Results are presented in this article for work carried out since 2006 at the Magnitogorsk Metallurgical Combine (MMK) blast-furnace workshop for improvement of automatic control systems created for blast-furnaces in the course of stagewise introduction of an automated system for control, optimization, and prediction of blast furnace smelting (ASCOP BF). A list of local systems within the ASCOP BF and stage of their introduction are given in Table 1. Results of introducing them have been described in [1, 2].

Table 1. Composition of a Group of ASCOP BF Systems Introduced into MMK Blast Furnaces
Full size table

Features of the work are stagewise introduction of local ASCOP BF systems; connection of systems to the automatic control system created using engineering facilities; transfer of calculated results of a system to master workstations of all furnaces and workshop management; functioning of the systems developed in an on-line regime with automatic delivery of recommendation for improvement of blast-furnace smelting; use of experience of creating automated systems for blast furnaces and their upper level in the USSR, CIS countries, and overseas, obtained during operation at VNIPI SAU (Moscow), Siemens (Germany), and at AKOMM (Moscow) created in 1992.

Group of Blast-Furnace Control systems. The tasks resolved in this group of systems are both of importance independently, and are used in calculations for groups of Optimization and Prediction systems. A system for monitoring radial gas distribution in the blast-furnace bell equipped with a CCU (see 1.1 in Table 1) uses the following sensors:

  • stationary radial multipoint thermocouples (RMT) of AKOMM [1, 3]. At the request of a user on the screen furnace video-terminals Nos. 2, 4, 6, 9, and 10 temperature trends and average curves (Fig. 1) are provided for recent periods: “output” (furnace operating period from closure to closure of the iron tap-hole for adjacent tappings, a shift, day, and week for all furnaces fitted with RMT;

    • peripheral AKOMM thermocouples above the charging level (PTA), installed in the same place as RMT;

    • in the tapered row of bell protective plates (BPP). Their junctions are located at the same distance from the vertical part of the BPP, which are also the first points of the TMT; and

    • furnace gas outlet thermocouple (GOT).

Fig. 1.
figure 1

Gas temperature distribution over BF-6 diameter per day (00:00, 05.15.2016 – 00:00, 05.16.2016).

Full size image

The importance of the monitoring system for circumferential gas distribution in the furnace bell (see 1.2 in Table 1) arose particularly in view of introduction of the chute type CCU. In addition to the peripheral thermocouple beneath the BPP (PTP) and a thermocouple of the furnace gas outlet (FGO) in five furnaces, within which a CCU was installed, eight PTA were installed. For an example, data are given in Fig. 2 for the reduction in PTA readings after charging iron-ore material (IOM) to different point of the bell circumference. They point to non-uniform IOM distribution over the bell circumference in this case.

Fig. 2.
figure 2

Gas temperature below gas offtakes. Absolute values of BF-10 reading changes.

Full size image

It was assumed that for operative control of the circumferential gas distribution in the furnace hearth (see 1.3 in Table 1) it is most expedient to use information about thermal action of the tuyere location for an air tuyere [4]. This problem was resolved by means of sensors for water flow rate (in order to determine heat exchange) for cooling air tuyeres, installed for monitoring burnout, and a mathematical model (MM) developed for calculating the temperature at the tuyere location (T tl) [4]. Calculation algorithms make it possible for actual furnace operating parameters and data for the amount of heat removed from a tuyere with cooling water to determine the T tl for each tuyere location (Fig. 3).

Fig. 3.
figure 3

Gas temperature in tuyere locations: w.t. is western iron tap-hole; e.t is eastern iron tap-hole.

Full size image

The monitoring system for material and heat balances (see 1.4 in Table 1) is intended for operational, on the theme of melting process, calculation of material and heat balances (MHB). The aim of performing MHB calculations (Tables 2 and 3) is use of the amounts of their discrepancies and some items of balances for analyzing occurrence of melting processes and recognition of deviations from standards for earlier stages.

Table 2. Results of Calculating BF-2 Material Balance per Day (00:00, 05.16.2016 – 00:00, 05.17.2016)
Full size table
Table 3. Results of Calculating BF-2 Heat Balance per Day (00:00, 05.16.2016 – 00:00, 05.17.2016)
Full size table

In order to monitor furnace heating (see 1.5 in Table 1), MM are used for gas dynamic heat and mass transfer processes in the furnace two-layers zones [5, 6]: the tuyere zone, and the zone of final melted product composition formation and heating (zones III and IV in Fig. 4). The model considers the following data: formation and dimensions of tuyere locations (1); combustion of coke and its replacements (2); passage of molten products through the tuyere location and oxidation of previously reduced iron within them (3); distribution of heat liberated as a result of oxidation of iron between melt and gas products (4); transfer of heat radiated from gases in the shell of the tuyere location and melt (5); thermochemical processes occurring in the “finishing” zone, including secondary iron reduction, oxidized in tuyere locations, and direct reduction of iron impurities (6); physical heat loss, including with tuyere cooling water (7).

Fig. 4.
figure 4

Distribution of heat exchange zones (I–IV) material (t) and gas (T) temperatures over furnace height (from charging level to cast iron tap-hole): b) blast furnace thermal condition criteria; I, II, III, IV) heat exchange zones; +, –) processes occurring with heat liberation (absorption).

Full size image

The most complex problems for solution were 1, 3, 4, and 5. In order to calculate the extent of the tuyere location L tl (problem 1), an expression was used provided in [6]. The possibility of calculating L tl, and also solution of problems 3, 4 [7], and 5 [8] makes it possible to determine complex parameters (CP) specifying processes in tuyere locations. Among CP there are: temperature of molten product at the entry of the tuyere zone \( \left({T}_3^{\prime}\right) \), or at the outlet from the furnace (shaft) \( \left({T}_2^{\prime \prime}\right) \), since they are equal numerically; temperature of the tuyere location shell (t shell); balanced gas temperature in the tuyere location \( \left({T}_3^{\prime \prime}\right) \), entering into heat exchange with charge materials in the second stage, and differing significantly with respect to physical nature from the theoretical combustion temperature; temperature of melt product at the outlet from the tuyere location \( \left({T}_3^{\prime \prime}\right) \); temperature of melt products at the inlet to the “finish” zone \( \left({T}_4^{\prime}\right) \); temperature of melt products at the outlet from the “finish” zone \( \left({T}_4^{\prime \prime}\right) \).

Group of Blast-Furnace Smelting Optimization systems. Improvement of a blast furnace as a unit for smelting metal consists in that all the main (gas dynamic, mass and heat exchange, smelting) processes within it have optimum values and blast-furnace smelting efficiency will be better the closer they are to the optimum level. In order to resolve this problem, calculation models are used providing calculated restoration of optimum regimes.

The system for optimizing blast consumption (see 2.1 in Table 1) is intended for determining that amount of blast fed to a blast-furnace with which there is formation of the optimum volume of gas flow for a specific raw material and operating conditions. In this case, criteria for optimization may be either minimum coke consumption, or maximum furnace productivity. The choice of optimization criterion is made by workshop production personnel proceeding from workshop requirements and specifications for subsequent cast iron conversion. An example is given in Fig. 5 for determining the optimum blast consumption for BF-1.

Fig. 5.
figure 5

Dependence of operating parameters for days: a) BF-1 (05.09.2016–07.08.2016); b) BF-6 (06.01.2016–07.31.2016): ) starting data; ——) regression; ) extremum; Gi is actual iron smelted, ton/day; Gi(c) is calculated iron smelted ton/day; Gr is iron-ore material (IOM) smelted, tons IOM/day; CBC is cold blast consumption.

Full size image

The system for optimizing radial gas distribution (RGD) (see 2.2 in Table 1) occupies an important place in blast-furnace smelting management. In order to control RGD in the furnace bell, the uncooled AKOMM RMT is most suitable [1, 3]. Apart from monitoring radial gas distribution in the furnace bell RMT readings are used for controlling the shape of the melting zone (MZ) by means of a two-dimensional mathematical model (TMM) [9].

For quantitative evaluation of the RGD in the furnace bell from TMT readings, two radial inequality coefficients were used:

$$ \mathrm{CR}={t}_{\mathrm{p}}/{t}_{\mathrm{c}},\mathrm{fract}.\mathrm{one}\kern0.5em \mathrm{and}\kern0.5em \mathrm{CRR}={t}_{\mathrm{p}}/{t}_{\mathrm{r}},\mathrm{fract}.\mathrm{one}, $$

where t p = 0.4t 1 + 0.4t 2 + 0.2t 3 is gas temperature in the “peripheral” part of the bell, °C; t c = 0.2t 3 + 0.4t 4 + 0.4t 5 is gas temperature in the “central” part of the bell; t r = 0.2t 1 + 0.2t 2 + 0.2t 3 + 0.2t 4 + 0.2t 5 is temperature over the furnace radius, °C; t 1–t 5 are thermocouple readings at five equal rings over the area of the bell (in the furnace-circle center), reckoned from the periphery to the center.

In addition, another coefficient is used:

$$ \mathrm{CRS}={t}_{\mathrm{pa}}/{t}_{\mathrm{bg}},\mathrm{fract}.\mathrm{one}, $$

where t pa is average temperature of peripheral gas for readings of the values of TPU thermocouples, °C; t bg is average bell gas temperature from readings of values of TPG thermocouples, °C.

Comparison of coefficients for radial inhomogeneity (CR, CRR, and CRS) and parameters specifying MZ shape [9] with final smelting indices (furnace productivity and specific coke consumption) showed that for MMK conditions MZ parameters have the most effect on the latter [9].

A system for optimizing circumferential gas distribution (CGD) in the bell (see 2.3 in Table 1) for blast-furnaces equipped with a shute type CCU is especially important. Monitoring of the reduction in PDN indices after loading the iron-ore part of a charge (IOC) at different points of a bell circumference (see Fig. 2), makes it possible to approach CGD optimization.

The system for optimizing circumferential gas distribution in a furnace hearth (see 2.4 in Table 1) accomplishes levelling of the tuyere location temperatures t tl (see Fig. 3) by means of heat exchange sensors installed in all of the BF-2 air tuyeres and regulation of natural gas supply to tuyeres (hydrocarbon, fuel oil, etc.) (Table 4).

Table 4. BF-2 Tuyere Location Temperature Distribution
Full size table

The system for optimizing smelting reducing processes (see 2.5 in Table 1) is intended for determining the optimum ratio of indirect (R i ) and direct (R d ) iron reduction in a BF, with which the minimum coke consumption is achieved. Points of intersection of lines for the dependence of coke consumption on R d as both a reducing agent and heat carrier obtained as a result of calculations determine the minimum possible coke consumption (Fig. 6).

Fig. 6.
figure 6

Diagram of BF-10 direct reduction for days (00:00, 05.15.2016 – 00:00, 05.16.2016).

Full size image

A system for optimizing the furnace thermal condition (see 2.6 in Table 1) is aimed at resolving a more complex problem of controlling blast-furnace smelting. Of what does complexity of this system consist? It is well known that in order to resolve any optimization problem it is necessary to know its criteria. In view of features of heat exchange and reduction processes over a furnace height (from the charging level to the iron tap-hole) there are two criteria [3]: temperature of materials for input from a “charge” (shaft, bosch, and shoulders) of a furnace \( \left({T}_2^{\prime \prime}\right) \); temperature of melting products at the outlet from the furnace “hearth” \( \left({T}_4^{\prime \prime}\right) \). This also causes complexity for controlling the BF thermal condition. The presence of two criteria is determined by the fact that a blast-furnace from the charging level to the iron tap-hole is separated into four zone of heat- and mass-transfer (see Fig. 4) within which processes occur with different final thermal effects.

In order to determine the optimum thermal condition of a furnace, the system fulfils the following operations: continuous monitoring of current values of \( {T}_2^{\prime \prime } \) and \( {T}_4^{\prime \prime } \); determination of optimum values of \( {T}_2^{\prime \prime } \) and \( {T}_4^{\prime \prime } \); comparison of current and optimum values of \( {T}_2^{\prime \prime } \) and \( {T}_4^{\prime \prime } \); choice for level and sign and difference between current and optimum values for each criterion of control effects “from above” (change in coke consumption) and “from below” (change in blast parameters in order to achieve the optimum furnace thermal condition. Optimum values of \( {T}_2^{\prime \prime } \) are determined from results of operating the system after determining extremes for the dependences of coke consumption and furnace productivity in \( {T}_2^{\prime \prime } \) (Fig. 7).

Fig. 7.
figure 7

Dependence of BF-2 operating parameters for days from 06.13.2015 to 08.12.2015: ) starting data; ——) regression; ) extremum.

Full size image

Research by the authors has also shown that an important factor sometimes having a decisive effect on furnace operation is material temperature at the inlet to the tuyere location (t3″). Exceeding its value may lead to mass burning of air tuyeres [10].

Results of functioning of the parameter monitoring and optimization system for melting zones (see 2.7 in Table 1) in MMK BF are given in [9].

A system for predicting silicon content in cast iron (see 3.1 in Table 1) is intended for predicting cast iron silicon content of forthcoming tapping using a self-adjusting (from results of operation) statistical model for data about the temperature of all tuyere locations (see Fig. 3), average for the period from closing the preceding to closing the next tapping, and about silicon content in cast iron of the last two tappings. Results of this calculation are given in a video-chart “Charge in Si content in cast iron” (Fig. 8).

Fig. 8.
figure 8

Change in silicon content in BF-2 cast iron in period from 07.05.2016 to 07.08.2016: - - -) cast iron predicted Si content ([Sipr]), %; ——) cast iron actual Si content ([Si]), %.

Full size image

Conclusion. In blast furnaces of MMK, work has been carried out for technological satisfaction of the control system created with stagewise introduction of local monitoring systems, optimization and prediction of blast-furnace smelting indices, combined into an automated system for control, optimization, and prediction of blast-furnace smelting (ASCOP BF).

The ASCOP BF for each blast furnace using mathematical models and special sensors provides continuous automatic monitoring of radial and circumferential gas distribution in the bell, circumferential gas distribution in the furnace hearth, material and thermal balances of furnace smelting and heating; optimization is accomplished for a gas dynamic smelting regime, radial and circumferential gas distribution in the furnace hearth, melting reduction processes, furnace heat condition, and smelting zone parameters: it gives a prediction of cast iron silicon content.

This helps blast-furnace production personnel to control smelting effectively in order to achieve maximum productivity and minimum coke consumption with given smelting raw material and operating conditions.