12 Months of Real-Time Digital Chemistry in 3-Phase Flow-LessonsLearned and Plans Forward

摘要

By miniaturizing and ruggedizing equipment used for quantum paramagnetic spectroscopy,it is nowpossible to take a real-time chemical snapshot of molecules flowing through the wellhead or other surfacefixtures.The digital time-series captures unique chemical properties of the fluid,such as the percentageof asphaltene in the oil,the oil-water ratio and gas-oil ratio.That data can be transmitted via industry-standard cloud protocols and be monitored from a global service center.12 months of real-time data hasbeen collected from operations around the world and the real-time monitoring has enabled prompt feedbackfor upgrades in both hardware and software.In a three-phase well configuration that had high rates of bothwater(over 90%)and gas(~1 MMSCf/day),this feedback drove some significant hardware modificationsin order to optimize the consistency of asphaltene data.The heart of the system is a microwave resonator that was designed to receive fluid at wellheadconditions with minimal reduction from wellhead pressure and temperature.The parameters of the resonatorwere optimized to maximize microwave intensity for typical oilfield fluids.A tailor-made set-up of fluidaccumulator and control-valves upstream of the resonator ensured that the resonator could obtain samplesthat were mostly oil.By combining the resonator with a solenoid that created a large magnetic field acrossthe oil,the resulting system provided spectroscopic data similar to that available in chemical laboratories butin a smaller package and one that tolerates some gas and conductive water in the oil.The combined quantumdata is now provided continuously to the operator via a cloud or other communication architecture ofoperator choosing.It is anticipated that the resulting Internet of Things(IoT)system will make possible theoptimization of chemical program and asphaltene remediation by incorporating system data with integratedflow assurance management.Qualification for offshore is ongoing with 5ksi pressure certification alreadyachieved.It was not obvious before installation,but once the 3-phase system was installed and the data transmittingin real-time,it became clear that software to automatically extract asphaltene information from spectral dataneeded to be able to cope with sudden and large changes in both asphaltene level and water-cut/gas-oil ratiowhich in turn required building an adaptive software model.Asphaltene percentage at one producing wellwas seen to vary from 0.3% to 3% in a single day.It was also discovered from the cloud-based monitoringthat daily temperature variation introduced a phase variation in the shape of the sensor response.Correct derivation of spectral voltages was achieved through the combination of machine learning,model-basedanalysis and additional diagnostic data such as the quality factor of the resonator and its resonance frequency.As a consequence,the AI-based software could extract the not only the asphaltene percentage but the oil-water cut in the resonator and its gas-oil ratio.For the first time,it is now possible to make a change in,say injected chemicals,look at the times-seriesdata for the corresponding change in asphaltene and then adjust the chemicals accordingly.Such frequencyof sampling(and volume of data)would be too much to handle with samples collected by hand.This devicelays the platform for a multiplicity of chemical sensors to be connected to the cloud in real-time and in turnsets the stage to take the hardware offshore and eventually to subsea.