Objective: types of fault. Condition monitoring, can be used

Objective:

This
work aims to introduce an approach to centrifugal pump fault diagnosis based on
Industrial big Data Analytics using artificial intelligence, machine learning and
multi-sensor data fusion.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Introduction:

With
the increasing degrees of automation and intelligence of industrial equipment,
along with the development and spreading of the application spectrum of related
cutting-edge technologies, data started to grow exponentially. It becomes
important to process and
analyze collected industrial big data in order to obtain a great diagnostic
value using it. However, corresponding problems such as computational overload,
widespread uncertainty, the lack of fault samples, and the difficulty of
predicting fault types, have prevent the applications of previous fault
diagnosis methods. Therefore, the question of how to realize rapid analysis and
deeper mining of massive monitoring data has become a study focus in the big data
era.

Centrifugal
pumps are complex machines, which can experience different types of fault.
Condition monitoring, can be used in centrifugal pump fault detection through
vibration analysis for mechanical and hydraulic forces. Vibration analysis
methods have the potential to be combined with artificial intelligence systems
where an automatic diagnostic method can be approached. An automatic fault
diagnosis approach could be a good option to minimize human error and to
provide a precise machine fault classification.

Vibration
monitoring and analysis are applicable for fault detection of pumps and
rotating machines as all machines vibrate accelerometers can be used to extract
vibration signals from machines, which are then analyzed using software. A
vibration signal is processed to present useful information about the condition
of the machine.

Automatic
fault detection methods make use of Artificial Intelligence (AI) which seeks to
replicate mental capabilities with the support of computational systems. The
Artificial neural network (ANN) was first introduced by McCulloch and Pitts ,
and Fuzzy logic was first introduced by Zadeh. Artificial intelligence systems
have been applied for centrifugal pump fault diagnosis using different methods
for the feature extraction, starting from a simple method of statistical
analysis and a wavelet transform has been applied using time frequency method.
Proposed ANN with machine learning to diagnose pump faults. Then, Zouari et al.
applied ANN and a fuzzy neural network to diagnose centrifugal pump faults;
statistical methods of time and spectral analysis were used for the feature
extraction.

Literature
survey:

1.    Industrial
Big Data for Fault Diagnosis: Taxonomy, Review, and Applications by Yan Xu,
Yanming Sun, Jiafu Wan, Xiaolong Liu and Zhiting Song,IEEE
Access (2017)- in this work they have proposed different fault diagnosis and detailed
description of each work has been given.

2.    Experimental
Set-Up for Investigation of Fault Diagnosis of a Centrifugal Pump by Maamar Ali
Saud Al Tobi, Geraint Bevan, K. P. Ramachandran, Peter Wallace, David Harrison,
International Journal of Mechanical and Mechatronics Engineering (2017)-in this
work fault diagnosis of centrifugal pump using big data analytics has been
done. Condition monitoring for vibrational analysis for mechanical and
hydraulic forces.

3.    Concentration
measurement of three-phase flow based on multi-sensor data fusion using
adaptive fuzzy inference system by XiaoxinWanga,HongliHua, AiminZhang (2014).-
bio mass, coal and air are the three parameters whose flow has been measured
using multi sensor data fusion and adaptive fuzzy inference system.

 

Proposed
work with methodology:

The
procedure consists of three main stages, namely, data collection, data
pre-processing and extraction, and fault classification. Classification and
diagnosis of the centrifugal pump condition will be implemented using
artificial intelligence. Classifiers consist of two main processes: training of
data, and testing, where an automatic classification will be implemented for
the different conditions. The performance of the machine learning techniques will
be measured according to the classification accuracy rates (%). The output
layer will be testing the centrifugal pump conditions, and with the current
proposed conditions.Here the concept of big data analytics and processing with
multi-sensor data fusion has been introduced.

Implementation:

Implementations
of vibration artificial intelligence techniques that for the first time in the
machinery condition-monitoring field; heuristic algorithm would be investigated
as a
learning algorithm for the neural networks.

Centrifugal
pump with  sensors for measuring
vibration measurements are used. Under various conditions i.e. normal and fault
conditions we get the data. Pre-processing of this data has been done using
machine learning techniques. And finally, analysis of this work is done with
the help of artificial neural networks.

For
the development of test bed environment, a noisy motor operating environment is
chosen for experiment where other equipment is also running in parallel. The
fault is created in motor artificially by damaging the bars in motor. For the
online test, an oscilloscope is used to record the current signal data through
current probe and capture the current signal and store in USB drive for
analysis. The signal is analysed by using MATLAB/Simulink model. For the
creation of complexity in analysis, two types of load levels (no load and full
load) are applied to observe the abnormal behavior of current signal because,
at no-load level, it is very difficult to identify significant sidebands and
decide the condition of motor.

 

Work
plan:

Month1:
Literature survey and Identification of model hardware and purchasing

Month
2: Experimental setup erection and data collection

Month3:
Data collection and Multi-sensor data fusion concept execution.

Month4:
Data preprocessing and extraction.

Month5:
Fault classification by AI techniques

Month6:
Industrial Big Data Analytics application in proposed work.

 

 

Expected outcomes:

The fabrication of the setup i.e. with centrifugal pump and
sensors setup for fault diagnosis. The health of the industrial machine will be
found and faulty condition will be detected based on which the fault
anticipation is achieved.

Applications:

To
anticipate and detect problems in industrial machines. Failures in machines can
incur high maintenance or replacement costs, or if neglected, may cause
catastrophic accidents leading to production downtime and potential failure to
supply, hence affecting profitability. Consequences include loss of
availability, cost of spares, cost of breakdown labor, cost of secondary damage
and risk of injury to people and the environment. Any company that seeks
optimum production at the lowest cost has to adopt a reliability function
rather than a repair function for the maintenance strategy; the proposed work
is a reliability function, which can lead to optimum function at efficient way
implementing advanced machine learning techniques in industrial machine health
monitoring

Conclusion:

Many experimental works can be conducted using the
centrifugal pump experimental setup including the investigation of the
vibrational behaviors, which can be measured, monitored and analyzed through
the vibration analysis methods by which machine health condition will be
monitored.  This work presents a novel
experimental setup, which is made specifically to study mechanical and
hydraulic vibrations of the centrifugal pump, and it offers different vibrational
monitoring options.  Work would be
implementation of vibration artificial intelligence techniques for the first
time in the machinery condition-monitoring field; machine learning algorithms
will be investigated as a learning algorithm for the neural networks.

x

Hi!
I'm Josephine!

Would you like to get a custom essay? How about receiving a customized one?

Check it out