Two-stage Scheme Research Articles

Two-stage model fusion scheme based on knowledge distillation for stragglers in federated learning

Two-stage model fusion scheme based on knowledge distillation for stragglers in federated learning

Hydrothermal precipitation of copper from leaching solutions of metallurgical dusts

In this work, we aim to establish the main regularities of hydrothermal precipitation of copper from the previously unstudied sulfuric-nitric acid and sulfate solutions of atmospheric and autoclave processing of metallurgical dusts. Pyrite was used as a sulfidizer. The elemental composition of the products was determined by X-ray fluorescence, inductively coupled plasma atomic emission spectrometry, and atomic absorption analysis. It was found that copper precipitation by the proposed technology enables its recovery at a level of more than 95%. For sulfate solutions with a high arsenic concentration, a two-stage scheme of oxidation-hydrothermal treatment with the following parameters was proposed: temperature 180°С, duration 2 h, oxygen consumption 0.026 dm3/g pyrite (for the 1st stage), and 200°С and 2 h (for the 2nd stage). Extraction in the cake amounted to 95.4% of copper (in the form of Cu2S) and 91.4% of arsenic (in the form of FeAsO4), which allowed further separation of these metals by flotation. An autoclave treatment of a Cu–Zn–Fe–As–NO3 solution in the presence of pyrite at 180–220°С produced the activation energy values (kJ/mol) corresponding to the kinetic regime calculated by the two methods: 64.6 by the conventional method and 60.5 by the shrinking core model. The kinetic parameters of CuSO4–H2SO4–FeS2–H2O and CuNO3–H2SO4–FeS2–H2O systems were also determined. Flotation enrichment of copper autoclave precipitation cake was shown to result in a high degree of Cu and As separation, with the recovery amounting to Cu ˃ 95% and As ˂ 5%. Precious metals contained in pyrite are transferred to a flotation concentrate by 92.7% (Au) and 96.5% (Ag). The composition of the resulting flotation concentrate comprised (%): 12 Cu; 37 Fe; 38 S; 13 other elements. It is shown that, in order to obtain a product with the required content of copper, flotation concentrate should be separated into pyrite and copper concentrates followed by an additional flotation of primary copper concentrate in an alkaline medium in the presence of lime. Hence, our study has established the main regularities of hydrothermal precipitation of copper from the sulfuric-nitric acid and sulfate solutions of atmospheric and autoclave processing of metallurgical dusts.

A dual-standby power source configuration scheme for micro grid in Pelagic Island

Realizing clean energy supply is important for the sustainable development of pelagic island. The volatility and randomness features of renewable energy generation may reduce power supply reliability, thus backup units is necessary in micro-grid in pelagic island. This paper proposed a dual-standby configuration scheme based on Sequence Operation Theory to achieve reliable power standby. Firstly, characteristics of standby units, micro gas turbine and diesel generator, are analyzed. Next, the probability distribution function models of wind, photovoltaic and load demand in pelagic island are established, and the reliability index of power shortage probability is calculated by Sequence Operation Theory. Then, the two-stage dual-standby control scheme to make full use of the complementary standby units is introduced and object function to minimize equivalent annual cost on the premise of satisfying probabilistic reliability index is modeled. Hybrid integer particle swarm algorithm is used to solve the problem. Finally, the proposed model for dual-standby configuration is applied to a pelagic island. The simulation results show that the proposed method can meet the load demand in both daily and extreme weather and achieve superior economy.

Multi-objective optimization of a triple-eccentric butterfly valve considering structural safety and sealing performance

The structural safety and sealing performance of a triple-eccentric butterfly valve are crucial technical indicators that influence its reliability and service life. In this study, a new multi-objective optimization strategy is proposed to realize a lightweight design of valve trims, reduce the maximum equivalent stress, and reasonably distribute the sealing-specific pressure. A two-stage optimization scheme is designed by combining topology optimization (TO) and response surface methodology optimization (RSM). The topology optimization is employed to allocate the material distribution of the valve trims and provide the parameters for the response surface optimization, while the response surface methodology optimization conducts a further revision and optimization of the structural parameters of the valve trims. The results of the simulation experiments indicate that the maximum equivalent stress of the lightweight designed valve trims is reduced from 290.85 MPa to 99.88 MPa, and the maximum sealing-specific pressure of the sealing surface is reduced from 197.78 MPa to 77.83 MPa. Additionally, a novel approach is presented for assessing the sealing performance using the clearance of the fitting surface. This method can intuitively evaluate the state of metal sealing and guide the design of the fitting tolerance by analyzing the sensitivity of the dimensional deviation to the sealing-specific pressure. The findings demonstrate that the optimized valve exhibits good structural safety and sealing performance.

Gas extraction generation of gas mixtures of polar organic compounds at trace concentration levels

The patterns of generating gas mixtures of polar organic compounds at the level of their MAC (several µg/m³) were studied using the method of continuous gas extraction from aqueous solutions with a known concentration in combination with dilution by a stream of diluent gas. The necessary distribution coefficients of phenol, isomeric cresols, nitrobenzene, and alkanols С4–С6 between the aqueous and gas phases (nitrogen) were determined. A two-stage generation scheme is proposed, based on saturating the sorbent (activated carbon) with a flow of extraction gas containing the target components at a specified concentration, followed by desorption by a flow of pure extraction gas.

Двоетапний метод розв’язання задачі комівояжера на основі генетичного алгоритму

One of the main tasks in logistics is to find the most efficient route in the traveling salesman problem on a given transportation network, allowing for the servicing of the maximum number of customers while considering certain criteria. In the typical traveling salesman problem, the objective function is most often the route length or duration. However, such a formulation does not account for the subjectivity in evaluating the travel durations at different stages, which may be influenced by various objective and subjective factors. The problem of the traveling salesman with fuzzily defined travel durations between network nodes is considered. The approach to solving the problem is based on a genetic algorithm with enhanced mutation procedures and population diversity generation. Trapezoidal fuzzy numbers are used to formalize fuzzy values, presented in the generalized case based on the application of Gaussian distribution and corresponding characteristics. A two-stage scheme for solving the optimization problem of the traveling salesman on a given transportation network is proposed, combining the genetic algorithm and Ward's clustering method. In the first stage, a clustering procedure is conducted using Ward's method. In the second stage, based on the obtained set of clusters and the identified optimal inter-cluster distances, route duration optimization is performed within the clusters. To test the efficiency of the developed method, two types of models were generated: one with a limited number of available paths and one with a fully connected topology. Computational experiments were conducted to apply the developed method to solving the traveling salesman problem with travel durations between cities represented as fuzzy numbers. A comparison of numerical calculation results was performed. The obtained results demonstrated the effectiveness of preliminary clustering when using the genetic algorithm to find the best local optima under the same input parameters.

Development of a GIS-module for analog search using an analytical no-code platform

The article is devoted to the development of GIS-applications for searching analogs of oil and gas objects. Depending on the tasks to be solved and data sources used in oil and gas industry different methods of analogs search are applied. Program products developed on the basis of private methods can be used only to a limited extent, which raises the question about the expediency of resources spent on their development. For efficient creation of programs for analogues search we propose a two-stage scheme including the stage of development of a formalized search model based on analytical no-code platform and the stage of traditional final product development based on the formalized model. It is shown that analytical no-code platforms are an effective and accessible for non-programming specialists means of creating formalized models as a basis for software product development. The expediency of the development of applications for the search of analogs on the basis of geoinformation systems is substantiated. An example of implementing two-stage scheme in the development of module for searching analogs based on free geoinformation system QGIS with the use of Russian analytical no-code-platform Loginom is presented.

Reducing CO[formula omitted] emissions, energy consumption, and decarbonization costs in manganese production by integrating fuel-assisted solid oxide electrolysis cells in two-stage oxide reduction

Manganese is one of the most consumed metals due to its use in iron and steel making, which generates between 7%–9% of all CO2 emissions produced globally per annum. For the first time, this work explores the potential of a novel process concept to reduce CO2 emissions, energy consumption, and decarbonization costs in manganese production, which involves integrating fuel-assisted solid oxide electrolysis cells (FASOECs) in a two-stage scheme for the reduction of raw manganese ores. In this scheme, higher oxides that are present in raw manganese ores (MnO2, Mn2O3, Mn3O4) are pre-reduced using CO, yielding manganese monoxide (MnO) that is further reduced to Mn in a submerged arc furnace (SAF) using coke and electricity. In the proposed FASOEC integration concept, off-gas from the manganese ores after pre-reduction (a mixture of H2, CO, and CO2) is supplied to the FASOECs’ anode as fuel, in order to produce high-purity H2 in the cathode that is directed to the first stage of manganese reduction. H2 has enhanced reaction kinetics for reducing higher manganese oxides in comparison to CO; therefore, integrating FASOECs can improve manganese oxide conversion rates during pre-reduction, resulting in lower consumption rates of coke and energy in the SAF. The off-gas supplied to the FASOECs’ anode also lowers the amount of energy required for H2 production in the cathode (can also generate electricity, simultaneously) and produces anode exhaust gas containing only CO2 and steam. Since steam can be easily condensed from this stream, the integration of FASOECs also enables efficient decarbonization of the manganese production process, thus eliminating the need for a designated CO2 capture system. The techno-economic analysis performed herein demonstrates that directing the full supply of off-gas produced by the SAF to the FASOECs’ anode at 800 °C reduces the overall energy consumption of manganese production by up to 18% in comparison to conventional processes. This results in decarbonization cost reductions by as much as 3%–15% and a corresponding decarbonization price range of $4-32 per ton of manganese product for plant capacities of 50 and 200 kt, respectively.

Two-stage terrestrial laser scan planning framework for geometric measurement of civil infrastructures

Terrestrial laser scanning plays a crucial role in measuring geometry shapes of civil infrastructures. However, its practical application often encounters challenges such as insufficient model quality and acquisition efficiency. To address these issues, this study proposes a two-stage planning scheme that optimises both scan positioning and single-station scan parameters. Recognising the need to consider both structures and background objects, UAV scanning is employed to capture a three-dimensional point cloud model of the actual scenario. A visibility analysis method based on spherical projection is proposed to enhance computational efficiency of virtual scan simulation. Additionally, an improved particle swarm algorithm is utilised to optimise the scan positioning. During onsite deployment, the scanner’s pose is measured to align structural components with the scanner’s coordinate system, allowing for automated scanning based on occupied angular ranges. The effectiveness of this method is demonstrated through geometry measurements of a small outdoor structure and a suspension bridge.

A Machine Learning-Based Wrapper Method for Feature Selection

This paper presents a two-stage feature selection scheme using machine learning techniques. In the first stage a wrapper method is adopted to select various combinations of subsets of features from the original dataset. The performance of the model is evaluated by three classifiers: K-Nearest Neighbor (KNN), Support Vector Machines (SVM), and Random Forest (RF). In the second and final stage, a sequential backward feature selection Method is applied. The proposed method is demonstrated on eighteen datasets and the average classification accuracy of eighteen datasets achieved is 89.81%, 87.55%, and 89.82% using the KNN, SVM, and RF classifiers, respectively with a maximum reduced size of the subset being ten only. Comparing the proposed method to eight other feature selection methods, the former achieves better classification accuracy in terms of selecting the most useful but a smaller number of features.

COMBINED SUPERCHARGING SYSTEM OF DOMESTIC HIGH-PRESSURE TWO-STROKE SPECIAL PURPOSE DIESEL

Modern trends in engine construction are inextricably linked to the improvement of power units by modernizing existing models. The introduction of a forced to 1100 kW diesel engine of the DN type allows to increase the efficiency and bring the equipment in line with modern requirements in a short time and with lower costs. The key method of improvement is the modernization of the supercharging system, which increases power without increasing the dimensions and weight of the engine, as well as improves technical and economic characteristics. A combined supercharging system with a low-pressure turbocharger (LPT) and a high-pressure drive compressor (HPT) is proposed, which work together or separately depending on the operating mode. The use of a mathematical model of the workflow made it possible to carry out a computational and experimental study of the proposed system in the modes of maximum power and maximum torque, confirming its high efficiency. The use of modern technologies and materials, as well as the introduction of innovative design and modeling methods, allows to create more reliable and durable power units. Modernization of the supercharging system is one of the most important directions in the development of modern engine construction, which allows to achieve an increase in the level of specific power, improvement of fuel economy and environmental characteristics of engines by ensuring optimal conditions for their operation in various operating conditions and in a wide range of operating modes. The paper presents the research results of the proposed two-stage supercharging scheme of the DN type diesel engine, which makes it possible to achieve a power of 1100 kW. It is shown that this scheme has advantages over the traditional one, as it makes it possible to provide a higher boost pressure in a wider range of crankshaft RPM, as well as to reduce power consumption from the crankshaft to drive the compressor at modes close to the nominal ones. These factors collectively lead to a significant improvement in the characteristics of the diesel due to the introduction of the two-stage supercharging system.

Efficient and Effective One-Step Multiview Clustering.

Multiview clustering algorithms have attracted intensive attention and achieved superior performance in various fields recently. Despite the great success of multiview clustering methods in realistic applications, we observe that most of them are difficult to apply to large-scale datasets due to their cubic complexity. Moreover, they usually use a two-stage scheme to obtain the discrete clustering labels, which inevitably causes a suboptimal solution. In light of this, an efficient and effective one-step multiview clustering (E2OMVC) method is proposed to directly obtain clustering indicators with a small-time burden. Specifically, according to the anchor graphs, the smaller similarity graph of each view is constructed, from which the low-dimensional latent features are generated to form the latent partition representation. By introducing a label discretization mechanism, the binary indicator matrix can be directly obtained from the unified partition representation which is formed by fusing all latent partition representations from different views. In addition, by coupling the fusion of all latent information and the clustering task into a joint framework, the two processes can help each other and obtain a better clustering result. Extensive experimental results demonstrate that the proposed method can achieve comparable or better performance than the state-of-the-art methods. The demo code of this work is publicly available at https://github.com/WangJun2023/EEOMVC.

Melon: metagenomic long-read-based taxonomic identification and quantification using marker genes

Long-read sequencing holds great potential for characterizing complex microbial communities, yet taxonomic profiling tools designed specifically for long reads remain lacking. We introduce Melon, a novel marker-based taxonomic profiler that capitalizes on the unique attributes of long reads. Melon employs a two-stage classification scheme to reduce computational time and is equipped with an expectation-maximization-based post-correction module to handle ambiguous reads. Melon achieves superior performance compared to existing tools in both mock and simulated samples. Using wastewater metagenomic samples, we demonstrate the applicability of Melon by showing it provides reliable estimates of overall genome copies, and species-level taxonomic profiles.

A novel wideband control scheme for uncertain multi-mass systems and its application to drivetrain benches

In multi-mass systems, torsional vibration is a common and annoying phenomenon. Effective vibration suppression and robustness to wide-range parameter variations are essential for a sound motion system. However, most control methods focus on the primary resonance mode, and the high-order resonance modes are not actively treated in the control design, resulting in the control bandwidth not being high enough and limiting the control performance. This paper proposes a novel two-stage design scheme to realize a wideband control to improve control performance. First, a hybrid uncertainty model is tailored for multi-mass systems, which uses an equivalent and uncertain spring constant to describe the variation of the primary mode and a dynamic uncertainty to cover the other resonance modes. This hybrid model strikes a better balance between the model conservatism and the feasibility of a less conservative design. Then, the passivity of the parameter uncertainty is utilized to conduct a phase compensation on the nominal system. After the phase compensation, all uncertainties are converted into norm-bounded ones, and the robust performance design is carried out. This method is applied to vehicle drivetrain benches, and its superiority is validated through simulation comparisons and experiments on two typical types of drivetrain benches.

Remaining useful life prediction of Lithium-ion batteries using spatio-temporal multimodal attention networks

Lithium-ion batteries are widely used in various applications, including electric vehicles and renewable energy storage. The prediction of the remaining useful life (RUL) of batteries is crucial for ensuring reliable and efficient operation, as well as reducing maintenance costs. However, determining the life cycle of batteries in real-world scenarios is challenging, and existing methods have limitations in predicting the number of cycles iteratively. In addition, existing works often oversimplify the datasets, neglecting important features of the batteries such as temperature, internal resistance, and material type. To address these limitations, this paper proposes a two-stage RUL prediction scheme for Lithium-ion batteries using a spatio-temporal multimodal attention network (ST-MAN). The proposed ST-MAN is to capture the complex spatio-temporal dependencies in the battery data, including the features that are often neglected in existing works. Despite operating without prior knowledge of end-of-life (EOL) events, our method consistently achieves lower error rates, boasting mean absolute error (MAE) and mean square error (MSE) of 0.0275 and 0.0014, respectively, compared to existing convolutional neural networks (CNN) and long short-term memory (LSTM)-based methods. The proposed method has the potential to improve the reliability and efficiency of battery operations and is applicable in various industries.

D2D-assisted two-stage model aggregation scheme based on over-the-air computation

D2D-assisted two-stage model aggregation scheme based on over-the-air computation

An approach to robot welding path autonomous planning of the intersection weldment based on 3D visual perception

To improve the accuracy, an approach to robot welding path planning of the intersection weldment based on 3D visual perception is proposed. It adopts the two-stage location scheme of “global 3D reconstruction + local scanning correction”. First, this paper proposes an improved K-means clustering method based on the geometric features of workpiece point-cloud, to achieve accurate and fast point-cloud segmentation. Second, the intersection parameters of workpiece can be quickly calculated by the proposed parameter identification method. Third, this paper plans the scanning path of laser scanner according to the identified parameters, and puts forward an algorithm of weld path points recognition combined with the actual weldments state. Then, the pose of robot welding torch is obtained by using spline fitting algorithm and welding torch attitude estimation algorithm. Finally, the experiments show the path planning error is about 0.3 mm, which has increased a lot than the published methods.

Unsupervised Test-Time Adaptation Learning for Effective Hyperspectral Image Super-Resolution With Unknown Degeneration.

Fusing a low-resolution hyperspectral image (HSI) with a high-resolution (HR) multi-spectral image has provided an effective way for HSI super-resolution (SR). The key lies on inferring the posteriori of the latent (i.e., HR) HSI using an appropriate image prior and the likelihood determined by the degeneration between the latent HSI and the observed images. However, in scenarios with complex imaging environments and various imaging scenes, the prior of HSIs can be prohibitively complicated and the degeneration is often unknown, which causes it difficult to accurately infer the posteriori of each latent HSI. To tackle this problem, we present an unsupervised test-time adaptation learning (UTAL) framework for HSI SR under unknown degeneration. Instead of directly modeling the complicated image prior, it first implicitly learns a content-agnostic prior shared across different images through supervisedly pre-training a mutual-guiding fusion module on extensive synthetic data. Then, it adapts the shared prior to those private characteristics in the latent HSI for posteriori inference through unsupervisedly learning a self-guiding adaptation module and a degeneration estimation network on two observed images in the test phase. Such a two-stage learning scheme models the complicated image prior in a divide-and-conquer manner, which eases the modeling difficulty and improves the prior accuracy. Moreover, the unknown degeneration can be estimated properly. Both of these two advantages empower us to accurately infer the posteriori of the latent HSI, thereby increasing the generalization performance in real applications. Additionally, in order to further mitigate the over-fitting in coping with more challenging cases (e.g., degenerations in both spectral and spatial domains are unknown) and speed up, we propose to meta-train UTAL on extensive synthetic SR tasks and solve it using an alternative optimization strategy such that UTAL learns to produce good generalization performance in real challenging cases with a small number of gradient descent steps. To verify the efficacy of UTAL, we evaluate it on HSI SR tasks with different unknown degenerations as well as some other HSI restoration tasks (e.g., compressive sensing), and report strong results superior to that of existing competitors.

Underwater Fish Object Detection with Degraded Prior Knowledge

Understanding fish distribution, behavior, and abundance is crucial for marine ecological research, fishery management, and environmental monitoring. However, the distinctive features of the underwater environment, including low visibility, light attenuation, water turbidity, and strong currents, significantly impact the quality of data gathered by underwater imaging systems, posing considerable challenges in accurately detecting fish objects. To address this challenge, our study proposes an innovative fish detection network based on prior knowledge of image degradation. In our research process, we first delved into the intrinsic relationship between visual image quality restoration and detection outcomes, elucidating the obstacles the underwater environment poses to object detection. Subsequently, we constructed a dataset optimized for object detection using image quality evaluation metrics. Building upon this foundation, we designed a fish object detection network that integrates a prompt-based degradation feature learning module and a two-stage training scheme, effectively incorporating prior knowledge of image degradation. To validate the efficacy of our approach, we develop a multi-scene Underwater Fish image Dataset (UFD2022). The experimental results demonstrate significant improvements of 2.4% and 2.5%, respectively, in the mAP index compared to the baseline methods ResNet50 and ResNetXT101. This outcome robustly confirms the effectiveness and superiority of our process in addressing the challenge of fish object detection in underwater environments.

USE OF A TWO-STAGE COMBINED SCHEME FOR HEAT EXCHANGERS CONNECTION IN HEATING POINTS FOR HOT WATER SUPPLY OF INSULATED BUILDINGS

The article addresses the features of the functioning of centralised heat supply systems of residential micro-districts considering the improvement of thermal insulation of existing buildings. It also analyses the performance characteristics of hot water supply heaters installed on individual heat points of insulated buildings. The authors compared the heat transfer surface area of heaters and network water flow rate through the heating point for one-stage and two-stage combined schemes of connection of heat exchangers and heat networks. The analysis of calculation results showed that using a two-stage combined connection scheme reduces the consumption of network water through the individual heat supply unit compared to the one-stage scheme. However, the two-stage scheme causes the necessity to increase the heat transfer surface area compared to the one-stage connection scheme. We generalised the results by formulas for calculating water flow rate from distribution heat networks and heat transfer surface area. It is possible to use the obtained results to develop a strategy for reforming micro-district centralised heating systems. Using a two-stage combined connection scheme for hot water supply heat exchangers installed on individual heating points (IHP) of additionally insulated buildings reduces network water consumption through the individual heating point by 8–16% compared to a one-stage connection scheme. The device of a two-stage combined scheme in comparison with a one-stage scheme causes the increase of heat transfer surface area of heat exchangers of the hot water supply of individual heating points from one and a half to two times depending on the accepted design temperature of heating of cold water at the first stage of the heating unit. The results of calculations are approximated by equations, allowing, with acceptable accuracy, to compare the indicators of two-stage combined and one-stage schemes of connection of water heaters for individual heating points of insulated buildings. Keywords: centralised heat supply, individual heating point, water heaters, hot water supply.