Data Driven Virtual Flow Metering: Leveraging AI and Edge Computing for Continuous Multiphase Flow Measurement

This paper describes the use of data-driven virtual flow metering (VFM) for continuous multiphase flow measurement, which has been developed and tested in an oil field well pilot in Austria. 12 ESP (Electric Submersible Pump) wells have been modelled and fine-tuned within the pilot. Hardware-based test separators were used to conduct quality control evaluations on the predicted production rates and calibrate the well models as required. For the practical deployment of VFM systems, we have addressed the need for optimized learning and scalability of the artificial intelligence (AI) models by means of what we call soft-sensing and will explain how to successfully deploy this technology on wells with artificial lift. Notably, the application of this software-based, soft-sensing VFM in combination with hardware-based multiphase flow measurement bears the potential to significantly reduce the CAPEX cost for future metering infrastructure investments and even reduce the OPEX of existing metering hardware by extending the duration of metering cycles. This makes data-driven VFM an economical option even for low-producing wells. Details of the well pilot project conducted with OMV in Austria will be provided. The use of soft-sensing VFMs via cloud computing for continuous multiphase flow measurement is a step toward the closed-loop, fully autonomous operation of oil fields.

Position Soft-Sensing of Direct-Driven Hydraulic System Based on Back Propagation Neural Network

Automated operations are widely used in harsh environments, in which position information is essential. Although sensors can be equipped to obtain high-accuracy position information, they are quite expensive and unsuitable for harsh environment applications. Therefore, a position soft-sensing model based on a back propagation (BP) neural network is proposed for direct-driven hydraulics (DDH) to protect against harsh environmental conditions. The proposed model obtains a position by integrating velocity computed from the BP neural network, which trains the nonlinear relationship between multi-input (speed of the electric motor and pressures in two chambers of the cylinder) and single-output (the cylinder’s velocity). First, the model of a standalone crane with DDH was established and verified by experiment. Second, the data from batch simulation with the verified model was used for training and testing the BP neural network in the soft-sensing model. Finally, position estimation with a typical cycle was performed using the created position soft-sensing model. Compared with the experimental data, the maximum soft-sensing position error was about 7 mm, and the error rate was within ±2.5%. Furthermore, position estimations were carried out with the proposed soft-sensing model under differing working conditions and the errors were within 4 mm, but the periodically cumulative error was observed. Hence, a reference point is proposed to minimize the accumulative error, for example, a point at the middle of the cylinder. Therefore, the work can be applied to acquire position information to facilitate automated operation of machines equipped with DDH.