Neuro-fuzzy system of continuous control of density asphalt concrete mixtures

Abstract

Improving the quality and increasing the service life of asphalt concrete pavements of roads is a national economic problem, which is solved through measures related to improving the regulatory framework, improving the properties of road materials, automating its control and management of the compaction process. Existing automated systems for monitoring and controlling the density of road rollers are based on artificial intelligence methods. A feature of the construction of the upper layers of asphalt concrete road surfaces in the Russian Federation is a significant impact on their quality of the results of the work of asphalt pavers, which provide several technological operations - acceptance, laying and compaction of asphalt concrete mixtures. The use of an automatic density control system for asphalt concrete mixtures in the process of laying them will eliminate many defects in road surfaces during their operation. The aim of the work is to build a system of continuous density control in the process of laying and compacting asphalt concrete mixtures by pavers based on artificial intelligence methods. The article presents the results of the development of a new system for intelligent control of the density of the asphalt concrete mixture by pavers. It is proposed to use the structure of a neuro-fuzzy network of the ANFIS type. Training of a neuro-fuzzy system of the ANFIS type was performed on the basis of a combination of methods of least squares and a decreasing gradient on an array of variables obtained on the basis of the results of experimental studies performed by VNIIStroydormash, SoyuzDorNII, MADI. Automation of density control is aimed at improving the quality of asphalt concrete pavements of roads.