A preview of this full-text is provided by IOP Publishing.
Content available from Engineering Research Express
This content is subject to copyright. Terms and conditions apply.
Eng. Res. Express 6(2024)025516 https://doi.org/10.1088/2631-8695/ad4004
PAPER
Fault diagnosis of three-cylinder mud pump based on transfer
learning
Chang Yan
1
, Zhiliang Liu
1
, Feilong Liao
2
, Jiyang Zhang
1
and Menghang Dai
1
1
School of Mechanical and Electrical Engineering, University of Electronic Science and Technology of China, Chengdu, People’s Republic
of China
2
Institute of Safety and Environmental Quality Supervision and Testing, CNPC Chuanqing Drilling Engineering Company Limited,
Guanghan, People’s Republic of China
E-mail: Zhiliang_Liu@uestc.edu.cn
Keywords: mud pump, fault diagnosis, transfer learning, multiple operating conditions
Abstract
Mud pumps serve as vital components within the circulating system of oil drilling platforms, primarily
facilitating the circulation of drilling fluid. With the rapid advancement of deep learning technology,
there has been a growing focus on fault diagnosis techniques for mud pumps based on deep learning
methodologies. However, existing deep learning approaches often struggle with fault diagnosis of
mud pumps under varying operational conditions, as adjustments to the working conditions are
necessary in real-time based on drilling depth. To address this challenge, this study introduces an
enhanced transfer learning method for diagnosing faults in mud pumps across different operating
conditions. Initially, the collected vibration data undergoes resampling to standardize frequency,
followed by the utilization of the short-term autocorrelation method to discern phase information of
signal impact. Leveraging this phase information, the signal is segmented into distinct segments with
uniform phases, thereby minimizing distribution discrepancies between the source and target
domains. Subsequently, the transformer is employed as a feature extractor for the model. Finally, a
deep sub-domain adaptation network is employed to facilitate transfer from the source domain to the
target domain. Validation of the proposed method was conducted using an experimental dataset, with
results demonstrating its efficacy compared to other contemporary approaches.
1. Introduction
The mud pump is an essential equipment for ensuring the required flow rate of the circulation system during
drilling. It extracts the drilling fluid from the mud tank and injects it into the well bottom through the drill rod,
removing the debris produced during the drilling process to the surface [1]. Due to a large amount of sediment
usually present in the mud and the high-pressure condition of the mud pump, the intake valve, discharge valve,
and piston of the mud pump are prone to failure, causing leakage of fluid and reducing the pressure generated by
the mud pump, thereby affecting the normal progress of drilling work. Although mud pumps are critical rig
equipment, their health monitoring currently still relies on human observation. This approach often fails to
detect pump damage at an early stage, resulting in non-productive time (NPT)and increased well construction
costs when pumps go down unexpectedly and catastrophically [2]. Therefore, fault diagnosis is of great
significance to improve production efficiency and reduce accident rates for complex mechanical systems [3].
Fault diagnosis of critical parts of the mud pump is of great significance for oil drilling [4].
To realize accurate fault diagnosis of the mud pump, many methods have been proposed. Under the
complex drilling working conditions, it is difficult to fully describe the working environment of the drilling
pump fluid end using a single signal source for achieving high fault diagnosis accuracy. Catalin Teodoriu et al [5]
utilize sound signals to detect Mud Pumps. Gang Li et al [6]propose a dual-source gramian angular (DS-GAF)
method to fuse vibration-strain signals to provide more fault information. By using the DS-GAF method, the
vibration-strain signals at the fluid end are fused and converted into a fault diagnosis image dataset. Fault
RECEIVED
7 December 2023
REVISED
24 March 2024
ACCEPTED FOR PUBLICATION
17 April 2024
PUBLISHED
29 April 2024
© 2024 IOP Publishing Ltd