Fault diagnosis method of submersible screw pump based on random forest

Abstract

The difficulty in directly determining the failure mode of the submersible screw pump will shorten the life of the system and the normal production of the oil well. This thesis aims to identify the fault forms of submersible screw pump accurately and efficiently, and proposes a fault diagnosis method of the submersible screw pump based on random forest. HDFS storage system and MapReduce processing system are established based on Hadoop big data processing platform; Furthermore, the Bagging algorithm is used to collect the training set data. Also, this thesis adopts the CART method to establish the sample library and the decision trees for a random forest model. Six continuous variables, four categorical variables and fault categories of submersible screw pump oil production system are used for training the decision trees. As several decision trees constitute a random forest model, the parameters to be tested are input into the random forest models, and various types of decision trees are used to determine the failure category in the submersible screw pump. It has been verified that the accuracy rate of fault diagnosis is 92.86%. This thesis can provide some meaningful guidance for timely detection of the causes of downhole unit failures, reducing oil well production losses, and accelerating the promotion and application of submersible screw pumps in oil fields.

Intelligent Fault Diagnosis Methods for Hydraulic Piston Pumps: A Review

“…This method effectively filtered signals and extracted fault features from the frequency domain. Jiang et al [37] studied an approach based on the balanced random forest algorithm under imbalanced data set conditions, which improved the accuracy of axial piston pump fault classification. Chao et al [38] proposed an adaptive decision fusion method, which combines multiple sensors to increase the fault identification performance for HPPs using multi-channel vibration signals.…”

Current Status and Applications for Hydraulic Pump Fault Diagnosis: A Review

“…Lakshmanan et al [110] proposed a hydraulic pump fault diagnosis method that takes the pressure signal, flow signal, and torque signal of the pump as original real-time data for feature extraction, and inputs them into SVM after CWT. Jiang et al [111] used the decision tree to build a random forest model, trained six continuous variables of the hydraulic screw pump system as input characteristics, and built a hydraulic pump fault diagnosis method based on the random forest model. Hu et al [112] built a multi-fault diagnosis system based on data fusion according to the D-S evidence theory and used DMM to build a fault diagnosis feature with a basic probability assignment function, ensuring the objectivity of reliability distribution evaluation.…”

A Survey on Industry 4.0 for the Oil and Gas Industry: Upstream Sector

“…The application of AI for accurate sand production prediction in wells [140] and for shale well production [141] have also shown promising results. The use of discrete event simulation DT in studying the operational risk in oil sand mining and processing of bitumen in response to geological uncertainty was shown to be a good coordination tool [142], enhancement of drilling operations using IoT and AI models [143] and detection of fault in submersible screw pumps in oil wells [144]. The use of AI in the detection of casing damage due to non-uniform in-situ stress has been explored in [145]- [147].…”