Rod-pumping is the most common artificial lift method used to move oil to the surface in low pressure resevoirs. A variety of mechanical problems can occur with this system. The adjustment of the pumping capacity to the reservoir production rate is another source of error compromising pumping efficiency. The need to identify these problems quickly and accurately is essential in attempts to minimize operating costs and maximize production. A dynamometer is attached in general to the polished rod of the pumping unit, and a plot of load vs. position, known as (downhole) dynamometer card (DC), is obtained for the purpose of rod load monitoring. Altough DC is a very important piece of information, other data may be required to support complex decision making about the actual rod-pumping condition. For this purpose, the engineer uses also information about the characteristics of the well, the type of oil being pumped, etc., besides taking into consideration the DC shape, maximum and minimum load values, etc. A new type of neuron is used to built neural nets having powerful numeric and symbolic processing capabilities, besides permiting knowledge to be encoded not only on the wiring of the net, but also on the selection of the types of neurons and synapsis composing the net. This new type of neural nets was used to develop SICAD, a hierarchical neural system whose purpose is the intelligent control of rodpumping. SICAD is composed by two famillies of neural nets specialized, respectively, in pattern recognition (PRN) and expert reasoning (ERN). Different modes of interactions 97 between ERN and PRN define different pumping control strategies.

No abstract

Summary This article clarifies some concepts of artificial intelligence (AI), discusses some of its applications, and demonstrates its potential applicationin the petroleum potential application in the petroleum industry. AI is dividedinto two levels: the psychological, where it attempts to represent knowledgeexplicitly, and the intuitive, where explication of knowledge is not importantand the emphasis is on brain architecture. Expert systems, which implementexplicit knowledge, are discussed in more detail. A brief discussion of use of AI in Brazil, particularly at Petrobras, is presented. Petrobras, is presented. Introduction AI has been, and will be, a controversial subject because it involves suchtopics as human intelligence, knowledge, brain architecture, and things that wehumans do not yet comprehend. Furthermore, the knowledge of this science, itstechniques, and its applications is dominated by a few people. Therefore, AIgenerally appears like a science fiction story to outsiders. AI has beendefined in several ways, but each definition aims at the same thing: thecomputer has to be more intelligent. This means that the computer has to beable to imitate the human being performing what we consider 'intelligent tasks. This attempt to duplicate human intelligence in a machine requires anenrichment of the actual ability of the computer to perform fast calculationswith the reasoning and learning mechanisms common in our natural intelligence. These mechanisms are very complex, and we know very little about them. Althoughthere have been great advances in AI techniques, the intended reproduction ofnatural intelligence in its abundance is very difficult, if not impossible. Because of this complexity, AI research is proceeding in two directions. One isthe pure science concerned with the pure science concerned with theunderstanding and the perfect reproduction of the true mechanisms of naturalintelligence. The other is the engineering side that seeks to devise datastructures and algorithms that, although they do not copy the naturalmechanism, are good enough to reproduce some aspects of natural intelligence. These data structures and algorithms are used to perform some "learningability" and to perform some "learning ability" and to representreasoning through a limited and specific knowledge base. The latter approachhas shown more practical and profitable results. With this in mind and withoutbeing pessimistic, I believe that the reproduction of pessimistic, I believethat the reproduction of human intelligence is far from being achieved. Onlytime will reveal how and when this could happen, but there are those among uswho visualize the unknown more quickly and will accelerate achievement of thegoal of AI. Duda gives a more academic definition of AI: "AI is the subfield ofcomputer science concerned with the use of computers in tasks that are normallyconsidered to require knowledge, perception, reasoning, learning, understandingand similar cognitive abilities." Thus, the goal of AI is a qualitativeexpansion of computer capabilities. If the parameters Duda uses to define AIare accepted as adequate (humans have them all at a high level), then thesefactors, when incorporated into computer programs, reveal intelligence. Amongthe most-cited applications of AI are computer vision, intelligent robots, natural language interpretation, game theory, automatic theorem proving, andexpert systems. Among the problems to be solved by AI are diagnosis, planning, design, prediction, interpretation, monitoring, debugging, prediction, interpretation, monitoring, debugging, repair, control, and instruction. AI History In 1956, 10 scientists in a conference at Dartmouth C. delineated what todayis called artificial intelligence (AI). Those scientists assumed thatintelligence was based primarily on reasoning techniques and that human beings, because of their intelligence, would easily reproduce it in a computer. Theypredicted that in 25 years we would be involved only in recreationalactivities, while computers would be doing all the hard work. AI has proved tobe more complex than originally expected. However, the AI efforts of thatperiod were not without merit. Many new things were learned that havecontributed to AI's success today. It was learned that knowledge is veryimportant to the intelligence. Perception, both visual and in language, isbased on knowledge that is found to be cumulative, voluminous, and hard tocharacterize. An example is "common sense, found to be simple reasoningbased on a great amount of experimental knowledge. During the first 15 years, AI had few successes. Automatic machine translation was attempted, but withlittle success. JPT P. 1306

Rod-pumping is the most common artificial lift method used to move oil to the surface in low pressure resevoirs. A variety of mechanical problems can occur with this system. The adjustment of the pumping capacity to the reservoir production rate is another source of error compromising pumping efficiency. The need to identify these problems quickly and accurately is essential in attempts to minimize operating costs and maximize production. A dynamometer is attached in general to the polished rod of the pumping unit, and a plot of load vs. position, known as (downhole) dynamometer card (DC), is obtained for the purpose of rod load monitoring. Altough DC is a very important piece of information, other data may be required to support complex decision making about the actual rod-pumping condition. For this purpose, the engineer uses also information about the characteristics of the well, the type of oil being pumped, etc., besides taking into consideration the DC shape, maximum and minimum load values, etc. A new type of neuron is used to built neural nets having powerful numeric and symbolic processing capabilities, besides permiting knowledge to be encoded not only on the wiring of the net, but also on the selection of the types of neurons and synapsis composing the net. This new type of neural nets was used to develop SICAD, a hierarchical neural system whose purpose is the intelligent control of rodpumping. SICAD is composed by two famillies of neural nets specialized, respectively, in pattern recognition (PRN) and expert reasoning (ERN). Different modes of interactions 97 between ERN and PRN define different pumping control strategies.

OBJECTIVE: To compare clinical outcomes after fresh embryo transfer on laser-assisted zona pellucida opening (LAO) versus thinning (LAT) according to maternal age in patients with repeated implantation failure (RIF).DESIGN: A retrospective study of 509 (n ¼ 458 patients) in vitro fertilization/intracytoplasmic sperm injection cycles was investigated from January 2013 to July 2017. MATERIALS AND METHODS: We compared whether LAT and LAO affect the clinical outcomes in young maternal age (YMA, <38 years) and old maternal age (OMA, R38 years) patient groups with R2 of RIF. The cycles with an oocyte donation, oocyte activation, genetic diagnosis, and that used surrogate mothers were excluded. Participants were divided into 4 groups according to maternal age and the two types of laser-assisted hatching (YMA: LAT, n ¼ 119 vs. LAO, n ¼ 179 and OMA: LAT, n ¼ 72 vs. LAO, n ¼ 139). LAO was opened using 3-4 laser shot in the zona pellucida. The laser thinning was performed by making 3-4 holes without reaching the inner membrane at a depth of 60%-80% of the zona pellucida thickness. Laser-assisted hatching was performed 2 hours before the embryo transfer.RESULTS: The characteristics of patients did not differ significantly among the groups (p > 0.05), with the exception of mixed factor infertility, which was more common in the LAT group than in the LAO group among patients <38 years of age (10.1% vs. 2.8%, p ¼ 0.008). We also observed similar rates of clinical pregnancy (27.7% vs. twin pregnancy (5.0% vs. 5.6%, p ¼ 0.498; 0.0% vs. 2.2, p ¼ 0.553) between LAT and LAO in the YMA or the OMA group. CONCLUSIONS: Clinical outcomes were similar between LAT and LAO in the YMA or the OMA group. However, the OMA group who underwent LAO tended to have a lower abortion rate. Further study is necessary to confirm these results in a larger population. P-274 Tuesday, October 9, 2018 6:30 AM PIEZO-ICSI AS ALTERNATIVE TOOL TO IMPROVE OOCYTE ACTIVATION IN IN VITRO MATURED BOVINE OOCYTES MODEL.OBJECTIVE: Evaluate the fertilization rates in a bovine oocyte activation failure model of Piezo-ICSI compared to conventional ICSI. DESIGN: We aimed to evaluate the participation of Piezo-pulse and sperm species in oocyte activation. In vitro matured bovine oocytes were randomly allocated to one of four groups: 1-conventional ICSI with bull sperm, 2-Piezo-ICSI with bull sperm, 3-conventional ICSI with human sperm, 4-Piezo-ICSI with human sperm.MATERIALS AND METHODS: Cumulus-oocyte complexes were obtained from bovine ovaries from slaughterhouse and matured in vitro. Cryopreserved sperm from bulls and humans with proven fertility were used. Conventional ICSI was performed with bevelled spiked micropipettes. Piezo-ICSI was performed with flat-tipped micropipettes. Fertilization rates were evaluated 18 h after ICSI by fixing and staining the presumptive zygotes with Hoechst 33342. Cleavage rates were evaluated 48 h after ICSI.RESULTS: Fertilization rate (2 pn) was higher with Piezo-ICSI (22.3%) than conventional ICSI (5.9%) using bull sperm (P<0.05). The sam...