JPH02182812A - Method for operating blast furnace - Google Patents

Abstract

PURPOSE:To quickly and accurately make the decision of the operation and to improve the operational level by making a knowledge base of hierarchical structure for executing inferences of rule group for protection, rule group for attack and rule group for improving distribution, in order. CONSTITUTION:At first, action 'yes' or 'no' 3 is decided with inference result of 'operational variation temporary inference' 2 in the rule group for protection, and in the case of the action 'yes', 'retreating action instruction' 4 is executed to make 'inference stop' 12. In the case of the action 'no' with the decision, action 'yes' or 'no' 6 is decided with inference result of 'operational margin temporary inference' 5 in the rule group for attack and in the case of the action 'yes', 'attack action instruction' 7 is executed to make 'inference stop' 12. Then, in the case of the action 'no', desirably, action 'yes' or 'no' 9 is decided with inference result of 'distribution- improving temporary inference' 8 in the rule group for improving distribution and in the case of the action 'yes', 'distribution-improving action instruction' 10 is executed to make 'inference stop' 12. In the case of the action 'no' with the decision, 'actual condition-keeping instruction' 11 is executed to make 'inference stop' 12. By this method, decision of the operation is quickly and accurately executed and the operational level is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method of operating a blast furnace using a knowledge engineering system.

[Conventional technology]

Blast furnace operation is made up of a large number of interrelated operating factors, and it is difficult to directly visually monitor the inside of the furnace due to equipment conditions, etc., so it is necessary to maintain and improve the operating level. In order to achieve this goal, it is necessary to comprehensively judge information from sensors installed in the blast furnace and control it accurately.

For this reason, the experience and knowledge of operators is still important for the daily operational management of blast furnaces.

Knowledge engineering systems can incorporate such human know-how into computers and process it.
Knowledge engineering systems are being introduced into blast furnace operation management as shown in Japanese Patent Application Laid-open No. 270708 and Japanese Patent Application Laid-Open No. 62-270712. By systemizing operational management, problems such as oversight of information and mistakes in judgment will be eliminated, and operational management will be optimized and standardized.

[Problem to be solved by the invention]

JP-A-62-270708 and JP-A-62-270
In the knowledge engineering system disclosed in Publication No. 712, the results obtained by inference include prediction of blowout and slippage,
A knowledge base has been established for partial items of phenomena inside the blast furnace, such as the control of furnace heat, and inferences are made independently for each item. However, in-furnace phenomena such as ventilation, loading unloading, and furnace heat occur in an interconnected manner as one in-blast furnace process, and a blast furnace operation management system should comprehensively handle these individual phenomena. It is necessary to make decisions and link them to action. Furthermore, actions in actual operation include retreat actions (escape actions) such as increasing the fuel ratio and decreasing air flow to avoid furnace condition abnormalities, and escape actions such as reducing the fuel ratio depending on the stability of the operation. There are reversal actions and operational level improvement actions (offensive actions) to improve the operational level, but conventional systems have a problem in that they cannot cover all of these daily operational management.

Therefore, the present invention has been made to solve these problems, and its purpose is to obtain a blast furnace operating method based on daily operation management that covers the entire daily operation using a knowledge engineering system. .

[Means to solve the problem]

The blast furnace operating method according to the present invention includes (1) a method of inferring the internal situation of a blast furnace using knowledge engineering and taking operational actions based on the results; the first step is to analyze various blast furnace operating fluctuations; The situation inside the reactor is judged based on a set of defensive rules that predict and derive retreat actions such as increasing the fuel ratio or decreasing the amount of air flow, and if a situation that requires action occurs using the set of defensive rules, instructs the action and makes inferences. If the defensive rule group does not lead to the occurrence of an action, the second step is to judge the situation inside the reactor using an offensive rule group that derives actions to improve the operational level, such as reducing the fuel ratio, depending on the stability of the operation. , based on a knowledge base configured such that when a situation that requires an action occurs in the attack rule group, an action instruction is given and inference is stopped, and when the attack rule group does not lead to the occurrence of an action, the status quo is maintained. A blast furnace operating method and (
2) If the attack rules do not result in an action, the third step is a distribution improvement rule group that determines the possibility of improving reaction efficiency, reducing furnace body heat dissipation, etc., and derives an appropriate charge distribution action. The situation inside the reactor is judged based on the distribution improvement rule group, and if a situation that requires an action occurs, an action is indicated and the inference is stopped, and if the distribution improvement rule group does not lead to the occurrence of an action, the status quo is maintained. The method described in item 1 above is configured as follows.

[Effect]

In the present invention, after a defensive rule group is used to determine an escape action to avoid reactor condition abnormalities, an offensive rule group is used to determine an offensive action to reverse the escape action and improve the operational level. Therefore, it is desirable to further cover the entire day-to-day operation by determining the distribution improvement action to make the operational situation suitable for taking aggressive action based on the distribution improvement rule group.

〔Example〕

Embodiments of the present invention will be described below based on the drawings. 1st
The figure is a schematic flow diagram of inference based on the knowledge base of the present invention. When inference starts 1, operation fluctuation hypothesis inference 2 is first executed. It is determined whether there is an action or not (3) based on the inference result, and if there is an action, a backward action instruction (4) is issued and the inference is stopped (12).

As a result of the determination of the presence or absence of action 3 based on the inference result, if there is no action, operation margin hypothesis inference 5 is executed. It is determined whether there is an action or not (6) in the inference result, and if there is an action, an attack action instruction (7) is issued and the inference is stopped (12). If there is no action based on the determination of the presence or absence of an action 6 based on the inference result, it is possible to maintain the status quo and return to the previous state, but desirably, the distribution improvement hypothesis inference 8 is further executed.

It is determined whether there is an action or not (9) in the inference result, and if there is an action, a distribution improvement action instruction (10) is issued and the inference is stopped (12). If there is no action as determined by the action presence/absence 9 based on the inference result, the status quo maintenance instruction 11 is issued and the inference is stopped 12. The retreat action instruction 4, attack action instruction 7, distribution improvement action instruction 10, and status quo maintenance instruction 11 may be output as messages to a display terminal, or may be sent as control data to a process computer.

Each inference will be explained in more detail below.

FIG. 2 is an explanatory diagram of the contents of operation fluctuation hypothesis inference 2 in FIG. 1. When the central gas flow is weak, the ventilation resistance in the furnace increases, and the furnace brick temperature increases, the central flow insufficient operation fluctuation 15 occurs, and when the peripheral gas flow is weak, the heat level decreases, and the furnace brick temperature decreases. Based on the operational fluctuation hypothesis 14 that knowledgeably expresses the phenomenon that operational fluctuation 16 due to insufficient peripheral flow occurs, and data 13 such as the detection end, inference 17 is performed to determine whether the current situation inside the reactor corresponds to the above hypothesis, and retreat is performed. Final judgment 1
Do step 8.

FIG. 3 is an explanatory diagram of the contents of operation margin hypothesis inference 5 in FIG. 1. The gas flow distribution is in proper condition and the heat level is high (
, the operating margin hypothesis 19, which expresses the state that there is a margin in the operating level when the ventilation resistance is low and stable, the furnace brick temperature is high, and the loading is stable, and the detection end etc. Based on the data 13, a current state inference 17L and a final attack judgment 20 are made.

FIG. 4 is an explanatory diagram of the contents of distribution improvement hypothesis inference 8 in FIG. 1. When the center gas flow is strong, it is determined that the center gas flow can be suppressed by increasing the amount of ore charged to the center to improve the reaction efficiency in the center. When , the intermediate gas flow can be suppressed 23, and when the surrounding gas flow is strong,
Based on the distribution improvement hypothesis 21 configured to determine that the surrounding gas flow can be suppressed 24 and the data 13 of the detection end, etc., an inference 17 is executed and a distribution improvement final determination 25 is made.

An example of blast furnace operation using the system configured as described above is shown in Figure 5. From 9 o'clock on the 1st, the distribution situation was insufficient in peripheral flow, and the ore charging position was shifted toward the center. Actual reactor action was carried out in accordance with the instructions for adjusting the peripheral flow distribution due to the internal ore shift. From then on, the gas flow distribution situation became appropriate from 13°. - Due to a temporary lack of peripheral flow, the temperature of the furnace bricks has decreased from around 10 degrees, but because we have adjusted the peripheral flow distribution, the decrease was stopped at 13 degrees, and it has since started to rise. . At around 13 degrees, a decrease in the furnace heat is predicted by the group of protective rules, and actions are taken in accordance with the instruction for zero heat increase due to a fuel ratio increase of 5 kg/l-p. As a result, the drop in hot metal temperature was 15
It stopped at ° and has since recovered. Due to the recovery of the furnace heat status and the stabilization of other operating conditions, the 23:00 attack rule group instructed us to reduce the heat to 0 by reducing the fuel ratio by 5 kg/l-p as a return to the heat increase action, and we are implementing the action. .

Furthermore, at 8:00 on the 2nd, it was determined that the reactor heat status had sufficient margin according to the attack rule group, and actions were taken in accordance with the instructions for heat reduction O by reducing the fuel ratio by 2 kg/l-p. 1
At 7°, ventilation conditions and furnace heat conditions are not in a condition to carry out retreat or joint attack actions, but according to the distribution improvement rules, it is determined that there are stray bullets in the surrounding area, and the surrounding area due to the outer area of the ore Flow suppression distribution improvement ○ instruction has been given and action is being taken. As a result, the gas flow distribution situation has returned to being appropriate again.

In this embodiment, inference based on the knowledge base of the present invention is performed every 10 minutes, and O-O in the figure indicates the occurrence of an action, and all other inference results remain as they are.

Effects of invention C] When operating a blast furnace using a knowledge engineering system, the knowledge base has a hierarchical structure in which inference is executed in the order of a group of defense rules, a group of attack rules, and a group of rules for improving distribution, so that only abnormality prediction can be performed. Not return after retreat action,
Most of the daily operational decisions, such as actions to improve the operational level, can be incorporated into the system in a highly readable structure, and by using this system, the operational level can be improved by making operational decisions faster and more accurate. is planned.

[Brief explanation of the drawing]

Figure 1 is a schematic flow diagram of inference based on the knowledge base of the present invention. Figure 2 is an explanatory diagram of the content of the operation fluctuation hypothesis inference in Figure 1. Figure 3 is the content of the operation margin hypothesis inference in Figure 1. FIG. 4 is an explanatory diagram of the content of the distribution improvement hypothesis inference in FIG. 1, and FIG. 5 is an explanatory diagram showing an example of blast furnace operation by the system of the present invention. Applicant Nippon Steel Corporation Patent Attorney Minoru Aoyagi Figure 4 Figure 1

Claims (1)

[Claims] 1. In a method of inferring the internal situation of a blast furnace using knowledge engineering and taking operational actions based on this result, the first step is to predict the operational fluctuations of various blast furnaces, and to determine the fuel ratio. The situation inside the reactor is judged based on a group of defensive rules that derive retreat actions such as an increase in air flow or a decrease in air flow, and if a situation that requires action occurs using the group of defensive rules, an action instruction is given and reasoning is stopped. If the defensive rule group does not lead to an action, the second step is to judge the situation inside the reactor using an attack rule group that derives actions to improve the operational level, such as reducing the fuel ratio, depending on the stability of the operation, and then execute the attack rule. When a situation that requires action occurs in a group, an action instruction is given and inference is stopped, and if the attack rule group does not lead to the occurrence of an action, the status quo is maintained.Inferring the situation inside the reactor. A method for operating a blast furnace characterized by performing operation actions based on the results. 2. If the attack rules do not result in an action, the third step is a distribution improvement rule group that determines the possibility of improving reaction efficiency, reducing furnace body heat dissipation, etc., and derives an appropriate charge distribution action. The situation inside the reactor is judged based on the rules, and if a situation that requires an action occurs with the distribution improvement rule group, an action instruction is given and the inference is stopped, and if the distribution improvement rule group does not lead to the occurrence of an action, the status quo is maintained. 2. The method of operating a blast furnace according to claim 1, wherein the blast furnace is configured as follows.