Continuous compaction control (CCC) offers a promising alternative to conventional spot tests for earthwork quality assurance. However, establishing reliable correlations between intelligent compaction measurement values (ICMVs) and traditional stiffness metrics, such as the deformation modulus (E_v2) from static plate load tests (PLT), remains a challenge. This study investigated the influence of PLT plate size and vibratory roller type on these correlations through a series of full-scale compaction tests conducted on a granular soil bed. Two plate diameters (300 and 600 mm) and two rollers with different static linear loads (36 and 65 kg/cm) were employed. Sixteen ICMVs spanning acceleration-, mechanics-, and energy-based categories were computed and correlated with E_v2. The results showed that the 600-mm plate, despite producing lower E_v2 values, consistently yielded stronger correlations with the representative ICMVs. The lighter roller produced stronger correlations than the heavier roller, likely due to reduced contact loss and better alignment of measurement depths. Mechanics-based ICMVs performed best, with the vibration modulus (E_vib) during the loading phase showing the strongest correlation (R up to 0.8). These findings confirm that E_vib is a suitable indicator for compaction quality control, suggest that the larger plate is preferable for calibration purposes, and underscore the need for machine-specific calibration.

Jordpackning är en av de mest grundläggande och kritiska jordförstärkningsmetoderna, som utförs i nästan alla typer av byggprojekt, med särskild betydelse inom infrastruktursektorn. En väl utförd packning är avgörande för en konstruktions funktion och livslängd, och med tanke på dess betydande miljöpåverkan är det av yttersta vikt att optimera packningsarbetet. För känsliga konstruktioner som järnvägar, där minimala sättningar är ett krav, är optimal packning av yttersta vikt. Detta gäller i synnerhet vid byggnation av moderna järnvägssystem med ”slab-track”, där undergrundens stabilitet är avgörande för att minimera sättningar. Idag finns ett stort behov av att effektivt kunna verifiera packningsresultat. Nuvarande metoder, såsom plattbelastningsförsök (PLT), är tidsödande och ger endast diskreta mätpunkter, vilket begränsar den yttäckande kontrollen. Därför är det av stor vikt att etablera tillförlitliga korrelationer mellan intelligenta packningsmått (ICMV) – som möjliggör yttäckande packningskontroll (YPK) – och konventionella punktprov.

Denna studie adresserar detta behov genom att presentera fullskaliga, kontrollerade laboratorieförsök. Syftet var att systematiskt utvärdera hur plattstorlek (300 och 600 mm) och vält-typ (Dynapac CA3500D och CA6500D med statiska linjelaster på 36 och 65 kg/cm) påverkar deformationsmoduler från PLT samt 16 olika ICMV, vilka spänner över kinematiska, mekanikbaserade och energibaserade kategorier.

Resultaten visar att 600 mm-plattan, trots något lägre deformationsmodul E_v2, gav konsekvent förbättrade korrelationer med ICMV, sannolikt tack vare dess djupare influensvolym. Den tyngre välten producerade högre E_v2 och lägre variabilitet än den lättare. Analys av ICMV-samband avslöjade stark redundans och tre tydliga kluster (harmonisk-, styvhets- och energibaserade), vilket stöder användningen av en enskild representant från varje kluster vid kalibrering. Vibrationsmodulen E_vib och harmonisk-baserade index presterade bäst i korrelation med PLT, medan energibaserade index var mindre lämpliga.

Sammanfattningsvis förbättrar större PLT-plattor och mekanikbaserade ICMV korrelationen mellan YPK och PLT, men vältspecifik kalibrering är fortsatt nödvändig. Dessa resultat ger praktisk vägledning för mer tillförlitliga YPK-baserade arbetsflöden för kvalitetssäkring.

Soil compaction is one of the most fundamental and critical ground improvement methods, performed in almost all types of construction projects, with particular significance within the infrastructure sector. Well-executed compaction is crucial for the performance and service life of a structure, and given its significant environmental impact, optimizing compaction work is of the utmost importance. For sensitive structures such as railways, where minimal settlement is a requirement, optimal compaction is vital. This applies in particular to the construction of modern railway systems using “slab-track”, where subgrade stability is decisive for minimizing settlement. Today, there is a great need to effectively verify compaction results. Current methods, such as the plate load test (PLT), are time-consuming and provide only discrete measurement points, limiting area-wide control. Therefore, it is highly important to establish reliable correlations between intelligent compaction measurement values (ICMV) – which enable continuous compaction control (CCC) – and conventional spot tests.

This study addresses this need by presenting full-scale, controlled laboratory experiments. The aim was to systematically evaluate how plate size (300 and 600 mm) and roller type (Dynapac CA3500D and CA6500D with static linear loads of 36 and 65 kg/cm) affect deformation moduli from PLT as well as 16 different ICMVs, spanning kinematic, mechanics-based, and energy-based categories.

The results show that the 600 mm plate, despite a slightly lower deformation modulus E_v2, yielded consistently improved correlations with ICMVs, likely due to its deeper influence volume. The heavier roller produced higher E_v2 values and lower variability than the lighter roller. Analysis of relationships among ICMVs revealed strong redundancy and three distinct clusters (harmonic-, stiffness-, and energy-based), supporting the use of a single representative from each cluster during calibration. The vibration modulus E_vib and harmonic-based indices performed best in correlation with PLT, while energy-based indices were less suitable.

In conclusion, larger PLT plates and mechanics-based ICMVs improve the correlation between CCC and PLT, but roller-specific calibration remains necessary. These results provide practical guidance for more reliable CCC-based workflows for quality assurance.

Jordpackning är en av de mest grundläggande och kritiska jordförstärkningsmetoderna och utförs i nästan alla typer av byggprojekt. En väl utförd packning är avgörande för en konstruktions funktion och livslängd, exempelvis för järnvägar och vägar där minimala sättningar är ett absolut krav. Traditionellt verifieras packningskvaliteten med hjälp av statiska plattbelastningsförsök (PLT), där deformationsmodulen från den andra lastcykeln (E_v2) mäts. PLT är dock arbetsintensivt och tidsödande, och ger endast diskreta mätpunkter, vilket lämnar stora delar av den packade ytan okontrollerad.

Yttäckande packningskontroll (YPK), även kallat intelligent packning, använder accelerometrar monterade på vibrationsvältar för att leverera spatialt kontinuerliga mätvärden i realtid. Trots att tekniken introducerades för flera decennier sedan är YPK fortfarande inte allmänt accepterat som en fristående metod för kvalitetssäkring, eftersom man hittills inte lyckats etablera tillförlitliga och konsekventa korrelationer mellan YPK-värden och konventionella PLT-resultat.

Detta projekt adresserar denna kritiska utmaning genom att genomföra en serie fullskaleförsök i laboratoriemiljö på en testbädd av grus. Arbetet kvantifierar systematiskt hur två faktorer påverkar den uppmätta styvheten och korrelationerna mellan YPK-värden och E_v2: PLT-plattans diameter (300 mm och 600 mm) samt typen av vibrationsvält (statiska linjelaster på 36 respektive 65 kg/cm). Totalt genomfördes 116 plattbelastningsförsök, och 16 olika YPK-värden utvärderades, vilka spänner över accelerationsbaserade, mekanikbaserade och energibaserade kategorier. De viktigaste resultaten är följande:

• Trots att 600 mm-plattan gav något lägre E_v2-värden, producerade den konsekvent starkare korrelationer med YPK-värdena än den normala 300 mm-plattan. En praktisk tolkning är att den större plattan mobiliserar en djupare influenszon som bättre överensstämmer med vältens effektiva mätdjup, vilket minskar känsligheten för lokal heterogenitet vid ytan.
• Den tyngre välten uppnådde en högre styvhet snabbare och uppvisade generellt lägre variabilitet vid ett måttligt antal överfarter, men den lättare välten gav ofta starkare korrelationer mellan YPK-värden och E_v2. Detta beror på minskad intermittent kontaktförlust samt en bättre överensstämmelse i mätdjup.
• Mekanikbaserade YPK-värden, i synnerhet vibrationsmodulen E_vib1 under lastningsfasen, visade den starkaste korrelationen med E_v2 (R upp till 0,8), medan energibaserade index var de minst lämpliga.
• Statistisk analys visade på en betydande redundans bland de analyserade YPK-värdena, vilka kan grupperas i tre tydliga kluster. Detta stöder strategin att endast använda en representativ indikator per kluster vid kalibrering.

Sammantaget ger resultaten praktisk vägledning för mer tillförlitliga YPK-baserade arbetsflöden för kvalitetssäkring. Studien identifierar också den primära utmaningen för att driva YPK-tekniken framåt: att utveckla maskinoberoende YPK-värden för universell kvalitetssäkring. Syftet med projektet är att utveckla en ny typ av YPK-värde som på sikt kan ersätta plattbelastningsförsök och således effektivisera kvalitetskontrollen.

The high-value reuse of recycled aggregates from construction and demolition waste has emerged as an important technological imperative for the construction of permeable road bases. However, the mechanisms of the fatigue crack generation and accumulation of micro damage to internal structure in cement-treated recycled aggregate materials(CRAM) remain unclear. Hence, it is necessary to analyze of the fatigue cracking behavior of CRAM and to explore the correlation mechanism between their internal microstructural characteristics and the macroscopic progression of fatigue damage. A damage mechanics model based on peridynamic theory was established accounting for the microstructural characteristics of CRAM. The fatigue damage mechanisms of these mixtures from meso-scale were revealed and the effects of recycled aggregate content and stress levels were explored. The results show that the peridynamic fatigue model is capable of accurately capturing the complete damage evolution process in CRAM. Notably, the stress concentration located at the interface between the aggregate and the cement matrix serving as the principal factor in the initiation and propagation of fatigue cracks. As the recycled aggregate content increases, the occurrence of fatigue cracks in dense cement-treated mixtures rises, with cracks exhibiting a straight morphology. Unlike dense cement-treated mixtures, the internal fatigue cracks in permeable cement-treated mixtures are influenced by stress concentration at the pore edges, resulting in fewer fatigue cracks. Their fatigue life exhibits a pronounced semi-logarithmic linear relationship with the stress ratio coefficient. Based on this finding, the derived fatigue equation can further predict the fatigue life of permeable cement-treated mixtures with varying recycled aggregate content at different stress levels, providing a theoretical foundation for the material optimization design of construction waste recycled cement-treated mixtures.

Unbound aggregate materials (UAMs) with large air voids are increasingly used to construct pavement base and subbase layers as part of the initiative to develop sponge cities and improve drainage performance. However, the motion of particles and the spatial distribution of kinetic energy during the particle rearrangement process induced by vibratory loading remain unclear. This study presented the results of laboratory vibratory plate compaction tests conducted on UAM specimens under various combinations of vibratory parameters and different levels of Gravel to Sand ratio (G/S). SmartRock (SR) sensors were embedded within the specimens to monitor real-time particle motion, while kinematic energy and its spatial distribution were analyzed from the acceleration time-history signals collected by the SR sensors. Based on high-precision industrial X-ray computed tomography (XCT) and three-dimensional (3D) reconstruction technology, the motion and migration characteristics of coarse particles were analyzed. A new compaction index was proposed based on particle motion and kinematic energy to evaluate the compaction quality of the specimens. The findings of this study reveal that vibratory compaction can be divided into three distinct stages. During the first stage, coarse particles primarily move vertically, while energy dissipation occurs mainly through the compression of air voids but does not form a dense skeleton structure. In the second stage, coarse particles translate horizontally while rotating vertically, resulting in a tendency of the particles to align horizontally with their long axes. During this stage, dissipated kinematic energy is primarily used to fill air voids, leading to the formation of a densely packed skeleton structure. Kinematic energy indices and particle movement in the middle of the specimens can be used to evaluate the compaction stage and quality. Additionally, the lateral particle motion within the specimens transitions from continuous ascent to gradual descent (i.e., culminating in minimal kinematic energy), thus indicating a relatively dense compaction state. Reducing the void ratio and increasing the contact area between particles within the size ranges of 4.75–9.5 mm and 2.36–4.75 mm can significantly increase the compaction density and improve the stability and deformation resistance of unbound pavement base/subbase layers. The results of this study provide valuable insights into the mechanisms of vibratory compaction and can be used to optimize compaction methods and improve pavement performance.

Jordpackning är den absolut vanligaste jordförstärkningsmetoden som används i princip i alla byggprojekt, och där är packningskvaliteten avgörande för hållbar infrastruktur. Traditionella metoder för kvali­tetskontroll, såsom plattbelastningsförsök (PLT), brister dock i effektivitet och täckning. Yttäckande packningskontroll ger en kontinuerlig och yttäckande uppföljning, men dess korrelation med plattbelast­ningsförsök har varit svag. Utvecklingen av yttäckande packnings­kontroll har under lång tid varit begränsad, men med nya tekniska möjligheter har intresse återigen väckts för att utveckla tekniken. Som en fortsättning på en serie tidigare utförd forskning, syftar detta doktorandprojekt till att vidareutveckla yttäckande packningskontroll med användning av moderna mät- och analysverktyg. I ett första steg korreleras PLT mot gamla och nya värden från yttäckande pack­ningskontroll. Det slutgiltiga målet är dock att utveckla den yttäckande packningskontrollen så att behovet av statiska plattbelastningsförsök försvinner.

Frictional sleepers have been shown to enhance lateral resistance effectively. However, the current design of the bottom grooves in frictional sleepers lacks standardized specifications, and the fundamental mechanisms responsible for enhancing lateral resistance remain unclear. To explore the underlying mechanisms behind the influence of different bottom groove patterns on lateral resistance, five distinct groove shapes for frictional sleepers were designed, and a series of hybrid discrete element method-finite difference method (DEM-FDM) numerical simulations were conducted by using the single sleeper pull-out test (SSPT) as the basis. The results show that in comparison with the standard sleepers, there exists a considerable variation in the degree of increase in lateral resistance among different types of frictional sleepers, ranging from 35.1% to 80.2%. Specifically, the lateral resistance increases logarithmically as the groove side area increases. The analyses show that the effective contact area (i.e., the contact area between ballast particles and the sleeper's bottom surface) was increased due to the constraint effect of the bottom grooves, simultaneously leading to increasing strength of the horizontal force chains near the grooves. Using discontinuous grooves can diminish the initial interference with the ballasts near the sleepers. Frictional sleepers primarily demonstrate anisotropy in the horizontal plane, which constitutes one of the mesoscale factors contributing to the enhancement of lateral resistance. In addition, the displacement and rotation of particles around the sleepers are constrained, whereas the most significant constraint was observed on the ballast shoulders. It is suggested to apply the frictional sleepers with discontinuous grooves and consider patterns that aim to maximize the groove side area in the design and maintenance of ballasted railway tracks.

As a typical multi-phase composite, the permeable cement-treated recycled aggregate base material (CRAM) faces critical technical challenge in balancing mechanical properties with ecological functionalities. To tackle the challenge, this study aimed to explore the meso-scale fracture and damage mechanisms of CRAMs via peridynamics (PD), providing valuable insights into and theoretical recommendations for the optimal design of CRAMs towards potential engineering applications. A mesoscopic heterogeneous numerical model of CRAMs was thus established, featuring the realistic consideration of the controllable pore structures and random aggregate distribution. The multi-scale damage mechanisms of CRAMs subjected to unconfined uniaxial compression were investigated, whereas the effects of porosity, interfacial transition zone (ITZ) properties, and the content of recycled aggregates on the unconfined compressive strength and fracture behavior of the CRAM specimens were further analyzed. The results show that the PD theory can accurately simulate the interactions within the CRAM specimens and naturally capture the initiation and propagation of cracks. The effective yield strength, ultimate compressive strength, and effective modulus of CRAM specimens all positively correlate with the matrix filling ratio Sm, with the ultimate compressive strength showing a clear linear relationship with Sm. A significant turning point in the failure mode of the specimens occurs within Sm=0.76–0.88, transitioning from pronounced aggregate crushing to cement matrix failure. The interface characteristics mainly affect the effective yield strength of the specimens: as the interface is enhanced, the effective yield strength increases, and the failure mode of the specimens gradually changes from interface debonding to ductile fracture of the cement matrix. As the replacement rate of recycled aggregates increases, the breakage of recycled aggregates becomes more severe, particularly for the recycled red brick aggregates, resulting in a decrease in the overall compressive strength and fracture performance of the CRAM specimens. In view of the impacts of these factors, the following technical measures were recommended to ensure higher strength of the CRAMs: the mixing and molding processes should be improved to enhance the homogeneity of the mixture, as the spatial distribution of aggregates and pores significantly influences the internal stress state of the specimens and ultimately the fracture performance of the mixture; furthermore, the proportion of red bricks in recycled aggregates should be minimized to limit the excessive breakage of aggregates.

Current practices in compaction quality assessment rely heavily on traditional spot tests, such as the static plate load test (PLT), which are labor-intensive, time-consuming, and limited in spatial coverage. This study develops a machine-learning framework to estimate deformation moduli from continuous compaction measurements. A small dataset of 60 samples from full-scale trials was used to train multi-output regression models, with deformation moduli from the first and second loading cycles (E_v1 and E_v2) as the targets. Two feature representations were compared: a five-feature set of selected intelligent compaction measurement values (ICMVs) and an eight-feature set of curated vibration response characteristics extracted from acceleration signals and contact force–displacement curves. Six algorithms were evaluated, including multiple linear regression, random forest, XGBoost, support vector regression, k-nearest neighbors, and the tabular foundation model TabPFN. The results show that curated response characteristics improved E_v2 prediction relative to the selected ICMV feature set. For Feature Set 1, SVR achieved the best E_v2 prediction with R² = 0.34 and RMSE = 7.1 MPa. For Feature Set 2, TabPFN achieved the best E_v2 prediction with R² = 0.45 and RMSE = 6.5 MPa, despite requiring no hyperparameter tuning. The simultaneous prediction of E_v1 and E_v2 enabled estimation of the E_v2/E_v1 ratio, with the lowest RMSE = 0.34 achieved by SVR on Feature Set 2. SHAP analysis indicated that second-harmonic acceleration amplitude, loading and unloading stiffness, and displacement amplitude were key features, providing physical interpretability for the data-driven assessment.

The nonlinear vibration response of the coupled roller-soil system remains difficult to interpret because it is governed by soil state, roller properties, operating parameters, and evolving drum-soil contact. This study develops a two-dimensional dynamic finite element model of vibratory compaction to examine the nonlinear vibration response and roller-integrated compaction measurement values (RICMVs). Four constitutive descriptions are compared under identical operating conditions: linear elasticity, Mohr-Coulomb (MC) plasticity, modified Drucker-Prager/Cap (DPC) plasticity, and hypoplasticity with intergranular strain (Hypo+IGS). The results show that the soil model strongly affects predicted settlement and stress-strain histories. The MC model captures yielding but produces only shear dilation, and the DPC model provides limited volumetric contraction. Hypo+IGS gives the most consistent response across initially loose to dense soil states because it captures void-ratio reduction, density-dependent stiffness, and distinct loading-unloading behavior. The depth of influence depends on the chosen response quantity, ranging from 1.0 to 1.3 m based on settlement and from 1.5 to 2.4 m based on density-related criteria. The vibration moduli, E_vib1 and E_vib2, are more closely related to final-state density measures than CMV and OMEGA. The normalized void-ratio reduction, (e₀−e)/(e₀−e_d), yields the most consistent relation across the evaluated ICMVs. These findings clarify the mechanisms controlling nonlinear roller response and support FEM-based interpretation of continuous compaction measurements.