Tesi etd-06302025-094423 |
Link copiato negli appunti
Tipo di tesi
Tesi di laurea magistrale
Autore
CALDARI, ROBERTO
URN
etd-06302025-094423
Titolo
Highway Pavement Maintenance in Italy: Traffic–Temperature Correlation Analysis and Improvement Strategies
Dipartimento
INFORMATICA
Corso di studi
DATA SCIENCE AND BUSINESS INFORMATICS
Relatori
relatore Nanni, Mirco
Parole chiave
- business analysis
- data analysis
- italian highway
- pavement
- temperature
Data inizio appello
18/07/2025
Consultabilità
Non consultabile
Data di rilascio
18/07/2095
Riassunto
Italian motorways have accumulated a maintenance backlog of roughly €12 billion since 2006, and many sections now exhibit rutting, cracking and potholes that compromise safety and ride quality. Two accelerating stressors drive this deterioration. First, freight demand has expanded the heavy-vehicle fleet by more than fifty per cent since 2000, so that by 2023 truck flows on tolled routes exceeded pre-pandemic 2019 levels by about four per cent, even though car traffic only just returned to its former volume. Second, summers are becoming markedly hotter: maximum air temperature along the network is rising by about half a degree per decade and the number of days above 35 °C has climbed everywhere, reaching roughly thirty-five days a year on the southern Adriatic corridor A14, about eighteen in central Italy on A1, and—significantly—around ten even in the traditionally temperate Po Valley on A13. Asphalt binders soften when surface temperatures exceed their design grade, so extreme heat now regularly triggers wheel-path rutting and bleeding, while intense rainfall events weaken base layers and accelerate fatigue.
The thesis analyses a thirteen-year record (2010-2022) of traffic counters and weather stations on A1, A13 and A14 and shows that annual average daily traffic rose steadily until 2019, collapsed to sixty–seventy per cent of baseline during the 2020 lockdown, then rebounded to at least ninety-five per cent by 2022. A1, the nation’s main north–south artery, consistently carries about ninety thousand vehicles a day, whereas A13 and A14 carry lower but still substantial volumes; truck share is highest on A13 at roughly thirty per cent, twenty-five on A1 and twenty on A14. Because pavement damage scales roughly with the fourth power of axle load, even a modest rise in heavy traffic multiplies structural fatigue. During the pandemic trough, car flows vanished far more than truck flows, so the percentage of heavy vehicles briefly spiked and demonstrated that pavement loading does not fall linearly with headline AADT.
Climate records over the same window reveal a clear rise in heat extremes. Southern sections already exceed 35 °C on roughly one summer day in three; central Italy now does so on one in five; and the north, once almost immune, experiences sequences of ten such days a year. Heatwaves in 2012, 2017–2019 and, most dramatically, 2022 produced pavement surface temperatures well above 50 °C, a threshold at which standard binders lose stiffness and rapid plastic deformation can occur. Time-series forecasting with ARIMA indicates that, all else equal, heavy-truck flows could reach thirty thousand a day on A1, twenty thousand on A14 and eleven thousand on A13 by 2030, while the annual count of very hot days is likely to climb toward forty-plus in central Italy and perhaps exceed fifty in extreme years. Thus the two external drivers—mechanical loading and thermal softening—are poised to intensify in parallel over the coming decade.
To integrate these orthogonal stressors with the age of the pavement surface, the study defines a Composite Pavement Stress Index. Each year’s PSI sums a traffic factor (heavy-vehicle volume normalised to a reference of thirty thousand trucks per day), a thermal factor (hot-day count normalised to forty days per year) and a maintenance-age factor (years since the last resurfacing normalised to a ten-year design life). Whenever a new overlay is applied, the age component resets, so PSI can drop sharply before climbing again as loads and heat accumulate. On A1, a major resurfacing in 2015 reset PSI to near zero, but by 2022 the index had surged to around eighty, the highest in the dataset, because the route combines the largest freight flow with the recent succession of hot summers. A13 and A14 show similar but slightly lower profiles, each interrupted by maintenance events in 2016 or 2018. Roughness measurements crowdsourced by the SmartRoadSense smartphone platform confirm that higher PSI aligns with poorer ride quality: plotting PSI against the platform’s Pavement Performance Estimate yields a near-linear relationship with an R² of roughly 0.78, meaning that three-quarters of observed variation in road smoothness can be explained by the composite stress estimate.
Because 2022 saw both near-record truck volumes and the most extreme heatwave on record, it stands out as a natural stress test for Italian highways. Field reports note wheel-path rutting and bleeding on heavily trafficked grades and toll-plaza approaches, particularly where vehicle speeds are low and shear stresses rise. The data therefore suggest that traditional eight-to-twelve-year overlay cycles will become insufficient: on critical freight corridors the composite index now climbs fast enough to breach degradation thresholds in six to eight years. Preventive strategies should therefore include shorter resurfacing intervals on segments with PSI growth rates above a calibrated alarm level, the adoption of polymer-modified or higher-grade binders capable of resisting surface temperatures above seventy degrees, thicker or rigid layers where freight concentration peaks, and the reinforcement of drainage and waterproofing in zones prone to intense precipitation or freeze–thaw swings.
Movyon’s Evolutive Pavement Management System already digests lane-level geometry, historical inspection indices and work orders; the thesis proposes extending it with continuous feeds from inductive loops, radar traffic detectors, weather stations and crowdsourced smartphone roughness, as well as integrating machine-learning deterioration models that refine ARIMA baselines by learning non-linear interactions among load, climate and material properties. Pilot tests show that coupling ARIMA with supervised learning can predict two-year International Roughness Index evolution with a mean absolute error below six per cent, and life-cycle optimisation studies indicate that data-triggered maintenance could cut long-term expenditure by fifteen to twenty per cent compared with fixed-interval resurfacing while reducing unplanned closures. The ultimate vision is a closed feedback loop in which periodic or real-time data update deterioration curves, the system projects each segment’s PSI trajectory with confidence bands, and dashboards flag the forecast year in which stress will cross a target threshold. Engineers can then simulate alternative schedules or material upgrades and select the minimum-cost path that keeps PSI, and thus roughness, within acceptable bounds.
In sum, Italy’s main motorways face a compounding challenge: freight volumes continue to rise and climate trends now expose pavements to heat levels for which many older surfaces were not designed. Historical analysis, stress-index synthesis and predictive modelling together show that reactive maintenance cannot keep pace; instead, agencies must shift to anticipatory asset management that blends denser sensing, climate-resilient mix design and data-guided intervention timing. Doing so promises smoother roads, fewer emergency repairs and a more sustainable allocation of scarce maintenance funds in the hotter, heavier-traffic era that lies ahead.
The thesis analyses a thirteen-year record (2010-2022) of traffic counters and weather stations on A1, A13 and A14 and shows that annual average daily traffic rose steadily until 2019, collapsed to sixty–seventy per cent of baseline during the 2020 lockdown, then rebounded to at least ninety-five per cent by 2022. A1, the nation’s main north–south artery, consistently carries about ninety thousand vehicles a day, whereas A13 and A14 carry lower but still substantial volumes; truck share is highest on A13 at roughly thirty per cent, twenty-five on A1 and twenty on A14. Because pavement damage scales roughly with the fourth power of axle load, even a modest rise in heavy traffic multiplies structural fatigue. During the pandemic trough, car flows vanished far more than truck flows, so the percentage of heavy vehicles briefly spiked and demonstrated that pavement loading does not fall linearly with headline AADT.
Climate records over the same window reveal a clear rise in heat extremes. Southern sections already exceed 35 °C on roughly one summer day in three; central Italy now does so on one in five; and the north, once almost immune, experiences sequences of ten such days a year. Heatwaves in 2012, 2017–2019 and, most dramatically, 2022 produced pavement surface temperatures well above 50 °C, a threshold at which standard binders lose stiffness and rapid plastic deformation can occur. Time-series forecasting with ARIMA indicates that, all else equal, heavy-truck flows could reach thirty thousand a day on A1, twenty thousand on A14 and eleven thousand on A13 by 2030, while the annual count of very hot days is likely to climb toward forty-plus in central Italy and perhaps exceed fifty in extreme years. Thus the two external drivers—mechanical loading and thermal softening—are poised to intensify in parallel over the coming decade.
To integrate these orthogonal stressors with the age of the pavement surface, the study defines a Composite Pavement Stress Index. Each year’s PSI sums a traffic factor (heavy-vehicle volume normalised to a reference of thirty thousand trucks per day), a thermal factor (hot-day count normalised to forty days per year) and a maintenance-age factor (years since the last resurfacing normalised to a ten-year design life). Whenever a new overlay is applied, the age component resets, so PSI can drop sharply before climbing again as loads and heat accumulate. On A1, a major resurfacing in 2015 reset PSI to near zero, but by 2022 the index had surged to around eighty, the highest in the dataset, because the route combines the largest freight flow with the recent succession of hot summers. A13 and A14 show similar but slightly lower profiles, each interrupted by maintenance events in 2016 or 2018. Roughness measurements crowdsourced by the SmartRoadSense smartphone platform confirm that higher PSI aligns with poorer ride quality: plotting PSI against the platform’s Pavement Performance Estimate yields a near-linear relationship with an R² of roughly 0.78, meaning that three-quarters of observed variation in road smoothness can be explained by the composite stress estimate.
Because 2022 saw both near-record truck volumes and the most extreme heatwave on record, it stands out as a natural stress test for Italian highways. Field reports note wheel-path rutting and bleeding on heavily trafficked grades and toll-plaza approaches, particularly where vehicle speeds are low and shear stresses rise. The data therefore suggest that traditional eight-to-twelve-year overlay cycles will become insufficient: on critical freight corridors the composite index now climbs fast enough to breach degradation thresholds in six to eight years. Preventive strategies should therefore include shorter resurfacing intervals on segments with PSI growth rates above a calibrated alarm level, the adoption of polymer-modified or higher-grade binders capable of resisting surface temperatures above seventy degrees, thicker or rigid layers where freight concentration peaks, and the reinforcement of drainage and waterproofing in zones prone to intense precipitation or freeze–thaw swings.
Movyon’s Evolutive Pavement Management System already digests lane-level geometry, historical inspection indices and work orders; the thesis proposes extending it with continuous feeds from inductive loops, radar traffic detectors, weather stations and crowdsourced smartphone roughness, as well as integrating machine-learning deterioration models that refine ARIMA baselines by learning non-linear interactions among load, climate and material properties. Pilot tests show that coupling ARIMA with supervised learning can predict two-year International Roughness Index evolution with a mean absolute error below six per cent, and life-cycle optimisation studies indicate that data-triggered maintenance could cut long-term expenditure by fifteen to twenty per cent compared with fixed-interval resurfacing while reducing unplanned closures. The ultimate vision is a closed feedback loop in which periodic or real-time data update deterioration curves, the system projects each segment’s PSI trajectory with confidence bands, and dashboards flag the forecast year in which stress will cross a target threshold. Engineers can then simulate alternative schedules or material upgrades and select the minimum-cost path that keeps PSI, and thus roughness, within acceptable bounds.
In sum, Italy’s main motorways face a compounding challenge: freight volumes continue to rise and climate trends now expose pavements to heat levels for which many older surfaces were not designed. Historical analysis, stress-index synthesis and predictive modelling together show that reactive maintenance cannot keep pace; instead, agencies must shift to anticipatory asset management that blends denser sensing, climate-resilient mix design and data-guided intervention timing. Doing so promises smoother roads, fewer emergency repairs and a more sustainable allocation of scarce maintenance funds in the hotter, heavier-traffic era that lies ahead.
File
Nome file | Dimensione |
---|---|
La tesi non è consultabile. |