Abstract：To avoid the safety hazards caused by structural damage during the operation of derrick steel structures, a method of derrick steel structure damage identification based on principal component analysis and random forest algorithm is proposed by combining time-domain multi-parameter information fusion with machine learning algorithms based on the displacement response signal of derrick steel structures. The acceleration response signal of the derrick steel structure under the impact load is extracted using the acceleration sensor, and the displacement response signal is obtained by processing the acceleration response signal, and three time-domain features containing the impulse factor, margin factor and cliffness are extracted; the three features are fused into a new comprehensive time-domain feature using principal component analysis, and most of the information contained in the original signal is retained; the processed data is fed into the random forest algorithm model for derrick steel structure damage identification, the whole process only needs to collect the post-damage signal, not the pre-damage signal. The above damage identification model was used to simulate the finite element model of the derrick steel structure, and the experimental results of the ZJ70 derrick steel structure laboratory model were also analysed. The analysis results show that the above damage identification method can accurately and efficiently determine the single or multiple damage of the derrick steel structure, and the correct rate of judgment is over 90%.
SHI P M, ZHANG H C, YI S Y, et al. An improved adaptive multivariate modal decomposition bearing fault signal feature extraction method ［J］. Acta Metrologica Sinica, 2022, 43(10): 1326-1334.
WANG X L, LIU J X, ZHANG R L, et al. Structural damage identification of simply supported girder bridges based on improved D-S evidence theory ［J］. Railway Construction, 2022, 62(7): 5-10.
ZHU J X, L B X, QIAO S, et al. An intelligent diagnosis method for ultrasonic flowmeter faults based on principal component analysis and multidimensional Gaussian Bayes ［J］. Acta Metrologica Sinica, 2020, 41(12): 1494-1499.
WANG C H, CAI J H, ZENG J S. Research on rolling bearing fault diagnosis based on empirical modal decomposition and principal component analysis ［J］. Acta Metrologica Sinica, 2019, 40(6): 1077-1082.
WANG P, GONG P, FENG D, et al. A soft measurement method for downhole crude oil water content based on random forest algorithm ［J］. Acta Metrologica Sinica, 2019, 40(5): 835-841.
WANG M, YANG J L, WANG X, et al. Identification of mud shale petrographic logs based on random forest algorithm ［J］. Earth Science ,48(1): 130-142.
HAO S B, ZHANG B, LIU Y. Application of direct analysis method in the analysis of overall stability of derricks for oil drilling rigs ［J］. Mechanical Engineer, 2022(8): 82-84.
XIONG H B, LI Z Q. Current status of research on structural health monitoring ［J］. Structural Engineer, 2006(5): 90-94.
SHI P B, LI R, LI M K, et al. Error estimation and fault detection of smart meters based on decision tree and clustering algorithms ［J］. Acta Metrologica Sinica, 2022, 43(8): 1089-1094.
LUO B, DU J, ZHANG G S. Design of wireless strain testing system for oil derricks ［J］. Computer Knowledge and Technology, 2010, 6(19): 5352-5354.
WANG Q, LEI Y Q, WANG H, et al. Development and application of telescopic portal derricks for continuous tubular drilling ［J］. Petroleum Machinery, 2022, 50(8): 1-8.
［8］
YANG J, ZHANG D D, FRANGI A F, et al. Two-dimensional PCA: A new approach to appearance-based face representation and recognition ［J］. IEEE Transactionson Pattern Analysis and Machine Intelligence, 2004, 26(1): 131-137.
CHANG H T, CAI J, WEN Y, et al. A high-frequency noise filtering based on principal component analysis for TDLAS ［J］. Acta Metrologica Sinica, 2022, 43(10): 1285-1290.
2024, Vol. 45 
