Research On Key Technologies Of Grouting Pressure Control System

Pressure grouting is an ordinary process in which pressure is an important factor to the quality of dam. Traditionally, the low-precision manual control style can not guarantee the security, economical efficiency of grouting project. Aiming at the nonlinear, uncertainty, strong coupling and time-varied of grouting process, research on key technologies of pressure control system has high theoretical value and applied significance.Supported by the key scientific project of Guizhou(2005J-069), the author carried out the study in four areas:the design of pressure reference curve, soft sensing of grouting pressure, optimal control and developing the control system and realized the automatic operation for grouting system. The main content of study and innovation are ordered as follows:1. An new intelligent identification method was developed based on hybird support vector Takagi-Sugeno)(SVMTS) for grouting stratum.Based on the study of SVM theory, SVM kenerl function, SVM parameters optimization algorithm, T-S model and analysis the dis-tributing rules of water pressure test statistical data for different stratum, a new method of fuzzy-rule decision through SVM learning was proposed using pressurized water experimental data. Firstly, fuzzy rules through optimizational RBF SVM training model was obtained, next the "black-box" SVM decision function was expressed as the T-S fuzzy rules model and t-uninorm was adopted. Simulation results show that the new method improves identification accuracy, computing speed and decision process security, which is superior to RBF-NN method. The method has been applied to real project, which demonstrates the method can provide more scientific grouting pressure curve for control system.2. Grouting pressure soft-sensor model was established based on an adaptive T-S fuzzy method.T-S fuzzy soft-sensor model of pressure grouting was established to resolve the deep hole pressure measurement. The major auxiliary were selected using the finite element analysis software to study fluid dynamics pressure friction loss factors, combined with simplified mathematical model. The numbers of fuzzy rules are matched to the actual identification accuracy, and the parameters identification of fuzzy model are adopted as Kalman recursive algorithm. Simulation results show that the adaptive fuzzy algorithm is superior to fuzzy clustering algorithm and the error feedback learning in the modeling accuracy and computational speed. The soft sensing model has been by real project, which can provide the effective feedback pressure parameter.3. A new control strategy was proposed based on adaptive particle swarm optimization(APSO) for multi-models of the grouting pressure control system.On the basis of pressure soft sensing, a hybrid control strategy was proposed, which the controller parameters were optimized by APSO on multi-model of control system. Cascade control, feedforward contol and optimization were integrated in the new control method to avoid pressure fluctuation and parameter re-tuning. The outer loop was adopted fuzzy controller, inner loop adopted PID controller, grouting density is tuned by fuzzy feedforward method on the purpose of saving the project cost. controller parameters are optimized with APSO. Simulation results show that the control strategy obtain good tracking control effect in different stratum environment.4.Hardware system and software system of control system were firstly developed to realize the pressure automatic control.Research on the high-precision measurement ways of grouting pressure, flow and density parameters of control system, and design pressure sensor with high-pressure plastic isolator and non-contact density detector, which provide accurate data for control system. The algorithm above was firstly used to the grouting control experimental system, the results demonstrate the control strategy is superior to manual control ways in control accuray, stability and antijamming capability.