Process Control Algorithms Research Articles

The whole process control algorithm of inflatable dustproof shed construction in substation engineering based on graph neural network

Abstract During the construction of an inflatable dust-proof shed in substation engineering, it is one of the key issues in power system research to use scientific whole-process control algorithms to complete substation engineering decisions. At present, the control algorithm has some problems, such as low control accuracy and poor convergence speed. Therefore, the whole process control algorithm of inflatable dust shed construction in substation engineering based on a graph neural network is proposed. The control indexes of the whole process of inflatable dust-proof shed construction in substation engineering are divided into the highest level, the middle level, and the lowest level. The discriminant matrix is established to obtain the single-factor fuzzy evaluation matrix of the middle-level index. Thus, the control indexes are constructed. According to this index, the basic operation flow of the graph neural network is constructed, the objective function of the graph neural network is set, the control steps of the whole process of inflatable dust-proof shed construction in substation engineering is analyzed, and the algorithm research is realized. The experimental results show that the convergence steps are greatly reduced, the control accuracy is high, and the convergence speed is significantly improved after applying the algorithm proposed in this paper.

A novel sim2real reinforcement learning algorithm for process control

While reinforcement learning (RL) has potential in advanced process control and optimization, its direct interaction with real industrial processes can pose safety concerns. Model-based pre-training of RL may alleviate such risks. However, the intricate nature of industrial processes complicates the establishment of entirely accurate simulation models. Consequently, RL-based controllers relying on simulation models can easily suffer from model-plant mismatch. On the one hand, utilizing offline data for pre-training of RL can also mitigate safety risks. However, it requires well-represented historical datasets. This is demanding because industrial processes mostly run under a regulatory mode with basic controllers. To handle these issues, this paper proposes a novel sim2real reinforcement learning algorithm. First, a state adaptor (SA) is proposed to align simulated states with real states to mitigate the model-plant mismatch. Then, a fix-horizon return is designed to replace traditional infinite-step return to provide genuine labels for the critic network, enhancing learning efficiency and stability. Finally, applying proximal policy optimization (PPO), the SA-PPO method is introduced to implement the proposed sim2real algorithm. Experimental results show that SA-PPO improves performance in MSE by 1.96% and in R by 21.64% on average for roasting process simulation. This verifies the effectiveness of the proposed method.

Analysis and prediction of sputtering yield using combined hierarchical clustering analysis and artificial neural network algorithms

Sputtering is a crucial technology in fields such as electric propulsion, materials processing and semiconductors. Modeling of sputtering is significant for improving thruster design and designing material processing control algorithms. In this study we use the hierarchical clustering analysis algorithm to perform cluster analysis on 17 descriptors related to sputtering. These descriptors are divided into four fundamental groups, with representative descriptors being the mass of the incident ion, the formation energy of the incident ion, the mass of the target and the formation energy of the target. We further discuss the possible physical processes and significance involved in the classification process, including cascade collisions, energy transfer and other processes. Finally, based on the analysis of the above descriptors, several neural network models are constructed for the regression of sputtering threshold E th, maximum sputtering energy E max and maximum sputtering yield SY max. In the regression model based on 267 samples, the four descriptor attributes showed higher accuracy than the 17 descriptors (R 2 evaluation) in the same neural network structure, with the 5×5 neural network structure achieving the highest accuracy, having an R 2 of 0.92. Additionally, simple sputtering test data also demonstrated the generalization ability of the 5×5 neural network model, the error in maximum sputtering yield being less than 5%.

Development and Application of Control Concepts for Twin-Screw Wet Granulation in the ConsiGmaTM-25: Part 1 Granule Composition

Continuous manufacturing of pharmaceuticals offers several benefits, such as increased production efficiency, enhanced product quality control, and lower environmental footprint. To fully exploit these benefits, standard operation mode (production processes with no or minimal disturbance mitigation measures) should be supported by adopting novel quality-by-control (QbC) methodologies. The paper at hand is the first part of a study focused on developing QbC algorithms for optimizing twin-screw wet granulation in the industrial manufacturing line ConsiGmaTM-25, specifically addressing granule composition.This work relies on previously established process-analytical-technology (PAT) equipment for real-time monitoring of the granule composition, i.e., the active pharmaceutical ingredient (API) and liquid content in wet granules. The developed control platform integrates model-based process control algorithms that aim to keep the API- and liquid content at target values through real-time adjustments of the process parameters. Furthermore, the platform integrates a digital operator assistant, which aims to detect and classify granulation disturbances and provides messages and instructions for the plant operator.The present manuscript systematically outlines all design steps from the development phase in the simulation environment to the final real system application and validation. The control platform’s performance is demonstrated through selected test scenarios on the ConsiGmaTM-25 manufacturing line. The obtained results indicate improved disturbance robustness and an increase in intermediate/final product quality (compared to conventional operating modes): The process control algorithms successfully maintained the API- and liquid content at target values despite process disturbances. Furthermore, realistic disturbances (feeder, pump, and material) were accurately detected and classified by the digital assistant algorithm. The information was provided through a user interface, offering real-time support for plant personnel.

АЛГОРИТМИЧЕСКОЕ ОБЕСПЕЧЕНИЕ ИНФОРМАЦИОННО-ИЗМЕРИТЕЛЬНОЙ И УПРАВЛЯЮЩЕЙ СИСТЕМЫ ПРОЦЕССА ГИДРИРОВАНИЯ АЦЕТИЛЕНА

Relevance The ethylene market in Russia has been undergoing significant changes in recent years due to a combination of internal and external factors. An important trend in the Russian ethylene market has been the expansion of production capacities. Russia is actively investing in new enterprises and modernization of existing complexes within the framework of the policy of import substitution and economic diversification. In the era of information technology, it is important to develop not only technological aspects in the petrochemical industry, but also to pay attention to the development of information measuring and control systems of technological processes. Such solutions will improve the quality of products at the stages of process management. Aim of research Improving the quality of ethylene production, and in particular improving one of the stages of ethylene production – acetylene hydrogenation. Research methods The research used literature and scientific papers on the improvement of acetylene hydrogenation. Regression methods were used to develop a virtual analyzer model. During the development of the process control algorithm, classical approaches of the theory of automatic control were used. Artificial intelligence technologies were used to search for coefficients of the virtual analyzer model. Results A model of a virtual acetylene conversion analyzer using a genetic algorithm has been developed, a process control algorithm based on the resulting virtual analyzer has been developed.

Development of a fuzzy controller for adaptive control of the adsorption process

In this paper presents the development of a fuzzy controller for adaptive control of the parameters of the activated carbon adsorption process. The main objective of the developed control system is to increase the efficiency of the control of the technological process of adsorption by using intelligent technologies. It should be noted that the methodological basis for the design of fuzzy controllers is the concept of fuzzy logic. Also, for the development of fuzzy controllers for an adaptive control system, a model of the absorption process in the adsorber is constructed, taking into account multiple relationships between the parameters of the technological process. The conceptual model of the control object was developed taking into account the inherent internal relationships between the parameters of the technological regime and external disturbing factors. That is, fuzzy adaptive control allows you to adapt to changes in the technological parameters of the adsorption process by changing the control algorithm of the technological process. The expediency of using for adaptive control of technological parameters of the adsorption process by using fuzzy controllers is substantiated. The development of an adaptive control system based on fuzzy controllers was carried out using MatLab software, where a control model of the adsorption technological process was built, and the results were obtained and analyzed. Thus, fuzzy controllers evaluate critical situations and also regulate the parameters of the adsorption process based on logical rules. Keywords: adsorption; activated carbon; ethanol; fuzzy controller; system control; process parameters; adaptive control; fuzzy logic.

Optimizing the switching operation in monoclonal antibody production: Economic MPC and reinforcement learning

Monoclonal antibodies (mAbs), crucial in medicine and biopharmaceuticals, require optimized large-scale production to meet high clinical dosage demands. Most of the processes for industrial mAb production rely on fed-batch operations, resulting in significant downtime. Transitioning to fully continuous and integrated processes holds the potential to boost product yield, enhance quality, and reduce storage costs for intermediate products. The integrated continuous mAb production process can be divided into the upstream and downstream processes. One crucial aspect that ensures the continuity of the integrated process is the switching of the capture columns, which are typically chromatography columns operated in a fed-batch manner downstream. Due to the discrete nature of the switching operation, advanced process control algorithms such as economic MPC (EMPC) are computationally difficult to implement. This is because an integer nonlinear program (INLP) needs to be solved online at each sampling time.This paper introduces two computationally-efficient approaches for EMPC implementation, namely, a sigmoid function approximation approach and a rectified linear unit (ReLU) approximation approach. It also explores the application of deep reinforcement learning (DRL). These three methods are compared to the traditional switching approach which is based on a 1 % product breakthrough rule and which involves no optimization.

Short-Term Solar Insolation Forecasting in Isolated Hybrid Power Systems Using Neural Networks

Solar energy is an unlimited and sustainable energy source that holds great importance during the global shift towards environmentally friendly energy production. However, integrating solar power into electrical grids is challenging due to significant fluctuations in its generation. This research aims to develop a model for predicting solar radiation levels using a hybrid power system in the Gorno-Badakhshan Autonomous Oblast of Tajikistan. This study determined the optimal hyperparameters of a multilayer perceptron neural network to enhance the accuracy of solar radiation forecasting. These hyperparameters included the number of neurons, learning algorithm, learning rate, and activation functions. Since there are numerous combinations of hyperparameters, the neural network training process needed to be repeated multiple times. Therefore, a control algorithm of the learning process was proposed to identify stagnation or the emergence of erroneous correlations during model training. The results reveal that different seasons require different hyperparameter values, emphasizing the need for the meticulous tuning of machine learning models and the creation of multiple models for varying conditions. The absolute percentage error of the achieved mean for one-hour-ahead forecasting ranges from 0.6% to 1.7%, indicating a high accuracy compared to the current state-of-the-art practices in this field. The error for one-day-ahead forecasting is between 2.6% and 7.2%.

The use of vision system to determine lateral deviation from landing trajectory

PurposeThis study presents an image processing algorithm capable of calculating selected flight parameters requested by flight control systems to guide aircraft along the horizontal projection of the landing trajectory. The parameters identified based on the basics of the image of the Calvert light system appearing in the on-board video system are used by flight control algorithms that imitate the pilot’s schematics of control. Controls were generated using a fuzzy logic expert system. This study aims to analyse an alternative to classical solutions that can be applied to some specific cases.Design/methodology/approachThe paper uses theoretical discussions and breakdowns to create the basics for the development of structures for both image processing algorithms and control algorithms. An analytical discussion on the first stage was transformed into laboratory rig tests using a real autopilot unit. The results of this research were verified in a series of software-in-the-loop computer simulations.FindingsThe image processing method extracts the most crucial parameters defining the relative position of the aircraft to the runway, as well as the control algorithm that uses it.Practical implicationsIn flight control systems that do not use any dedicated ground or satellite infrastructure to land the aircraft.Originality/valueThis paper presents the original approach of the author to aircraft control in cases where visual signals are used to determine the flight trajectory of the aircraft.

Developing a unit testing framework for control system software implementation of a degassing vacuum pumping station

Research objective is to develop a test automation system for the process algorithm software implementation of the degassing vacuum pumping station control system operation at a coal mine. Relevance. The vacuum pumping station is an essential component of a mine degassing system. Degassing system operation depends on its correct functioning. A programmable logic controller with appropriate software is a key component of the vacuum pumping station automation system. Correct software implementation of the system operation process algorithm is of primary importance. To verify the compliance of the implementation with the requirements, manual testing is commonly used according to a pre-compiled test-plan, which is formed based on the requirements imposed on the system. Manual testing has a number of disadvantages. It is a long process where human errors are likely to occur, and the use of a test bench and/or an assembled control system sample is inevitable. Requirements for the algorithm can be changed in the course of program development, debugging, and initiation, which greatly complicates the testing process. Results. To solve the problem of process algorithm software implementation testing automation, the process control algorithm should be isolated from the software components responsible for the hardware. Access to the process control algorithm should be provided through standard interfaces (for example, OPC UA). Additional process equipment (valves, etc.) used in the automation system should be implemented in the form of emulators with an opportunity to assign their modes of operation. By creating an additional software layer of interaction with a process algorithm and emulator in a programming language with advanced test tools (for example, Python), it is possible to implement test scenarios from a test plan in this language. It will allow testing in automatic mode rather than manually. Conclusions. The proposed approach to process algorithm software implementation testing automation will greatly simplify software development for this type of automation and control systems

A Deep Learning Approach to Optimize Recombinant Protein Production in Escherichia coli Fermentations

Fermentation is a widely used process in the biotechnology industry, in which sugar-based substrates are transformed into a new product through chemical reactions carried out by microorganisms. Fermentation yields depend heavily on critical process parameter (CPP) values which need to be finely tuned throughout the process; this is usually performed by a biotech production expert relying on empirical rules and personal experience. Although developing a mathematical model to analytically describe how yields depend on CPP values is too challenging because the process involves living organisms, we demonstrate the benefits that can be reaped by using a black-box machine learning (ML) approach based on recurrent neural networks (RNN) and long short-term memory (LSTM) neural networks to predict real time OD600nm values from fermentation CPP time series. We tested both networks on an E. coli fermentation process (upstream) optimized to obtain inclusion bodies whose purification (downstream) in a later stage will yield a targeted neurotrophin recombinant protein. We achieved root mean squared error (RMSE) and relative error on final yield (REFY) performances which demonstrate that RNN and LSTM are indeed promising approaches for real-time, in-line process yield estimation, paving the way for machine learning-based fermentation process control algorithms.

A MECHANISM FOR IMPROVING PROFESSIONAL TRAINING OF FUTURE SPECIALISTS BASED ON AN INTEGRATIVE DEVELOPMENTAL APPROACH

The aim of the investigation is development of innovative strategyof quality control training of engineers and skilled workers (hereinafter – futurespecialists) in educational professional organizations on the principles of socialpartnership. Methods: Theoretic: theoretic and methodological analysis, polytheoreticsynthesis, modeling. Empirical: research and generalization of the system, processand competence – based approaches experience, experiment, observation,surveys, expert evaluation, SWOT-analysis as a method of strategic planningwhich is to identify the internal and external factors (socio-cultural) of the organizationsurrounding. Results: The strategy of the development of the process of quality controltraining in educational professional organizations and a predictive model of thesystem of quality control training for future engineers and workers have been createdon the analysis and synthesis of a quantitative specification of the quality, the obtained experience and success in control training of future specialists ineducational professional organizations in recent economic and educational conditions. Scientific novelty: There has been built a predicative model of quality controltraining of future specialists to meet modern standards and the principles of socialpartnership the control algorithm of the learning process, developed in accordancewith the standards (international) of quality ISO in the implementationof the quality control systems of the process approach (matrix-based responsibility,competence and remit of those responsible for the education process in theeducational organization, the «problem» terms and diagnostic tools for assessingthe quality of professional training of future specialists). The perspective directionsof innovation in the control of the quality of future professionals traininghave been determined the parameters of a comprehensive analysis of the state ofthe system to ensure the quality of their training on the basis of procedures forself-examination and questioning have been supported the improving systemtechnology to ensure the quality of vocational training have been developed theimplementation mechanism of the system of quality training control training offuture specialists in the professional organization have been specified. Practical Significance: The obtained results exploitation in educational practicecan improve the effectiveness of the innovative development of the qualitycontrol training system of future specialists in educational organizations.

Surface quality improvement of steel parts by combined laser-ultrasonic treatment: determination algorithm of technological parameters

To provide the quality of the surface layer and improve operational properties, a combined laser-ultrasonic surface hardening and finishing technology of steel products is proposed. This work is devoted to determining the range of rational parameters of laser heat treatment and ultrasonic impact treatment for enhancing the complex hardening process of AISI 1045 steel and AISI D2 steel. Laser surface transformation hardening was carried out with a constant temperature strategy using a fiber laser and scanning optics at a heating temperature of 1200–1,300 °C and a processing speed of 40–140 mm/min. Ultrasonic surface hardening and finishing were performed on technological equipment with an amplitude of ultrasonic vibration of 18 μm and a load of the ultrasonic tool of 50 N. The ultrasonic treatment duration varied from 60 to 180 s. The results showed that laser-ultrasonic treatment leads to an increase in the hardening intensity by more than 200 %, forming a hardening depth of 200–440 μm. Combined treatment leads to a significant increase in wear resistance due to the formation of ultrafine-grained martensitic microstructure with hardness (58–60 HRC5) in the near-surface layer. The combined laser-ultrasonic hardening process control algorithm for surface treatment of structural and tool steels is proposed, limiting the heating temperature, the duration of laser (ultrasonic) exposure, and the vibration amplitudes of the ultrasonic horn. Laser-ultrasonic treatment will allow the formation of a surface layer with a given set of properties, providing increased wear and corrosion resistance. The developed technology can be used for surface hardening and finishing of large-sized steel products in the mechanical engineering industry.

A novel imputation based predictive algorithm for reducing common cause variation from small and mixed datasets with missing values

Most process control algorithms need a predetermined target value as an input for a process variable so that the deviation is observed and minimized. In this paper, a novel machine learning algorithm is proposed that has an ability to not only suggest new target values for both categorical and continuous variables to minimize process output variation but also predict the extent to which the variation can be minimized.In foundry processes, an average rejection rate of 3%–5% within batches of castings produced is considered as acceptable and is considered as an effect of the common cause variation. As a result, the operating range for process input values is often not changed during the root cause analysis. The relevant available historical process data is normally limited with missing values and it combines both categorical and continuous variables (mixed dataset). However, technological advancements manufacturing processes provide opportunities to further refine process inputs in order to minimize undesired variation in process outputs.A new linear regression based algorithm is proposed to achieve lower prediction error in comparison to the commonly used linear factor analysis for mixed data (FAMD) method. This algorithm is further coupled with a novel missing data algorithm to predict the process response values corresponding to a given set of values for process inputs. This enabled the novel imputation based predictive algorithm to quantify the effect of a confirmation trial based on the proposed changes in the operating ranges of one or more process inputs. A set of values for optimal process inputs is generated from operating ranges discovered by a recently proposed quality correlation algorithm (QCA) using a Bootstrap sampling method. The odds ratio, which represents a ratio between the probability of occurrence of desired and undesired process output values, is used to quantify the effect of a confirmation trial.The limitations of the underlying PCA based linear model have been discussed and the future research areas have been identified.

Short-Term Prediction of the Wind Speed Based on a Learning Process Control Algorithm in Isolated Power Systems

Predicting the variability of wind energy resources at different time scales is extremely important for effective energy management. The need to obtain the most accurate forecast of wind speed due to its high degree of volatility is particularly acute since this can significantly improve the planning of wind energy production, reduce costs and improve the use of resources. In this study, a method for predicting the speed of wind flow in an isolated power system of the Gorno-Badakhshan Autonomous Oblast (GBAO), based on the use of a neural network with a learning process control algorithm, is proposed. Predicting is performed for four seasons of the year, based on hourly retrospective meteorological data of wind speed observations. The obtained wind speed average error forecasting ranged from 20–28% for a day ahead. The prediction results serve as a basis for optimizing the energy consumption of individual generating consumers to minimize their financial and technical costs. In addition, this study takes into account the possibility of exporting electricity to a neighboring country as an additional income line for the isolated GBAO power system during periods of excess energy from hydropower plants (March–September), which is a systematic vision of solving the problem of improving energy efficiency in the conditions of autonomous power supply.

Real-time operating system for autonomous drone control

Drones have become a transformative force across industries in the quickly changing world of autonomous technologies. Their applications are expanding, from precision agriculture to emergency response, thanks to developments in real-time operating systems. A key development is a "Real-time operating system for autonomous drone control" that reshapes the potential of unmanned aerial vehicles through the convergence of high-speed data processing, low-latency responsiveness, and complex control algorithms.Autonomous drones once imagined only in science fiction, are now a real possibility and are revolutionizing a wide range of industries. The integration of autonomous drones has created new opportunities for efficiency, safety, and scalability in a variety of fields, from agriculture and logistics to disaster relief and surveillance. The ability of a real-time operating system to process information with microsecond-level responsiveness is essential to the success of these initiatives because it enables drones to navigate dynamic environments, adapt to unexpected changes, and make quick decisions.

АНАЛИЗ И ОПТИМИЗАЦИЯ РАБОТЫ ИНФОРМАЦИОННЫХ СИСТЕМ С ПРИМЕНЕНИЕМ ВЕРОЯТНОСТНЫХ МЕТОДОВ

Probabilistic methods are one of the key tools for analyzing and optimizing the work of information systems in the modern world. They allow you to analyze and select optimal solutions for processing control algorithms and ensure an acceptable level of reliability of the system. When using probabilistic methods, it is important to take into account the influence of various factors on the results of the system. To do this, it is necessary to conduct a comprehensive analysis, which includes an assessment of the probability of malfunctions, an analysis of the consequences of these malfunctions, as well as an assessment of the probability of restoring the system after a malfunction. Comparison and selection of optimal solutions are also important steps when using probabilistic methods, this makes it possible to analyze various variants of control algorithms and choose the one that will provide the best results at the lowest cost.

Study of microwave and convective drying kinetics of pea pods (Pisum sativum L.): A new modeling approach using support vector regression methods optimized by dragonfly algorithm techniques

AbstractMachine learning and mathematical modeling techniques have been conducted to model the thin layer drying kinetics of pea pods, under either microwave or conventional air drying,. The effect of nine different microwave output powers (200–1000 W) and five different ventilated oven temperatures (40, 60, 80, 100, and 120°C) on drying kinetics was studied. The experimental drying rates were fitted to 11 literature semi‐empirical models to determine the kinetic parameters, finding the higher goodness‐of‐fit for the Midilli et al. model (average R2 = 0.999 for both drying methods). Moreover, the data were modeled using support vector machine (SVM) for regression which was optimized with dragonfly algorithm (DA) technique. The best result was obtained by Gaussian kernel with the optimal parameters σ, C, and ε values estimated as 0.2871, 78.45, and 0, respectively. The small root mean square error (RMSE = 0.0132) and the high determination coefficient (R2 = 0.9983) values proved how robust the SVM model is. DA‐SVM techniques can reliably be utilized to describe the thin layer drying kinetics of pea pods. It is useful to provide models that can assist in the development of food process control algorithms, and provided insights into complex processes, for the technological design of microwave or convective drying for pea pods preservation.Practical applicationsDrying of by‐products from pea processing industry was investigated as a critical step prior to their valorization. The drying of pea pods has never been investigated before which is the case of the present study whose objective was to study and model the microwave and convective drying kinetics of pea pods. Our research work reported that the Midilli et al. model was the most appropriate to describe the thin layer drying kinetics of pea pods for both drying methods, but mathematical drying models, although a useful tool, remains empirical in nature and product specific. Because of these limitations the new model DA‐SVM, developed using artificial intelligence techniques, can reliably be used to describe the nonlinear behavior of pea pods drying. These results could be further used for scale up calculation, which would further allow industrial scale preservation by microwave or convective drying of pea pods.

High-throughput screening of optimal process conditions using model predictive control.

Modern biotechnological laboratories are equipped with advanced parallel mini-bioreactor facilities that can perform sophisticated cultivation strategies (e.g., fed-batch or continuous) and generate significant amounts of measurement data. These systems require not only optimal experimental designs that find the best conditions in very large design spaces, but also algorithms that manage to operate a large number of different cultivations in parallel within a well-defined and tightly constrained operating regime. Existing advanced process control algorithms have to be tailored to tackle the specific issues of such facilities such as: a very complex biological system, constant changes in the metabolic activity and phenotypes, shifts of pH and/or temperature, and metabolic switches, to name a few. In this study we implement a model predictive control (MPC) framework to demonstrate: (1) the challenges in terms of mathematical model structure, state, and parameter estimation, and optimization under highly nonlinear and stiff dynamics in biological systems, (2) the adaptations required to enable the application of MPC in high throughput bioprocess development, and (3) the added value of MPC implementations when operating parallel mini-bioreactors aiming to maximize the biomass concentration while coping with hard constrains on the dissolved oxygen tension profile.

Machine Learning Technique for Recognition of Flotation Froth Images in a Nonstable Flotation Process

The paper is focused on the analysis of the relation between the stability of the flotation process and the efficiency of Machine Learning (ML) algorithms based on the flotation froth images. An ML process should enable researchers to construct Artificial Intelligence (AI) algorithms for flotation process control. The image of the flotation froth includes information characterizing the flotation process. The information can be extracted with the aid of the Image Recognition (IR) algorithms based on the ML. This enables construction of a flotation process control system in the mineral processing plant, which is based on the recognition of images of the flotation froth. The IR algorithms do not provide stable image recognition results and are not efficient in the situation where the parameters of the flotation process are highly unstable. The classification results were equal to 75.11% and 69.62% for a stable and unstable process, respectively. The experimental data collected at the Polish Pb/Zn mineral processing plant provided better insight to the relationships between the flotation process parameters and ML efficiency. These relationships were analyzed, and guidelines for the construction of the ML process for flotation process control have been formulated.