Browsing by Autor "M.. Villamizar"

Now showing 1 - 2 of 2

Application of Linear and Partial Correlation Techniques to Enhance the Waterflooding Surveillance Process
(2013) M.. Villamizar; Francklin Rivas; G. A. Carvajal; F.. Md Adnan; A.. Kheterpal; S. Knabe; José Antonio Rodríguez García; A.. Al-Jasmi; Bader Al-Saad; H. K. Goel
Abstract Linear correlation techniques (LCTs) and partial correlation techniques (PCTs) are well known statistical techniques useful in generating smart workflows for real-time surveillance and monitoring processes, such as waterflooding. In real-time environments, which require short-term analysis (30-day cycling), traditional simulation techniques are less effective and take a considerable amount of CPU time. As an alternative and considering LCTs have been widely applied in the oil industry, to understand relationships between producer and injector wells, LCTs can provide rapid results and even predict expected water breakthrough in producer wells. This paper describes the use of the Pearson correlation coefficient (PCC), a statistical measurement that is very sensitive to linear relationships between water injection and producing wells, including correlations where one parameter is a nonlinear function of the other. While this technique can show high correlation between some variables that are unrelated and only small correlations between highly related variables, the use of PCC, in many cases, returns measurements concerning the joint behavior of two variables that are important to the decision-making process. This paper describes how PCC was used to clean noisy data generated in LCTs. This paper also describes the use of the PCT to detect the influence of other variables on the LCT-based results. The study was conducted on a mature, carbonate black-oil reservoir in the Middle East under waterflood. The LCT/PCT results were compared with streamline simulation, and some similarities were found between the streamline simulation well allocation factors and the LCT/PCT results. Regardless of local reservoir conditions, the results showed that LCTs/PCTs are powerful tools that can be used to quickly assess inter-dependencies among producing wells and associated injectors. Use of these tools can enable engineers to make prompt decisions to help prevent water breakthrough and adjust water injection rates.
Short-Term Production Prediction in Real Time Using Intelligent Techniques
(2013) A.. Al-Jasmi; H. K. Goel; Hatem N. Nasr; M.. Querales; Jordani Rebeschini; M.. Villamizar; G. A. Carvajal; S. Knabe; Francklin Rivas; Luigi Saputelli
Abstract Intelligent digital oilfield operations collect real-time data from an operating asset and transform that raw data into information through intelligent, automated work processes, which assist engineers with key well operations and monitoring, improving their productivity and decision-making. A major oil and gas operator in the Middle East is developing a set of intelligent workflows for key activities and processes for its production operations, with the ultimate goal of improved asset performance. Real-time surveillance and monitoring of production operation processes have proven to be operationally and economically important for managing complex, high-cost reservoirs. However, predicting short-term production and production interruptions—for example, related to pump settings—has posed a tremendous challenge. While operators routinely forecast production for the next 60 to 90 days, sophisticated tools such as full-field numerical simulation models are of limited use in predicting short-term production of 30 days. Similarly, while nodal analysis can estimate current operating conditions, it cannot be used for prediction. Because of its simplicity, rapid training, and demonstrated results, a prediction technique using neural networks (NN) has emerged as a solution that can predict short-term well production behavior with acceptable accuracy. This paper presents a case study using NNs to predict liquid rate and water cut performance in a mature reservoir with more than 20% water cut. The NN was trained using available surface and down-hole, real-time production data, time-dependent data, and completion design data. The time-dependent data are included as time series configured to let users generate scenarios by changing well operations. This approach not only provides a base-case prediction but also simulates results after making adjustments in control variables, such as tubing head pressure (THP) and pump frequency. Changing THP and frequency lets users model production to predict and circumvent negative well pump events. This project was implemented in a mature carbonate oil reservoir under waterflood in the Middle East. Despite limited reservoir data, the results show that the NN is a powerful and rapid tool that predicts liquid rate and water cut with acceptable accuracy, helping engineers make prompt decisions to prevent and reduce downtime.