Chapter Four · failure evidence
What Friction & Wear Tribology got wrong, from 49 dissertations
The collected records detail experimental and computational investigations in friction and wear across materials, mechanical components, and contact mechanics models. Researchers frequently encountered problems where idealized friction laws failed under realistic contact conditions, unmodeled friction compromised state estimation and control, and novel surface treatments or lubricant formulations underperformed baselines. These records come from PhD theses at 18 institutions, 2021 to 2026. Each links to its thesis. They were extracted by language models reading the full text, so treat each as a lead to read, not a verdict.
Classical and analytical friction models fail under real contact conditions
Idealized friction formulations, empirical regressions, and classical contact mechanics laws frequently fail to capture phenomena such as non-constant gross slip, non-linear normal load dependency, or pitch-point elastohydrodynamics. In several cases, simpler baseline models or direct efficiency factors outperformed complex wedging, eversion, or bristle-based contact formulations that produced inaccurate or non-physical predictions.
Tried and failed
analytical contact mechanics with empirical friction fitting applied to tapered roller bearing wedging calculations. Reason: friction coefficient fitted to non-physical negative values to match empirical manufacturer catalog ratings
Planetary Traction Drive for Submarine Application withTapered Rollers · MIT
Tried and failed
spline-based friction modeling revision applied to multiplate clutch torque prediction. Reason: failed to sufficiently improve prediction fidelity of locking torque to an acceptable level
Optimal time and handling methods for motorsport differentials. · Cranfield
Tried and failed
Benedict and Kelley friction coefficient model applied to gear contact friction prediction. Reason: Fails to capture coefficient of friction drop near the pitch point under high-speed elastohydrodynamic conditions
Lost to a baseline
The detailed wedging friction model was beaten in fitting SKF bearing catalog calculation factor data by a simpler model omitting wedge friction and applying a direct efficiency factor (eta = 0.8, mu = 0.25).
Planetary Traction Drive for Submarine Application withTapered Rollers · MIT
Considered and rejected
Considered and rejected: Decided against constant friction models during gross slip in favor of wear-scar interaction models due to the observed non-constant friction force during gross slip.
Considered and rejected
Considered and rejected: Rejected theoretical friction formula (sqrt(Fx^2 + Fy^2)/Fz) due to real tire deviations from ideal brush model behavior (variable pressure distribution, compliance, camber, residual forces).
Hybrid Friction Estimation based on Intelligent Tires and Vehicle Dynamics · Virginia Tech
Considered and rejected
Considered and rejected: Rejected friction models based on bristle deflection due to complexity, discontinuous state variables, lack of velocity dependence, and computational inefficiency.
Non-smooth dynamics and sensitivity analysis of multibody systems with clearances and friction in differential variational inequality framework · Virginia Tech
Considered and rejected
Considered and rejected: Rejected standard statistical/empirical regression modeling for predicting friction from texture parameters due to small sample size and lack of physical interpretability.
CHARACTERIZATION AND MODELING DETERIORATION OF HIGH FRICTION SURFACE TREATMENT · Georgia Tech
Tried and failed
Uniform crack slip state analytical assumptions applied to frictional microcrack hysteresis modeling. Outcome: did not generalise. Reason: Inhomogeneous crack states and non-sticking zones during unloading invalidated uniform forward/reverse slip assumptions
Micromechanical model for hysteresis in fluid-saturated porous rocks under hydrostatic loading · Imperial
Tried and failed
Persson contact mechanics friction model applied to rubber sliding friction under varying pressure. Outcome: did not generalise. Reason: assumes complete contact rather than partial contact, making friction invariant to nominal pressure and missing load-induced temperature changes
Multi-Length Scale Modeling of Rubber Tribology For Tire Application · Virginia Tech
Lost to a baseline
Estimation of static friction coefficient using an averaged plain eversion model yielded 0.20, underestimating the force-sensor-measured baseline of 0.29 by 31%.
Modeling of Everted Tubes for Medical Applications · ResearchWorks
Considered and rejected
Considered and rejected: Rejected relying solely on slider friction test measurements to represent tire-ice interaction, because slider tools ignore tire-specific normal load, tread inflation pressure, and rubber compound dynamics.
Experimental Characterization and Modeling of Tire-Ice Interface · Virginia Tech
Considered and rejected
Considered and rejected: Rejected the analytical friction force derivative integration approach across the hysteresis cycle in favour of using the FFT algorithm on time-domain derivative signals, because the hysteresis cycle cannot be known a priori in real turbine applications.
Structural Dynamics Design of Steam Turbine Blades with Friction Contacts for Renewable Energy · IRIS - POLITO - prod
Tried and failed
Amontons-Coulomb friction model applied to untreated elastomeric polymer contact. Outcome: did not generalise. Reason: Tangential frictional force was non-linear with respect to normal force for raw elastomeric surfaces
Mechanics of Physical Knots: From Filaments in Contact to Surgical Suturing · EPFL
Parasitic mechanical friction and bearing drag degrade system efficiency
Excessive friction in rolling element bearings, radial supports, and sliding interfaces caused substantial power dissipation that prevented prototypes from matching baseline performance. In other setups, parasitic scrubbing increased cutting resistance or generated thermal transients that disrupted torque measurements.
Tried and failed
reducing tool rake angle applied to soil cutting and tillage assemblies. Outcome: worse than baseline. Reason: loss of cutting edge clearance caused parasitic frictional scrubbing and increased horizontal resistance
Optimising the performance of shallow, high-speed mouldboard ploughs · Cranfield
Lost to a baseline
Prototype turbocharger overall efficiency lost to baseline conventional hydrodynamic bearing radial turbochargers due to high axial bearing friction (ηm ~90% vs 95-98%).
Einstufige Abgasturboaufladung von Pkw-Motoren mit Luftüberschuss für einen realtitätsnahen Fahrzyklus Single-Stage Turbocharging of Lean-Burn Passenger Car Engines for Real-World Driving Cycles · Leibniz Universität Hannover Repository
Tried and failed
raceway control hypothesis for ball kinematics applied to rolling element bearing contact mechanics. Outcome: worse than baseline. Reason: violates friction torque equilibrium compared to virtual power minimization
Advances in contact mechanics and fluid dynamics for rolling contacts · Leibniz Universität Hannover Repository
Tried and failed
single ball bearing support for turbine rotor applied to fluid-driven rotary atomiser. Reason: Excessive bearing friction reduced rotational speed below atomisation threshold and suffered fluid contamination.
Fluid driven rotary atomiser for controlled droplet application of herbicides. · Cranfield
Tried and failed
using sealed pre-lubricated bearings to isolate drag applied to mechanical transmission rundown torque measurements. Outcome: unstable. Reason: internal frictional heating caused thermal transients that distorted excess torque measurements and broke steady-state agreement
Lost to a baseline
Unbalanced sliding vane ERD prototype achieved only 17% power savings vs. 62% theoretically achievable by an ideal ERD due to 121 W frictional losses (48% in radial bearings).
Lost to a baseline
Gear mechanism fixed-recovery ERD did not decrease SEC below a standalone ideal pump without energy recovery due to tip slip and bearing friction.
Considered and rejected
Considered and rejected: Rejected off-the-shelf roller bearings as transmissions because transmission ratios were <= 3, were prelubricated with grease, and roller cages held elements on their OD which causes excessive friction losses.
Planetary Traction Drive for Submarine Application withTapered Rollers · MIT
Omitting friction dynamics or assuming frictionless contact degrades model accuracy
Kinematic estimators, state observers, and simulation models experienced large prediction errors or filter tracking failures when friction and compliance were omitted. Assuming frictionless contact or neglecting internal motor and setup friction failed to represent real physical dynamics and distorted hardware validation.
Tried and failed
kinematics-based series elastic actuator force modeling applied to robotic exoskeleton force estimation. Reason: unmodeled friction and compliance/deformations in mechanical linkages caused high force prediction errors
Vision-Based Force Planning and Voice-Based Human-Machine Interface of an Assistive Robotic Exoskeleton Glove for Brachial Plexus Injuries · Virginia Tech
Lost to a baseline
Experimental cutoff frequency for all printed beam designs on hardware was approximately half of the predicted simulation value due to unmodeled friction and sensor noise
Application of Co-Design Principles for Design of Series Elastic Joints · Virginia Tech
Tried and failed
training regression models on frictionless simulation data applied to mechanical stiffness estimation. Outcome: did not generalise. Reason: frictionless simulation failed to capture real-world friction dynamics leading to catastrophic evaluation errors
ULTRASONIC IMAGING AND TACTILE SENSING FOR ROBOTIC SYSTEMS · Georgia Tech
Lost to a baseline
Ideal PTO model matched input force better than the developed circuit model at higher driving frequencies (e.g. 3-4 Hz), where the circuit model overestimated force by ~17% due to unmodeled backlash and Stiction/Stribeck friction.
Model, Design, and Control for Power Conversion in Wave Energy Converter System · Virginia Tech
Lost to a baseline
Removing modeled friction improved real-to-virtual similarity ratings for low-friction surfaces (Whiteboard, MDF, ABS Plastic, Silk 1, Tarp) due to uncompensated internal device motor friction.
Data-Driven Haptic Modeling and Rendering of Realistic Virtual Textured Surfaces · Penn
Considered and rejected
Considered and rejected: Measuring torque indirectly via DC motor current and torque constant KT, rejected due to winding imperfections, model variations, and setup friction.
DESIGN AND IMPLEMENTATION OF AN MRI-COMPATIBLE MOTOR TEST STAND · JScholarship
Considered and rejected
Considered and rejected: Rejected frictionless contact assumptions (zero shear traction) as neglecting shear interface transfer unrealistically inflates predicted lining utilization.
Daten-gestützte analytische Mechanik für die Neue Österreichische Tunnelbaumethode Data-driven analytical mechanics for the New Austrian Tunneling Method · DSpace-CRIS at TU Wien
Tried and failed
extended Kalman filter wheel odometry fusion applied to mobile robot track navigation. Outcome: unstable. Reason: unmodeled variable frictional slip between drive wheels and contact surfaces degraded state estimation
A novel railway maintenance robot for inspection and repair · Cranfield
Engineered coatings and solid materials suffer from severe wear and elevated friction
Solid lubricant multilayers and plasma jet coatings experienced brittle failure, spallation, and rapid degradation into stick-slip with high friction under mechanical load. Alternative coating and composite formulations also underperformed simpler baseline materials by exhibiting accelerated wear rates or higher friction coefficients.
Lost to a baseline
UHMWPE had higher friction coefficient against bovine cartilage (0.473 at 5 cycles and 0.776 at 600 cycles) than the hybrid/gyroid replacement device (0.133 at 5 cycles and 0.109 at 600 cycles).
Hybrid materials for meniscus replacement in the knee · Imperial
Tried and failed
high layer count solid lubricant multilayers applied to cutting tool protective coatings. Outcome: worse than baseline. Reason: brittle failure and coating spallation under load leading to high friction
Development of low friction coatings for alsi milling · Cranfield
Lost to a baseline
Al2O3-4wt%ZrO2 had a higher wear rate (6.0426 x 10^-4 mm3/Nm) and higher friction coefficient (0.7478) than monolithic pure Al2O3 (wear rate 4.4408 x 10^-5 mm3/Nm, COF 0.5874).
Lost to a baseline
Atmospheric plasma jet coating on upside-down mixed EUSD failed under tribological stress after ~70 cycles, degrading into stick-slip with high friction compared to commercial sliding varnish.
Einfluss der Wechselwirkungen auf die Substratanbindung von Atmosphärendruck-Plasmaschichten · Leibniz Universität Hannover Repository
Lost to a baseline
Non-commercial formulations F-RH and F-RHT exhibited rather unstable friction coefficient curves compared to optimized commercial materials.
Particulate Matter Emission Issues in Brake Systems · IRIS - UNITN - prod
Tried and failed
anti-icing surface coatings applied to aerospace substrate de-icing. Reason: insufficient mechanical durability causing complete wear-through under dry sliding friction cycles
A Hybrid De-icing Strategy Utilizing Tailored Anti-Icing Coatings Paired with Electro-Thermal and Electro-Mechanical Ice Protection Systems for Use on Rotary-Wing Aircraft · Carleton University Institutional Repository
Lost to a baseline
12 W Ta-DLC exhibited a higher coefficient of friction (0.26) than uncoated CoCrMo alloy (0.22) and undoped DLC (0.18).
Synthesis and characterization of Ta-DLC coatings over CoCrMo alloy · HARVEST
Novel lubricant additives and surface texturing increase friction or accelerate wear
Modifying contact surfaces with porosity led to wear track collapse, higher wear volume loss, and increased friction coefficients under liquid lubrication. Similarly, applying surfactant formulations or microencapsulated additives caused severe wear degradation or sudden friction spikes once protective capsules depleted under shear.
Tried and failed
lubrication to reduce friction applied to metal cutting at intermediate speeds. Outcome: worse than baseline. Reason: increased dissipated energy by 7-8% and increased peak tool temperature by 13%
Coupled phenomena in thin layers · UT Austin
Tried and failed
introducing controlled surface porosity to reduce friction applied to metal alloy bearing sliding surfaces. Outcome: worse than baseline. Reason: porous surfaces suffered wear track collapse and increased friction coefficients under lubricated reciprocating contact
Tried and failed
micellar surfactant lubricant formulations applied to boundary lubrication and wear reduction. Outcome: worse than baseline. Reason: friction decreased but wear performance was orders of magnitude worse than standard oil lubrication
Biomimetic water-based lubrication: an exploration into vesicle lubrication · Imperial
Tried and failed
microencapsulated functional liquid additives applied to lubricating oil friction reduction. Outcome: unstable. Reason: friction spiked rapidly after small-diameter capsule fraction depleted under shear cycles
Design of ionic liquid based additives for integration with modern lubrication systems · UT Austin
Tried and failed
introducing surface porosity to reduce friction applied to metal surfaces under liquid lubrication. Outcome: worse than baseline. Reason: higher porosity increased the friction coefficient and wear volume loss rather than improving lubrication
Lost to a baseline
Empty DSPC liposomes had higher friction (0.093) and wear area compared to multi-lipid natural lecithin liposomes.
Biomimetic water-based lubrication: an exploration into vesicle lubrication · Imperial
Friction formulations and parameter estimators suffer from dynamic and numerical instability
Introducing separating contact interactions or slip-weakening formulations created numerical instabilities, discontinuities, and joint luxation in finite element simulations. Parameter estimation routines and process noise filters also experienced divergence and noisy outputs during sudden traction loss or interface temperature shifts.
Considered and rejected
Considered and rejected: Rejected using a wet sliding friction coefficient of 0.2 with separating surface interactions in FE models because it caused numerical instability and complete joint luxation.
Overcorrected Implants for Total Shoulder Arthroplasty · EPFL
Tried and failed
multi-friction-surface mechanical clutch modeling applied to limited slip differential characterisation. Outcome: did not generalise. Reason: uncontrolled contact interface temperature drops altered friction behavior during testing
Optimal time and handling methods for motorsport differentials. · Cranfield
Considered and rejected
Considered and rejected: Rejected making the heteroscedastic process noise model dependent on the latent friction parameter l because incorrect contact state estimates caused filter divergence.
Combining Learning and Structure for Robotic Manipulation · Publikationssystem UB Tuebingen
Tried and failed
lumped-parameter viscous radiation damping approximations applied to dynamic friction slider systems. Reason: fixed damping coefficients either eliminate or excessively overpredict dynamic stress overshoot
Mathematical and Computational Modeling of Injection-induced Seismicity · MIT
Tried and failed
standard least squares parameter estimation applied to friction and adhesion parameter estimation. Outcome: unstable. Reason: produced noisy, inconsistent estimates during sudden traction loss without dynamic filtering
Optimization approaches for high-performance and efficient utilization of redundancy in robotic systems · UT Austin
Considered and rejected
Considered and rejected: Slip-weakening friction rejected due to numerical discontinuities and piece-wise handling of healing across multi-block cycle simulations.
Left open by the authors
Problems the authors named and did not get to.
Left open
Conduct a Latin Hypercube Monte-Carlo sensitivity study in Simscape incorporating translational joint compliance and friction for the folding aeroshell model. Blocker: Requires access to or recreation of the custom HATHOR Simscape multibody model developed in the thesis.
Left open
Calibrate the interactive friction model for other material-tool combinations using direct friction test data under machining loads. Blocker: Requires experimental direct friction testing apparatus and physical machining load data.
Machining informatics: A data guided research on interfacial tribology of machining · Imperial
Left open
Develop an algorithm modeling the friction coefficient's dependency on excitation frequency in wound electrical steel rotor press-fits. Blocker: Requires experimental tribological test rig data characterizing frequency-dependent friction coefficients between electrical steel laminations under vibrational/rotational excitation
Methodik zur Auslegung von gewickelten Rotoren für permanenterregte Axialflussmaschinen Methodics for design of wound rotors in permanently excited axial flux machines · DSpace-CRIS at TU Wien
Left open
Implement the calibrated interactive friction model into FEA software user-subroutines to enable dynamic tool life and wear prediction. Blocker: None
Machining informatics: A data guided research on interfacial tribology of machining · Imperial
Left open
Incorporate temperature dependence and inflection points into analytical viscosity-pressure models to predict viscometry directly from uncorrected friction data. Blocker: Requires uncorrected friction data from specific elastohydrodynamic contact experiments
Effect of polymers on lubricant rheology in an elastohydrodynamic contact · Imperial
Left open
Estimate tire force and moment characteristics using the Brush model coupled with the developed condition-dependent friction model. Blocker: None
Multi-Length Scale Modeling of Rubber Tribology For Tire Application · Virginia Tech
Left open
Optimize deep surface rolling finite element models through detailed studies of slit effects and friction interactions. Blocker: Missing baseline finite element model files and experimental validation data.
Left open
Develop a performance model to predict the rate of pavement friction deterioration over time. Blocker: Requires longitudinal pavement friction and texture monitoring datasets over time which are not publicly available
A network-level statistical model for friction estimation in Texas · UT Austin
Left open
Reduce the number of texture clusters and simplify the regression equations used for estimating pavement Grip Number. Blocker: Requires original Texas pavement 2D laser profile and friction field dataset which is not publicly available.
A network-level statistical model for friction estimation in Texas · UT Austin
Left open
Characterize and validate the efficiency-factor bearing model against the friction wedging model for tapered roller planetary traction drives. Blocker: Requires a physical planetary traction drive test rig or empirical experimental data for validation.
Planetary Traction Drive for Submarine Application withTapered Rollers · MIT
Checking a claim in this area?
We can run the same search on any method or claim. If nothing turns up, we will say so, and that proves nothing on its own.