Adhesive Property Research Of HPFL Reinforced Concrete Based On BP Neural Network

In recent years,the High Performance FerrocementLaminate(HPFL)technology is widelyappliedinthe engineering reinforcementfield widely applied.The reinforcement technique has many advantages such ashigh strength, good durability, convenient construction etc, and a broadapplication prospect.But in practical engineering application, the strippingdamage will appear between the reinforced layer mortar and the concrete,leading to the failure of reinforcement, so the adhesive property of HPFL layerwill influence the overall effect of the reinforcement or repairmen in a greatextent.In this paper, further study of adhesive property of HPFL layer and theconcrete is based on the experimental research which has completed, thisresearch especially focus on the adhesive property of the reinforced layer mortarand the concrete under the combined forced state, the main research contents areas follows:1. The construction process of HPFL reinforcement method and materialperformance of the main materials were introduced in this paper. The bondmechanism research status of HPFL reinforced layer at home and abroadin recent yearswere summarized, the adhesive property and the influence factors ofthe new-old concrete were analyzed, it can be considered thatthe researchesabout the adhesive propertyof HPFL reinforced layer come down to the basictheory of the adhesive propertyof the new-old concrete.2. A deep research has been done to the adhesive propertyof HPFLreinforced layer, the tensile bonding strength, shear bonding strength, failuremodel of the reinforced layer, influence factor, and the significance ofreinforcement layer mortar and concrete were analyzed. Also the research aboutHPFL reinforced layer was compared with the adhesive propertyand influencefactors of the new-old concrete, the similarities between two of them weresummarized.3. The basic model of artificial neural network and its classification, thenetwork characteristics, the structure of network, the network learning principles,etc. were completely summarized,and go into particulars the BP neural networkmodel and the algorithm of itsmath expression, model design and parameterdetermining methods, LM algorithm math expression.4. The BP neural network model about the prediction of the HPFL reinforced layer tensilebonding strength and shear bonding strength wereestablished, the training of the network was done by using the experimental dataas a data base, and successful predicted the tensilebonding strength and shearbonding strength of the HPFL reinforced layer and concrete.5. According to the experiment research of shear strength of the new-oldconcrete under the combined forced state, and using the experimental data of theHPFL reinforced layer, the prediction of the HPFL reinforced layertensilebonding strength and shear bonding strength were established withsatisfactory prediction results.6. The formula fitting of tensilebonding strength and shear bonding strengthof the HPFL reinforced layer and the concrete was done by using MATLABtoolbox functions, the formula calculation results is in good agreement with theneural network prediction results, which proved the validity of the formula, theresearch have a certain reference value to the HPFL reinforcement designmethod.