Oil & Gas EPC - Digitizing P&ID

Oil & Gas which was once was a lucrative investment, now struggles to ensure the same level of light. A lot of things have changed in the past few years which has pushed the Oil & Gas industry into a less profitable terrain for the investments. The slump in the oil market and the resulting fall in new investments has drastically changed the Oil & Gas EPC market. This has led to fewer interests from investors and projects for EPC clients implying a low number of deals in the market. This shift has led to a reduction in the number of "Proposals to Conversion" ratio. In order to sustain the same order every year, EPC companies are having to respond to almost twice the number of proposals. This has put the scalability of the proposal process to test.

Leveraging P&ID

One of the cumbersome and manual processes is to read the hundreds of P&ID diagrams and arrive at the correct MTO (Material take off). This process is highly time-consuming and incurs miscalculations due to manual errors. This could end up affecting overall profitability in the project. Also most of the times, EPC companies keep a high buffer in order to hedge the risk of last-minute change required because of incorrect MTO created manually by the engineers. This is a risky affair altogether as the process is monotonic and requires a lot of human-involvement, leaving a large room for standardization.

Fast Code Solution

With a leading EPC Oil & Gas player, Fast Code has automated the process of developing AutoMTO, thereby saving a huge amount of time and resources for the in terms of man-hours and costs for the EPC giant. Fast Code leveraged Artificial Intelligence techniques which automatically detects all the instruments, pipes, texts and codes. This step takes less than a minute to extract all the MTO information from a P&ID.

Extracting P&ID from a simple PDF

Reading of P&ID from pdf and image-based format: Computer vision techniques were employed to read P&ID files in PDF/image form and identify the text, lines and other components. Machine learning was leveraged to identify and locate all the Equipment, Instruments, and Valves. It also identifies T-joints, Spec Brakes, Reducers etc. involved in MTO generation. This complete process which takes an engineer 3-4 days of work, is finished by AutoMTO in 3 steps as shown in the above figure. The components were then associated with the lines and thus preparing the Line List, Equipment List and the full MTO automatically. The process improved the overall efficiency of reading and reduced the time taken to process per P&ID sheet by a factor of 15. Provision has also been provided to bring in a human in the loop to review and make necessary modifications to ensure 100% accuracy of MTO.

Toward 100% Accuracy

AutoMTO as a state of the art system lets achieve 100% accuracy whilst preparing MTO and this accuracy is then translated into better sales, lean buffer size and competitiveness in the bidding stage. The solution achieved 100% accuracy in the generation of MTO. This reduced engineering work drastically by leveraging one-shot and active learning AI techniques. The product has led to saving about 90% the time taken to generate the MTO. The version tracking mechanism has allowed more efficient identification of changes made to the input P&IDs.
Lean Buffer and Competitiveness: The solution ensured a more accurate MTO and thus provides a higher degree of confidence to the proposal owners. This has allowed the client to reduce the Material cost buffer, thus making the proposals more competitive. Fast Code enabled the EPC player to create MTO with leaner buffer size and in the process making the bids for EPC much more competitive.
Divisions like Sales, Proposals, Procurement get direct advantage in terms of this impact as AutoMTO ensures digitization benefits are distributed across the system 1) 14x Improvement in Efficiency in reading and processing P&ID sheets
2) 90% Reduction in time while reading and generating an MTO
3) 100% Accuracy generated from the solution
4) > $3m Cost saved per year (40 proposals/year)

Download Case Study