Simulated Failure Modes Research Articles

Experimental Study on the Seismic Performance of Insulated Single-Sided Composite Shear Walls under Different Shear Spans and Axial Compression Ratios

The new insulated single-sided composite shear wall (NISCSW) composition involves setting a precast wall panel on one side and an insulation panel on the other side, with a middle cavity for casting concrete. To investigate the seismic performance of NISCSW under different shear spans and axial compression ratios, eight specimens are made, including six composite and two cast-in-place walls. The shear span ratio is controlled at 1.2 and 1.9, and the axial compression ratio is controlled at 0.1, 0.3, and 0.4. The specimens are subjected to quasistatic tests to analyze failure modes, hysteresis characteristics, stiffness degradation, displacement ductility, and energy dissipation capacity and to compare the seismic performance of the composite and cast-in-place walls. Results show that for each composite specimen, under the same axial compression ratio, the large shear span ratio specimen has a lower ultimate bearing capacity and faster stiffness degradation but better ductility and postyield energy dissipation capacity. Under the same shear span ratio, the high axial compression ratio specimen had a higher ultimate bearing capacity, slightly worse ductility, and similar stiffness degradation and energy dissipation capacity compared to other specimens. Compared with the cast-in-place specimen with the same axial compression ratio, the composite specimen failure mode and hysteresis characteristics are similar, and the ductility and energy dissipation capacity are comparable to the cast-in-place shear wall specimen, indicating that NISCSW has similar seismic performance to the cast-in-place shear wall under conditions of a large shear span ratio and high axial compression ratio. Based on the test results, the program ABAQUS is used to simulate the specimens. Compared with the test results, the simulated specimen failure mode is consistent with the test results, and the hysteresis and skeleton curves are consistent with the test curve, indicating that the model is correct, reliable, and can be verified with test results.

Empirical Equations for Modelling Yarn–Mortar Debonding in TRM-Strengthened Masonry Walls Subjected to Out-of-Plane Loading

The effectiveness of textile-reinforced mortar (TRM) strengthening of masonry walls largely depends on the bond between the constituent materials. Finite element analysis (FEA) can provide valuable insights on the effect of the parameters affecting the bond; however, detailed FEA is computationally intensive. To alleviate this, we develop novel empirical equations to estimate effective textile fibre properties, thus implicitly accounting for yarn and mortar debonding. As a result, 3D finite element simulations of strengthened wall specimens are simplified and accelerated. The proposed scheme is calibrated using load–displacement paths derived from experimental data, and the simulated failure modes are compared against the experimental ones demonstrating perfect agreement. A parametric analysis is conducted, exploring the impact of the mechanical ratio of TRM reinforcement and the axial wall load on the effectiveness of TRM strengthening. We demonstrate that low values of mechanical reinforcement, corresponding to natural fibres, give rise to an 8-fold increase in the capacity of unreinforced walls. The findings draw conclusions about the efficacy of TRM strengthening in masonry structures, and provide valuable insights for optimising TRM reinforcement, considering different fibre materials and axial loads in masonry structures.

Macroscopic and microscopic mechanical characteristics and crack propagation behavior of sandstone-like samples with single cracks under freeze–thaw cycles: Experimental and numerical simulation

In permafrost regions, rock engineering projects inevitably suffer from the effects of freeze–thaw (F-T) cycles under complex geographical and climatic conditions. Under F-T cycles, the damage and deterioration of macro- and microdefects, such as joints and cracks, in the rock will reduce the structural stability of the rock and even lead to frost damage in rock engineering. In this study, through indoor F-T cycle tests, uniaxial/triaxial compression tests, acoustic emission monitoring, and discrete element method (PFC 3D) modeling, the macro/microscopic mechanical characteristics and crack propagation behavior of sandstone-like samples with single cracks under F-T cycles were investigated. The results show that the compressive strength of sandstone-like samples with single cracks tends to decrease and then increase with an increasing crack angle and exhibits a stable decreasing trend with increasing F-T cycles, indicating a significant crack angle effect and F-T deterioration effect. Additionally, the confining pressure has an inhibiting effect on the F-T deterioration of samples, and there is a mutual inhibition relationship between the F-T deterioration effect and the crack angle effect. The characteristics of the acoustic emission signals during the loading process of sandstone-like samples with single cracks show good consistency with the stress–strain curves. The analysis of fracture mechanisms based on rise time/amplitude (RA) and average frequency (AF) is consistent with the macroscopic failure modes observed in the experiments. Moreover, with an increase in the number of F-T cycles, the number of tension cracks in the sandstone-like samples with single cracks increases, while the number of shear cracks decreases. The crack propagation behavior of sandstone-like samples with single cracks obtained through discrete element simulations (PFC 3D) is mainly divided into four stages: Stage I - initial calm stage, Stage II - crack initiation stage, Stage III - crack propagation stage, and Stage IV - failure stage. The simulated mechanical properties and failure modes agree well with the experimental results, and five crack propagation modes of sandstone-like samples with single cracks under F-T cycles were proposed: direct tension (DT), relative tension (RT), shear-tension (ST), direct shear (DS) and relative shear (RS).

Seismic behaviors of CFT-column frame-four-corner bolted connected buckling-restrained steel plate shear walls using ALC/RAC panels

A Frame-Buckling Restrained Steel Plate Shear Walls (BRSPSWs) system has been designed, featuring a steel plate connected to frame elements, to withstand lateral loads like seismic or wind forces. This system incorporates recycled aggregate concrete (RAC) to create concrete-filled and concrete panels, promoting eco-friendly and sustainable construction materials. Two single-span, two-floor specimens were tested under cyclic quasi-static load. A four-corner bolted connection was used to connect the steel plate to the square concrete-filled steel tubes (CFTs) column and H-section beam, minimizing the potential deformation of the frame elements produced by the tension field of the steel plate after high-order bucking. The steel plate was sandwiched using either autoclaved lightweight concrete (ALC) or RAC panels. The study analyzed failure modes, load-displacement responses, and characteristic capacities. Test results inferred that BRSPSWs exhibited favorable cyclic behavior, with similar failure modes observed using both ALC and RAC panels. The buckling of steel plates was reduced, and the type of restrained panels had a negligible impact on buckling. Consequently, ALC panels can effectively replace RAC panels. Furthermore, bearing capacities and yield stiffness were primarily determined by the bolted connection forms. A finite element (FE) modeling was developed using ABAQUS software and validated against test results. This FE modeling method successfully simulated failure modes and load-displacement curves. Parametric analyses were conducted and revealed that bearing capacity and yield stiffness positively correlated with steel plate thickness, bolt diameter, and the number of bolts but negatively correlated with the axial compression ratio. In contrast, the panel thickness and span length had a minor impact due to shear deformation in the bolted connection. Based on the test results and parametric studies, an equation for the yield-bearing capacity of BRSPSWs was proposed and verified.

Lattice discrete particle modeling of the cycling behavior of strain-hardening cementitious composites with and without fiber reinforced polymer grid reinforcement

Strain-hardening cementitious composites (SHCC) has become increasingly prevalent in structural design. Compared to ordinary concrete, SHCC exhibits ultra-high tensile ductility, significant strain-hardening behavior, very fine multi-point cracking, high energy dissipation, and good durability. One robust model for simulating reinforced concrete behavior is the so-called Lattice Discrete Particle Model (LDPM), and specifically LDPM-F which includes the effect of fiber reinforcement. The present cycling constitutive law implemented in LDPM and the cycling fiber-bridging law in LDPM-F though cannot accurately capture the residual tensile plastic strain, loading–unloading path, and energy dissipation of SHCC during cyclic tension. To solve these concerns, the cycling tension–compression constitutive law and the nonlinear cycling fiber-bridging law were reformulated. Further, the new model was used to simulate the cycling tensile behavior of plain concrete, SHCC, and Fiber-Reinforced Polymer (FRP) grid reinforced SHCC (FRP-SHCC). Simulation results show that the multi-point cracking, crack widths, and ultra-high ductility properties are correctly captured by LDPM-F. In addition, LDPM-F with the modified cycling constitutive model can effectively simulate the cycling tension–compression behavior, accurately reproducing the stress–strain relationship, residual plastic strain, and overall dissipated energy. Finally, the simulated failure modes agree well with the actual fracture planes of these materials under cyclic tension.

Let us to the TWISST; Plan, Simulate, Study and Act.

Simulation-based Clinical Systems Testing approach includes simulated scenarios, Summarize, Anchor, Facilitate, Explore, Elicit debriefing, and Failure Mode and Effect Analysis. In iterative Plan-Simulate-Study-Act cycles, frontline teams explored work system inefficiencies, identified LSTs, and tested potential solutions. As a result, system improvements were hardwired through SbT. Finally, we present a case study example of the TWISST application in the Pediatric Emergency Department. TWISST identified 41 latent conditions. LSTs were related to resource/equipment/supplies (n = 18, 44%), patient safety (n = 14, 34%), and policies/procedures (n = 9, 22%). Work system improvements addressed 27 latent conditions. System changes that eliminated waste or modified the environment to support best practices mitigated 16 latent conditions. System improvements that addressed 44% of LSTs cost the department $11,000 per trauma bay. TWISST is an innovative and novel strategy that effectively diagnoses and remediates LSTs in a working system. This approach couples highly reliable work system improvements and training into 1 framework.

Artificial Neural Network-Based Predictive Model for Finite Element Analysis of Additive-Manufactured Components

Additive manufacturing is becoming one of the most utilized tools in an increasing number of fields from Industry 4.0 concepts, engineering, and manufacturing to aerospace and medical applications. One important issue with additive-manufactured components is their orthotropic behaviour where mechanical properties are concerned. This behaviour is due to the layer-by-layer manufacturing process and is particularly hard to predict since it depends on a number of factors, including the manufacturing parameters used during the manufacturing process (speed, temperature, etc.). This study aimed to create and train an artificial neural network-based predictive model using empirical tensile strength data obtained from additive manufactured test parts using the FDM method and PLA material. The predictive model was designed to predict mechanical characteristics for different orientation axis, which were used to set the material properties for finite element analysis. Results indicate a strong correlation between predicted finite element analysis behaviour and real-world tests on additive-manufactured components. The neural network model was trained to an accuracy of ~93% for predicting the mechanical characteristics of 3D-printed PLA material. Using the predicted mechanical characteristics for defining a custom orthotropic material profile in finite element analysis, the simulated failure mode and the behaviour of a complex geometry component agreed with the real-world test.

선박 추진용 2행정 저속엔진의 고장모드 데이터 개발 및 LSTM 알고리즘을 활용한 특성인자 신뢰성 검증연구

In the 4th industrial revolution, changes in the technological paradigm have had a direct impact on the maintenance system of ships. The 2-stroke low speed engine system integrates with the core equipment required for propulsive power. The Condition Based Management (CBM) is defined as a technology that predictive maintenance methods in existing calender-based or running time based maintenance systems by monitoring the condition of machinery and diagnosis/prognosis failures. In this study, we have established a framework for CBM technology development on our own, and are engaged in engineering-based failure analysis, data development and management, data feature analysis and pre-processing, and verified the reliability of failure mode DB using LSTM algorithms. We developed various simulated failure mode scenarios for 2-stroke low speed engine and researched to produce data on onshore basis test_beds. The analysis and pre-processing of normal and abnormal status data acquired through failure mode simulation experiment used various Exploratory Data Analysis (EDA) techniques to feature extract not only data on the performance and efficiency of 2-stroke low speed engine but also key feature data using multivariate statistical analysis. In addition, by developing an LSTM classification algorithm, we tried to verify the reliability of various failure mode data with time-series characteristics.

A new Distinct Element meso‐model for simulating the rocking‐dominated seismic response of RC columns

AbstractThe rocking‐dominated seismic response of reinforced concrete (RC) columns is a complex mechanism that entails P‐Δ effects, concrete crushing and spalling, rupture and dislocation of steel bars and stirrups, as well as combined localized failures causing splitting and shear‐flexure cracking. In this paper, we present a new meso‐scale numerical modeling strategy for the detailed simulation of the rocking response of slender RC columns under seismic actions. This work advances the current state‐of‐the‐art by exploring the unprecedented use of the Distinct Element Method (DEM), originally conceived for geotechnical studies and whose applicability in earthquake engineering has been primarily limited to the assessment of unreinforced masonry sub‐structures. To decrease analysis time and enable the use of DEM for RC problems, concrete is herein idealized as an assembly of solid macro‐block layers of size directly proportional to their distance from expected plastic hinges. These layers are connected by nonlinear interface springs to simulate tensile and shear cracking via a simplified Mohr‐Coulomb criterion. Crushing of concrete in deformable blocks is modeled using the Feenstra and De Borst strain‐softening linearized compression law, while confinement and combined failures are accounted for numerically through the explicit representation of longitudinal and transversal reinforcement bars, idealized as link elements. Previous experimental static and dynamic tests on full‐scale RC column specimens and results from fiber‐based Finite Element and sectional analysis models are used to evaluate DEM results and quantify the influence of key modeling parameters, including block number and deformability, as well as the amount of damping. The agreement between simulated failure modes and force and displacement capacities and their experimental counterparts demonstrate the applicability of the proposed approach, while demonstrating the pros and cons of simplified numerical methods.

Numerical Investigation of Brittleness Effect on Strength and Microcracking Behavior of Crystalline Rock

Brittleness has a significant influence on rock failure under compression; however, the mechanism is rarely comprehensively discussed. This study numerically investigates the brittleness effect on microcracking behavior of crystalline rock using a grain-based model implemented into a two-dimensional particle flow code, with a focus on the discussion of how rock brittleness affects the failure mechanism. The simulated failure mode changes from tension to shear with decreasing rock brittleness, which is consistent with previous laboratory test results. As the brittleness gradually decreases in the model, the grain boundary (GB) tensile crack to shear crack ratio increases, and the corresponding fractures change from vertical or subvertical to an angle about 45° along the vertical direction. The propagation and coalescence of generated microcracks result in a transition of failure pattern from splitting to shear under uniaxial compression with a decreasing brittleness level in the rock. A transition from GB tensile crack to shear crack is also observed under direct tension when the brittleness index gradually decreases. The tension to shear transition mechanism is closely related to the relative strength of the mineral grain and mineral bonding. The relative strength of mineral and mineral bonding could be used as a parameter to characterize rock brittleness from a microscale viewpoint.

Finite element analysis of the shear capacity of stainless-steel screw connections

Stainless-steel has been widely applied in engineering for its good corrosion resistance and low life-cycle cost. The use of screws in the thin sheet has the advantages of simple construction and tight connection, so it is widely used in open-air maintenance systems. However, the mechanical performance of stainless-steel screw connections is quite different from that of carbon steel due to the material difference, which needs further research. In this paper, finite element method was carried out to simulate the shear performance of the stainless-steel screw connections. The simulated failure modes were consistent with the experimental results, the simulated load–displacement curves were close to the experimental results, and the difference was within 10%. A variety of parametric studies utilizing the finite element method (FEM) had been performed on the shear capacity of stainless-steel screw connections, which included sheet thickness, spacing and connecting length. Results shown that the connecting length was the main factor affecting the shear capacity of the stainless-steel screw connections. For equal-thickness connection (t1 = t2 = 1.5 mm, 2.0 mm, 3.0 mm), when the connecting length was within 6d-60d, the shear capacity increased linearly with the increase of the connecting length. For unequal-thickness connection (t1 = 1.5 mm, t2 = 2.0 mm/t1 = 2.0 mm, t2 = 3.0 mm/t1 = 2.0 mm, t2 = 5.0 mm), when the connecting length was within 6d-45d, the shear capacity increased linearly with the increase of the connecting length. However, when the connecting length was within 45d-60d, the end screw occurred shear failure first, its shear capacity almost no longer increased. The change of screw spacing and the sheet thickness had few effect on the shear capacity. Based on reliability research, the resistance factor and design strength were proposed. For shear failure,γR=1.273, fvb=275MPa, and for bearing failure, γR=1.306, fvb=315MPa.

Flexural property evaluation of web reinforced GFRP-PET foam sandwich panel: Experimental study and numerical simulation

This work focus on the flexural property of polyethylene terephthalate (PET) foam-filled lattice composite sandwich panels subjected to four-point bending (FPB). The effects of different face sheet and core thicknesses on the flexural properties of the sandwich panels were analyzed. Experimental results indicated that the glass fiber-reinforced polymer (GFRP) webs can effectively prevent the sandwiches from catastrophic failure , the failure modes of the structures are dominated by foam shear, top face sheet compressive, and top face sheet-core debonding. Increasing the face sheet and PET foam thickness can effectively improve the ultimate load of the panels by 88.9% and 115.6%, respectively. An analytical model was employed to accurately predict the bending stiffness , ultimate load, and failure modes of the structures. The errors between theoretical predicted and experimental results are within 10% in terms of bending stiffness and ultimate load. Furthermore, a 3D numerical model was built to understand the failure modes and responses of the structures, a continuum damage material model for PET foam was imported into ABAQUS software by VUMAT subroutine, and Hashin failure criterion was used for GFRP. The simulated failure modes and load-displacement curves were in good agreement with experimental results.

Axial compressive behavior of ultra-high performance concrete confined by high-strength transverse reinforcements

The axial compressive performance of the confined system composed of the high-strength transverse reinforcements and ultra-high performance concrete (UHPC) was studied in this paper. The axial compression tests were carried out on 24 UHPC columns to investigate the effect of strength, spacing and configuration of the transverse reinforcements and steel fibers content on the ductility, toughness and maximum load of UHPC columns. The failure modes, axial load-axial strain behavior and strain of steel bars were also analyzed and discussed. The test results showed that high-strength transverse reinforcement could significantly increase the post-peak ductility of UHPC columns, compared with the ordinary-strength transverse reinforcements. In addition, the ductility, maximum load and toughness of the UHPC columns could be improved by reducing the transverse reinforcements spacing and using the L-type multiple hoops. Then, simulated axial load-axial strain curves and failure modes of UHPC columns were obtained by finite element analysis. Three performance indexes, i.e., the maximum axial load, strain corresponding to the maximum load and strain corresponding to 85% drop of the maximum load of UHPC columns were evaluated in the numerical model. The error between test value and simulated value of the three performance indexes was mostly within 15%. The results demonstrated that the three-dimensional numerical model of UHPC columns could effectively simulate the axial compressive performance and failure modes of the specimens.

Quantification of the resistance modeling uncertainty of 19 alternative 2D nonlinear finite element approaches benchmarked against 101 experiments on reinforced concrete beams

AbstractNineteen 2D nonlinear finite element analysis (NLFEA) solution strategies were benchmarked against a wide variety of 101 experiments on reinforced concrete beams failing in bending, flexural shear, or shear compression. The relatively high number of solution strategies was motivated by the conviction that choices for the constitutive models, the finite element kinematics and equilibrium settings will interact, and must therefore be tested in conjunction. Modeling uncertainty distribution parameters are presented for the 19 solution strategies, using all beams, and using beams with and without stirrups separately. The resulting statistics are discussed against the correctness of the simulated failure modes and failure loads, revealing that rotating crack models perform well for the relatively ductile failures in beams with stirrups, while fixed crack models perform better for the more brittle failures in beams without stirrups. The tailored solution strategies that predict failure modes correctly, imply a log‐normal distribution of the modeling uncertainty with relatively low coefficients of variation. The outlook is that these estimates of the statistical properties of the modeling uncertainties could serve as a basis within safety formats.

Dynamic numerical simulation of pre-cracked concrete samples under different mechanical parameters

Numerical simulations of pre-cracked concrete subjected to dynamic impact compressive stress wave are carried out via DRFPA (Dynamic Realistic Failure Process Analysis). This code is validated firstly by comparing experimental results. Then, the initiation, propagation and coalescence of wing cracks and secondary cracks are numerically investigated by varying elastic modulus, uniaxial compressive strength, and homogeneity index of some phases (bond and aggregate) of concrete. The numerically simulated concrete failure mode (tension/shear failure) and its development in both quasi-coplanar plane and concrete bridge are analyzed and discussed. Numerical results show that the tension failure occurs firstly whereas shear failure occurs shortly thereafter with continuous dynamic loadings and its reflected wave. Meanwhile, the evolution of wing cracks and secondary cracks are numerically investigated. In addition, the effect of elastic modulus, uniaxial compressive strength, and homogeneity index of concrete phases on the crack evolutions are also studied.

Compression responses of composite corrugated sandwich square tube: Experimental and numerical investigation

By means of re-bonding inner square tube, corrugated core and outer square tube produced by hot pressing molding method, the composite corrugated sandwich square tubes (CCSSTs) were manufactured. Axial and lateral compression experiments were carried out on CCSSTs with length of 50 mm and 25 mm respectively, and the mechanical responses such as load, displacement and mid-height strain during the loading process were measured. The composite progressive damage finite element models were used to simulate the compression failure processes of CCSSTs. The simulated failure modes and load–displacement curves were in good agreement with experimental results. An analytical model based on homogenization theory was provided to evaluate the regularity of axial and lateral compression stiffnesses of CCSSTs. The research showed that the CCSST mainly underwent face crushing, face debonding and local buckling under axial compression. The lateral compression bearing capacity of CCSST reduced mainly caused by the outer face debonding. Increasing the height of the core could improve the axial compression stiffness and specific axial compression stiffness. The lateral compression stiffness performance and light weight benefit were both satisfying when the corrugation angle reaches 90°.

The axial compressive behavior of stone-lightweight aggregate concrete-filled steel tubular stub columns

To study the axial compressive mechanical properties of stone-lightweight aggregate concrete-filled steel tubular (SLACFST) stub columns, five circular stone-lightweight aggregate concrete-filled steel tubular (C-SLACFST) and square stone-lightweight aggregate concrete-filled steel tubular (S-SLACFST) stub columns were made by taking the volume replacement ratio of natural gravel aggregate and ceramsite in lightweight aggregate concrete as a variation parameter. After observing their mechanical process and failure modes, the axial load–displacement curves of the specimens were obtained. The influence of gravel replacement ratio on the energy dissipation and stiffness of the specimens was analyzed, and a simplified model for predicting their stiffness degradation was proposed. The plastic damage model was then used to verify those specimens with a finite element, and this model was subsequently adopted in the whole process analysis of these specimens and in finding out the loading relationship between the steel tube and core stone-lightweight aggregate concrete (SLAC). The effects of varying parameters on the compressive behavior of the specimens were investigated, and the extent to which these varying parameters are related to the ultimate bearing capacity of SLACFST was assessed via a grey correlation analysis. Results show that the rigidity of the calculated load–displacement curves in the elastic and elastic–plastic stages is greater than that of the test curves, but the ultimate bearing capacity and peak displacements agree with the test values within an acceptable error. Moreover, the simulated failure modes of the control specimens are the same as the test failure modes. Before reaching the ultimate bearing capacity, the core SLAC made a significant contribution to the bearing capacity of the specimen due to the constraint of steel tube. The negligible influence of gravel replacement ratio on the ultimate bearing capacity of SLACFST specimen. In addition, the nature gravel can inhibit the floating of LAC to a certain extent, so as to achieve the purpose of using SLAC to make SLACFST, reducing the self-weight of the structure while having higher bearing capacity.

Damage Behavior Prediction of Repaired Composite Laminates Subjected to Oblique Penetrate Loading

To predict the effect of oblique impact on the damage and performance of perforation resistance of repaired composite laminates, a three-dimensional continuum damage model (CDM) is used to simulate and calculate the low-velocity impact process of repaired laminates. In the established finite element model, several different types of failure modes are considered. The damage model is based on the modified 3D-Hashin failure criterion to predict the initial damage of the fiber and matrix. The linear-exponential law and the exponential law are used to simulate the tensile and compressive softening process of the material. The triangle traction-separation law and mixed-mode fracture energy method are applied to simulate interface debonding damage between patch/lamina, patch/patch and lamina/lamina. A comparison of the numerical results of several different offsets with the experimental results shows that the simulated failure modes and impact responses are very close to the experiment, which verifies the validation of the finite element model (FEM). The increase of the impact angle can enhance the residual velocity of the projectile in the laminate and reduces the energy absorption. At the same time, it has a great influence on the failure modes, especially the interface debonding damage of the patch.

The application of machine learning to sensor signals for machine tool and process health assessment

Due to the latest advancements in monitoring technologies, interest in the possibility of early-detection of quality issues in components has grown considerably in the manufacturing industry. However, implementation of such techniques has been limited outside of the research environment due to the more demanding scenarios posed by production environments. This paper proposes a method of assessing the health of a machining process and the machine tool itself by applying a range of machine learning (ML) techniques to sensor data. The aim of this work is not to provide complete diagnosis of a condition, but to provide a rapid indication that the machine tool or process has changed beyond acceptable limits; making for a more realistic solution for production environments. Prior research by the authors found good visibility of simulated failure modes in a number of machining operations and machine tool fingerprint routines, through the defined sensor suite. The current research set out to utilise this system, and streamline the test procedure to obtain a large dataset to test ML techniques upon. Various supervised and unsupervised ML techniques were implemented using a range of features extracted from the raw sensor signals, principal component analysis and continuous wavelet transform. The latter were classified using convolutional neural networks (CNN); both custom-made networks, and pre-trained networks through transfer learning. The detection and classification accuracies of the simulated failure modes across all classical ML and CNN techniques tested were promising, with all approaching 100% under certain conditions.

Seismic cracking evolution for anti-seepage face slabs in concrete faced rockfill dams based on cohesive zone model in explicit SBFEM-FEM frame

It is critically importance to accurately locate vulnerable areas and quantitatively assess the damage of the face slab in the seismic safety evaluation of concrete faced rockfill dams (CFRDs). In this paper, the cohesive zone model (CZM) is augmented with generalized plastic models of the rockfill and state-dependent elasto-plastic model of the interface to investigate the seismic cracking evolution in the slab of CFRDs. An explicit coupled scaled boundary finite element method-finite element method (Explicit SBFEM-FEM) is developed to enable cross-scale analysis. The method considers the strain softening of rockfill and concrete after being damaged, and avoids negative stiffness and convergence problems that can be found during implicit analysis. A fine, cross-scale model of the CFRD is established based on the quadrature technique. Subsequently, simulations of the construction and impoundment processes are performed to facilitate dynamic analysis. In the first step, different damping ratios for the slab are simulated to optimize accuracy and efficiency. Then, the seismic cracking evolution of the slab is investigated in detail with consideration to the ground motion intensity and steel reinforcement. The results indicate that the computational efficiency can be significantly improved by decreasing the damping ratio of the concrete face slab in the explicit seismic analysis. Penetrating cracks are observed on the slab and the crack reaches maximum width during the earthquake. A residual width is retained after the earthquake. The simulated failure mode of face slab conforms to the characteristics of concrete. The developed method can help precisely locate weak areas of face slab, quantitatively determine the damage severity, and evaluate the ultimate seismic performance of the slab in the CFRDs. In addition, the method can be further employed for concrete cracking analysis involving soil-concrete interaction.