flow stress of mechanically stable austenitic stainless steels increases with increasing concentration of nitrogen in solution and with decreasing temperature. This phenomenon has been studied in a series of FeNiCrMo alloys with nitrogen contents between 0.04 and 0.36 wt% by measuring the flow stress and the thermal activation parameters for plastic flow as a function of stress, plastic strain and nitrogen concentration in stress relaxation and strain-rate change experiments. Care is taken, when analyzing the data, to distinguish between athermal and thermal effects. The significant increase of the athermal flow stress with increasing nitrogen concentration is attributed to short-range ordering of chromium and nitrogen atoms. The thermally activated component of the flow stress is also dependent on the nitrogen concentration and is thought to be due to localized, predominantly modulus interactions between lattice disturbances in the immediate vicinity of nitrogen atoms and slip dislocations. The thermally activated component is suitably described by Friedel's model of solid solution strengthening.

The development of dual-phase structures with different morphologies has been studied in detail by intercritical annealing specimens of a steel containing 0.11 pct C and 1.6 pct Mn with different microstructures before annealing. The kinetics of formation of the two-phase structure at the annealing temperature and the redistribution of substitutional solute elements were measured in specimens quenched from the intercritical annealing temperature. The structure before annealing was either banded ferrite-pearlite, homogenized ferrite-pearlite, lath martensite, spheroidal cementite dispersed in ferrite, or austenite. No measurable partitioning of silicon or molybdenum, present in the steel in small concentrations, was found. However, close to equilibrium partitioning of manganese occurred on annealing specimens with either a ferrite-pearlite or a lath martensite structure, but during the separation of ferrite from austenite in step-quenched specimens there was no partitioning. Surprisingly, measurements of manganese concentrations using an electron beam of 1 nm diameter at intervals of 5 nm or less revealed the presence of narrow spikes in the concentration profile at many ferriteaustenite interfaces in specimens with a ferrite-pearlite or martensite starting structure as well as in those step-quenched from austenite. In some instances, a minimum in the concentration profile was found in ferrite, adjacent to a maximum at an interface. Thus, adsorption of manganese at ferriteaustenite interfaces produces concentrations in excess of the concentrations indicated by the equilibrium diagram. The probable diffusion processes controlling the kinetics of transformation in the different microstructures are identified.

The effect of vanadium on the isothermal austenite-ferrite transformation, between 725 °C and 775 °C, of a hot-deformed microalloyed steel has been studied by examination of the microstructure and measurement of the volume fraction of ferrite in specimens quenched from the reaction temperature. The accompanying precipitation was studied by transmission electron microscopy of thin foils and carbon extraction replicas and by electron energy-loss spectroscopy. Very early in the transformation a continuous band of fine-grained ferrite forms at austenite grain boundaries. After some time some of these grains coarsen to form large equiaxed ferrite grains. It is found that vanadium has no effect on the time to the start of coarsening but thereafter accelerates the rate of formation of ferrite. Interphase precipitation of VN occurs throughout the transformation in the vanadium steels and this is thought to influence the rate at which the ferrite coarsens at the lower temperatures (750 ° and 725 °C) in the range studied.

The formation of duplex carbonitride precipitates in an ultra-low carbon Ti-Nb steel, either isothermally annealed or controlled rolled after a high temperature solutionizing treatment, has been studied experimentally and the relative phase stabilities of the precipitate and matrix phases, over the temperature range of the experiments, have been calculated. The austenite and ferrite phases were modelled using the Redlich-Kister-Muggianu interpolation formula and the (TiXNb1−X)(C1−YNY) phases, which were assumed to be stoichiometric, using the Hillert-Staffansson formalism. Two possible ways in which the duplex precipitates could have been formed were explored: Equilibrium phase separation within a (TiXNb1−X(C1−YNY) phase or two-step precipitation, the low-temperature precipitate being nucleated on a precipitate which had been formed earlier at a higher temperature. Comparison of the experimental and theoretical results clearly indicates that both the duplex precipitates formed by precipitation in austenite in specimens double-annealed at 1250C and 900C and those formed by controlled rolling specimens to 700C from a solution temperature of 1050C were formed by two-step precipitation processes.

Eshelby's continuum elastic model of the strain field of a coherent inclusion in an infinite matrix is adapted to describe the strain field associated with a circular plate of ellipsoidal cross-section transformed martensitically. Initially, the plate and the austenitic matrix are assumed to be elastically Isotropic, but with different moduli. The model is also used to calculate the strain field of a pair of parallel, ellipsoidal, martensite plates and of a single, isolated, square, flat plate. Finally, the treatment is extended to reveal the effects of elastic anisotropy. A criterion for thermoelastic growth of an isolated plate of martensite in terms of the permissible plastic yielding is proposed. Using the model of the strain fields, contours of the yield envelopes associated with an isolated plate of martensite formed in Fe3Pt are calculated as a function of the atomic order in the parent, austenitic phase. It is concluded that thermoelastic growth is to be expected when the long-range order, S, is greater or equal to 0.2, in agreement with experimental observations. The equilibrium thickness of an isolated thermoelastic plate at Ms and the associated chemical driving force are found to be much smaller than for plates formed irreversibly. Quantitatively, the predictions of the model applied to the ‘ordered’ alloy are in good agreement with experimental measurements. Clearly, Invar elastic softening is a major factor determining thermoelastic behavior in Fe3Pt. In addition, the model is used to explore the reduction in elastic accommodation strain which can be achieved by the interactions of two or more plates arranged edge-to-edge or in stacks of self-accommodating plates. The influence of the shape of the individual plates is also discussed. Calculations using the anisotropic model for an isolated plate in the form of an oblate spheroid show that the effects of the large elastic anisotropy, which develops on cooling partially ordered Fe3Pt, on the equilibrium thickness of the thermoelastic plate and the chemical driving force at Ms are surprisingly small.

The variation in the equilibrium compositions of austenite and titanium-niobium-carbonitride precipitates in Fe-Ti-Nb-C-N-Mn-Si-Mo is considered over a range of temperatures between 1050 and 1500K and for an alloy composition which is typical of ultra-low carbon titanium-niobium microalloyed steels. The Kohler temperature dependent-subregular solution model is used to describe the austenite phase. The (Ti,Nb)(C,N)x precipitate phase is described by the Hillert-Staffansson sublattice model for a four-component solution, with components mixed in pairs. The equilibrium between the two phases is explored while allowing for the nonstoichiometry of the precipitate phase. The results are shown to be in good agreement with experimental data on ultra-low carbon microalloyed steels.

The plastic flow behavior of Hadfield manganese steel in uniaxial tension and compression is shown to be greatly influenced by transformation plasticity phenomena. Changes in the stress-strain (σ−e) curves with temperature correlate with the observed extent of deformation twinning, consistent with a softening effect of twinning as a deformation mechanism and a hardening effect of the twinned microstructure. The combined effects give upward curvature to the σ−e curve over extensive ranges of plastic strain. A higher strain hardening in compression compared with tension appears to be consistent with the observed texture development. The composition dependence of stacking fault energy computed using a thermodynamic model suggests that the Hadfield composition is optimum for a maximum rate of deformation twinning. Comparisons of the Hadfield steel with a Co-33Ni alloy exhibiting similar twinning kinetics, and an Fe-21Ni-lC alloy deforming by slip indicate no unusual strain hardening at low strains where deformation is controlled by slip, but an unusual amount of structural hardening associated with the twin formation in the Hadfield steel. A possible mechanism of anomalous twin hardening is discussed in terms of modified twinning behavior (pseudotwinning) in nonrandom solid solutions.

It has been pointed out (Old 1972) that the common practice of reporting a company's research and development (R and D) expenditures as a ratio of net sales can be misleading. Instead two ratios are suggested: one, R and D expense divided by cash flow, and the other, technologybased capital investment less tax credit divided by cash flow. Typically, on this basis, a U.S. steel company might have an R and D investment ratio of 0.03 and a technology capital investment ratio greater than 0.60. Thus, about 65 % of the total cash flow is invested in technology. This is another way of saying that the steel industry is both technology and capital intensive. For a company's R and D programme to be cost-effective, it must develop technology in which it will prove profitable to invest at some not too distant time. The R and D investment ratio is so much smaller than the capital investment ratio that it is inevitable that much of the research done in a company's R and D programme will not contribute directly to a new or modified technology which will attract capital investment. In this sense, for many years research was considered to be relatively cheap and few serious attempts to assess its cost effectiveness or to direct it into profitable channels were made. In another sense, research of this kind was considered by many steel companies to be a luxury so that cuts in research budgets occurred early in periods when economies in operating costs were considered necessary. In recent years there has been a substantial change in outlook. Basic research is no longer relatively cheap and it is realized that in many cases it is more profitable to buy, at home or abroad, the technology in which large investments are to be made. The communication of technical information within the steel industry, nationally and internationally, is remarkably unrestricted (Hiestand 1974). In the United States this situation dates back to 1922 when, in settlement of a major antitrust prosecution, the U.S. Steel Corporation agreed to make technological information available to its competition to help them expand. The selection of a technology in which a company decides to invest is influenced greatly by the local conditions: its geographical location in relation to ore and fuel sources, the nature of the products required by its market, the technology in which capital is already invested, the local environmental

The equilibrium compositions, at 1073 and 1123K, of the coexisting austenite (f.c.c.) and (Nb,V)(C,N) precipitate phases in steels containing 1.4 w/o Mn, 0.02 w/o N, up to 0.2 w/o C, 0.2 w/o V, 0.03 w/o Nb are computed using a Kohler, temperature dependent, subregular solution model for the component binary systems and an ideal solid solution model for (Nb,V)(C,N). It was found to be necessary to introduce certain simplifying approximations and these are set out in some detail. The computer method used involves reducing six equations with six unknowns to two nonlinear equations with two unknowns which are then solved using the Newton-Raphson iteration method. The calculated compositions of the precipitate phase are compared with the compositions measured earlier by quantitative X-ray microanalysis.