Abstract: The aim of this research is to effectively prevent water diversion tunnel from collapse during construction. According to previous research results and engineering practice, an index system for assessing the collapse risk of water diversion tunnel was established. The preference ratio method (PRM) and anti-entropy weight (AEW) were employed to analyze the weight of each index factor. Moreover, in consideration of the uncertainty of risk factors and the fuzziness of assessment rating, a risk assessment model for diversion tunnel collapse was constructed based on entropy theory and extension cloud model. The collapse risks of nine typical segments of the single line diversion tunnel of Dongguan branch in the Pearl River Delta water resources allocation project were assessed as case study. Results demonstrated that: the collapse risk level of diversion tunnel GS19+590-GS20+050, GS20+050-GS21+340, GS22+226-GS22+553 and GS24+911-GS26+035 segments was grade II, implying moderate risk; the risk level of GS21+340-GS22+226,GS22+553-GS23+788,GS23+788-GS24+911,GS26+035-GS27+700, and GS27+700-GS28+061 segments is grade III, indicating high risk. The assessment results are consistent with the actual situation of the project, which verifies the practicability and effectiveness of the method in tunnel collapse risk assessment.

Key words: tunnel collapse, preference ratio method(PRM), anti-entropy weight(AEW) method, extension cloud model, risk assessment