講演情報
[AI-07]Development and technical verification of a temperature-inference type compaction management system
*Satoru Kajiwara1, Gorou Ushikubo1, Hisashi Aita1 (1. NIPPO CORPORATION (Japan))
キーワード:
asphalt pavement、compaction、temperature prediction、machine Learning
Ensuring the structural integrity and long-term performance of asphalt pavement requires rigorous control of the compaction level of asphalt mixtures. Among the governing factors, the temperature of the mixture during placement and compaction plays a critical role in achieving the target density. Conventional construction practice assigns dedicated personnel to measure and manage temperature in real time; however, such manual approaches pose challenges in terms of efficiency, labor demand, and on-site safety. These limitations underscore the need for an automated and reliable temperature management method.
This study investigates an autonomous on-site temperature management framework utilizing a machine-learning-based predictive model for estimating internal temperatures of asphalt mixtures during the laying process. Field data-including mixture temperature, meteorological conditions, and construction parameters-were collected from active paving sites and analyzed to construct a predictive equation for internal temperature behavior. The derived model was subsequently integrated into an enhanced compaction management system.
The improved system incorporates three major automated functions: (1) real-time acquisition and recording of spreading and compaction temperatures; (2) automated guidance for roller operators, providing optimal compaction zones and timing based on predicted thermal profiles; and (3) automatic generation of temperature management documents, thereby reducing administrative workload and human error.
This paper details the architecture and operational features of the developed system, as well as the findings from its field trials. The results demonstrate that the system effectively improves temperature-based quality control, enhances operational safety by eliminating the need for manual temperature observation, and contributes to greater consistency and efficiency in asphalt paving processes.
This study investigates an autonomous on-site temperature management framework utilizing a machine-learning-based predictive model for estimating internal temperatures of asphalt mixtures during the laying process. Field data-including mixture temperature, meteorological conditions, and construction parameters-were collected from active paving sites and analyzed to construct a predictive equation for internal temperature behavior. The derived model was subsequently integrated into an enhanced compaction management system.
The improved system incorporates three major automated functions: (1) real-time acquisition and recording of spreading and compaction temperatures; (2) automated guidance for roller operators, providing optimal compaction zones and timing based on predicted thermal profiles; and (3) automatic generation of temperature management documents, thereby reducing administrative workload and human error.
This paper details the architecture and operational features of the developed system, as well as the findings from its field trials. The results demonstrate that the system effectively improves temperature-based quality control, enhances operational safety by eliminating the need for manual temperature observation, and contributes to greater consistency and efficiency in asphalt paving processes.
