Automated Classification of Well Test Responses in Naturally Fractured Reservoirs Using Unsupervised Machine Learning

Understanding the impact of fractures on fluid flow is fundamental for developing geoenergy reservoirs. Pressure transient analysis could play a key role for fracture characterization purposes if better links can be established between the pressure derivative responses (p′) and the fracture properties. However, pressure transient analysis is particularly challenging in the presence of fractures because they can manifest themselves in many different p′ curves. In this work, we aim to provide a proof-of-concept machine learning approach that allows us to effectively handle the diversity in fracture-related p′ curves by automatically classifying them and identifying the characteristic fracture patterns. We created a synthetic dataset from numerical simulation that comprised 2560 p′ curves that represent a wide range of fracture network properties. We developed an unsupervised machine learning approach that can distinguish the temporal variations in the p′ curves by combining dynamic time warping with k-medoids clustering. Our results suggest that the approach is effective at recognizing similar shapes in the p′ curves if the second pressure derivatives are used as the classification variable. Our analysis indicated that 12 clusters were appropriate to describe the full collection of p′ curves in this particular dataset. The classification exercise also allowed us to identify the key geological features that influence the p′ curves in this particular dataset, namely (1) the distance from the wellbore to the closest fracture(s), (2) the local/global fracture connectivity, and (3) the local/global fracture intensity. With additional training data to account for a broader range of fracture network properties, the proposed classification method could be expanded to other naturally fractured reservoirs and eventually serve as an interpretation framework for understanding how complex fracture network properties impact pressure transient behaviour.

Alkaline air: changing perspectives on nitrogen and air pollution in an ammonia-rich world

Ammonia and ammonium have received less attention than other forms of air pollution, with limited progress in controlling emissions at UK, European and global scales. By contrast, these compounds have been of significant past interest to science and society, the recollection of which can inform future strategies. Sal ammoniac (nūshādir, nao sha) is found to have been extremely valuable in long-distance trade (ca AD 600-1150) from Egypt and China, where 6-8 kg N could purchase a human life, while air pollution associated with nūshādir collection was attributed to this nitrogen form. Ammonia was one of the keys to alchemy-seen as an early experimental mesocosm to understand the world-and later became of interest as 'alkaline air' within the eighteenth century development of pneumatic chemistry. The same economic, chemical and environmental properties are found to make ammonia and ammonium of huge relevance today. Successful control of acidifying SO₂ and NO_x emissions leaves atmospheric NH₃ in excess in many areas, contributing to particulate matter (PM_2.5) formation, while leading to a new significance of alkaline air, with adverse impacts on natural ecosystems. Investigations of epiphytic lichens and bog ecosystems show how the alkalinity effect of NH₃ may explain its having three to five times the adverse effect of ammonium and nitrate, respectively. It is concluded that future air pollution policy should no longer neglect ammonia. Progress is likely to be mobilized by emphasizing the lost economic value of global N emissions ($200 billion yr^-1), as part of developing the circular economy for sustainable nitrogen management. This article is part of a discussion meeting issue 'Air quality, past present and future'.

Is there a gender gap in chemical sciences scholarly communication?

Characterisation of gender differences throughout peer-review publication process as revealed by thorough analysis of Royal Society of Chemistry submissions, publications and citation data.

The Royal Society of Chemistry is committed to investigating and addressing the barriers and biases which face women in the chemical sciences. The cornerstone of this is a thorough analysis of data regarding submissions, review and citations for Royal Society of Chemistry journals from January 2014 until July 2018, since the number and impact of publications and citations are an important factor when seeking research funding and for the progression of academic career. We have applied standard statistical techniques to multiple data sources to perform this analysis, and have investigated whether interactions between variables are significant in affecting various outcomes (author gender; reviewer gender; reviewer recommendations and submission outcome) in addition to considering variables individually. By considering several different data sources, we found that a baseline of approximately a third of chemistry researchers are female overall, although this differs considerably with Chemistry sub-discipline. Rather than one dominant bias effect, we observe complex interactions and a gradual trickle-down decrease in this female percentage through the publishing process and each of these female percentages is less than the last: authors of submissions; authors of RSC submissions which are not rejected without peer review; authors of accepted RSC publications; authors of cited articles. The success rate for female authors to progress through each of these publishing stages is lower than that for male authors. There is a decreasing female percentage when progressing through from first authors to corresponding authors to reviewers, reflecting the decreasing female percentage with seniority in Chemistry research observed in the “Diversity landscape of the chemical sciences” report. Highlights and actions from this analysis form the basis of an accompanying report to be released from the Royal Society of Chemistry.

Improving the learning of chemical-protein interactions from literature using transfer learning and specialized word embeddings

In this paper, we explore the application of artificial neural network ('deep learning') methods to the problem of detecting chemical-protein interactions in PubMed abstracts. We present here a system using multiple Long Short Term Memory layers to analyse candidate interactions, to determine whether there is a relation and which type. A particular feature of our system is the use of unlabelled data, both to pre-train word embeddings and also pre-train LSTM layers in the neural network. On the BioCreative VI CHEMPROT test corpus, our system achieves an F score of 61.51% (56.10% precision, 67.84% recall).

Clinical outcome of metastatic melanoma involving the CNS with or without (w/o) intracranial hemorrhage (IHC)

8577

Background: Melanoma has a high rate of ICH associated with CNS metastasis (mets). There is no data on the clinical outcome of these patients (pts) in the literature. Methods: A retrospective review of 52 pts with melanoma CNS mets diagnosed from 11/2001 to 09/2006 at our institution. Two pts were excluded due to a second malignancy. The clinical, radiological, and pathological characteristics of these pts were reviewed. Results: 39 pts were diagnosed after onset of neurological symptoms and 11 were diagnosed during staging. Of these, 14 (28%) were solitary, 23 (46%) were = 4, and 13 (26%) were multiple brain mets. Mean age at brain mets diagnosis was 55 years (range 24–88). Mean follow-up was 25 months (mos) (range 3–61) and 64% were male. ICH was found in 22 pts (44%) with 13 pts (26%) having frank hemorrhage on cranial CT and/or MRI scan. ICH was less common (9%) in pts with brain mets diagnosed on staging. 11 pts (8 with ICH, 3 w/o ICH) underwent CNS mets resection through craniotomy followed by whole brain radiation (WBRT) and/or stereotactic radiotherapy (STR). 7 pts received best supportive care. The remaining pts received WBRT and/or STR. The median survival (MS) for the entire group from CNS mets diagnosis was 7.6 mos. For pts with ICH, overall MS was 9.6 mos with MS of 21.2 mos for pts who underwent craniotomy vs 2.9 mos for pts w/o craniotomy. One pt refused therapy and later died from frank ICH. For pts w/o ICH, overall MS was 6.0 mos with MS of 21.3 mos for pts who underwent craniotomy vs 4.2 mos for pts w/o craniotomy. Pts who underwent craniotomy followed by WBRT and/or STR fared the best with MS of 21.2 mos. Of these, 2 pts with solitary brain mets (1 with ICH, 1 w/o ICH) were alive with no relapse at 58 and 53 mos respectively. The cause of death was disease progression in almost all pts and 85% of pts died from CNS disease progression. One pt died from post-operative complications of craniotomy. Conclusions: ICH by itself is not associated with a negative clinical outcome with appropriate clinical management. Aggressive surgical resection followed by radiotherapy in selected pts improves clinical outcome with prolonged survival possible in a minority of pts.

No significant financial relationships to disclose.

Management of acute complications associated with sickle cell disease

Sickle cell disease is the commonest haemoglobinopathy within the United Kingdom. Although the majority of patients will present to hospitals within major cities, this is not invariably the case. It is therefore important that all physicians on acute medical take are familiar with the acute management of sickle cell disease. This review encompasses the initial management which is subdivided into analgesia, investigations and supportive care. In addition the more severe complications of sickle cell, including the acute chest syndrome and stroke are covered. It should be remembered that close collaboration is required with the haematology department, particularly in those patients with respiratory distress or stroke, so that prompt arrangements can be made if exchange transfusion is required.

Premature hypergonadotropic ovarian failure: clinicopathological study of 19 cases

During the 5-year period 1977-1982, 57 patients below 35 years of age with secondary amenorrhea were assessed for hypergonadotropic (primary) ovarian failure. The histological findings within the ovaries as well as pertinent clinical and laboratory correlates are described. Nineteen had diagnostic ovarian biopsies performed. The importance of this technique is stressed. The ovaries of 14 patients showed absence of primordial follicles (true premature menopause); three others showed "resistant ovary syndrome" characterized by the presence of primordial follicles but little or no follicular development (including a case of galactosemia, in which the associated ovarian failure has been ascribed to follicular atresia). The remaining two revealed florid chronic perifollicular inflammatory reactions in the presence of both primordial and also developing follicles--one lymphoplasmacytic and the other granulomatous. The former has been previously suggested as evidence of an autoimmune process, but the latter has not hitherto been reported.