Spatial-Temporal Separation Based on the Dynamic Recurrent Wavelet Neural Network Modelling for ASP Flooding

Alkali-surfactant-polymer flooding (ASP) is one of most attractive chemical EOR techniques in view of incremental recovery upon waterflooding. Emulsification mechanism in ASP flooding is very important but not well understood. Effects of emulsification in ASP flooding is reviewed based on reported field tests as well as laboratory studies, especially on progress and understanding of ASP flooding in China, where the only commercial ASP flooding has been conducted. The main mechanism of ASP flooding can be summarized to the displacement efficiency improvement due to the ultra-low interfacial tension (IFT) between oil and water and the sweep efficiency increase due to mobility control technique by polymer viscosifying and emulsification effect. Emulsification is crucial in ASP flooding since all ASP flooding pilots were seen emulsification with different extent. Oil emulsifying and emulsion profile controlling was regarded as important in ASP flooding mechanisms. Laboratory tests showed that emulsification increase the oil recovery by 5%-6% when emulsified compared with not emulsified. Experience from Daqing oilfield in China reported that contribution of emulsifying ability of ASP system to oil displacement efficiency can be as high as 30%. Factors affecting emulsification included the properties of oil and water, type and concentration of chemical, water cut, external force applied and permeability. Effects of alkali to ASP system and oil emulsification was carefully studied. IFT was an important but not crucial factor to emulsification. Lower IFT at higher alkali concentration promoted easier emulsification, while too low IFT was detrimental to emulsion stability due to the competitive adsorption of in-situ surfactants and added surfactant in oil/water interface. Addition of polymer was beneficial to the stability of emulsion and the effect of associate polymer recent was obvious. ASP field tests in Daqing oilfield verified the emulsify ability of NaOH was almost the same as Na2CO3, which was quite different from laboratory studies. In all development stages of ASP flooding, emulsification was seen. Injection pattern and water cut affected emulsification. Separate injection of alkali and surfactant as one system, while polymer as the other make higher degree of O/W type emulsion. Emulsification in main slug and vice slug showed difference characteristics, which was attributed to the relative content of surfactants in different water cut stage.

Recurrent neural networks unlike feed-forward networks are able to process inputs with time context. The key role in this process is played by the dynamics of the network, which transforms input data to the recurrent layer states. Several authors have described and analyzed dynamics of small sized recurrent neural networks with two or three hidden units. In our work we introduce techniques that allow to visualize and analyze the dynamics of large recurrent neural networks with dozens units, reveal both stable and unstable points (attractors and saddle points), which are important to understand the principles of successful task processing. As a practical example of this approach, dynamics of the simple recurrent network trained by two different training algorithms on context-free language anbnwas studied.

Alkali-surfactant-polymer (ASP) flooding is very promising chemical enhanced oil recovery (EOR) technology which can make an incremental oil recovery factor (IORF) of 30% original oil in place (OOIP). How to choose alkali in ASP flooding remains a question for a long time. As the world’s only and largest ASP flooding application place, Daqing Oilfield has always adhered to the strategy of parallel development of strong alkali ASP flooding (SASP) and weak alkali ASP flooding (WASP), but SASP is in a dominant position, indicated by more investments and more project numbers. This leaves an impression that SASP is better than WASP. However, WASP is drawing more interest than SASP recently. Moreover, as the ASP flooding in Daqing went from field tests to commercial applications since 2014, how to comprehensively consider the benefit and cost of ASP flooding has become a new focus at low oil prices. This paper compares two typical large-scale field tests (B-1-D SASP and B-2-X WASP) completed in Daqing Oilfield and analyzes and discusses the causes of this difference. The injection viscosity and interfacial tension (IFT) for the two field test areas are substantially equivalent under the conditions of Daqing Oilfield, and WASP is better than SASP when reservoir geological conditions are considered. WASP exhibits the same IORF of 30% as SASP while having a much better economic performance. For the SASP field test, the injected strong alkali NaOH makes the test behave unlike a typical strong ASP flooding due to the presence of CO2 in the formation fluid, which well explains why IORF is much higher than all the other SASPs but scaling is less severe than others. This paper confirms that under Daqing Oilfield reservoir conditions, it is the alkali difference that caused the performance difference of these two tests, although some minor uncertainties exist. WASP is better than the SASP providing the same conditions . In addition, the detailed information of the two ASP field tests provided can give reference for the implementation of ASP flooding in other oilfields. After all, the study of ASP flooding enhanced oil recovery technology under low oil prices requires great foresight and determination.

To reduce the computational complexity and improve the performance of the recurrent wavelet neural network (RWNN), a novel modular recurrent neural network based on the pipelined architecture (PRWNN) with low computational complexity is presented in this paper. Its modified adaptive real-time recurrent learning (RTRL) algorithm is derived on the gradient descent approach. The PRWNN comprises a number of RWNN modules that are cascaded in a chained form and inherits the modular architectures of the pipelined recurrent neural network (PRNN) proposed by Haykin and Li. Since those modules of the PRWNN can be performed simultaneously in a pipelined parallelism fashion, it would result in a significant improvement in computational efficiency. And the performance of the PRWNN can be also further improved. Computer simulations have demonstrated that the PRWNN provides considerably better performance compared to the single RWNN model for nonlinear dynamic system identification.

This is a continuation of a recent study (Doruk RO, Zhang K. Fitting of dynamic recurrent neural network models to sensory stimulus-response data. J Biol Phys 2018; 44: 449-469), where a continuous time dynamical recurrent neural network is fitted to neural spiking data. In this research, we address the issues arising from the inclusion of sigmoidal gain function parameters to the estimation algorithm. The neural spiking data will be obtained from the same model as that of Doruk and Zhang, but we propose a different model for identification. This will also be a continuous time recurrent neural network, but with generic sigmoidal gains. The simulation framework and estimation algorithms are kept similar to that of Doruk and Zhang so that we can have a solid base to compare the results. We evaluate the estimation performance in two different ways. First, we compare the firing rate responses of the original and the estimated model. We find that responses of both models to the same stimuli are similar. Secondly, we evaluate variations of the standard deviations of the estimates against a number of samples and stimulus parameters. They show a similar pattern to that of Doruk and Zhang. We thus conclude that our model serves as a reasonable alternative provided that firing rate is the response of interest (to any stimulus).

In this paper, a dynamic recurrent wavelet neural network observer and tracking control strategy is presented for a class of uncertain, nonaffine systems. In proposed scheme a dynamic recurrent wavelet network is used to design a nonlinear observer .Adaptation laws are developed for the online tuning of wavelet parameters. Based on the estimated states, a state feedback control law is derived to achieve the desired tracking performance. The stability of closed loop system and ultimate upper boundedness all closed loop signals is proven in Lyapunov sense. Effectiveness of proposed scheme is demonstrated through numerical simulation.

Character level and word level embedding with bidirectional LSTM – Dynamic recurrent neural network for biomedical named entity recognition from literature

In ASP (Alkali-Surfactant-Polymer) flooding processes, surfactants help to enhance oil recovery by lowering the interfacial tension between the oil and water. However, due to the high cost of surfactants and the stability of the emulsion that varies with surfactant concentration, it is necessary to optimize the surfactant concentration in ASP flooding. In this study, we combined numerical simulation and physical experimental research to solve this problem. In order to screen for the optimal surfactant concentration in the main and vice slugs, CMG (Computer Measurement Group) numerical simulation software was used to change the surfactant concentration in the injected compound system and the oil recovery factor and the recovery percent of reserves were compared. The physical experiments were also carried out with different surfactant concentrations and the results verified the simulation results. It shows that the recovery factor increases with the surfactant concentration. The optimal surfactant concentration in the main and vice slug are 0.3% and 0.15%, respectively. As for improving the recovery factor, it is more efficient to increase the mass fraction of the surfactant in the vice slug than in the main slug. It demonstrates that the amount of surfactant in the main slug plays a more important role in displacing oil from the formation.

Tertiary recovery techniques such as alkali–surfactant–polymer flooding (henceforth ASP) and foam flooding have been widely administered in production operations of oil reservoirs which have led to the higher value of oil recovery factor, especially in heavy-oil reservoirs. The objective of this extensive study is to provide a comparative study between numerical simulation which was used by the commercial software implicit pressure and explicit composition (henceforth IMPEC) and experimental investigation for different injectivity scenarios of foaming agent and ASP solution after water and gas flooding on the oil recovery enhancement. To do this, simulation parameters for each scenario were considered in the simulator to investigate the results of each scenario on the oil recovery enhancement. Due to the presence of a gas phase in the realistic performances, the experiments were performed in the presence of three-phase fluids to match the results with the operational circumstances. Moreover, regarding the lack of information about the considerable influence of in situ gas which is in the trapped and continuous phase, the dissolved gas is taken into the evaluation. To validate the simulation analysis, an experimental investigation was performed for different scenarios and selected the optimum injectivity scenario. Consequently, regarding the presence of gas phase, ASP flooding after gas and water flooding had the maximum oil recovery factor and was considered as the optimum technique to produce initial oil in place. The reason for this phenomenon is elaborated to the more feasibility of gas phase which had led to the more mobilization of the oleic phase in the porous medium.

A dynamic recurrent neural network for multiple muscles electromyographic mapping to elevation angles of the lower limb in human locomotion

An adaptive identification method of simple dynamic recurr ent neural network (SRNN) for nonlinear dynamic systems is presented in this paper. This method based on the theory that by using the inner-states feed-back of dynamic network to describe the nonlinear kinetic characteris tics of system can reflect the dynamic ch aracteristics more directly, deduces the recursive prediction error (RPE) learning algorithm of SRNNe and improves the algorithm by studying topological structure on recursion layer without the weight values. The simulation results indicate that this kind of neural network can be used in real-time control, due to its less weight values, simpler learning algorithm, higher identification speed, and higher precision of model. It solves the problems of intricate in training algorithm and slow rate in convergence caused by the complicate topologi cal structure in usual dynami c recurrent neural network. Keywords: non-linear system, system identification, dynamic recurrent neural network, RPE algorithm

According to the formation conditions of alkali/surfactant/polymer (ASP) flooding in Daqing oilfield, the effects of the concentrations for polymer and alkali in ASP solution on oil recovery efficiency were studied by means of oil displacement experiments and numerical simulation. The effects on improving oil recovery by different ASP solution have been investigated from the aspects of interfacial characteristics and rheological behavior. Only the apparent viscosity of ASP solution reaches the critical viscosity, the ultralow interfacial tension can play an important role in enhanced oil recovery. Both displacement experiments and numerical simulation shows that ASP solution has a critical viscosity to certain heterogeneity.

A field-programmable gate array (FPGA)-based recurrent wavelet neural network (RWNN) control system is proposed to control the mover position of a linear ultrasonic motor (LUSM) in this study. First, the structure and operating principles of the LUSM are introduced. Since the dynamic characteristics and motor parameters of the LUSM are nonlinear and time-varying, an RWNN controller is designed to improve the control performance for the precision tracking of various reference trajectories. The network structure and its on-line learning algorithm using delta adaptation law of the RWNN are described in detail. Moreover, an FPGA chip is adopted to implement the developed control algorithm for possible low-cost and high-performance industrial applications. Finally, the effectiveness of the proposed control system is verified by some experimental results.

An optimization for injection of alkali/surfactant/polymer (ASP) flooding in oil recovery is considered in this paper. This optimal control problem (OCP) is formulated as a parameter identification system, where the objective function is revenue maximization and the governing equation is multiphase flow in porous media. We use the ASP concentrations and slug size as the control and give the pointwise constraint for the control. An improved particle swarm optimization (IPSO) which is a particle swarm optimization (PSO) with second-order oscillatory in velocity, is applied to solve the OCP. Finally, an example of the OCP for ASP flooding is exposed and the results show that the IPSO method is effective and feasible.

Effects of crude oil composition on the ASP flooding: A case from Saertu, Xingshugang and Lamadian Oilfield in Daqing