Year
Publisher
Journal
Institution
1
Institution Country
Publication Type
Field Of Study
Topics
Open Access
Language
Filter 1
Year
Publisher
Journal
Institution
1
Institution Country
Publication Type
Field Of Study
Topics
Open Access
Language
Filter 1
Export
Sort by: Relevance
Auto-Updated Virtual Multiphase Flow Metering System Using Neural Networks and Edge Computing

Abstract In oil and gas production, the multiphase flow rate is one of the most important measurements at the wellhead because it is essential for production allocation, flow assurance, and production surveillance. Multiphase flowmeter devices (MPFM) provide accurate measurements. However, this performance comes at high capital and operating costs that are not economically viable for all wells within aging or on mature fields. A real-time platform based on current technologies offers methods to deploy and automatically update the reduced-order models (ROM) to edge field devices. This approach is used specifically to provide a solution to estimate flow rates in real time at the field level, with and accuracy performance close to what the MPFM delivers, but at a reduced capital and operating costs. Edge field devices can provide the required computing power to run data-driven ROM models on the field side using neural network algorithms, trained to calculate well multiphase flow rate with acceptable accuracy. Periodic updates are performed on the data-driven models on the cloud, and the updated models are downloaded to the edge field device. In this way, the models based on neural networks, deployed and running at the edge field device level, are automatically adapted to changes of the well flow regime over the well's life cycle. An auto-trainable, machine-learning-based, multiphase estimated flow metering system running at the wellsite has been deployed and implemented. The system uses real-time and well test data, mathematical well models, neural-network-based data-driven models, which are automatically trained and updated on the cloud. The solution includes implementation in a real-time digital oil field automation system, which provides the technology platform required by this virtual flow metering system. The system offers a cost-effective way to obtain well flow rate measurement estimates with an acceptable accuracy. This virtual metering system can be used on hundreds or thousands of oil wells simultaneously using real-time data. The system provides oil and gas operators with good real-time flow metering estimates of their wells, accurate enough for production surveillance and allocation.

Read full abstract
On Partial Discharge/Corona Considerations for Low Voltage Switchgear and Controlgear

Electrical field and partial discharge (PD, corona) considerations usually are not in the foreground with low voltage switch- and control gear as per the product standards of IEC, UL and CSA. The design is mainly based on creepage and clearance dimensions, material categories described through CTI and RTI values. IEC product standards refer to the subject of insulation coordination as per the horizontal IEC 60664-family in their normative references. These standards, in principle, require PD measurements for rated voltages exceeding 500 VAC at room temperature. IEC 60664-2 gives some rules for consideration and electrical dimensioning of solid insulation. Miniaturization accompanied by increased power density and raised operational voltage as well as (non-sinusoidal) power supply generated by pulse width modulated (PWM) converters require to reconsider the relevance of electrical field stress in low voltage equipment. With variable frequency drive (VFD) fed motors reflected voltage waves were observed, causing partial discharge effects and insulation damages. Meanwhile motors and cables are available as VFD-proof versions. PD measurements on commercial manual motor controllers with sinusoidal voltages at both room temperature and a more severe elevated temperature related to the maximal operational temperature of the internal components (bimetal trip units) are presented. Conclusions are drawn and recommendations for future designs and testing are proposed.

Read full abstract