Data mining and analytics for performance optimization

Exploring data analysis techniques for easy, cost-effective solutions in the oil and gas industry.

By Bert Baeck, TrendMiner October 9, 2016

Courtesy: Rick Ellis, Oil and Gas Engineering, CFE MediaThe oil and gas industry is under huge pressure to alleviate cost burdens. Crude oil prices have been low and while the market may be stabilizing, this is the perfect time for these companies to streamline operations to stay competitive.

Drilling operations are one of the best areas for improvement. IDC Energy Insights, a market research provider, reported that while drilling costs represent nearly half of well expenditures, only 42% of drilling operations’ time is actually spent drilling. The majority of time is spent on problems, rig movement, defects, and waiting periods. 

Improving production with data analytics

One of the ways to improve production is by using data analytics that help:

  • Enable operators to correlate real-time downhole drilling data with production data of nearby wells to optimize drilling strategy. According to industry experts, this practice can help oil and gas companies increase production by between 6 and 8%.
  • Avoid potential problems by identifying deviations from signature profiles; for example, ensuring the downhole equipment and reservoir are protected when bringing wells online to avoid costly mistakes.
  • Predict downhole tool failures and immediately determine which parameters may need adjusting by analyzing real-time data relative to past performance or events. 

Identifying problems early with data analysis

Wells have a type-curve that profiles expected production over their lifetime. Decline curve analysis (DCA) is a common graphical procedure used for understanding declining production rates and forecasting performance.

Over time, well production rates typically decrease from a loss of reservoir pressure or changing volumes of the produced fluids. The DCA concept involves fitting a line through the performance history then assuming this same trend will continue in future forms.

While helpful, these curve analysis trends can be made more precisely through the use of analytics.

For example, data analysis can quickly identify when flow profiles deviate from the historical or type-curves. This information can alert an operator of a potential problem.

Managing assets by using data analytics

Data analytics can improve a company’s bottom line by prolonging equipment life, increasing asset availability, and extending maintenance windows. ARC Advisory Group reported that of all asset failures that occur, only 18% of them show any kind of perceptible pattern due to any increased use or age. By identifying changes in system behavior well before traditional operational alarms do, analytics software gives decision-makers time to analyze the necessary data and take corrective action.

For example, data mining and analytics can make a big difference in pressure testing blow-out preventers (BOPs). According to the Code of Federal Regulations (CFR), pressure tests must be conducted every two weeks for every drilling operation worldwide to ensure BOPs can withhold a well control event.

Analytics software can track the actual component performance data and look at profiles to improve BOP system uptime, reduce unnecessary maintenance, and help deliver better cost forecasting. It can provide information to onshore engineers for better decision-making, thus reducing downtime associated with accessing historical BOP data while optimizing maintenance and reducing unnecessary parts replacements. 

Data mining techniques

One of the barriers to identifying and monitoring various interconnected elements in production cycles is cost. Having data scientists build models to determine how minor changes to certain areas of operations can yield big productivity gains, is expensive and time consuming. With low oil prices and tight profit margins, many oil and gas companies believe they cannot afford the investment to implement traditional analytics solutions.

There is a new method to uncovering areas for improvement. This approach is based on pattern recognition within the enormous pool of data. Extremely sophisticated software can read trends to determine where similar patterns have occurred in the past. Moreover, this software can be installed and deployed in as little as 2 hours with just another couple of hours needed for training operators and other authorized users.

Analytics software can uncover hidden patterns in data to find an optimal sequence, which is repeatable. It can identify and send alerts on any deviations (in profile and duration) in both real time and also in post-run analysis.

For instance, these data mining techniques can be applied to a comprehensive data set to identify potential correlations between drilling activity and incremental rate of penetration (ROP). It can help calculate drilling performance in real time, under specific conditions and constraints.

Data mining techniques can also greatly help with cementing operations. Cementing casing strings into drilled holes is a critical operation in well construction. History indicates that many of the cement jobs performed in the industry did not achieve quality metrics; however the work is usually deemed acceptable because there was no means to analyze the quality until now. Today’s advanced software can perform that level of analysis in real time or immediately after as an evaluation technique.

Meeting oil and gas industry challenges with data analysis

Today’s oil and gas companies face many challenges. In order to be successful, they must ensure all areas of operations are performing optimally. It is essential to have the right information at all levels of the organization in order to make the best decisions.

  • At the production unit level, operators need to know how equipment is performing in real time. This is where analytics can be used to improve uptime, efficiency, and throughput.
  • At the facilities level, managers must be able to make educated decisions about procurement, production scheduling, and shipping without having to spend a lot of time and money on modeling and data scientists.
  • At the enterprise level, executives need real-time, accurate data to relate production to the larger business context and understand the impact of fluctuating costs, changing market conditions, and asset performance.

Although crude oil prices may have reached a bottom, oil and gas producers are still under considerable pressure to optimize all areas of production. In the past, expensive data analytics solutions could be used to uncover hidden areas of improvement, but these solutions required data scientists and modeling to work effectively.

Now oil and gas companies have a quick, easy, and affordable solution for optimizing performance through pattern-recognition software that can quickly sift through billions of time series data points to find instances where events have occurred. By using this new approach to data analytics, oil and gas companies gain valuable insight into operations and systems behavior—often in just a few hours after installation-to discover new areas for improvement.

Bert Baeck is CEO of TrendMiner, a research company for data mining in the process industry.