Data mining and squashing
The first step of Solution Seeker’s data processing pipeline is to automatically and continuously mine all raw production data.
Oilfield operations capture large amounts of raw data, and make it available through real-time databases or historians. During the last decade, most large oil companies have made significant investments to enable such data capture. Simultaneously, data analytics, in particular within machine learning and mathematical optimization, has experienced much success in many sectors. However, the oil and gas industry has not been able to keep pace with the development and deployment of these powerful technologies. Today, stored operational data is not utilized to its full potential for production enhancement. This is especially true for daily production optimization, which often consists of manual and work-intensive processes. Solution Seeker’s ProductionCompass AI overcomes these challenges through its innovative data compression and analytics algorithms, its compact database technology, and its innovative visualization capabilities.
The first step of ProductionCompass AI’s data processing pipeline is to automatically and continuously mine all raw production data from thousands of sensors, using machine learning techniques and advanced statistical analysis to filter, sort, clean, compress and prepare information sets for the AI engine. The processed production data is stored in our proprietary compressed database in real-time, retaining all operationally relevant information at less than 1% of the initial data volume. This compressed, high quality data set then fuels the machine learning algorithms and ensures the best possible learning, estimation, prediction and optimization.
Currently it is common for engineers to visually inspect data, with analysis being limited to time spans where information is known to be rich and relevant (such as during well test campaigns). This is only natural, since manual data inspection is such a time-consuming task. This is especially true for fields without a dedicated test separator, like subsea tie-in fields. The downside is that large quantities of data lies untouched. The opposite is true with the ProductionCompass AI; a constant stream of refined and processed data enables an "everything is information" paradigm, which leads to a much higher level of data utilization. Thus, even before introducing advanced analytics and optimization algorithms, our AI-analyzed data holds value on its own.