Research On Real-time Monitoring Of Filling And Rolling Quality And Prediction Model Of Compactness For High Fill Channel

South-to-North Water Diversion Middle Route Project has a length of1432km,and the length of high fill channel segment, with a filling height larger than6m, is139.5km. As a typical linear engineering, high fill channel project is characterized bywide distribution, huge earthwork volume, tight schedule and complex managementmode, which have brought a severe challenge to management and control of theconstruction quality. Filling and rolling quality control is one of the key links ofconstruction quality control for high fill channel. How to realize meticulous andall-weather real-time monitoring and analysis of filling and rolling construction of thechannel, meanwhile, how to realize real-time and on-line prediction of the compactionquality, are important issues to be considered during engineering construction. Aimingat the study on real-time monitoring of filling and rolling quality and prediction modelof compactness for high fill channel, this paper has obtained some achievements asfollows.(1) Aiming at the key scientific problems of filling and rolling quality control forhigh fill channel, index system and control criterion of real-time monitoring for fillingand rolling quality control is established, mathematical model of real-time monitoringis proposed and visual expression of filling and rolling construction is realized. Basedon the theoretical research, real-time monitoring system of filling and rolling qualitycontrol for high fill channel is researched and developed. Then the system issuccessfully applied to the high fill channel of South-to-North Water DiversionMiddle Route Project, which realizes meticulous, all weather and real-timemonitoring and analysis for filling and rolling process of high fill channel andimproves the quality control level and efficiency during construction of the high fillchannel.(2) Based on the real-time monitoring technology of filling and rolling qualitycontrol for high fill channel, rolling parameters during construction process, includingcompaction passes, compaction thickness, rolling speed and vibration status, arereal-timely acquired and monitored. Combined with moisture content andcompactness obtained by field test, correlation between compactness and its maininfluencing factors, namely rolling parameters and moisture content, is analyzed.Then based on the correlation analysis, prediction model of compactness isestablished applying multiple linear regression analysis method. Verified by measured data, the model is of high precision.(3) Characterized by high self-learning ability and self-adaptive, high robustnessand fault-tolerant ability, strong ability of generalization and treatment of complexnonlinear issues, artificial neural network method is applied in establishing predictionmodel of compactness as well. After evaluating performance of the network,prediction precision of the model is compared with that of the model established bymultiple linear regression method. Establishment of the prediction models ofcompactness make it possible to obtain compactness real-timely.Aiming at the problem that traditional control methods of construction qualitymay be greatly affected by human factors, all-weather, meticulous and real-timeconstruction quality monitoring technology is applied to carry out real-time andonline prediction of compactness, overcoming the considerable limitations thatcompactness obtained by field test can only evaluate compaction quality of somerandom positions. The study has provided decision basis for filling and rollingconstruction quality control for high fill channel, and accumulated many precioustechnical data. This successful case provides an innovative method for filling androlling construction quality control for high fill channel.