Pipes are the lifeline of every industry. Any damage to these pipes due to manufacturing defects or wear and tear can cause loss of resources and obstruction to normal activities. Detection of these damages at an earlier stage by human inspection methods is impossible. There is a need to find an effective method to detect pipe damage. This study aims at damage detection in pipe structures by using the image classification technique. A machine learning model is developed to detect the early-stage damages in the pipe. Machine learning (logistic regression, SVM, KNN, and random forest) and deep learning model (CNN) are used for developing the model. It is identified that the deep learning algorithm CNN has greater accuracy than any of the machine learning models, and hence, it could be used in real-time damage detection of pipes.
Top1. Introduction
Pipelines are like arteries of the world. They are critical structures employed in many real-time applications such as agricultural irrigation, oil, gas, water supply, drainage, heat supply, and various types of industrial equipment (Du et al., 2016). They are the best-preferred mode of transportation for liquid and gas. The pipeline transportation industry is divided into three subsectors such as pipeline transportation of crude oil (NAICS 4861), pipeline transportation of Natural gas (NAICS 4862), and other pipeline transportation (NAICS 4869), which covers around 3,500,000 km of pipeline in 120 countries of the world. The transportation of water, oil, and natural gas are being done through millions of miles of pipelines worldwide. Pipelines play a very important role in the transfer of oil and natural gas, which are the most used energies globally, contributing to 58% of global energy consumption (Goss, 1983).
Piplines are like lifelines for human lives. Schorr et al. (2012) argues that the pipelines need to protected from erosion, corrosion, scaling and fouling. Water is a very important primary source on our planet used in many industries. This water is transmitted regularly using underground pipelines. Water pipelines loses on an average of 30% of water transmitted to various industries. There are multiple reasons for water loss, such as leakage, theft, and metering errors (Hawari, 2015). The loss is expected to increase in under-managed networks of pipelines (Van Zyl & Clayton, 2007). The damage in water pipes would have dangerous impacts on industries and human beings. Hence there is a serious requirement on minimising their impact by dealing with damages of pipes as early as possible.
There are two main causes of damage detection in oil and natural gas pipes: intentional like vandalism and unintentional like pipe surface damages due to ageing (Ajao et al., n.d.). The recent statistics show the major causes of damages in the pipeline are many factors, such as corrosion, human negligence, installation, and erection, and manufacturing process (Bolotina et al., 2018). The statistics show in figure no.1 shows the incidents of pipe leakage, which are hard to avoid and the also the causes of damage are several. Hence, it is very important to continuously monitor the pipelines and detect these causes of damage in pipelines at an earlier stage to reduce the loss rate and prevent the environmental consequences due to pipeline failure.
The supply network of oil, gas and watr in any industry, is prone to many risks (Technology Status Report on Natural Gas Leak Detection in Pipelines, 2005). To detect these risks and evaluate the safety of pipelines have become significant in pipeline integrity management (Pan et al., 1999). In many nations, including India, manual techniques are executed by human examiners, which are tedious and many times faulty. Corroded areas can be disregarded in pipeline frameworks that are hard to reach and notice. Hence there is a high need to devise a more useful and exact technique for pipe condition survey.