ÐÓÑÑÊÈÉ       ÓÊÐÀ¯ÍÑÜÊÀ      

ABSTRACT OF THESIS OF MASTER'S WORK


On main    Library    References   

DONETSK NATIONAL TECHNICAL UNIVERSITY

Physics-metallurgical faculty
Department "Electrometallurgy"
Ìåò-04ì
Speciality:"Metallurgy of black metals"


Master's work:
"The design and optimization of raskyslenya and alloying became at the issue from DSP and vnepechnoy treatment"


Author: Alexandr Dikalenko.
Scientific leader: S.Khrapko


Intensification of steel-smelting production results in a necessity creations of the systems of automatic conduct of melting, in the function of which not only automation but also optimization of executable enters technological operations. And if at production of poluprodukta in basic the burst performance is required, at the receipt set chemical composition became is used dear enough raskyslytely, ferro-alloys and modifiers, that destroys on the first plan systems of minimization of prime price of vnepechnoy treatment.

We will consider system of optimization for the process control of alloying, taking into account non-linearity of task in most simple case. We consider, that the mechanical losses of metal, slag and materials are absent. It is necessary to expect the amounts of materials possessing minimum possible by a cost, addition of which in the system gives the required composition of metal. The set problem is the task of the constrained optimization, I.e. search such values of the optimized variables (the masses of the Xj materials), at which is achieved the minimum of having a special purpose function (total cost of materials), on condition of implementation of the set limitations on optimized pere- mennye.

We will mark, that dependence of composition of metal on the masses of materials has difficult nonlinear character which is related to physical and chemical cooperation elements and their redistribution between a metal and slag, therefore limitations are nonlinear functions in relation to the optimized variables. Universal methods of decision of such tasks in a general view or does not exist, or they have so low efficiency, that their application in given one case (the number of the optimized variables is calculated in tens) it is impossible. The linear of having a special purpose function means at the same time, that decision (if it exists) it is on the border of the possible region limited by present inequalities. The methods of linear are the very effective method of decision of such tasks programing (sympleks-method and his modifications), however in a clean kind they are applicable only in that case, when limitations are set in a linear kind. The possible method of application of neiieaen-method is in examined case yteratsyonnoe decision of task by the method of progressive approximations, at which on every yteratsy lynearyzatsya of all limitations in a current point is produced and search of minimum of having a special purpose function by a neiieaen-method for got linear limitations with the subsequent reiteration of yteratsyy to establishment of permanent values of variables of search (the masses of materials).

We will consider some current (initial or intermediate) yteratsyyu, which some vector of the masses of the materials X0, set in the system, is known on (on first yteratsyy, I.e. at the beginning of calculation, when in the system nothing is set yet, pry-nyat' X0=0 is needed). Described higher thermodynamics model of distributing elements between phases allows on the known composition and amounts of materials To expect X0 the masses of all elements in a metal, slag and gas phase and, consequently, compositions and amounts of metal and slag, and also all DKU (in further exposition overhead by the index "zero" the sizes calculated in the current point X0 are marked). Got composition metal can fall short of to inequalities, or for the present set materials can be not executed condition of minimization . For the decision of put tasks it is necessary to change the masses of materials, that will cause change of the masses of elements in a metal and composition of metal. Mass of element of I in a metal by difficult appearance is determined by the masses of all elements being in the system, and last - in the turn - arrival in the system of materials. For lynearyzatsyy of dependence on Xj we will decompose a function in a row (in the current point X0) in relation to X=X-X0 and we will cast aside elements higher than 1th order.

We will consider systems of inequalities is the znakoperemennost' sizes Xj, that complicates the use of neiieaen-method (requires doubling of number of variables).

E system, in particular, requirements to minimum and/or maximal mass of the got metal, to general mass (or volume) entered materials, to mass or stake (in relation to general mass) of separate materials (or groups of the materials incorporated on a certain sign), to composition, amount and properties of slag, and also limitation of presence of materials on storage and etc

We will mark system of inequalities by a neiieaen-method formed thus gives a new set of materials , possessing a minimum cost. As the used values DKU are derivative, calculated in a current point (at current oxidizing potential), and the change of the masses of materials in the system causes change of okyslennosty metal and slag, the composition of metal and slag got as a result will not exactly correspond to required. The repeated calculation DKU in a new point allows to conduct correction of the masses of materials (to increase or decrease their amounts). As a result of a few progressive approximations will be got sought after the masses of materials, providing popada-nye in the set chemical composition (and also implementation of any other additional limitations) and possessing a minimum cost. Yteratsyy is halted, and a calculation is considered complete.

We will mark, that in the process of decision can both be multiplied calculation the masses of materials and be diminished, and also addition, exception or replacement of one materials, is possible other, that is related to the mutual influencing of materials (more precisely, composition of metal and, especially, oxidizing potential in the system) on mastering of elements from materials and, consequently, degree of "advantage" of one materials as compared to other.

Interestingly, that matrices and after the insignificant count enable additionally to take into account the factors of vagueness of chemical compositions of the used materials, mass and composition of initial metal and slag, constants of distributing of elements between a metal and slag and etc, that allows to operate by the no point values of sizes, and their interval estimations (by lower and top limits), as a result is achieved the substantial increase of universality, community and reliability of calculation during practical realization of algorithm.

Possible cases being interest for practice, when the decision specifies on impossibility in the set terms of the assured hit in the required composition (absence of necessary ferro-alloys, error in weighing, etc.). In such cases the system in advance gives out the proper report and due to ability to analyse the reasons of possible declination offers list of operative decisions: to reduce the vagueness of the preset parameters (to specify composition of one or another ferro-alloy, to alloy mass of the set materials and o.a.); in two receptions with the intermediate test of metal; to translate melting in other brand in accordance with the present brief-case of orders or in any standard brand. On all possible variants a new set of materials, cost of alloying and expected composition of metal, is reported. In the automatic mode is made decision the system independently within the limits of the beforehand set list and priorities. At the receipt of test of metal information about the actually got composition of metal is automatically used for statistical adaptation of parameters of model, which take into account the degree of deviation from an equilibrium and the technological features of every aggregate and pe-ryoda melting of this brand became.

Developed system enters in the complement of the package "ORACLE", inculcated at the Moldavian metallurgical plant, where is used for the decision of the following tasks: operative process control of melting in the mode of "adviser" (gives out recommendations to the personnel) and in composition to the ACE TP (gives out managing signals on metering devices) [9].

Like models, developed system can be used for the decision of the following tasks:

1.operative process control of melting in autonomous behavior (gives out the necessary reports and recommendations to the personnel) and in composition to the ACE TP (gives out uprav-lyayuschye signals on the metering devices of materials);

2.short-term and perspective planning of work of workshop (calculation of potreb-nosty materials and estimation of efficiency of smelting of staley different brands);

3.choice preferable materials on a present nomenclature, prey-skurantam and other

Conclusions. Offered approach allows on principle to extend voz-mozhnosty optimizations of alloying, raskyslenyya and affinage of metal. Influencing of the elements contained in a metal is thus taken into account not only, but also influencing of shla-ka, and also moment of input of materials. Optimization of not only amounts but also sequence of input of materials becomes possible (including shlakoobrazuyuschykh). These tasks are decided in none of other existent systems.


On main    Library    References