Research On The Control Theory And Method Of Construction Proeess Quality For Ballastless Turnout

With the construction of high speed railways at massive scale, ballastless turnout becomes widely adopted and rapidly developed. Because of the complicity in its construction, high level of accuracy requirement, and multiple influential factors for ballastless turnout quality, a robust and effective quality control methodology needs to be developed to ensure the construction quality. Based on related researches of this topic conducted domestically and at overseas, general process quality control theory and method was applied for ballastless turnout construction in this study and a preliminarily ballastless turnout quality control system was established. Major research work conducted includes:(1) Establishing preliminary process quality control system of construction process of ballastless turnoutUsing slab ballastless turnout on subgrade as an example, the study analyzed the construction process of ballastless turnout in detail, summarizing ballastless turnout construction characteristic. Through in-depth analysis of multiple influential factors that may affect construction quality, including constructor, machinery, material, method, and environment etc., necessity of quality control in ballastless turnout construction is illustrated. Furthermore, a framework of quality control for ballastless turnout construction is established.(2) Proposing ballastless turnout construction process recognition methodBased on the characteristic of ballastless turnout construction process and the unique contribution of each construction phase and construction sub-process to overall construction quality, the important construction phase and sub-process were identified and analyzed to determine the quality parameters of the process output.According to the features of diversity, many types of distribution and a small sample of output parameters of ballastless turnout construction process, calculation method of ballastless turnout construction process capability index is established. Muti-process index calculation method was developed in this study to recognize critical sub-process, using realistic ballastless turnout construction quality datum data as basis. (3) Enabling quality surveillance and diagnosis for the ballastless turnout construction processBased on Grubbs calculation theories, we have improved Grubbs calculation process, and compared simulation data with existing Grubbs method. The results show that the improved Grubbs method is better, and this method is used on data screening when establish ballastless turnout construction sub-process surveillance model. Muti-sub-process surveillance method on ballastless turnout construction was established using screened quality data, T2control chart, and generalized variance control chart. Combining the main component analysis method and x-R chart, the unusual factor diagnosis method for muti-sub-process of ballastless turnout construction was developed to realize ballastless turnout construction muti-sub-process surveillance and diagnosis.(4)Establishing quality assessment method of ballastless turnout construction processBased on comprehensive assessment on current ballastless turnout construction practice, the weakness of existing assessment method was identified, and the principle and idea of establishing assessment method for ballastless turnout construction process quality were put forward. This study also established the assessment model and confirmed the assessment index for this process. The weighting coefficients in the model were obtained through AHP, and three levels of quality assessment results of the sub-process, the phase and the whole process of ballastless turnout construction were obtained and the analytical methods of assessment results were put forward. A practical implementation procedure for the proposed assessment method of ballastless turnout construction is successfully developed.(5) Establishing quality forecast model for construction process of ballastless turnoutCombing gray GM(1,1) forecast model with BP neural network forecast model, GM-BP forecast model was proposed in this study. When compared together, the GM-BP model, GM model, and BP neural network model had relative errors of forecast of0.9%,2.5%and3.8%respectively. Therefore we concluded that GM-BP model combing model has the highest accuracy in its forecast results. Using GM-BP as forecast model and utilizing ballastless turnout construction quality datum in some construction site, the forecasts for quality of the sub-process, the phase and the whole process of ballastless turnout construction were implemented and the premise for prior intervention of quality of ballastless turnout construction process was provided.