Exploring acoustic and linguistic information embedded in spontaneous speech recordings has proven to be efficient for automatic Alzheimer's dementia detection. Acoustic features can be extracted directly from the audio recordings, however, linguistic features, in fully automatic systems, need to be extracted from transcripts generated by an automatic speech recognition (ASR) system. We explore two state-of-the-art ASR paradigms, Wav2vec2.0 (for transcription and feature extraction) and time delay neural networks (TDNN) on the ADReSSo dataset containing recordings of people describing the Cookie Theft (CT) picture. As no manual transcripts are provided, we train an ASR system using our in-house CT data. We further investigate the use of confidence scores and multiple ASR hypotheses to guide and augment the input for the BERT-based classification. In total, five models are proposed for exploring how to use the audio recordings only for acoustic and linguistic information extraction. The test results on best acoustic-only and best linguisticonly are 74.65% and 84.51% respectively (representing a 15% and 9% relative increase to published baseline results).
Horizontal well completions in low permeability formations with multistage fracturing have advanced greatly over the last decade. However, achieving an optimal balance between operational and cluster efficiency remains challenging. Several studies across unconventional basins have shown less than 70% productive perforation clusters in plug-and-perf (PnP) completions, highlighting a need to improve cluster efficiency without sacrificing operation efficiency. This paper presents a case study of Wolfcamp horizontal shale wells utilizing degradable diverter particulates to successfully improve cluster efficiency and well production.
Degradable diverter was implemented in five of eleven wellsacross three separate padsfor direct comparison. The diverter particulates were pre-tested in the laboratory with source water and formation cuttings samples to determine the dissolution rate and reservoir compatibility. Concentrations and deployment rate of the diverter "pill" were optimized from pressure responses during the job execution to achieve both the desired number of perforations covered and corresponding pressure increase as a leading indicator of improved cluster efficiency. Surface microseismic survey was acquired to further evaluate diverter effectiveness as compared to the offset non-diverter wells.
Initial engineering design/modeling targeted 50% to 65% of perforations for diverter coverage. All diverter frac stages pumped to the expected frac design with no screen outs. Post treatment analysis were run between each pad to optimize diverter integrity for further displacement and enhancement of diversion efficiency based on observed pressure build-up. Significant pressure increases pre-and post diverter were observed in 75% of stages. Surface microseismic results measured in the first pad indicated a 50% increase in the number of microseismic events in the well with diverter along with subtle shifts in both frac geometry and orientation. In 90% of stages a noticeable correlation was perceived in surface pressure responses to microseismic events. Wider event distribution post-diversion was also noted in stages with larger surface pressure responses. Production results show wells with diverter average 10% incremental cumulative barrels of oil equivalent (BOE) production at nine totwelve months as compared to offset non-diverter wells. There is a higher prevalence of elevated GOR with the diverter wells. Average incremental oil production during the six to twelve-month time frame is 9%. Incremental impact on individual pads range from neutral to +20% at the same time frame.
This paper shares the effective testing strategy to trial intra-stage diversion, engineering design work, application, analysis of diagnostic data and performance of degradable particulates in new unconventional horizontal wells. This paper also incorporates the lessons learned and best practices from field execution, real-time pressure responses, microseismic data, and production signpost results.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.