No abstract

The refractory lining of most of the thermal units is a structural monolith since the individual components are mortar bonded~ Different types of stress states are developed in various elements of the lining because of the specific features of the design and the technological process.The stress--strain state of the lining is determined to a large extent by the mechanical properties of the mortars [i]. However, these properties have not been studied in detail up to now.Some information regarding the strength of mortars was given in the earlier publications [2, 3]. More recently, the creep of the chamotte, the mullite-silica, and the dinas-based mortars was evaluated [4]. This paper presents the experimentally determined thermomechanical properties of a dinas based mortar used for coke oven lining.We used the MDK-I mortar produced by the Krasnogorovsk Refractories Plant.It possesses standard properties and meets the specification OST 14-23-183-86 completely.An aqueous solution was made from the mortar with aluminum sulfate addition.In order to evaluate the technological suitability of the solution, its lining characteristics were controlled using an AzNII Kr-i cone and a special vibrating stage.In order to carry out compressive creep tests, cylindrical specimens measuring 36 mm in diameter and 50 mm in height were made in specially selected molds using the vibration-casting (vibrocasting) method.The specimens were subjected to drying at 120~ for a period of I h. The apparent density of the specimens was 1.81 g/cm 3 and their open porosity amounted to 29%. Before carrying out creep measurements, the end faces of the specimens were polished.

A rheological model was worked out in [1, 2] for describing the creep of elastoviscoplastie media~ In the framework of this model the deformation kinetics with an isothermal schedule may be presented in the form of generalized s~-Z~ diagrams for the relationship between the deformation e or the parameter Z = exp(~ Q/RT) and the function r =f 02dTwhere Q is the apparent energy of activation of creep; R, universal gas constant; T, temperature; a, applied stress; r, time; a, model parameter; ~(T) = ~o expt(-~ Arrhenius function. The use of these models for treating the results of measurements of high-temperature creep on mullite-corundum 113, 4] and magnesia [5] refractories showed that the generalized ~-and Z~-diagrams in logarithmic coordinates are linear, and the values of a are constant for the stated type of refractory: 'IgZ=const+ctlgqP.This enables us from the analytical expressions (3) and (4) to calculate the creep limits aTt,~ or sTa,r for each form of refractory, and thereby to predict its plastic deformation in time during variations in e, a, and T in the range of values attained in the laboratory experiment. In the above researches a similar study was made on dinas refractories for coke furnaces. Tests for creep were made on specimens cut from wall products grade 6194 or 6196 produced at the Krasnoarmeisk (specimens KA), the Krasnogorovsk (KG), and the Pervourals (P) factories. The phase composition of the specimens, data on their density, and porosity are given in [6]. The creep was measured with monoaxial compression in air at 1200, 1250, 1300, and 1400"C. The test method is described in more detail in [6]. At 1400"C tests were done with a staged loading cycle (Fig. 1) on specimens that had been prefired at 1450"C for 50 h in order to avoid additional expansion, which would distort the results of the creep measurements. At the remaining temperatures the tests were done at constant stress (a = 5 N/mm2), excluding the possibility of the development of additional expansion of the dinas [6]. The maximum test time was about 100.h. Using the same conditions the tests were reproduced on 3-4 specimens.The results obtained are shown in the form of s~-diagrams in Fig. 2. As we see, despite the significant spread in the dinas inherent in dinas during the tests of identical specimens the diagrams for all temperatures are linear, and the slope of the lines is approximately the same, i.e., r, = const. Using the averaged values of s for the section ff = 1014 (N/mZ)2.h, we constructed an Arrhenius graph in the coordinates lg (sC~'")--l/T (Fig. 3). For c~ we took the average value (~ = 0.32) of the angles of inclination of the straight lines in Fig. 2. Since e~-~-= koexp ~ --c~Q / R T ) 9(5) then the linearity of the Arrhenius graph indicates the constancy of the apparent activation energy of creep. The processing of the data in Fig. 3 by the least-squares method gives the value Q ~, (500 +_ 15) kJ/mole. The magnitude obtained for the factor is much less than the apparent energy of activation of creep in mullite-co...