Large Machines Research Articles

Develop corpora and methods for cross-lingual text reuse detection for English Urdu language pair at lexical, syntactical, and phrasal levels

In recent years, Cross-Lingual Text Reuse Detection (CLTRD) has attracted the attention of the research community because large digital repositories and efficient Machine Translation systems are readily and freely available, which makes it easier to reuse text across the languages and very difficult to detect it. In the previous studies, the problem of CLTRD for the English-Urdu language pair has been explored at the sentence/passage and document level, and benchmark corpora and methods have been developed. However, there is a lack of benchmark corpora and methods for the CLTRD for the English-Urdu language pair at the lexical, syntactical, and phrasal levels. To fulfill this research gap, this study presents three large benchmark corpora for detecting the Cross-Lingual Text Reuse (CLTR) at three levels of rewrite (Wholly Derived (WD), Partially Derived (PD), and Non Derived (ND)). The CLEU-Lex, CLEU-Syn and CLEU-Phr corpora contain 66,485 (WD = 22,236, PD = 20,315 and ND = 23,934), 60,267 (WD = 20,007, PD = 16,979 and ND = 23,281) and 60,106 (WD = 23,862, PD = 15,878 and ND = 20,366) CLTR pairs respectively. As a secondary major contribution, we have applied the Cross-Lingual Word Embedding (CLWE), Cross-Lingual Semantic Tagger (CLST), and Cross-Lingual Sentence Transformer (CLSTR) based methods on our three proposed corpora for the CLTRD. Our extensive experimentation showed that for the binary classification task, the best results on the CLEU-Lex corpus were obtained using the cross-lingual sentence transformer ( $$F_{1}$$ = 0.80). For the CLEU-Syn and CLEU-Phr corpora, the best results were obtained using the cross-lingual sentence transformer and a combination of the CLWE, CLST and CLSTR methods ( $$F_{1}$$ = 0.92 on CLEU-Syn and $$F_{1}$$ = 0.94 on CLEU-Phr). For the ternary classification task, the best results on the CLEU-Lex corpus were obtained using the cross-lingual sentence transformer method ( $$F_{1}$$ = 0.69). For the CLEU-Syn corpus, the best results were obtained using a combination of the CLWE, CLST, and CLSTR methods ( $$F_{1}$$ = 0.82). For the CLEU-Phr corpus the best results were obtained using cross-lingual sentence transformer and combination of CLWE, CLST, and CLSTR methods ( $$F_{1}$$ = 0.78). To foster and promote research in Urdu (a low-resourced language) all the three proposed corpora are free and publicly available for research purposes.

Evaluation of a High-Temperature Pre-Heating System Design for a Large-Scale Additive Manufacturing System

Additive Manufacturing (AM) of titanium (Ti6Al4V) material using Selective Laser Melting (SLM) may generate significant residual stresses of a tensile nature, which can cause premature component failure. The Aeroswift platform is a large volume AM machine where a high-temperature substrate preheating system is used to mitigate high thermal gradients. The current machine platform is unable to achieve a target build-plate temperature of 600 °C. This study focuses on the analysis of the preheating system design to determine the cause of its inefficiency, and the experimental testing of key components such as the heater and insulation materials. A Finite Element Analysis (FEA) model shows the ceramic heater achieves a maximum temperature of 395 °C, while the substrates (build-plates) only attain 374 °C. Analysis showed that having several metal components in contact and inadequate insulation around the heater caused heat loss, resulting in the preheating system’s inefficiency. Additionally, experimental testing shows that the insulation material used was 44% efficient, and a simple insulated test setup was only able to obtain a maximum temperature of 548.8 °C on a 20 mm thick stainless steel 304 plate, which illustrated some of the challenges faced by the current pre-heating design. New design options have been developed and FEA analysis indicates that a reduction in heat loss through improved sub-component configurations can obtain 650 °C degrees above the substrate without changing the heating element power. The development and challenges associated with the large-scale preheating system for AM are discussed, giving an insight into improving its performance.

Explainable AI for Bearing Fault Prognosis Using Deep Learning Techniques.

Predicting bearing failures is a vital component of machine health monitoring since bearings are essential parts of rotary machines, particularly large motor machines. In addition, determining the degree of bearing degeneration will aid firms in scheduling maintenance. Maintenance engineers may be gradually supplanted by an automated detection technique in identifying motor issues as improvements in the extraction of useful information from vibration signals are made. State-of-the-art deep learning approaches, in particular, have made a considerable contribution to automatic defect identification. Under variable shaft speed, this research presents a novel approach for identifying bearing defects and their amount of degradation. In the proposed approach, vibration signals are represented by spectrograms, and deep learning methods are applied via pre-processing with the short-time Fourier transform (STFT). A convolutional neural network (CNN), VGG16, is then used to extract features and classify health status. After this, RUL prediction is carried out with the use of regression. Explainable AI using LIME was used to identify the part of the image used by the CNN algorithm to give the output. Our proposed method was able to achieve very high accuracy and robustness for bearing faults, according to numerous experiments.

A New Large Scale SVM for Classification of Imbalanced Evolving Streams

Classification from imbalanced evolving streams possesses a combined challenge of class imbalance and concept drift (CI-CD). However, the state of imbalance is dynamic, a kind of virtual concept drift. The imbalanced distributions and concept drift hinder the online learner’s performance as a combined or individual problem. A weighted hybrid online oversampling approach,”weighted online oversampling large scale support vector machine (WOOLASVM),” is proposed in this work to address this combined problem. The WOOLASVM is an SVM active learning approach with new boundary weighing strategies such as (i) dynamically oversampling the current boundary and (ii) dynamic weighing of the cost parameter of the SVM objective function. Thus at any time step, WOOLASVM maintains balanced class distributions so that the CI-CD problem does not hinder the online learner performance. Over extensive experiments on synthetic and real-world streams with the static and dynamic state of imbalance, the WOOLASVM exhibits better online classification performances than other state-of-the-art methods.

Improving the robustness of industrial Cyber–Physical Systems through machine learning-based performance anomaly identification

We propose a versatile and fully data-centric methodology towards anomaly detection and identification in modern industrial Cyber–Physical Systems (CPS). Our motivation behind this move is the ever-growing computerisation in these systems, in the form of complex distributed computing nodes, running complex distributed software. Industrial CPS also demonstrate heavy deployment of hardware sensors, as well as an increasing role for software. We observe the insufficiency and costliness of design-time measures in prevention of anomalies. As our main contribution, our methodology is taking advantage of this data-rich environment by means of Extra-Functional Behaviour (EFB) monitoring, analytics pipelines and Artificial Intelligence (AI). Specifically, we demonstrate the use of compartmentalisation of execution timelines into distinct units, i.e., execution phases. We introduce the generation of representations for these phases, i.e., behavioural signatures and behavioural passports, as our way of behavioural fingerprinting. Composed using regression modelling techniques, signatures as the representation of ongoing behaviour, are compared to passports, representing reference behaviour. The comparison is done by means of goodness-of-fit scores, creating quantifiable measures of deviation between different recorded behaviours. We have used both partially synthetic and real-world traces in our experiments, depending on the use-case. We have also followed both white box and black box approaches for our use-cases, with discussions on the pros and cons of each.The effectiveness of our data-centric methodology is demonstrated by two proofs-of-concept from the industry, to represent the two ends of the industrial CPS complexity spectrum, with one being a large semiconductor photolithography machine, while the other is an image analysis platform. Each use-case comes with its own characteristics and limitations, confirming the flexibility of our methodology and the relevance of its integral steps in the approach towards the initial analysis and data transformations. The results of anomaly classification show overall high accuracies, as high as 99% in certain set-ups. These results show the capability of our data-centric methodology, suiting the presented modern industrial CPS designs.

FEM-Based Power Transformer Model for Superconducting and Conventional Power Transformer Optimization

There were many promising superconducting materials discovered in the last decades that can significantly increase the efficiency of large power transformers. However, these large machines are generally custom-made and tailored to the given application. During the design process the most economical design should be selected from thousands of applicable solutions in a short design period. Due to the nonlinearity of the task, the cost-optimal transformer design, which has the smallest costs during the transformers’ planned lifetime, is usually not the design with the highest efficiency. Due to the topic’s importance, many simplified transformer models were published in the literature to resolve this problem. However, only a few papers considered this preliminary design optimization problem in the case of superconducting transformers and none of them made a comparison with a validated conventional transformer optimization model. This paper proposes a novel FEM-based two-winding transformer model, which can be used to calculate the main dimension of conventional and superconducting transformer designs. The models are stored in a unified JSON-file format, which can be easily integrated into an evolutionary or genetic algorithm-based optimization. The paper shows the used methods and their accuracy on conventional 10 MVA and superconducting 1.2 MVA transformer designs. Moreover, a simple cost optimization with the 10 MVA transformer was performed for two realistic economic scenarios. The results show that in some cases the cheaper, but less efficient, transformer can be the more economic.

How do career promotion policies affect research publications and open access?

We present a novel dataset which enables quantitative analysis of the relationship between institutional support for Open Science and research performance. We analysed promotion, review, and tenure policies (PRT) from institutions originating from seven countries and combined them with bibliographical data from the outputs generated by each institution. The data were normalised and evaluated against Open Science and Responsible Research and Innovation (RRI) indicators, to enable comparisons and easy machine readable access. The significance of this dataset lies in its potential to answer a range of questions that are key to the understanding of what motivates academics with regards to their research practices and publishing behaviors, using various indicators, including the MoRRI. To our knowledge, this collection constitutes one of the first efforts in delivering a large machine readable dataset enabling quantitative analysis on these aspects, as much work in this area has been carried out only through surveys and qualitative analysis.

Hybrid feature selection method for SVM classification and its application for fault diagnosis of wear and peeling in journal bearing with a little muddy water using long-term real data

Rotating machines are widely used as components of various industries around the world, and its normal operation of rotating machines is important. Thus, condition monitoring and fault diagnosis of rotating machines have considerable attention in recent years. Industrial statistics illustrate that 40% of total large machine breakdowns happened due to broken bearings, while for small machines, the analogous number reaches up to 90%. This study aimed at researching fault diagnosis of journal bearings using the support vector machine (SVM) method. The experimental systems of vertical and horizontal rotating shafts were developed. There was no adding any initial artificial failure in the bearing, and a little muddy water was used, and long-term vibration data in both systems were obtained in the normal operation of the machines until bearing damages occurred in the journal bearing (3-hour tests were conducted repeatedly and total 128 datasets for vertical shaft and 24 datasets for horizontal shaft were obtained). A feature selection method is focused, and a hybrid feature selection method by combining Fisher score (FS) and a sequential forward selection (SFS) method was proposed. Its accuracy and efficiency was proved experimentally with 97.14% and 100% for vertical shaft and 100% for horizontal shaft. Furthermore, as a result of the SVM model and hybrid feature selection method, the most important feature for journal bearing of horizontal rotor system was clarified as mean value of RMS, and only this feature can give good diagnosis result. It is useful suggestion in selecting the features for the fault diagnosis of horizontal rotating machines.

EDEM Simulation Study on the Performance of a Mechanized Ditching Device for Codonopsis Planting

Codonopsis pilosula is cultivated mainly in sandy soils, especially in the Dingxi area of the Gansu province, northwest China. They are mainly planted in hilly areas, where large machines cannot reach easily. Codonopsis pilosula transplanting has been adopted with a conventional flat planting way, film mulching and seedling outcrops. While its planting requires opening a shallow ditch, short operation cycle, considerable labor intensity and is in large demand, a simulation analysis was performed according to specific tillage resistance, helpful in the optimization design in later stages and the improvement of a domestic ditching plow’s performance. This paper studies the simulation performance of an adjustable trench plough and analyzes the orthogonal test and Design-Expert response surface of the data. EDEM simulation software was used to analyze the traction resistance of the furrow ditching mechanism in the ditching process. The results show that the traction resistance increased from 1751.31 N to 2197.31 N as the simulation working speed increased from 0.9 m/s to 1.5 m/s, 1.25 times higher than the former. It indicates that speed had significant effects on traction resistance. With the increase in working speed, the furrow traction resistance and specific consumed power were increased considerably. Using the stability coefficient of the ditching depth and consistency coefficient of ditching bottom width as test indices, speed and angle of furrowing as impact factors, Box–Behnken orthogonal experiment design method, and establishing test indices and test factors, the regression model between the analysis of the influence of the test index, comprehensive agronomic requirements and MATLAB factor optimization of the experiment, the optimized operation parameter combination was obtained: the forward speed was 0.9 m/s, the angle of ditching was 35° and the stability coefficient of the ditching depth and consistency coefficient of ditch bottom width were 97.57% and 98.03%, respectively. The data after the simulation comparison test can provide a design reference for the domestic small trench operation.

Design of Hydrostatic Chassis Drive System for Large Plant Protection Machine

In order to meet the working performance of large plant protection machine and according to the actual working requirements, this paper proposes a design of hydrostatic chassis drive system for a large plant protection machine. The purpose of this study is to realize the anti-slip rotation function of the plant protection machine and improve the driving stability through the combination of a hydraulic drive system and shunt valve. In this study, a closed circuit with a single pump and four motors is used, and a diverter valve is used to prevent the wheels from skidding during the driving of the plant protection machine. The parameters of the main hydraulic components of the hydraulic drive system were firstly calculated and selected; then the AMESim software was used to model and simulate the hydraulic drive system. Finally, a test platform with anti-skid function is designed and built, and the test results are as follows: when the diverter valve is closed, the plant protection machine drives at 3 km/h and 6 km/h respectively, and the skid rate is 3.79% and 6.17%; when the diverter valve is open, the plant protection machine drives at 3 km/h and 6 km/h respectively, and the skid rate is 1.33% and 2.70% respectively. The test results show that the hydraulic chassis of the plant protection machine designed in this study has good driving stability and can effectively reduce the slip rate of the plant protection machine in the process of driving in the field, which provides an effective theoretical support for the design of the driving system of the hydraulic chassis of the plant protection machine.

Oil drag loss in oil-flooded, twin-screw compressors

Rotary twin screw compressors are widely used because of their high efficiency and reliability. Their most common mode of operation is as oil-flooded machines when delivering air and gases at moderate pressures and flow rates. In order to achieve the best performance, it is essential to be able to predict the optimum amount of oil, required for the oil injection process, accurately. Analytical procedures for the design and performance estimation of twin screw compressors are well developed and widely available, but the determination of oil drag losses, in oil-flooded machines is only guesstimated. This paper describes a more detailed and accurate procedure for estimating oil drag loss, using a combined Couette-Poiseuille flow model and gives the results of studies on three sizes of machines operating over a range of pressure ratios and speeds. To this end, a parametric analysis has been developed based on a combined Couette-Poiseuille flow model and has been used to estimate the individual effects of pressure ratio, the various clearances and the oil viscosity on the total drag loss, for different sizes of the compressor. It can be seen from the results that at pressure ratios of up to 8.5, the drag loss due to the discharge axial clearance gap is nearly 2/3rd of the total, while nearly 1/3rd is due to the radial clearance. At normal operating speeds, the loss due to the interlobe clearance is insignificant, but as the pressure ratio increases, this rises more rapidly than that due to the axial and radial losses. The gain in the drag loss due to greater oil viscosity becomes more significant as the compressor size is increased. In larger machines, when clearance values are increased, the radial and axial elements of the drag loss are reduced more rapidly than that due to the interlobe loss.

Three-dimensional structure-guided evolution of a ribosome with tethered subunits.

RNA-based macromolecular machines, such as the ribosome, have functional parts reliant on structural interactions spanning sequence-distant regions. These features limit evolutionary exploration of mutant libraries and confound three-dimensional structure-guided design. To address these challenges, we describe Evolink (evolution and linkage), a method that enables high-throughput evolution of sequence-distant regions in large macromolecular machines, and library design guided by computational RNA modeling to enable exploration of structurally stable designs. Using Evolink, we evolved a tethered ribosome with a 58% increased activity in orthogonal protein translation and a 97% improvement in doubling times in SQ171 cells compared to a previously developed tethered ribosome, and reveal new permissible sequences in a pair of ribosomal helices with previously explored biological function. The Evolink approach may enable enhanced engineering of macromolecular machines for new and improved functions for synthetic biology.

Pharmaceutical plant machine availability prediction using Artificial Neural Network

During the 2nd phase of COVID-19 pandemic, pharmaceutical plant industry is facing lot of production pressure and machine availability plays vital role in maximizing the manufacturing pharmacy product output. In this paper, Artificial Neural Networks (ANNs) based information processing algorithm has been used to provide a solution to this problem and it has been found suitable to predict machines availability as a prediction function. The considered pharmaceutical plants are dealing with production of medicines related common symptoms in case of COVID-19 (fever, coughing, and breathing problems). The pharmaceutical plant data corresponding to different values of repair and failure rates of different subsystems is collected from plant and analyzed with the help of validated neural network value of availability. This configuration of ANNs approach developed in this research allowed simplifying computational complexities of conventional approaches to solve a large plant machines availability problem. The ANNs methodology in the paper permitted making no assumption, no explicit coding of the problem, no complete knowledge of system configuration, only raw input and clean data found to be sufficient to determine the value of machine availability function for different value of failure and repair rates considered in the paper. The results obtained in the paper are useful for the plant leadership, as the value of failure and repair rates of various subsystems can be fine-tuned at a require clear-cut level to achieve higher availability, and avoid considerably loss of production, loss of man power, and by-pass complete breakdown of concerned system.

Distributed three-level QR codes based on visual cryptography scheme

Owing to the large storage and fast machine recognition, QR codes have been widely utilized in many fields such as mobile payment, website navigation and user identity authentication. However, any QR code reader can access to the message contained in the QR code, the security becomes a major challenge to QR codes for privacy usage scenarios. Moreover, the management of QR codes for users are also inconvenient, since the human vision is hard to distinguish a QR code from the others. To solve the security and management problems, we propose the three-level QR codes for a group of participants. The first-level management information and the second-level public information are recognizable for the human vision and QR code reader device, respectively. The third-level privacy information is protected using visual cryptography scheme, and can be decoded using simple and non-cryptography computations. Furthermore, the shares can be stored or transferred in not only e-format but also print-format and photo-format, leading to the broad applicability. Experimental results and analysis demonstrate that the proposed scheme can encode three-level information into several distributed QR codes, and has more advantages compared with the previous schemes.

A Differential Strand-Slot Inductance Model for Improved Compensation of Circulating Currents in the Core Part of Large AC Machines

The Roebel bar conventional design of large AC machines uses the classical strand-slot inductance model (CSSIM). Suitable alternatives are missing as the CSSIM is favored for its inherent simplicity based on the ideally permeable iron core hypothesis. However, saturated armature slots can lead to high variations of the slot inductance, where the CSSIM cannot represent this precisely. An accurate prediction of the strand inductances is crucial when optimizing the transpositions of large Roebel bars to be competitive on efficiency and low measurement tolerances. This fact is crucial in under-roebeling, having less than a 360-degree transposition over the active part. In the end, the goal is to compensate the winding overhang parasitic field with the slot-parasitic field. This paper proposes a differential strand-slot inductance model (DSSIM) based on the concept of differential inductance (DI). It is compatible with a circuital lumped-element model (LEM) that considers the strand topology, geometrical dimensions, saturation level, and small- scale effects. Numerical simulations showcase the performance improvement of the DSSIM against known models. Finally, a 20- strand prototype of a slot model with actual Robel bar strands corresponding to a simplified bar cross-section in a large AC machines slot demonstrates the presented DSSIMs precision.

Reduction of Starting Current in Large Induction Motors

Large induction motors can have a high inrush and run-up current during starting, often up to ten times the rated current. In weak supplies, this could be a problem, causing system stability issues and the voltage to dip below acceptable levels. In islanded systems, the capacity could be pulled below its maximum. There are several different starting methods possible, but they are often only suitable for smaller machines. One method not investigated is the use of parallel capacitor compensation during the starting because large induction motors are very inductive during the starting sequence, so that supplying reactive power may be more effective than supplying energy. This paper first investigates several different induction motors with increasing size and assesses their compensation requirements in terms of the reactive power requirement during starting. It is shown that they do generally become more inductive as they increase in size. It is illustrated, using simulations, that using parallel capacitors during starting can significantly reduce the starting current. The concept is tested using a small experimental motor. Series inductors are added to the motor to make it more inductive and look more like a larger machine. The experimental results confirm that the method works. It is also found that the parallel capacitors require series filters to prevent harmonic current, and these can be realized using inductors in series with the compensation capacitors. Point-on switching using solid-state relays was tested, and these can reduce the transient switch-on current over the first voltage cycle.

Modeling and classifying the in-operando effects of wear and metal contaminations of lubricating oil on diesel engine: A machine learning approach

The lubricating oil analysis may be used to verify an assessment of the engine's health and operational conditions, as well as the need for oil changes. The wide sight of oil characteristics allows for the detection of elemental contaminations, additives, and wear for maintenance monitoring. The goal of this work is to investigate the in-operando effects of oil metal pollution on diesel engine conditions using large datasets and Support vector machine (SVM) and Radial basis function (RBF) models. To do this, 1948 datasets from engine lubricant spectral analysis were used. The greatest accuracy for forecasting engine conditions was obtained using cubic Polynomial (poly 3) and rbf kernel function in SVM; and RBF-NN classifiers used two training methods, trainbr, and trainlm. For engine condition predictions, RBF and SVM classifiers had an average accuracy of 99 percent and 97 percent, respectively. The capacity of models to generalize was tested by varying the dataset size (10% to 90%), with RBF demonstrating the greatest performance. Following that, RBF-NN used the confusion matrix approach to diagnose the engine's critical state with a 99.38 percent accuracy. The presence of metal contaminations such as Cr, Si, and Fe has a significant impact on the accuracy of engine identification in normal, critical, and numerous classes, according to the RBF-NN sensitivity study. Finally, RBF-NN has been chosen to define and monitor diesel engine conditions since it outperforms the SVM model.

Analysis and test of conductive shaft of large horizontal NC ECM machine tool

In order to solve the problems of serious heating and easy ablation of conductive shaft in electrochemical machining of special-shaped deep hole with large aspect ratio, the two schemes of single-sided copper bar and symmetrical copper bar of conductive shaft are studied by ANSYS software. By establishing the thermoelectric coupling model of the conductive shaft, the distribution law of temperature field and thermal deformation of conductive shaft under different working conditions is analyzed. The conductive shaft structure based on symmetrical copper bar scheme is determined. Monitoring points are set at the key parts of the conductive shaft, and the temperature field distribution is verified by using the data acquisition card. The temperature acquisition results are consistent with the simulation results. Through the machining test of special-shaped inner spiral deep hole parts with large aspect ratio, the results show that under the conditions of working current of 15 000 A, feed rate of 5 mm/min, and continuous machining for 14 h, the forming accuracy of the workpiece is ± 0.15 mm and the surface roughness is better than Ra0.8 μm. The performance of the conductive shaft is stable and meets the actual processing requirements. The heat dissipation performance of the conductive shaft is optimized by providing air flow on the upper surface and side at the same time.

Engineering a Reduction of the Levelized Cost of Energy of Distributed Wind Turbines via Rotor and Control Enhancements

Distributed wind turbines (DWTs), with rated capacity less than 100 kW, are penalized by high balance-of-system costs that lead to a high levelized cost of energy (LCOE). As a result, to contain the capital expenditure and thus LCOE, traditional designs have avoided active control systems that are standard in larger machines, e.g., active yaw and pitch control. In this paper, we present the key results of a trade-off study between larger rotor diameter, loads abatement, and simplicity in the pitch design for a typical mid-sized DWT with the overarching goal of increasing performance and decreasing LCOE. We focused on the upgrade potential for a generic downwind, passive-yaw, stall-controlled turbine model with high-speed shaft (HSS) brake and rated capacity of ∼ 60kW.A rotor and control redesign proved successful in increasing power capture through the deployment of a blade-root extender and an innovative, low-cost, independent pitch system for overspeed protection. The improvements remove the need for a mechanical brake and yield a significant decrease in LCOE. The loads in all components were kept under the target threshold by a combination of optimum blade extender length and rotor rotational velocity. The failsafe and redundant aerodynamic braking system is economical and can lend itself to retrofit applications to other turbine models.

Spatial control over ultrasonic cleaning of mining equipment using a phased array technology

Purpose. To develop methods for spatial control over ultrasonic cleaning by using ultrasonic phased array of radiators. To simulate the cleaning process using the developed methods to prove their effectiveness. Methodology. Application of the ultrasonic array as a basic radiator for ultrasonic cleaning enables redistribution of intensity in the bath by increasing it in the most contaminated zones of the cleaned object. Geometric and physical laws provide analytically defined parameters of the beam. Findings. The authors determine basic parameters for the ultrasonic beam through considering input and output data of the 3-D fuzzy interval controller. The focus distance is calculated by means of the arrival time of the threshold signal considering distances between the sensor and the array. The azimuth is directed into the bath center and dependent on its height only. The zenithal angle is calculated as a ratio of intensities of the current arrays and the nearest adjacent ones towards the greatest one. By default, the beam is directed to the bath center for the phased array with the greatest intensity. The simulation reveals that the applied approach enables a 41.5% increase in intensity in the contamination zone, this improving energy efficiency of cleaning and reducing time required for ultrasonic treatment. Originality. The authors suggest new methods for forming control over ultrasonic cleaning, which enables considering spatial distribution of this process by optimizing energy losses. Practical value. The new approach to spatial control over ultrasonic cleaning enables redirecting intensity in the bath to the most contaminated zones, this allowing an increase in energy efficiency of large mining machines of complicated configuration.