#3 2022

CONTENTS №3

Dozoertsev V.M., Baulin E.S., Anosov A.A., Boronin A.B.  Plant-wide optimization: a real opportunity or a distant future

The paper overviews the evolution of two prevalent classes of industrial automation systems: production optimization and advanced process control. It emphasizes the longstanding gap between those systems caused by the complexity of their synchronization because of basically different level of detail of plant description and the lack of the necessary algorithms and software for such synchronization. The plant-wide optimization approach is formulated, which closes the gap and enables production planning in closed loop by means of the selected optimal plan with the help of cascade advanced process control systems. Recent plant-wide optimization solutions from key industrial automation vendors are outlined.

Keywords: optimal production panning, scheduling, linear and successive linear programming, analytical design of controllers, advanced process control, model predictive control, cascaded MPC, plant-wide optimization.

Discussing a Topic…

Automation of oil, gas, and petrochemical industries

Baulin E.S., Khokhlov A.S., Rat’kovskaya M.A., Ruchin R.V.  Evaluation procedures for crude oil and its distillates in production management tasks

The paper discusses the existing methods and offers new ones for obtaining detailed information about physicochemical properties of crude oil based on incomplete data about its image. The methods comprise the approximation problems for missing data recovery, oil fractions overlap processing per their boiling points, selection of analogs from a crude database under incomplete information about crude properties with subsequent assessment of key properties of the crude and its distillates based on analog’s properties. The results of numerical experiments are presented, which confirm the applicability of the proposed approaches. Recommendations are made on the preparation of initial data about the properties of crude oil and its narrow cuts for subsequent calculations in optimal planning systems.

Keywords: crude properties, crude analogs selection, missing data recovery, properties evaluation, model accuracy, planning models.

Rychkov D.A., Torgashov A.Yu.  Synthesis of robust quality control system for gas plant products

The paper discusses joined parametric optimization of PID controllers and decoupling devices in gas plant control system. The control philosophy is based on soft sensors with uncertain process model. The resulting control system compensates for the mutual influence of control loops and ensures the system’s optimal speed.

Keywords: soft sensor, performance criteria, robust PID controller, control system, decoupling device, relative gain array, successive quadratic programming.

Antokhov M.V., Antokhov S.V., Abramov D.G., Kodolov A.V., Pevchenko B.V., Popov A.F.  Automated process control system for dimethylacetamide production. Design features

The paper discusses the results of development and trial operation of a process control system for dimethylacetamide production of Angara-Reactiv, LLC (Angarsk, Irkutsk Region, Russia). The system was developed by Altai Federal Scientific and Production Center JSC. Dimethylacetamide (DMAA) is a highly effective aprotic solvent increasingly important for many segments of chemical industry. DMAA production process is a new one, it has no analogs in Russia in its automation level. The control system is based on commercial automation and computing hardware and supports the unified classification of measurement and control tools, unification of I/O signals, power characteristics, uniform accuracy and reliability requirements. The system’s trial operation has confirmed its required process control functionality, safety, and reliability, as well as product quality consistency.

Keywords: dimethylacetamide, petrochemical, process control system, hazardous industries, unification.

Kudriashov V.S., Ivanov A.V., Alekseev M.V., Kozenko I.A.  The search of optimal distillation column’s feed tray in ethylbenzene production under changing feed flowrate and composition

End product loss reduction is an important task in ethylbenzene manufacture. To that end, a process model is offered for searching the optimal feed tray by means of optimal parameters under changing feed flowrate and composition. The steady-state multicomponent distillation model describes the ethylbenzene separation process. Gradient methods are used to minimize the complex objective function developed by the authors; the algorithm delivers the optimal values of reflux rate and reboiler temperature and also enables feed tray optimization. Dynamic selection of feed tray compensates for feed rate and composition fluctuations and improves the overall distillation performance.

Keywords: mathematical model, distillation, process optimization, criterion minimization, control parameters, optimal feed tray, key distillation products.

Buryak Yu.I., Mikhal’chenko I.N., Petrov A.V.  Automated control technology for integrated maintenance, repairs, and logistics processes

The paper offers new approaches to the organization of aircraft equipment maintenance in regional service centers. The approach is based on the search of rational balance between the mechanical availability of the fleet and its maintenance costs. A software suite is developed for assessing optimal facilities of aircraft maintenance centers and the parameters of their supply system.

Keywords: maintenance, repairs, aircraft equipment, integrated logistic support.

Sinelnikov A.O., Kuznetsov E.V., Golyaev Yu.D., Kolbas Yu.Yu., Solovyova T.I.  Computer simulation of the operation of cavity perimeter control system for Zeeman laser gyroscope

Computer simulation of the processes in the regulatory control system of operation for cavity perimeter of Zeeman laser gyroscope (LG) enables the precise prediction of its single-mode operation time with consistently high accuracy. The paper discusses the simulation technique, which allows replacing hours of in-process testing by short one-hour’s tests, and further use the acquired data for the prediction of LG behavior by means of an approximating function. It presents the algorithm implementing the technique proposed and the results of the experimental research, which confirm a good match of the approximating function with the observed dependence. The application of the method results in multiple reduction of LG production time at the process phase and the availability of data for LG sorting subject to their characteristics. Altogether, this improves the organization of Zeeman LG manufacture and aligns it with lean manufacturing requirements, such as the elimination of useless procedures and the reduction of production step duration.

Keywords: laser gyroscope, computer simulation, Zeeman frequency base plate.

Terteyan A.S., Brovko A.V.  Automated control technology for integrated maintenance, repairs, and logistics processes

The paper presents a method for iterative-order optimization of multicriteria problems and an iterative-order optimization algorithm, which supports both local and global criteria importance. A decision support system is developed for checking the applicability of the method and algorithm. An example demonstrating the effectiveness of the solutions proposed is cited.

Keywords: multicriteria optimization problems, criteria importance, global and local importance, decision support systems (DSS).

Suskin V.V., Filimonov V.V.  An algorithm for synthesizing multiprocessor information handling devices

The problem of data management in multiprocessor computing devices is discussed. The objective is improvement of multiprocessor system speed by means of optimal memory layout. The available solutions are studied, and a new data distribution algorithm is developed. The values generated by the algorithm were compared with the precise heuristic ones both in the accuracy and computing speed. The importance of the influence of memory layout and data availability for processing on the computing speed is substantiated.

Keywords: computing system’s speed, optimization methods, multiprocessor computing systems, parallel computing, synthesis, memory layout.

Kabanov A.A.  Investigation of machine learning techniques in the problem of electromyography signal pattern classification based on raw data

The paper presents the results of a machine learning techniques study for the classification of electromyography signal patterns. A dataset, which included labeled raw data of four hand gestures, was used. The dataset was pre-processed, analyzed, and split into training and test samples. The following machine learning algorithms were applied: naïve Bayesian classifier, k-NN method, decision trees, random forest, and support vector machine for two kernels (polynomial and radial basis function). The models were assessed by means of error matrix construction and the calculation of accuracy, confidence, and F1 score.

Keywords: electromyography machine learning, signal processing, support vector machine.