Cable technology as a complex, multicomponent intelligent system providing high accuracy, reliability and safety

. Cable production consists of complex technological processes focused on the production of cable products with high operational reliability and fire safety. The article analyzes the main problematic issues that arise in the development of intelligent systems focused on information processing with advanced management functions, especially in the conditions of emerging technological uncertainties associated with the use of various production technologies. The classification of technological uncertainties of cable production is given and recommendations are developed for the selection of complex control systems integrated into existing technological processes to ensure modern requirements for the operational reliability of cable products. The main results on the development of a control system as a complex, multicomponent intelligent system that ensures high accuracy, reliability and safety of the production process of manufacturing cable and wire products are presented.


Introduction
Modern cable production is a combination of various technological processes for the manufacture of cable and wire products (CWP).Ideally, the production and technological process operating at a cable company is a closed technological cycle, which includes all technological operations starting from the production of a cable billet (copper or aluminum wire rod), followed by the manufacture of a finished cable product and ending with the recycling of cable production waste.Thus, it is a multidimensional production process, in which specialized cable making machines are engaged, ensuring the maintenance of a given technology and technological modes.At the same time, each individual technology is determined by various technological parameters (linear velocity, drawing, tension, pulling force, angular velocity, etc.) on the correctness of the choice, which determines the efficiency of process control and ultimately the quality of the finished cable product.Numerically, the value of these parameters is chosen in accordance with the regulatory, technical and design documentation for the production cycle under consideration.
The final result of the technological process is a finished cable product (cable, wire, cord), the structural elements of which are made of various materials and raw materials .
Each material and raw materials used (Table-1) have their own technical and technological parameters, presented numerically with an established permissible range of variation, according to the technology requirements and features of cable making machines, as well as the design of the cable product itself: material density; material mass; melting point; softening temperature; fluidity; viscosity, etc.Each cable making machine also has its own technical parameters, depending on the features of the technology and the operation of the electromechanical system (EMS): billet drawing speed, drawing, tension, pulling force, input and output diameters of the billet and the finished cable product, intermediate diameters and parameters of the cable semi-finished product, design parameters of the cable product, etc.
Based on the above, it can be concluded that the process of the CWP manufacturing is a complex production system consisting of many interrelated technological operations, which ensures high quality and operational reliability of the cable product produced.The goal set by the production service of the cable company is achieved by a well-coordinated and proven technology, as well as the coordinated work of all cable manufacturing equipment involved in the process (main and auxiliary), including ensuring strict convergence of transient (interoperable) technological parameters within the technological chain: "the beginning of the operation -the end of the operation" (Fig. 1) with an agreed range of regulation of production indicators (speed and time of manufacture, % of output, volume of production, design parameters of the cable product).At the same time, it should be noted that during the production process there is a high probability of various (possible) minor deviations of technical and theological parameters from the nominal (established within the norm) values caused by the technical condition of both cable making machines, their service life, and violation of the manufacturing technology of cable products.These include: random external disturbances and impacts; interference of various nature; noise induced by EMS of technological equipment and instrumentation; signal disturbances (emergency modes of equipment operation); disturbances formed during abnormal situations of the technological process (conditions of extreme operation / modes exceeding the maximum and minimum values).
Ensuring the reliability of the technological process is a multi-component task, the solution of which is aimed not only at achieving high quality of the finished cable product, but also minimizing cable production waste, unplanned downtime of equipment, which is inextricably linked with the reliability, durability, maintainability and preservation of cable making machines during their operation, which does not always go according to the planned scenario.An adequate assessment of the operation of cable manufacturing equipment involved in the technological route (chain) of manufacturing a certain nomenclature of cable or wire will reduce or even eliminate the costly mechanism for carrying out restoration and repair work during the execution of the production order and prevent the failure of its production deadlines.This is ensured by maintaining the necessary level of reliability not only of performing a specific technological operation, but also ensuring the troublefree operation of all coupled cable making machines involved in the manufacturing technology of the finished cable product.
The specifics of the work of cable production is based on the sequential manufacture of cable semifinished products by cable making machines involved in the manufacturing technology, which work within the technological chain, according to the manufacturing route, in strict accordance with the sequence of technological operations involved.It should be noted that no cable company has the ability to fulfill only one (limited) order on cable making machines sequentially assigned to the technology.As a rule, there is a distribution of the production order (production of a certain volume and grade of the CWP) for the development on a variety of machines with sequential transfer of the volume of the produced, unfinished semi-finished product to another technology (Fig. 1).For example, the technological operation -drawing, and consequently drawing machines, produce a drawn wire, which is the first stage (or the first structural element) in the production of any cable product, it is immediately supplied for insulation, stranding into a strand or for the manufacture of a conductive multiwire core, etc.Thus, the failure or unplanned shutdown of at least one element of the system (cable manufacturing equipment) disables the entire production chain under consideration or, even worse, makes it inoperable (Fig. 1), which leads to the disruption of the established production task.To analyze the reliability state of the research object (technological process), a method has been applied based on the study of an object having a logical scheme of the dependence of states and their interactions described by a mathematical model based on the characteristics of its reliability, including input and output parameters of raw materials and materials, cable billet, finished cable product, as well as internal parameters of technological equipment, EMS and disturbing effects.
Each technological unit -a cable making machine, has a complex automatic control system (ACS), with the main task -to ensure the smooth functioning of all its systems and subsystems, as well as working units, their EMS within the established technology modes, subject to the action and occurrence of various kinds of uncertainties.However, this control system (CS) is an integral part of a single technological machine, which we can consider as a complex, multidimensional object of research.The synchronization of the work of all cable manufacturing equipment assigned to a certain technology shall ensure the speed, reliability and consistency of the work of all production units, which in fact is an unsolvable production task, since centralized control of the manufacturing technology of the CWP is coordinated by the production service, where there is a human factor existing in production management, distribution of the order volume across production workshops, with a high probability of delay or the choice of an erroneous action in making a decision [2, 3].In this regard, the creation of a unified process control system (UPCS) for manufacture of the CWP will centrally ensure the execution of the production cycle, which includes all the main and auxiliary cable manufacturing equipment involved in the technological chain, including transitional (interoperation) processes, which confirms the relevance of the task we are solving through modern management theory [2, 4,5].At the same time, the UPCS shall be built as an intelligent control system (ICS) that ensures strict maintenance of the accuracy of the output design parameters of the finished CWP, all technological modes and deadlines, in various production conditions, taking into account the conditions of uncertainty, reliability and speed to various dynamic characteristics that arise when performing the technology [6,7].
The solution of this issue begins with the design of the UPCS and is complicated by external and internal factors that operate under conditions of various definite, uncertain and inaccurate knowledge of the dynamic properties of the technological system we are considering: -input and output parameters of the control system, both for a single cable making machine and for the entire technological operation; -monitoring of input and output data and disturbances on the functioning of the control system of complex technological objects, their uncertainty, non-linearity, a small percentage of predictability.
Taking into account the wide variety of cable technologies connected into a single production process aimed at the manufacture of high-quality CWP, the variety of their operating modes and operating conditions, the emerging uncertainties in technology during the operation of cable units, the presence of uncontrolled external disturbances caused by changes in the properties of dynamic systems (DS), the subject of our research are models of complex multidimensional nonlinear DS optimization and control, which will allow to achieve the necessary level of reliability and stability by applying adaptive laws of the DS operation control.Many world scientific and practical laboratories are working on the problem of building complex, multicomponent intelligent systems that ensure high accuracy, reliability and safety of technology implementation, and the results obtained have been widely implemented (Table -2).So currently, according to the analysis conducted by "Marketsand Markets", the global Smart Manufacturing Market will grow from $ 214.7 billion in 2020 to $384.8 billion in 2025, with an average annual increase of 12.4%.The conducted research "IT in Industry" (CNews Analytics) showed that the overall value of digitalization for the industrial sector of Russia will increase by 14 times by 2030 [1].

Solution methods
The technological process of the CWP production, in fact, is a complex, rapidly changing DS, consisting of the equipment involved in the technology and its control system.The development of an intelligent process control system, as a technical system, cannot be represented as a solution of differential equations, and therefore have the form of a system described by a traditional mathematical modeling apparatus, since it is not able to cover the change in all technological and technical parameters of the system under consideration: raw materials, materials, cable manufacturing equipment (main and auxiliary), technological equipment and technological tools, as well as cable billets, semi-finished (unfinished) and finished cable products.
It should be noted that traditional neural network (NN) models describe such systems as a "black box", where the values of input and output parameters, including disturbing effects, are the resulting values of the experimental study of the technology.All this negatively affects the results of modeling, because it is a significant limitation in achieving the required level of complexity of the simulated multicomponent system.In the case we are considering, it is advisable to apply a hybrid neural network mathematical model, which includes differential equations that jointly describe the dynamics of the technology and the operation of all the cable manufacturing equipment involved.
Such an integrated approach will allow combining all the positive aspects of traditional models and compensating for the existing negative factors: ensuring adaptability, increasing the threshold of complexity of the task under consideration, as well as identifying the characteristics of a dynamically changing, multicomponent system.
The primary stage of the development of the automated control system for the CWP manufacture is based on working out the levels of interaction of all models of the system with the external environment, as well as disturbances of various nature and interference.However, in addition to possible disturbances and interference that act on the ICS, there are control actions formed by analyzing the results of system (regulatory) data and knowledge of the operation of cable making machines, technological modes and design parameters of cable products embedded in the regulatory, technical and design documentation accompanying a single technology (Fig. 2).All CAs are "non-mathematical" data, but complex knowledge obtained in the process of manufacturing a cable product of a certain brand.When entering them into the ICS database (logical output) for subsequent associative memorization, the signals are initially encoded and decoded.At the same time, an algorithm of numerical, multi-valued (fuzzy) or symbolic form is being developed for the subsequent compilation of a mathematical model [2, 4, 5, 8].
The task of the mathematical model is to process the available information by implementing the main results of the corresponding numerical calculations.The working variants of the incoming information to the ICS, which are subject to system analysis, are divided into elements: accumulated knowledge; generalized and analyzed data (Fig. 3).The developed control algorithm is focused on estimating dynamic parameters based on the formed production and technological basic values [8, 9,10,11,12,13,14,15].The developed functional structure of the cable production control system is based on the multilayer structure of an artificial neural network (Fig. 4).
The initial stage of building a neural network model (NNM) of the technological process of manufacturing cable products is based on the development of a mathematical model that includes data from the analysis of the physical parameters of both technology and technological equipment, and is a system of differential equations that is solved by effective analytical and numerical methods.
It is a very difficult task to describe the generalized structure of the CWP production technology, because it consists of many working mechanisms and shall include specific parameters of the working systems of cable manufacturing equipment, including the initial values of variables, and difficult unobservable quantities that are derivatives of the joint operation of equipment, tooling and processed cable billet, including raw materials and materials involved in the technology.
One of the ways to solve the problem of creating a dynamic model of a neural network is its joint use with "classical" differential equations -objective functions (Fig. 5).
The construction of the NN reflecting the production process of CWP manufacturing includes neurons and their interconnection, the work of which, including the work of interneuronal connections, is set by means of a specially developed program.The functioning of the NN begins with a formal neuron, in which the initial moment of time (current time) along the dendrites receives a processed input signal coming from the structural units of the lower level and which are output signals for them.Next, the input multiplies the incoming signal to the neuron by the corresponding weighting factor and sums it with other signals, followed by multiplying it by the weighting factor of the available inputs, where the output signal is formed, which can be represented as follows: ∑   , (1) where, n -the number of neuron inputs, x i -the value of the «I» neuron input; w i -the weight of the «I» signal.
The mathematical model of a complex, multicomponent intelligent system of the technological process of cable product production has the form: (2) where, Х 1 , Х 2 -input signals determined by the parameters and properties of raw materials and materials, according to the design of the cable product; W -a generalized signal of the system, including elements: both technological parameters and operating parameters of the cable manufacturing equipment involved in the technological process, including its technical condition and moral wear; Y -an output signal showing the quantitative indicators of the finished cable product.

Research results
The conducted complex of research works allowed to obtain the NNM DS of a complex, multicomponent, nonlinear, intelligent structure of the technological process of cable product production (Fig. 5) in real time, having the form of a connected (a) and optimized structure (b).The work of the NN is carried out by processing the corresponding input signals coming to some neurons and converting them into an output response signal determined by all the weighting coefficients of the corresponding technological process.The NN training process is based on the search with subsequent optimization of the values of weighting coefficients, the use of which minimizes the control error.
Fig. 6 shows the results of mathematical modeling for a neural network DS with a fully connected (a) and optimized structure (b) for a complex, multicomponent intelligent system of the technological process of cable product production.The given characteristics show that with random values of the flow of the technological process of cable product production, the ICS can work as unstable (Fig. 6, a), and in the case of optimization of technologies operating within the production process, it behaves as a stable system (Fig. 6, b).As an evaluation criterion, a technological parameter has been selectedtension, which is relevant for all types of cable manufacturing equipment and is defined as the force acting inside the object (cable product) that arose when the cable billet descended from the pay-off, passed through the working body of the cable making machine and received on the receiver as a finished (finished for this technology) product.

Conclusion
Based on the above, it can be concluded that the proposed intellectualized system is adapted to the technological process of the CWP production.It is able to accumulate the information received, generalize and analyze the knowledge gained, solve complex problems of evaluating dynamic processes by using combined simulation models based on a combination of collected knowledge from the relevant subject area with practical experience and methods of teaching NN: technical processes, cable making machines with different operating principles, the design of cable products according to mark positions and mass of materials, from which structural elements of the cable product are manufactured.
The use of a hybrid NN model for cable production ensured the fulfillment of the research task by solving the problem of dimensionality of quantities and ensuring the specified accuracy of modeling by adapting the technology involved in real time.All this makes it possible to effectively evaluate cable production as a nonlinear controlled DS, capable of repeatedly increasing the accuracy of modeling, ensuring the reliability and safety of multicomponent CWP manufacturing technology.
Thus, there is a prospect of creating technical systems focused on cable technology, having expanded technical, technological and production capabilities.

Fig. 1 .
Fig. 1.The scheme of production routes of the technological process of manufacturing cable and wire products

Fig. 2 .
Fig.2.The primary structure of the ICS of the technological process of CWP manufacturing, built on information indicators: EE -the external environment; ICS -an intelligent control system; I -interference of the information and measuring systems and control and measuring system of cable making machines; D1 -disturbances, by technology and technological nature; D2 -disturbances caused by the operation electromechanical system and control system of cable making machines; D3 -disturbances caused by a deviation in the design of the cable product; D4disturbances caused by pre-emergency situations in the operation of cable making machines; D5 -other disturbances.

Fig. 3 .
Fig.3.ICS structure: X1, X2 -input signals determined by the parameters and properties of raw materials and materials, according to the design of the cable product; F -disturbances and interference; Y -output signal.

Fig. 5 .
Fig.5.Structural model of a neural network dynamic system: a) a neural network with a fully connected structure; b) a neural network with an optimized structure

Fig. 6 .
Fig.6.Results of mathematical modeling for NN DS with a fully connected structure (a) -not steady-state mode and with an optimized structure (b) -steady-state mode.

Table 1 .
Distribution of materials and raw materials in structural elements of a cable product