Abstract: To solve the problems of ambiguity and multifactor interaction in the geological adaptability assessment of a tunnel boring machine(TBM), a multidimensional relationship forward normal cloud model evaluation method based on set pair analysis coupled with the multidimensional cloud model is proposesd. The characteristics of 12 TBMbased geological adaptability cases in 106 sections of tunnels in China are analyzed. In addition, 10 typical indicators, including equipment parameters, geological conditions, and adverse geological problems, are selected to establish an evaluation index system. Each evaluation index is used as a onedimensional variable of a multidimensional relational forward normal cloud model. According to the principle of isotropy and cloud theory, three appropriate cloud digital feature values are obtained to generate cloud droplets at each adaptive level. Then, according to the comprehensive certainty of the TBM geological adaptability for each adaptive level, the principle of maximum certainty is used to 〖HJ1.6mm〗determine the geological adaptability grade of a TBM. This comprehensively reflects the relationship between certainty and uncertainty between multiple factors in the evaluation. In addition, based on the comprehensive constant weight determined by the IFAHP, improved ICRC, and improved game theory combinatorial empowerment, the variable weight theory based on the factor space idea is used to determine the total level ranking weight of the indicator. Furthermore, the active intervention of indicators in the evaluation process is realized. The results show that the variableweight multidimensional contact normal forward cloud model is feasible for evaluating the geological adaptability tendency of TBMs, and the evaluation accuracy is 90%. Compared with conventional multidimensional cloud models and other calculation methods, the degree of agreement is higher and the calculation is more concise.

Key words: TBM geological adaptability, TBM adaptive tendency, multidimensional contact cloud model, variable weight, set pair analysis, evaluate forecasting method, cloud model