-
Date submitted2023-03-14
-
Date accepted2023-06-20
-
Date published2023-07-19
Forecasting planned electricity consumption for the united power system using machine learning
The paper presents the results of studies of the predictive models development based on retrospective data on planned electricity consumption in the region with a significant share of enterprises in the mineral resource complex. Since the energy intensity of the industry remains quite high, the task of rationalizing the consumption of electricity is relevant. One of the ways to improve control accuracy when planning energy costs is to forecast electrical loads. Despite the large number of scientific papers on the topic of electricity consumption forecasting, this problem remains relevant due to the changing requirements of the wholesale electricity and power market to the accuracy of forecasts. Therefore, the purpose of this study is to support management decisions in the process of planning the volume of electricity consumption. To realize this, it is necessary to create a predictive model and determine the prospective power consumption of the power system. For this purpose, the collection and analysis of initial data, their preprocessing, selection of features, creation of models, and their optimization were carried out. The created models are based on historical data on planned power consumption, power system performance (frequency), as well as meteorological data. The research methods were: ensemble methods of machine learning (random forest, gradient boosting algorithms, such as XGBoost and CatBoost) and a long short-term memory recurrent neural network model (LSTM). The models obtained as a result of the conducted studies allow creating short-term forecasts of power consumption with a fairly high precision (for a period from one day to a week). The use of models based on gradient boosting algorithms and neural network models made it possible to obtain a forecast with an error of less than 1 %, which makes it possible to recommend the models described in the paper for use in forecasting the planned electricity power consumption of united power systems.
-
Date submitted2022-08-01
-
Date accepted2022-11-17
-
Date published2023-02-27
Use of machine learning technology to model the distribution of lithotypes in the Permo-Carboniferous oil deposit of the Usinskoye field
- Authors:
- Denis V. Potekhin
- Sergei V. Galkin
Permo-Carboniferous oil deposit of the Usinskoye field is characterized by an extremely complex type of the void space with intense cross-sectional distribution of cavernous and fractured rock. In this study, for this production site, the process of 3D geological modeling has been implemented. At the first stage, it provided for automated identification of reservoir volumes by comparing the data of core and well logging surveys; at the second stage, identification of rock lithotypes according to Dunham classification is performed on the basis of comparison of thin sections examination and well logging data. A large array of factual information enables the use of machine learning technology on the basis of Levenberg – Marquardt neural network apparatus toward achievement of our research goals. The prediction algorithms of reservoir and rock lithotype identification using well logging methods obtained on the basis of the training samples are applied to the wells without core sampling. The implemented approach enabled complementing the 3D geological model with information about rock permeability and porosity, taking into account the structural features of the identified lithotypes. For the Permo-Carboniferous oil deposit of the Usinskoye field, the volumetric zoning of the distribution of different rock lithotypes has been established. Taking into account the lithotypes identified based on machine learning algorithms, density and openness of fractures were determined, and fracture permeability in the deposit volume was calculated. In general, during the implementation, the machine learning errors remained within 3-5 %, which suggests reliability of the obtained predictive solutions. The results of the research are incorporated in the existing 3D digital geological and process model of the deposit under study.
-
Date submitted2021-05-13
-
Date accepted2022-11-28
-
Date published2022-12-29
Reproduction of reservoir pressure by machine learning methods and study of its influence on the cracks formation process in hydraulic fracturing
Hydraulic fracturing is an effective way to stimulate oil production, which is currently widely used in various conditions, including complex carbonate reservoirs. In the conditions of the considered field, hydraulic fracturing leads to a significant differentiation of technological efficiency indicators, which makes it expedient to study in detail the crack formation patterns. For all affected wells, the assessment of the resulting fractures spatial orientation was performed using the developed indirect technique, the reliability of which was confirmed by geophysical methods. In the course of the analysis, it was found that in all cases the fracture is oriented in the direction of the development system element area, which is characterized by the maximum reservoir pressure. At the same time, reservoir pressure values for all wells were determined at one point in time (at the beginning of hydraulic fracturing) using machine learning methods. The reliability of the used machine learning methods is confirmed by high convergence with the actual (historical) reservoir pressures obtained during hydrodynamic studies of wells. The obtained conclusion about the influence of the formation pressure on the patterns of fracturing should be taken into account when planning hydraulic fracturing in the considered conditions.
-
Date submitted2022-05-13
-
Date accepted2022-09-24
-
Date published2022-11-03
Rapid detection of coal ash based on machine learning and X-ray fluorescence
Real-time testing of coal ash plays a vital role in the chemical, power generation, metallurgical, and coal separation sectors. The rapid online testing of coal ash using radiation measurement as the mainstream technology has problems such as strict coal sample requirements, poor radiation safety, low accuracy, and complicated equipment replacement. In this study, an intelligent detection technique based on feed-forward neural networks and improved particle swarm optimization (IPSO-FNN) is proposed to predict coal quality ash content in a fast, accurate, safe，and convenient manner. The data set was obtained by testing the elemental content of 198 coal samples with X-ray fluorescence (XRF). The types of input elements for machine learning (Si, Al, Fe, K, Ca, Mg, Ti, Zn, Na, P) were determined by combining the X-ray photoelectron spectroscopy (XPS) data with the change in the physical phase of each element in the coal samples during combustion. The mean squared error and coefficient of determination were chosen as the performance measures for the model. The results show that the IPSO algorithm is useful in adjusting the optimal number of nodes in the hidden layer. The IPSO-FNN model has strong prediction ability and good accuracy in coal ash prediction. The effect of the input element content of the IPSO-FNN model on the ash content was investigated, and it was found that the potassium content was the most significant factor affecting the ash content. This study is essential for real-time online, accurate, and fast prediction of coal ash.
-
Date submitted2021-09-22
-
Date accepted2022-03-24
-
Date published2022-04-29
Predicting dynamic formation pressure using artificial intelligence methods
Determining formation pressure in the well extraction zones is a key task in monitoring the development of hydrocarbon fields. Direct measurements of formation pressure require prolonged well shutdowns, resulting in underproduction and the possibility of technical problems with the subsequent start-up of wells. The impossibility of simultaneous shutdown of all wells of the pool makes it difficult to assess the real energy state of the deposit. This article presents research aimed at developing an indirect method for determining the formation pressure without shutting down the wells for investigation, which enables to determine its value at any time. As a mathematical basis, two artificial intelligence methods are used – multidimensional regression analysis and a neural network. The technique based on the construction of multiple regression equations shows sufficient performance, but high sensitivity to the input data. This technique enables to study the process of formation pressure establishment during different periods of deposit development. Its application is expedient in case of regular actual determinations of indicators used as input data. The technique based on the artificial neural network enables to reliably determine formation pressure even with a minimal set of input data and is implemented as a specially designed software product. The relevant task of continuing the research is to evaluate promising prognostic features of artificial intelligence methods for assessing the energy state of deposits in hydrocarbon extraction zones.
-
Date submitted2019-03-13
-
Date accepted2019-05-07
-
Date published2019-08-23
Modern Mathematical Forecast Methods of Maintenance and Support Conditions for Mining Tunnel
- Authors:
- S. A. Ignatyev
- A. E. Sudarikov
- A. Zh. Imashev
The research focuses on mathematical methods of mining pressure forecast to develop rational support patterns for mining tunnels and to ensure safety of mining operations. The purpose of research is to develop the methodology of applying advanced calculation methods and software solutions based on neural networks to reduce dispersion of factors influencing stability of mining tunnels, as well as to define rational parameters of mining tunnel support. The authors review the algorithm of geomechanical process examination, which is divided into several stages. First of all, it is proposed to use cluster analysis to examine location conditions of man-made outcrops, which allows to divide all the diversity of existing conditions for mining tunnel construction. Cluster analysis first allows to reduce the dispersion of factors that influence the stability of mining tunnels in various clusters, and then to determine rational parameters of tunnel support in each cluster. After the problem of cluster analysis is solved, it is proposed to use software programs that allow to study geomechanical processes in each cluster. At this stage, both standard methods (normative techniques, numerical modelling, analogies use, etc.) and the most advanced methods – neural networks – can be applied. Described algorithm of solving geomechanical problems, which utilizes advanced numerical methods and a software package based on neural networks, ensures an individual approach to estimation of mining pressure under varying conditions of man-made outcrop location in the rock mass.
-
Date submitted2018-07-02
-
Date accepted2018-09-04
-
Date published2018-12-21
Strategic Planning of Arctic Shelf Development Using Fractal Theory Tools
- Authors:
- V. S. Vasiltsov
- V. M. Vasiltsova
The paper justifies the necessity to utilize new methods of strategic planning in oil and gas field exploitation in the Arctic shelf during the implementation of high technology diversified model of development for oil and gas companies (OGC) based on principles and tools of fractal theory. It has been proved that despite its challenging conditions the Arctic represents not only resource potential of the country and a guarantee of national safety, but also a key driver of market self-identification and self-organization of OGCs. Identified and analyzed problems in institutional procurement of shelf development and utilized methods of strategic planning and project management, both on the levels of state and corporate governance, demonstrate that reductive approach of the fractal theory allows to take into account diversification of heterogeneous multicomponent project models, which can be reduced to a single management decision with inverse iterations of neural network modelling. Suggested approach is relevant for strategic planning not only on the stage of investment portfolio justification, but also for identification and assessment of project risks; ranking of projects according to the order of their implementation; back and - forth management (monitoring and supervision) and project completion. It has been detected that such basic properties of the fractal as self-similarity, recurrence, fragmentation and correlation between all fractal dimensions allow to systematize chaotically changing values of market parameters in the Arctic shelf development project, which provides an opportunity to forecast market development with minimal prediction errors.
-
Date submitted2009-10-26
-
Date accepted2009-12-27
-
Date published2010-09-22
Support of geodynamic safety in mining of the Khibini deposits
The paper deals with the problems of geodynamics in mining of the Khibini deposits. Description is given to the complex of organizational-technical arrangements for provision of geodynamic safety at the Apatit Co and to principal trends of its development.
-
Date submitted2009-08-01
-
Date accepted2009-10-08
-
Date published2010-02-01
Advance of the metallurgical limestone shaft kilning process control system
- Authors:
- N. I. Koteleva
Today at management system engineering by metallurgical processes used to special methods of the control theory such as optimal, neuro-fuzzy and adaptive methods. First of all, it is connected with increase of problems complexity maintained in control process. In article possibility of application of neural networks is considered at improvement of a control system by process of mine roasting of limestone, are described the neural network scheme controls and the basic stages of its construction.