Growing demand, increasing exploration costs, global competition and enhanced regulatory requirements are driving energy companies to focus on production optimization and return on investment, while managing their risk portfolios.
Your organization continues to amass vast amounts of exploration, production, and distribution data. There is a tremendous opportunity for your firms to tap into this data and uncover insights that can result in new, billion-dollar revenue streams. However, this complex data has far too many attributes for even your most capable scientists and engineers to effectively query.
Ayasdi’s solution provides your domain experts with a more accessible approach to advanced analytics. The solution combines advanced machine learning techniques with topological data analysis to automatically analyze thousands of attributes simultaneously and surface patterns from complex data. With Ayasdi’s solution, your scientists and engineers can extract insights from your data without having to know which questions to ask or write a single line of code.
Leaders in the energy sector are using Ayasdi’s advanced analytics solution to optimize energy exploration and recovery, improve uptime and reduce risk through predictive maintenance, discover break-through events, improve trading performance, operationalize trader surveillance, and complement existing signal processing methods.
Upstream oil and gas activities such as exploration, appraisal and field development often require significant capital resources. With Ayasdi’s solution, firms can glean new insights from subsurface geological, geophysical, petrophysical, and engineering production data.
For instance, an upstream exploration company used Ayasdi’s analytics solution to segment more than 6,000 unconventional wells into good, moderate and low producing regions. The analysis of multi-modal attributes, including rock properties, completion parameters, and production measures, identified features (proppant size and fluid viscosity) in a sub-population of high producing wells in one region where production levels were generally mediocre. The insight provided the company with an immediate opportunity to adjust the protocols for its wells across the entire region to increase recovery.
In another instance, a major oil and gas company used Ayasdi’s solution to create a topological network of core samples. Within this network, two sub-groups of core samples were identified that correlated with high production. The analysis revealed that wells that are geographically close can have subtly different geologic signatures, and that there can be more than one geologic profile that correlates with high estimated ultimate recovery (EUR). This finding is promising, considering that geologic features can change over relatively short distances, and knowing that there are multiple geologic profiles to consider increases the locations within a play where production can be optimized. The analysis also showed that these groups could be identified early in the production of the shale, with very little production information.
Deepwater expenditure is set to grow by almost 130% compared to the preceding five-year period, totaling $260 billion between 2014 and 20181. It is estimated that companies engaging in deepwater exploration incur approximately $5 million in drilling and exploration costs every day. By positively identifying underwater oil reserves, your firm can reduce the costs of exploration and accelerate time to market.
Analyzing multi-modal measurements (e.g., seismic data, sensor readouts, and geophysical data) can be challenging using standard techniques. For instance, rock sample characteristics such as density, calcium content, water content, and porosity have a direct correlation with positive oil reserve identification. Ayasdi’s advanced analytics solution can combine these data sets and shed new light on where to target drilling operations.
For example, one of the world’s largest independent oil and gas exploration and production companies used Ayasdi’s solution to identify key factors that indicate deepwater oil reserves, which translated into significant financial savings for the company. Specifically, petrophysicists found new correlations in an analysis of petrofacial rock sample data, which, when overlaid with seismic and sensor data, resulted in the identification of high-value target regions for deepwater drilling. The large volumes of complex data had previously surpassed the ability of the company's specialized software applications, and left their petrophysicists struggling to determine the factors that would make a drill site profitable.
Using Ayasdi’s deep analytics, your companies can redirect their drilling initiatives, saving millions of dollars in what can otherwise be fruitless efforts, and accelerate time to value.
1World Deepwater Market Forecast 2014-2018 - http://www.douglas-westwood.com/shop/shop-infopage.php?longref=1266#.U-PdFYBdWxw
There is an opportunity to further exploit existing signal processing techniques using Ayasdi’s advanced analytics solution.
Traditional approaches to signal processing tend to discard large fractions of the data as “noise”, focusing instead on very tightly defined signals. Ayasdi’s advanced analytics solution has been used to quickly distinguish between traditional “signal” and “noise”, while also identifying additional signal within the noise. The ability to extract additional signal from data increases the amount of information that existing sensors can deliver. While this technique has been demonstrated on acoustic signals, it can also be applied to temperature as well as other sensor data, potentially reducing the need to deploy more sensors in hard-to-reach areas.
For instance, using Ayasdi’s solution, oil and gas companies can analyze acoustic, temperature and seismic readings to identify a signature that represents an oil trap. Ayasdi’s advanced analytics recovers information that is generally ignored or processed out of sensor or seismic data analysis.
Ayasdi’s analytics solution can help your organization examine the various attributes that characterize equipment condition and use that information to predict when a problem is likely to occur. With that insight, you can schedule maintenance for the equipment at the right time, before the problem affects process or equipment performance. This significantly reduces unexpected downtime, repair costs, and the potential risk of fines from health and safety issues.
For instance, a leading exploration and production (E&P) company leveraged Ayasdi’s analytics solution to predict equipment failure by examining sensor data. The solution ingested system measurements such as the temperature and pressure at multiple points. It was able to validate key indicators of impending failure. More importantly, the solution revealed two distinct failure modes by examining variance of these indicators. This additional signal in the data, identified within a few hours of analysis, had previously gone undetected. It provided a better way to predict equipment failure thereby preventing unnecessary downtime and maintenance procedures. It is estimated that the cost of suboptimal maintenance (20% replacement value) can be 10x that of proactive maintenance (2% replacement value). Predicting machine failure can result in significant savings through increased uptime and reduced repair costs.
A major oil and gas company turned to Ayasdi to help them identify indicators of breakthrough events that impact oil recovery using the steam assisted gravity drainage (SAGD) technique.
Ayasdi’s analytics solution took in a series of measurements (e.g., temperature and pressure) at multiple well depths and automatically segmented them. It identified clusters of measurements leading up to, during and after the breakthrough event. The team was then able to delve into specific attributes (and their values) and compare them across groups of measurements. This helped the company quickly identify indicators of breakthrough events that impact oil recovery.
Trading at some of the largest oil companies has evolved into very sophisticated profit centers. Your trading desks are now able to tell refineries to increase capacity to produce beyond customer demand so they can take the overage to trade. Crude oil, natural gas, diesel, and refined fuels can be sold several times over in transit between source, pipeline, and final retail location. However, trading divisions’ performance can be highly volatile and risky.
Using Ayasdi’s analytics solution, energy producers can uncover insights from disparate sources of data to gain a better understanding of macro-economic conditions and trends and improve trading performance. It can also automatically uncover patterns of fraud which helps firms operationalize trading surveillance best practices and deter rogue behavior.