Abstract: As one of the main factors of evaluating the anchoring quality, bolt anchoring strength in the concrete is affected by many factors which have complex interrelationships with each other. Concrete age, bolt diameter and anchorage length are used as the input column vector to build sample set, and radial basis kernel function is adopted for an prediction model of bolt anchoring strength in concrete based on support vector machine for regression. There are 30 groups of experimental data of anchoring strength. Two-thirds of those groups are selected as training samples, and the others are selected as predicted samples to forecast anchorage strength. Then the predicted results are compared with the test result and BP network result. Comparison suggest that this model has high prediction accuracy and good generalization ability, and the prediction results are credible. It is reasonable and effective to apply SVM to predict the anchorage strength in concrete.

Key words: concrete, bolt, anchoring strength, support vector machine, regression model