In a context of mobile radio transmissions a statistical model of propagation channels is needed in order to evaluate the performance of the link. Many models involve an underlying Markov chain where the number of states and the associated probability law are a priori unknown. We propose a method to extract these parameters from measurements of instantaneous received power. It is applied to model the channel in the case of downlink transmissions in the L-band, from a non-geostationary satellite to a receiver on Earth in various terrestrial environments. The results are displayed for two categories of receivers, corresponding to vehicular and pedestrian users, and reveal difference between both models.
One technique useful in the testing and development of drilling automation system is to use synthetic data. A good drilling time series simulator can enrich a dataset for testing, and enable the inference of the states of drilling in real time. However, conventional simulators do not generate the "warts" of real data (noise, gaps, etc.). The proposed solution is a model that learns from real data, characterizes the different drilling responses and conditions the data with a deep neural network (DNN) approach, and generate realistic drilling time series dataset. To simulate a drilling-time series dataset, a DNN can model physical properties of the formation, rig, and sensors, and generates data with realistic curve patterns when it is trained on actual measurements, e.g. block position, hook load, standpipe pressure, and surface torque. The neural network has multiple convolutional, recurrent, and fully-connected layers. The model, trained with wellsite recorded data, captures the spatio-temporal distributions among data channels, and then uses a windowed input to predict the next data points, which are then fed back into the network to generate the simulated data sequence recursively. An actual sensor drilling-time series dataset containing various channels are input into the DNN. The networks contain eight convolutional layers with three max-pooling layers, three recurrent layers, and four fully-connected layers. The time window used in the input contains 512 samples for each channel, while the output is 1 sample for each channel. After training the network for 200 epochs, the network can successfully simulate time series data recursively. The simulated time series preserve the features of the original training data, while maintaining the data distribution of multiple channels. For example, the network shows a consistent "inslips" pattern in the hook load channel when the block position moves quickly from bottom to top. Currently the simulation is autonomous based on the training data, and does not take input as controls, which is our future steps. The proposed DNN model is a low-cost, robust model that simulate drilling-time series datasets containing complex spatio-temporal patterns. Our proposed algorithm is the first known simulator of drilling time series datasets that models the nontrivial physics laws and properties, including formation, rig, and sensors, and generates data containing realistic curve patterns with a deep neural network approach. The simulator greatly helps the inference component of automation systems with the enrichment of datasets that are available for testing.
Drilling automation design can benefit from simulated data. A drilling timeseries simulator can enrich the dataset for testing, and enable the inference of the drilling states in real time; however, conventional simulators based on simplified physical models usually lack the vivid patterns seen in real datasets. In this paper, we present a solution that uses a deep neural network-based Kalman filter. This system learns from real data, characterizes the different drilling responses, and conditions the data. The dynamics of a rig drilling system can be modeled as G(P, E, U) → Z, where P is the plant of the rig mechanical system, E the environment (i.e., the earth, formations, etc.), U the rig controls (e.g., the top drive RPM, the flow rate, etc.), and Z the measurement (e.g., the hookload). The dynamics can be described with a non-linear Kalman filter. The Kalman filter contains components of internal states, inputs, and measurements, and transfer functions for the internal states, inputs, and measurements. We model these transfer functions with deep neural networks, and solve these transfer functions by training with real data. The implemented deep neural network model contains components of convolutional neural network layers, long short-term memory layers, and deconvolutional layers. We trained the neural network with a real drilling datasets. Two experiments are presented. In the first one, the inputs are time-windowed multiple channels of RPM, block position, slip status, flow rate, and weight on bit, and the outputs are the response channels of torque, ROP, hookload, standpipe pressure, etc. The training accuracy achieved 99.5% after 50 epochs. In the second experiment, the network is trained to predict 16 rig states. In summary, the introduced deep neural network is a generic model for generating simulated time series and classification tasks.
Today, directional drilling is considered a mix between art and science only performed by experts in the field. In this paper, we present an autonomous directional drilling framework using an industry 4.0 platform that is built on intelligent planning and execution capabilities and is supported by surface and downhole automation technologies to achieve consistently performing directional drilling operations accessible for easy remote operations. Intelligent planning builds on standard planning activities that are needed for directional drilling applications and advances them with rich data pipelines that feed predictive and prescriptive machine-learning (ML) models; this enables more accurate BHA tendencies, operating parameters, and trajectory plans that ultimately reduce executional risk and uncertainty. Intelligent execution provides technologies that facilitate decision-making activities, whether they be from the wellsite or town, by leveraging the digital-drilling program that is generated from the intelligent planning activities. The program connects planning expectations, real-time execution data from the surface and downhole equipment, and generates insights from data analytics, physics-based simulations, and offset analysis to achieve consistent directional drilling performance that is transparent to all stakeholders. This new framework enables a self-steering BHA for directional drilling operations. The workflow involves an automated evaluation of the current bit position with respect to the initial plan, automated evaluation of the maximum dogleg capability of the BHA, and the capability to examine the health of the BHA tools and, if needed, an automated re-planning of an optimized working plan. This is accomplished on a system level with interdependencies on the different elements that make up the complete workflow. This new autonomous directional drilling framework will minimize operational risk and cost-per-foot drilled; maximize performance, procedural adherence, and establish consistent results across fields, rigs, and trajectories while enabling modern remote operations.
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.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.