Artificial intelligence (AI) has permeated every fabric of society from toys to oil rigs. The COVID-19 pandemic is — in a way — paving the way for AI to be integrated in a more comprehensive manner in different industrial and technological sectors, such as trenchless technology.

The AI incorporation is inevitable considering the vulnerability of human beings to sickness and the ensuing absence from work.

But, can AI be effectively incorporated into the trenchless sector for pipeline risk assessment?

We believe that AI and human intelligence can be combined to create an intelligent risk assessment for pipeline management.

The Pipeline Maintenance Issue

The problem with pipeline maintenance is an ongoing issue that costs millions of dollars every year in repairs, wasted resources and wasted business time.

It is also called a ticking time bomb, and rightly so, considering the condition of sub-surface pipelines, some of which are far past their serviceable life and far out of reach, lost in the underground labyrinth.

The solution to the problem is not simple and it is unfair to demand answers based on the available data in a short period of time.

How AI Can Be the Solution to Pipeline Risk Assessment

Advanced pipeline inspection tools have made it easier to inspect pipelines as they can be operated remotely and can detect defects that may be overlooked by the human eye. (QUIZ: All About Robotic Crawlers — 12 Questions to Test Your Pilot Prowess.)

However; analyzing the data is a time-consuming process especially when a significant amount of data is concerned.

According to Mike Russin, General Manager at WinCan, CCTV inspection data can be reviewed at greater speed and accuracy using AI and it can assist the operators in the field to make more accurate observations during the inspection process using the software.

AI is trained to recognize patterns that indicate problems which will need human attention and can analyze a large amount of data at once.

AI for pipeline inspection has the ability to recognize defects in pipelines by scanning the entire available data from CCTV inspections.

The defect detection using AI takes much lesser time and is more accurate compared to the relatively small amount of data that can be analyzed by an expert analyst after investing a significant amount of time.

Since the analysis procured using AI is complete, it helps pipeline maintenance inspectors and operators to pinpoint the exact location of the defect and prevent costly errors.

Phil Cannon, CEO of POSM Software notes that AI will make the identification and coding of defects much more consistent by reducing human subjectivity from the defect identifying and coding process.

The AI technology should also be capable of replicating the specific coding system developed by NASSCO, known as the Pipeline Assessment Certification Program (PACP).

The coding system provides standardization, convenience, and consistency to the methods for identifying, recording, evaluating and managing problems with pipes such as deterioration, defects, and deformities.

Incorporating AI automates the risk assessment process by identifying defects, annotating them in the video, grading the pipeline health and generating reports as per PACP requirements.

Risk Assessment Technology Using Artificial Intelligence

To understand how AI benefits risk assessment, we will mention one such platform.

It will give an insight into how AI can increase the capability of an organization by allowing them to deal with multiple aspects in one platform.

Risk Intelligence Platform (RIPL) is a software developed by American Innovations that supports compliance and risk management functions by integrating huge amounts of data from disparate sources such as PODS (Pipeline Open Data Standard), GIS (Geographic Information System), ArcGIS, and other relational geospatial databases.

It also helps facilitate regulatory compliance and allows for active pipeline risk management by providing in-depth risk analysis.

According to their datasheet, the key benefits of RIPL include (written verbatim):

Data Integration

Accommodates a wide variety of distribution and operator pipeline data and minimizes time spent acquiring, organizing and validating information. Brings volumes of data from all sources into a single data repository.

Compliance Solution

Supports regulatory compliance and provides an audit path from source data to final results. Facilitates regulatory compliance with Preventive and Mitigative Measures (P&MM) requirements and provides for active risk management through various risk reduction “what if?” scenario capabilities.

Data Alignment

Manages inspection data through time. Visualization of multiple anomaly features with reports, linear graphing and GIS. Stores and aligns ILI (in-line inspection) data and allows comparison with other critical integrity data.

Risk Assessment and Analysis

Threat screening analysis and quantitative/ qualitative risk rankings for prioritization and monitoring of inspection activities.

Powerful Direct Assessment Tools

For external corrosion, internal corrosion and stress corrosion assessment. Integrates data to determine HCA’s (habitat conservation assessments) and assessment regions, classifies assessment data, prioritizes excavation sites and analyzes data to determine remaining live and/or re-assessment intervals.

Relentless Support, Training, and Implementation

Supports RIPL & System Analyzer implementation, configuration and data analysis.

Final Thoughts

Now that AI is making giant strides in almost all types of industries, it is time for trenchless pipeline contractors to incorporate it into their system as well.

The ultimate purpose of AI for pipeline risk assessment is to get to the problem before the problem gets to us. Replacing humans for AI should not be the goal as human input is invaluable.

To completely submit to AI technology will take many years and in fact, may not even be a reality. After all, AI is a machine and it does not have its own intelligence.

The intelligence it uses is human, and after all, humans make mistakes.