The Engineering Sciences Center at Sandia National Laboratories provided an independent peer review of the structural analysis supporting the National Transportation Safety Board investigation of the August 1, 2007 collapse of the I-35W Bridge in Minneapolis. The purpose of the review was to provide an impartial critique of the analysis approach, assumptions, solution techniques, and conclusions. Subsequent to reviewing numerous supporting documents, a SNL team of staff and management visited NTSB to participate in analysis briefings, discussions with investigators, and examination of critical elements of the bridge wreckage. This report summarizes the opinion of the review team that the NTSB analysis effort was appropriate and provides compelling supporting evidence for the NTSB probable cause conclusion.
The primary goals of the present study are to: (1) determine how and why MEMS-scale friction differs from friction on the macro-scale, and (2) to begin to develop a capability to perform finite element simulations of MEMS materials and components that accurately predicts response in the presence of adhesion and friction. Regarding the first goal, a newly developed nanotractor actuator was used to measure friction between molecular monolayer-coated, polysilicon surfaces. Amontons law does indeed apply over a wide range of forces. However, at low loads, which are of relevance to MEMS, there is an important adhesive contribution to the normal load that cannot be neglected. More importantly, we found that at short sliding distances, the concept of a coefficient of friction is not relevant; rather, one must invoke the notion of 'pre-sliding tangential deflections' (PSTD). Results of a simple 2-D model suggests that PSTD is a cascade of small-scale slips with a roughly constant number of contacts equilibrating the applied normal load. Regarding the second goal, an Adhesion Model and a Junction Model have been implemented in PRESTO, Sandia's transient dynamics, finite element code to enable asperity-level simulations. The Junction Model includes a tangential shear traction that opposes the relative tangential motion of contacting surfaces. An atomic force microscope (AFM)-based method was used to measure nano-scale, single asperity friction forces as a function of normal force. This data is used to determine Junction Model parameters. An illustrative simulation demonstrates the use of the Junction Model in conjunction with a mesh generated directly from an atomic force microscope (AFM) image to directly predict frictional response of a sliding asperity. Also with regards to the second goal, grid-level, homogenized models were studied. One would like to perform a finite element analysis of a MEMS component assuming nominally flat surfaces and to include the effect of roughness in such an analysis by using a homogenized contact and friction models. AFM measurements were made to determine statistical information on polysilicon surfaces with different roughnesses, and this data was used as input to a homogenized, multi-asperity contact model (the classical Greenwood and Williamson model). Extensions of the Greenwood and Williamson model are also discussed: one incorporates the effect of adhesion while the other modifies the theory so that it applies to the case of relatively few contacting asperities.
This work demonstrates a polycrystalline silicon surface-micromachined inchworm actuator that exhibits high-performance characteristics such as large force ({+-}0.5 millinewtons), large velocity range (0 to {+-}4.4 mm/sec), large displacement range ({+-}100 microns), small step size ({+-}10, {+-}40 or {+-}100 nanometers), low power consumption (nanojoules per cycle), continuous bidirectional operation and relatively small area (600 x 200{micro}m{sup 2}). An in situ load spring calibrated on a logarithmic scale from micronewtons to millinewtons, optical microscopy and Michelson interferometry are used to characterize its performance. The actuator consists of a force-amplifying plate that spans two voltage-controlled clamps, and walking is achieved by appropriately sequencing signals to these three components. In the clamps, normal force is borne by equipotential rubbing counterfaces, enabling friction to be measured against load. Using different monolayer coatings, we show that the static coefficient of friction can be changed from 0.14 to 1.04, and that it is load-independent over a broad range. We further find that the static coefficient of friction does not accurately predict the force generated by the actuator and attribute this to nanometer-scale presliding tangential deflections.
The Microsystems Subgrid Physics project is intended to address gaps between developing high-performance modeling and simulation capabilities and microdomain specific physics. The initial effort has focused on incorporating electrostatic excitations, adhesive surface interactions, and scale dependent material and thermal properties into existing modeling capabilities. Developments related to each of these efforts are summarized, and sample applications are presented. While detailed models of the relevant physics are still being developed, a general modeling framework is emerging that can be extended to incorporate evolving material and surface interaction modules.
This report summarizes experimental and test results from a two year LDRD project entitled Real Time Error Correction Using Electromagnetic Bearing Spindles. This project was designed to explore various control schemes for levitating magnetic bearings with the goal of obtaining high precision location of the spindle and exceptionally high rotational speeds. As part of this work, several adaptive control schemes were devised, analyzed, and implemented on an experimental magnetic bearing system. Measured results, which indicated precision positional control of the spindle was possible, agreed reasonably well with simulations. Testing also indicated that the magnetic bearing systems were capable of very high rotational speeds but were still not immune to traditional structural dynamic limitations caused by spindle flexibility effects.
A MEMS test structure capable of measuring friction between polysilicon surfaces under a variety of test conditions has been refined from previous designs. The device is applied here to measuring friction coefficients of polysilicon surfaces under different environmental, loading, and surface conditions. Two methods for qualitatively comparing friction coefficients (μ) using the device are presented. Samples that have been coated with a self-assembled monolayer of the lubricating film perfluorinated-decyltrichlorosilane (PFTS) have a coefficient of friction that is approximately one-half that of samples dried using super-critical CO2 (SCCO2) drying. Qualitative results indicate that μ is independent of normal pressure. Wear is shown to increase μ for both supercritically dried samples and PFTS coated samples, though the mechanisms appear to be different. Super critically dried surfaces appear to degrade continuously with increased wear cycles, while PFTS coated samples reach a steady state friction value after about 105 cycles.
The force exerted on the rotor by an active magnetic bearing (AMB) is determined by the current flow in the magnet coils. This force can be controlled very precisely, making magnetic bearings a potential benefit for grinding, where cutting forces act as external disturbances on the shaft, resulting in degraded part finish. It is possible to achieve precise shaft positioning, reduce vibration of the shaft caused by external disturbances, and even damp out resonant modes. Adaptive control is an appealing approach for these systems because the controller can tune itself to account for an unknown periodic disturbance, such as cutting or grinding forces, injected into the system. In this paper the authors show how one adaptive control algorithm can be applied to an AMB system with a periodic disturbance applied to the rotor. An adaptive algorithm was developed and implemented in both simulation and hardware, yielding significant reductions in rotor displacement in the presence of an external excitation. Ultimately, this type of algorithm could be applied to a magnetic bearing grinder to reduce unwanted motion of the spindle which leads to poor part finish and chatter.
A project to develop non-intrusive active sensors that can be applied on existing aging aerospace structures for monitoring the onset and progress of structural damage (fatigue cracks and corrosion) is presented. The state of the art in active sensors structural health monitoring and damage detection is reviewed. Methods based on (a) elastic wave propagation and (b) electro-mechanical (NM) impedance technique are sighted and briefly discussed. The instrumentation of these specimens with piezoelectric active sensors is illustrated. The main detection strategies (E/M impedance for local area detection and wave propagation for wide area interrogation) are discussed. The signal processing and damage interpretation algorithms are tuned to the specific structural interrogation method used. In the high-frequency EIM impedance approach, pattern recognition methods are used to compare impedance signatures taken at various time intervals and to identify damage presence and progression from the change in these signatures. In the wave propagation approach, the acoustic-ultrasonic methods identifying additional reflection generated from the damage site and changes in transmission velocity and phase are used. Both approaches benefit from the use of artificial intelligence neural networks algorithms that can extract damage features based on a learning process. Design and fabrication of a set of structural specimens representative of aging aerospace structures is presented. Three built-up specimens, (pristine, with cracks, and with corrosion damage) are used. The specimen instrumentation with active sensors fabricated at the University of South Carolina is illustrated. Preliminary results obtained with the E/M impedance method on pristine and cracked specimens are presented.
A project to develop non-intrusive active sensors that can be applied on existing aging aerospace structures for monitoring the onset and progress of structural damage (fatigue cracks and corrosion) is presented. The state of the art in active sensors structural health monitoring and damage detection is reviewed. Methods based on (a) elastic wave propagation and (b) electro-mechanical (E/M) impedance technique are cited and briefly discussed. The instrumentation of these specimens with piezoelectric active sensors is illustrated. The main detection strategies (E/M impedance for local area detection and wave propagation for wide area interrogation) are discussed. The signal processing and damage interpretation algorithms are tuned to the specific structural interrogation method used. In the high-frequency E/M impedance approach, pattern recognition methods are used to compare impedance signatures taken at various time intervals and to identify damage presence and progression from the change in these signatures. In the wave propagation approach, the acousto-ultrasonic methods identifying additional reflection generated from the damage site and changes in transmission velocity and phase are used. Both approaches benefit from the use of artificial intelligence neural networks algorithms that can extract damage features based on a learning process. Design and fabrication of a set of structural specimens representative of aging aerospace structures is presented. Three built-up specimens (pristine, with cracks, and with corrosion damage) are used. The specimen instrumentation with active sensors fabricated at the University of South Carolina is illustrated. Preliminary results obtained with the E/M impedance method on pristine and cracked specimens are presented.
The ultimate limitation in obtainable resolution and sensitivity for space-based imaging systems is the size of the optical collecting aperture. Large collecting apertures are at odds with maintaining low launch costs and with current launch vehicle configurations. Development of a deployable mirror is one approach being considered to satisfy these conflicting requirements. The focus of this research is to develop fundamental technology toward the realization of deployable electron-gun-controlled piezoelectric thin films mirrors as shown below. A bimorph layer of film will bend in response to an applied electric field and can therefore be deformed into desirable shapes using a scanning electron gun. Surface curvature measurements govern the electron gun scanning strategy, yielding distributed shape corrections.
Due to extreme surface to volume ratios, adhesion and friction are critical properties for reliability of Microelectromechanical Systems (MEMS), but are not well understood. In this LDRD the authors established test structures, metrology and numerical modeling to conduct studies on adhesion and friction in MEMS. They then concentrated on measuring the effect of environment on MEMS adhesion. Polycrystalline silicon (polysilicon) is the primary material of interest in MEMS because of its integrated circuit process compatibility, low stress, high strength and conformal deposition nature. A plethora of useful micromachined device concepts have been demonstrated using Sandia National Laboratories' sophisticated in-house capabilities. One drawback to polysilicon is that in air the surface oxidizes, is high energy and is hydrophilic (i.e., it wets easily). This can lead to catastrophic failure because surface forces can cause MEMS parts that are brought into contact to adhere rather than perform their intended function. A fundamental concern is how environmental constituents such as water will affect adhesion energies in MEMS. The authors first demonstrated an accurate method to measure adhesion as reported in Chapter 1. In Chapter 2 through 5, they then studied the effect of water on adhesion depending on the surface condition (hydrophilic or hydrophobic). As described in Chapter 2, they find that adhesion energy of hydrophilic MEMS surfaces is high and increases exponentially with relative humidity (RH). Surface roughness is the controlling mechanism for this relationship. Adhesion can be reduced by several orders of magnitude by silane coupling agents applied via solution processing. They decrease the surface energy and render the surface hydrophobic (i.e. does not wet easily). However, only a molecular monolayer coats the surface. In Chapters 3-5 the authors map out the extent to which the monolayer reduces adhesion versus RH. They find that adhesion is independent of RH up to a threshold value, depending on the coating chemistry. The mechanism for the adhesion increase beyond this threshold value is that the coupling agent reconfigures from a surface to a bulk phase (Chapter 3). To investigate the details of how the adhesion increase occurs, the authors developed the mechanics for adhesion hysteresis measurements. These revealed that near-crack tip compression is the underlying cause of the adhesion increase (Chapter 4). A vacuum deposition chamber for silane coupling agent deposition was constructed. Results indicate that vapor deposited coatings are less susceptible to degradation at high RH (Chapter 5). To address issues relating to surfaces in relative motion, a new test structure to measure friction was developed. In contrast to other surface micromachined friction test structures, uniform apparent pressure is applied in the frictional contact zone (Chapter 6). The test structure will enable friction studies over a large pressure and dynamic range. In this LDRD project, the authors established an infrastructure for MEMS adhesion and friction metrology. They then characterized in detail the performance of hydrophilic and hydrophobic films under humid conditions, and determined mechanisms which limit this performance. These studies contribute to a fundamental understanding for MEMS reliability design rules. They also provide valuable data for MEMS packaging requirements.
A technological break through for supporting rotating shafts is the active magnetic bearing (AMB). Active magnetic bearings offer some important advantages over conventional ball, roller or journal bearings such as reduced frictional drag, no physical contact in the bearing, no need for lubricants, compatibility with high vacuum and ultra-clean environments, and ability to control shaft position within the bearing. The disadvantages of the AMB system are the increased cost and complexity, reduced bearing stiffness and the need for a controller. Still, there are certain applications, such as high speed machining, biomedical devices, and gyroscopes, where the additional cost of an AMB system can be justified. The inherent actuator capabilities of the AMB offer the potential for active balancing of spindles and micro-shaping capabilities for machine tools, The work presented in this paper concentrates on an AMB test program that utilizes the actuator capability to dynamically balance a spindle. In this study, an unbalanced AMB spindle system was enhanced with an LMS (Least Mean Squares) algorithm combined with an existing PID (proportional, integral, differential) control. This enhanced controller significantly improved the concentricity of an intentionally unbalanced shaft. The study included dynamic system analysis, test validation, control design and simulation, as well as experimental implementation using a digital LMS controller.
As part of a collaborative effort between Sandia National Laboratories and the University of Kentucky to develop a deployable mirror for remote sensing applications, research in shape sensing and control algorithms that leverage the distributed nature of electron gun excitation for piezoelectric bimorph mirrors is summarized. A coarse shape sensing technique is developed that uses reflected light rays from the sample surface to provide discrete slope measurements. Estimates of surface profiles are obtained with a cubic spline curve fitting algorithm. Experiments on a PZT bimorph illustrate appropriate deformation trends as a function of excitation voltage. A parallel effort to effect desired shape changes through electron gun excitation is also summarized. A one dimensional model-based algorithm is developed to correct profile errors in bimorph beams. A more useful two dimensional algorithm is also developed that relies on measured voltage-curvature sensitivities to provide corrective excitation profiles for the top and bottom surfaces of bimorph plates. The two algorithms are illustrated using finite element models of PZT bimorph structures subjected to arbitrary disturbances. Corrective excitation profiles that yield desired parabolic forms are computed, and are shown to provide the necessary corrective action.
The conflicting demands for finer features and increased production rates in integrated circuit manufacturing have emphasized the need for improved wafer positioning technology. In this paper we present operational test results from a magnetically levitated platen with structurally integrated piezoelectric actuators. The strain based actuators provide active damping of the platen's flexible body modes, enabling increased bandwidth on the mag-lev positioning system. Test results reveal a dramatic reduction in steady state positioning error and settling time through implementation of active vibration control.
This paper summarizes the results of cutting tests performed using an actively damped boring bar to minimize chatter in metal cutting. A commercially available 2 inch diameter boring bar was modified to incorporate PZT stack actuators for controlling tool bending vibrations encountered during metal removal. The extensional motion of the actuators induce bending moments in the host structure through a two-point preloaded mounting scheme. Cutting tests performed at various speeds and depths of cuts on a hardened steel workpiece illustrate the bar`s effectiveness toward eliminating chatter vibrations and improving workpiece surface finish.
This paper summarizes the design, modeling, and initial evaluation of a hinged structure for friction measurement in surface micromachining technology. While the area requirements are small, the present structure allows a much larger velocity and pressure range to be evaluated as compared to comb drive structures. The device consists of a cantilevered driver beam connected to a friction pad through a strategically located hinge. AC excitation of the beam flexure forces axial sliding of the friction pad due to beam foreshortening. Normal force is controlled by DC voltage on wings adjacent to the friction pad. While the achievable slip is small (10-30 nm), it is sufficient to disengage contacting asperities and engage new points of contact, and thus should be representative of frictional processes. Furthermore, the design enables the friction pad contact area to remain relatively constant over the excitation cycle. Computer simulation results are provided to mimic on-going experimental work. Increased friction forces are shown to enhance the size of hysteresis loops relating beam deflection to driver voltage.
This paper summarizes results of metal cutting tests using an actively damped boring bar to suppress regenerative chatter. PZT stack actuators were integrated into a commercially available two-inch diameter boring bar to suppress bending vibrations. Since the modified tool requires no specialized mounting hardware, it can be readily mounted on a variety of machines. A cutting test using the prototype bar to remove metal from a hardened steel workpiece verifies that the actively damped tool yields significant vibration reduction and improved surface finish as compared to the open-loop case. In addition, the overall performance of the prototype bar is compared to that of an unmodified bar of pristine geometry, revealing that a significant enlargement of the stable machining envelope is obtained through application of feedback control.
Recent advances in smart materials have renewed interest in the development of improved manufacturing processes featuring sensing, processing, and active control. In particular, vibration suppression in metal cutting has received much attention because of its potential for enhancing part quality while reducing the time and cost of production. Although active tool clamps have been recently demonstrated, they are often accompanied by interfacing issues that limit their applicability to specific machines. Under the auspices of the Laboratory Directed Research and Development program, the project titled {open_quotes}Smart Cutting Tools for Precision Manufacturing{close_quotes} developed an alternative approach to active vibration control in machining. Using the boring process as a vehicle for exploration, a commercially available tool was modified to incorporate PZT stack actuators for active suppression of its bending modes. Since the modified tool requires no specialized mounting hardware, it can be readily mounted on many machines. Cutting tests conducted on a horizontal lathe fitted with a hardened steel workpiece verify that the actively damped boring bar yields significant vibration reduction and improved surface finishes as compared to an unmodified tool.
The development and initial evaluation of a prototype boring bar featuring active vibration control for increased chatter immunity is described. The significance of active damping both normal and tangential to the workpiece surface is evaluated, indicating the need for two axis control to ensure adequate performance over expected variations in tool mounting procedures. The prototype tool features a commercially available boring bar modified to accommodate four PZT stack actuators for two axis bending control. Measured closed-loop dynamics are combined with a computer model of the boring process to simulate increased metal removal rate and improved workpiece surface finish through active control.
This paper describes an experimental investigation of workpiece surface waviness that stems from poor single point diamond wheel dressing procedures in cylindrical grinding. If done improperly, single point dressing can produce a thread on the grinding wheel surface that is then imposed on the workpiece during machining. The circumferential waviness exhibited by the threaded workpiece is similar to that resulting from one per rev vibrations of the grinding wheel. In order to differentiate between these two sources of waviness, a geometrical approach to predicting the circumferential and axial waves produced during grinding is presented. The concepts presented are illustrated through a series of plunge grinding tests incorporating dressing procedures of varying quality. Test results verify that dressing induced circumferential waviness is similar to waviness resulting from one per rev type vibrations of the grinding wheel. The two sources can be distinguished, however, through examination of the workpiece waviness in the axial direction.
An experimental investigation into active control of bending vibrations in thick plate-like structural elements is described. This work is motivated by vibration problems encountered in manufacturing processes that require greater control authority than is available from conventional surface mounted PZT patches or PVDF films. The focus of this experiment is a surrogate photolithography platen in which PZT stacks are mounted in cutouts on the platen top surface. These actuators provide significant vibration control authority by generating moments in the platen through their compressive loads. A Positive Position Feedback control law is used to significantly augment the damping in the first two bending modes. The implications of the experimental results for photolithography machines are discussed.
An experimental investigation into active control of bending vibrations in thick bar and plate-like structural elements is described. This work is motivated by vibration problems in machine tools and photolithography machines that require greater control authority than available from conventional surface mounted PZT patches or PVDF films. Focus of this experiment is a cantilevered circular steel bar in which PZT stacks are mounted in cutouts near the bar root. Axially aligned and offset from the neutral axis, these actuators control the bending vibrations by generating moments in the bar through their compressive loads. A Positive Feedback control law is used to significantly augment the damping in the first bending mode. Implications of the experimental results for machine tool stability enhancement are discussed.
This report summarizes two approaches to time-optimal control of a nonlinear magnetically levitated platen. The system of interest is a candidate technology for next-generation photolithography machines used in the manufacture of integrated circuits. The dynamics and the variable peak control force of the electro-magnetic actuators preclude the direct application of classical time-optimal control methodologies for determining optimal rest-to-rest maneuver strategies. Therefore, this study explores alternate approaches using a previously developed computer simulation. In the first approach, conservative estimates of the available control forces are used to generate suboptimal switching curves. In the second approach, exact solutions are determined iteratively and used as a training set for an artificial neural network. The trained network provides optimal actuator switching times that incorporate the full nonlinearities of the magnetic levitation actuators. Sample problems illustrate the effectiveness of these techniques as compared to traditional proportional-derivative control.
This paper describes an application of artificial neural networks to the problem of time-optimal control of a magnetically levitated platen. The system of interest is a candidate technology for advanced photolithography machines used in the manufacturing of integrated circuits. The nonlinearities associated with magnetic levitation actuators preclude the direct application of classical timeoptimal control methodologies for determining optimal rest-to-rest maneuver strategies. Instead, a computer simulation of the platen system is manipulated to provide a training set for an artificial neural network. The trained network provides optima switching times for conducting one dimensional rest-to-rest maneuvers of the platen that incorporate the full nonlinear effects of the magnetic levitation actuators. Sample problems illustrate the effectiveness of the neural network based control as compared to traditional proportional-derivative control.
Excessive deceleration forces experienced during high speed deployment of parachute systems can cause damage to the payload and the canopy fabric. Conventional reefing lines offer limited relief by temporarily restricting canopy inflation and limiting the peak deceleration load. However, the open-loop control provided by existing reefing devices restrict their use to a specific set of deployment conditions. In this paper, the sensing, processing, and actuation that are characteristic of adaptive structures form the basis of three concepts for active control of parachute inflation. These active control concepts are incorporated into a computer simulation of parachute inflation. Initial investigations indicate that these concepts promise enhanced performance as compared to conventional techniques for a nominal release. Furthermore, the ability of each controller to adapt to off-nominal release conditions is examined.
This paper focuses on the development of an approximate time-optimal feedback strategy for conducting rest-to-rest maneuvers of a magnetically levitated table. Classical switching curves are modified to account for the complexities of magnetic actuation as well as the coupling of the rigid body modes through the control. A smooth blend of time-optimal and proportional-derivative controls is realized near the destination point to correct for inaccuracies produced by the approximate time-optimal strategy. Detailed computer simulations of the system indicate that this hybrid control strategy provides a significant reduction in settling time as compared to proportional-derivative control alone.
This paper examines the role of the control objective and the control time in determining fuel-optimal actuator placement for structural vibration suppression. A general theory is developed that can be easily extended to include alternative performance metrics such as energy and time-optimal control. The performance metric defines a convex admissible control set which leads to a max-min optimization problem expressing optimal location as a function of initial conditions and control time. A solution procedure based on a nested Genetic Algorithm is presented and applied to an example problem. Results indicate that the optimal locations vary widely as a function of control time and initial conditions.