JP5042878B2 - Container wall state management method, apparatus, and computer program - Google Patents

Description

  The present invention relates to a container wall state management method, apparatus, and computer program, and in particular, thermal history information that has a temperature difference between an inner wall surface and an outer wall surface and changes in temperature over time, and the thermal history. The present invention relates to a technique suitable for use in estimating and managing a container wall state by estimating a future temperature of each region in a container whose thickness information at the time when the information is obtained is known.

  When managing the operation of reaction vessels with high-temperature gas reactions or liquid reactions such as converters, degassing furnaces, coal gasification reactors, and containers that transport molten iron, such as kneading cars, hot metal ladle, molten steel ladle, etc. It is necessary to monitor the situation (for example, the state of wear) of the container wall that handles the substance and manage the situation.

  Conventionally, the wear state of the container wall has been managed by a human visually observing the state of the inner wall surface when no high-temperature substance is present in the container.

  However, even when a high temperature substance is not present in the container, the surface of the refractory is heated to a high temperature of 500 ° C. or higher, and it is extremely difficult to grasp the wear state as a quantitative value by visual observation as described above. Yes, it must be qualitative management.

Moreover, since the management under the condition that the high temperature substance does not exist in the container is unavoidable, the situation of the outflow of the internal high temperature substance during operation cannot be managed.
Moreover, based on the time transition data of a certain period of temperature measured for a part of the material, knowing the temperature information of the whole material and predicting the future behavior of the material temperature over time is, for example, When measuring the iron skin temperature with infrared thermography for a process that moves in a steel factory using a crane, the extent to which the iron skin temperature rises is estimated from the measurement data for a certain period of time, and molten steel leaks. It is extremely important when rapidly detecting abnormal melting of a refractory that leads to

  On the other hand, a method has been proposed in which the outer wall surface temperature of an operating container is measured using a radiation thermometer or infrared thermography, and the wear state of the container wall is managed from the measured value of the outer wall surface temperature. There has been proposed a method of determining abnormal melting of a refractory by the temperature of a container wall when a container containing molten iron passes through an infrared thermography installation position (see, for example, Patent Document 1).

  However, in the determination method disclosed in Patent Document 1, since the heat capacity of the refractory constituting the container wall is large, it is extremely rare that the temperature of the refractory becomes a steady state, and molten iron is charged. Since the time for the container to pass through the position of the infrared thermography later is not constant, the container wall temperature changes remarkably even if the refractory on the container wall is in the same molten state, as described in Patent Document 1 In the method of determining the abnormal refractory damage of the refractory only by the temperature of the container wall, there is a problem that it is extremely difficult to quantitatively evaluate the amount of refractory with high accuracy.

  On the other hand, considering the heat conduction phenomenon in the container wall as an unsteady heat conduction inverse problem, the unsteady heat is measured based on the temperature data measured by one or two temperature measuring means such as a thermocouple installed on the container wall. There has been proposed a method for estimating the container wall thickness by calculating the temperature inside the container wall using the inverse conduction problem and searching for a position where the temperature of the container wall matches the solidification temperature of the molten iron (for example, Patent Document 2). See).

  However, since the method disclosed in Patent Document 2 must assume the initial temperature distribution of the refractory, the temperature calculation result is not stable immediately after charging the high temperature material into the container, and the container wall thickness The estimation accuracy was reduced. In addition, since one-dimensional is assumed, it is possible to know only the refractory wear state at one location near the temperature measurement point, and it is impossible to estimate the two-dimensional or three-dimensional molten state. there were. Furthermore, in Patent Document 2, the confidence interval of the temperature calculation result by the inverse problem analysis is a period in which the temperature measurement data exists, and the future temperature of the object to be measured is estimated only for a part of the period of the temperature measurement data. There was a problem that it was difficult.

JP-A-3-169474 JP 2001-234217 A

The worn state of the container wall can be determined by the thickness of the container wall. For example, when the wear can be approximated to a one-dimensional shape with a uniform shape, if the container wall is in a thermally steady state, the temperature distribution inside the container wall is linear, and the thickness L of the container wall is the outer wall of the container wall. Using the heat flux Q, the thermal conductivity k _x in the thickness direction of the container wall, the container inner wall temperature T _in and the container outer wall temperature T _out can be estimated from the following equation.
L = k _x · (T _in −T _out ) / Q

  However, since the actual temperature of the container wall varies depending on the operating / non-operating time cycle, the value of the heat flux Q measured on the outer wall of the container wall also changes unsteadily. In addition to this, since the temperature distribution inside the container wall changes unsteadily in a curved shape, estimating the thickness of the container wall with the above equation causes a large error.

  As described above, even if there is no temperature measurement data inside the wall, it is necessary to obtain the temperature or heat flux inside the wall that changes unsteadily. Further, the above method cannot be applied when a three-dimensional melted shape is handled.

The present invention has been made in view of the above points, and based on data obtained by measuring a predetermined area on the outer wall surface of the container, the container wall state of the inner wall of the container wall corresponding to the predetermined area (for example, the wear state) ) Can be accurately estimated over a wide range, and the container wall condition can be managed with high accuracy, and the container wall condition can be managed with high accuracy even if the temperature measurement data is only for a partial period. Objective.
In particular, there is a process that has a high temperature material inside like a molten steel pan, there is a temperature difference between the inner surface and the outer surface, the material temperature changes unsteadily, the thickness of the container wall inner wall changes due to wear etc. The purpose is to estimate the behavior of the future wall temperature from a relatively short temperature measurement period, and to manage the container wall condition with high accuracy without being limited to the period during which temperature measurement data exists. And

（但し、Ｎは基底関数φｉ（ｔ）の個数、ｉはそのインデックス、ｗｉはＮ個の基底関数を使って予測式γ(t)を記述する際の基底関数φｉ（ｔ）の重み係数、ｔは任意の時間を示す。）
着目する領域ｉ＝ｃについて、サーマルコントラスト実測値ｄ（ｔ）＝ｕ _c （ｔ）−ｕ _a （ｔ）を求め、前記予測式構成ステップにおいて構成した予測式「γ（ｔ）」における「係数ｗｉ」を、以下の式を用いて決定する係数決定ステップと、

（但し、φｉ（ｔ ₁ ），φｉ（ｔ ₁ ），φｉ（ｔ ₁ ）（ｉ＝１，２，・・・、Ｎ）はｔ＝ｔ ₁ ，ｔ ₂ ，・・・ｔ _m における基底関数、ｄ（ｔ ₁ ），ｄ（ｔ ₂ ），・・・，ｄ（ｔ _m ）は着目する領域ｉ＝ｃについてｔ＝ｔ ₁ ，ｔ ₂ ，・・・ｔ _m におけるサーマルコントラスト実測値、ｍはサーマルコントラストの実測値の個数を示す。）
前記係数決定ステップにおいて決定した「係数ｗ１〜ｗＮ」を、前述した（２式）に代入して、Ｔ０（ｔ）＋γ（ｔ）により計測している被測定物の将来予測曲線を決定する予測曲線決定ステップと、
を備えることを特徴とする。
また、本発明の容器壁状態の管理方法は、前記係数決定ステップにおいて、Ｎ≦ｍであることを特徴とする。
また、本発明の容器壁状態の管理方法は、前記係数決定ステップにおいて、Ｎ＝ｍであることを特徴とする。 The container wall state management method of the present invention has a temperature difference between the inner wall surface and the outer wall surface, the thermal history information whose temperature changes with time, and the thickness information when the thermal history information is obtained. Is a container wall state management method for estimating the future temperature of each region in a container where the container is known and managing the container wall state,
A procedure for measuring the temperature u _i (t) at time t of each region i obtained by dividing the analysis area set on the outer wall surface of the container into N pieces over time by thermography;
A thermal contrast calculation procedure for calculating a thermal contrast d (t) that is a difference between the temperature u _i (t) of each region i and the average temperature u _a (t) of the analysis area;
The peak time of the thermal contrast in the focused area i = c, and the peak time acquisition procedure for obtaining estimates the total measuring whether we overall temperature changes in a predetermined time t1 to tm,
A peak time distribution obtaining procedure for obtaining a thermal contrast peak time distribution in the analysis area based on a thermal contrast peak time in the region of interest i = c obtained in the peak time obtaining procedure;
Before Kipi over click time acquisition procedure,
A thermal history calculation step of increasing or decreasing the thickness of the container in the area i to be measured, and calculating a thermal history for a plurality of thicknesses by a thermal history calculation means;
From the thermal history, the temperature Ti (t) of each region i at time t and the analysis area for the thickness of the object to be measured that is known and the thickness set to a value around that thickness, respectively. A thermal contrast calculation step of calculating a thermal contrast φi (t) that is a difference from the average temperature T0 (t) from “φi (t) = Ti (t) −T0 (t) 1 ≦ i ≦ N”;
Prediction formula construction step for constructing a prediction formula “γ (t)” as shown below using a plurality of thermal contrasts φi (t) calculated in the thermal contrast calculation step as basis functions for the region of interest i = c When,

(Where N is the number of basis functions φi (t) , i is an index thereof, wi is a weighting coefficient of basis functions φi (t) when describing the prediction formula γ (t) using N basis functions, t represents an arbitrary time.)
For the region of interest i = c, the thermal contrast measured value d (t) = u _c (t) −u _a (t) is obtained, and the “coefficient” in the prediction formula “γ (t)” constructed in the prediction formula construction step is calculated. a coefficient determination step for determining wi "using the following equation:

_{(However, φi (t 1), φi} (t 1), φi (t 1) (i = 1,2, ···, N) is t = t _1, t _2, basis functions in · · · t _{m , d (t 1), d} (t 2), ···, d (t m) is t = t ₁ for the focused area i = c, t _2, thermal contrast Found in · · · t _m, m Indicates the number of measured values of thermal contrast.)
Prediction for determining the future prediction curve of the measurement object measured by T0 (t) + γ (t) by substituting the “coefficients w1 to wN” determined in the coefficient determination step into the above-described (Equation 2). A curve determination step ;
It is characterized by providing.
The container wall state management method of the present invention is characterized in that N ≦ m in the coefficient determination step.
The container wall state management method of the present invention is characterized in that N = m in the coefficient determination step.

（但し、Ｎは基底関数φｉ（ｔ）の個数、ｉはそのインデックス、ｗｉはＮ個の基底関数を使って予測式γ(t)を記述する際の基底関数φｉ（ｔ）の重み係数、ｔは任意の時間を示す。）
着目する領域ｉ＝ｃについて、サーマルコントラスト実測値ｄ（ｔ）＝ｕ _c （ｔ）−ｕ _a （ｔ）を求め、前記予測式構成ステップにおいて構成した予測式「γ（ｔ）」における「係数ｗｉ」を、以下の式を用いて決定する係数決定ステップと、

（但し、φｉ（ｔ ₁ ），φｉ（ｔ ₁ ），φｉ（ｔ ₁ ）（ｉ＝１，２，・・・、Ｎ）はｔ＝ｔ ₁ ，ｔ ₂ ，・・・ｔ _m における基底関数、ｄ（ｔ ₁ ），ｄ（ｔ ₂ ），・・・，ｄ（ｔ _m ）は着目する領域ｉ＝ｃについてｔ＝ｔ ₁ ，ｔ ₂ ，・・・ｔ _m におけるサーマルコントラスト実測値、ｍはサーマルコントラストの実測値の個数を示す。）
前記係数決定ステップにおいて決定した「係数ｗ１〜ｗＮ」を、前述した（２式）に代入して、Ｔ０（ｔ）＋γ（ｔ）により計測している被測定物の将来予測曲線を決定する予測曲線決定ステップと、
を行うことを特徴とする。
また、本発明の容器壁状態の管理装置は、前記係数決定ステップにおいて、Ｎ≦ｍであることを特徴とする。
また、本発明の容器壁状態の管理装置は、前記係数決定ステップにおいて、Ｎ＝ｍであることを特徴とする。 The container wall state management device of the present invention has a temperature difference between the inner wall surface and the outer wall surface, the thermal history information whose temperature changes with time, and the thickness information when the thermal history information is obtained. Is a container wall state management device that estimates the future temperature of each region in a container of which is known and manages the container wall state,
The difference between the temperature u _i (t) at the time t of each region i obtained by dividing the analysis area set on the outer wall surface of the container into N pieces, which is measured over time by thermography, and the average temperature of the analysis area Thermal contrast calculation means for calculating a certain thermal contrast;
The peak time of the thermal contrast in the focused region, the peak time obtaining means for obtaining estimates a temperature change of the entire measurement or al in a predetermined time t1 to tm,
Based on the thermal contrast peak time in the interest region i = c determined by the peak time acquiring unit, and a peak time distribution obtaining means for obtaining a peak time distribution of the thermal contrast in the analysis area,
Before Kipi over click time acquisition means,
A thermal history calculation step of increasing or decreasing the thickness of the container in the area i to be measured, and calculating a thermal history for a plurality of thicknesses by a thermal history calculation means;
From the thermal history, the temperature Ti (t) of each region i at time t and the analysis area for the thickness of the object to be measured that is known and the thickness set to a value around that thickness, respectively. A thermal contrast calculation step of calculating a thermal contrast φi (t) that is a difference from the average temperature T0 (t) from “φi (t) = Ti (t) −T0 (t) 1 ≦ i ≦ N”;
Prediction formula construction step for constructing a prediction formula “γ (t)” as shown below using a plurality of thermal contrasts φi (t) calculated in the thermal contrast calculation step as basis functions for the region of interest i = c When,

(Where N is the number of basis functions φi (t) , i is an index thereof, wi is a weighting coefficient of basis functions φi (t) when describing the prediction formula γ (t) using N basis functions, t represents an arbitrary time.)
For the region of interest i = c, the thermal contrast measured value d (t) = u _c (t) −u _a (t) is obtained, and the “coefficient” in the prediction formula “γ (t)” constructed in the prediction formula construction step is calculated. a coefficient determination step for determining wi "using the following equation:

_{(However, φi (t 1), φi} (t 1), φi (t 1) (i = 1,2, ···, N) is t = t _1, t _2, basis functions in · · · t _{m , d (t 1), d} (t 2), ···, d (t m) is t = t ₁ for the focused area i = c, t _2, thermal contrast Found in · · · t _m, m Indicates the number of measured values of thermal contrast.)
Prediction for determining the future prediction curve of the measurement object measured by T0 (t) + γ (t) by substituting the “coefficients w1 to wN” determined in the coefficient determination step into the above-described (Equation 2). A curve determination step ;
It is characterized by performing.
The container wall state management device of the present invention is characterized in that N ≦ m in the coefficient determination step.
The container wall state management device of the present invention is characterized in that N = m in the coefficient determination step.

（但し、Ｎは基底関数φｉ（ｔ）の個数、ｉはそのインデックス、ｗｉはＮ個の基底関数を使って予測式γ(t)を記述する際の基底関数φｉ（ｔ）の重み係数、ｔは任意の時間を示す。）
着目する領域ｉ＝ｃについて、サーマルコントラスト実測値ｄ（ｔ）＝ｕ _c （ｔ）−ｕ _a （ｔ）を求め、前記予測式構成ステップにおいて構成した予測式「γ（ｔ）」における「係数ｗｉ」を、以下の式を用いて決定する係数決定ステップと、

（但し、φｉ（ｔ ₁ ），φｉ（ｔ ₁ ），φｉ（ｔ ₁ ）（ｉ＝１，２，・・・、Ｎ）はｔ＝ｔ ₁ ，ｔ ₂ ，・・・ｔ _m における基底関数、ｄ（ｔ ₁ ），ｄ（ｔ ₂ ），・・・，ｄ（ｔ _m ）は着目する領域ｉ＝ｃについてｔ＝ｔ ₁ ，ｔ ₂ ，・・・ｔ _m におけるサーマルコントラスト実測値、ｍはサーマルコントラストの実測値の個数を示す。）
前記係数決定ステップにおいて決定した「係数ｗ１〜ｗＮ」を、前述した（２式）に代入して、Ｔ０（ｔ）＋γ（ｔ）により計測している被測定物の将来予測曲線を決定する予測曲線決定ステップとを有することを特徴とする。
また、本発明のコンピュータプログラムは、前記係数決定ステップにおいて、Ｎ≦ｍであることを特徴とする。
また、本発明のコンピュータプログラムは、前記係数決定ステップにおいて、Ｎ＝ｍであることを特徴とする。
なお、本発明でいう「内壁表面と外壁表面とで温度差を有する容器」とは、転炉、脱ガス炉、石炭ガス化反応炉等の高温のガス反応又は液体反応を伴う反応容器及び混銑車、溶銑鍋、溶鋼鍋等の溶鉄を運搬する容器等をいう。 The computer program of the present invention has a temperature difference between the inner wall surface and the outer wall surface, heat history information whose temperature changes with time, and thickness information when the heat history information is obtained. A computer program for estimating the future temperature of each region in the container and managing the state of the container wall,
The difference between the temperature u _i (t) at the time t of each region i obtained by dividing the analysis area set on the outer wall surface of the container into N pieces, which is measured over time by thermography, and the average temperature of the analysis area Thermal contrast calculation processing for calculating a certain thermal contrast;
The peak time of the thermal contrast in the focused region, the peak time obtaining processing to obtain estimates the total measuring whether we overall temperature changes in a predetermined time t1 to tm,
Based on the thermal contrast peak time in the region i = c to the focus determined by the peak time acquisition process is performed and a peak time distribution obtaining process of obtaining the peak time distribution of the thermal contrast in the computer in the analysis area,
Before Kipi over click time acquisition process,
A thermal history calculation step of increasing or decreasing the thickness of the container in the area i to be measured, and calculating a thermal history for a plurality of thicknesses by a thermal history calculation means;
From the thermal history, the temperature Ti (t) of each region i at time t and the analysis area for the thickness of the object to be measured that is known and the thickness set to a value around that thickness, respectively. A thermal contrast calculation step of calculating a thermal contrast φi (t) that is a difference from the average temperature T0 (t) from “φi (t) = Ti (t) −T0 (t) 1 ≦ i ≦ N”;
Prediction formula construction step for constructing a prediction formula “γ (t)” as shown below using a plurality of thermal contrasts φi (t) calculated in the thermal contrast calculation step as basis functions for the region of interest i = c When,

(Where N is the number of basis functions φi (t) , i is an index thereof, wi is a weighting coefficient of basis functions φi (t) when describing the prediction formula γ (t) using N basis functions, t represents an arbitrary time.)
For the region of interest i = c, the thermal contrast measured value d (t) = u _c (t) −u _a (t) is obtained, and the “coefficient” in the prediction formula “γ (t)” constructed in the prediction formula construction step is calculated. a coefficient determination step for determining wi "using the following equation:

_{(However, φi (t 1), φi} (t 1), φi (t 1) (i = 1,2, ···, N) is t = t _1, t _2, basis functions in · · · t _{m , d (t 1), d} (t 2), ···, d (t m) is t = t ₁ for the focused area i = c, t _2, thermal contrast Found in · · · t _m, m Indicates the number of measured values of thermal contrast.)
Prediction for determining the future prediction curve of the measurement object measured by T0 (t) + γ (t) by substituting the “coefficients w1 to wN” determined in the coefficient determination step into the above-described (Equation 2). A curve determining step.
The computer program of the present invention is characterized in that N ≦ m in the coefficient determining step.
The computer program of the present invention is characterized in that N = m in the coefficient determining step.
The “vessel having a temperature difference between the inner wall surface and the outer wall surface” as used in the present invention means a reaction vessel and kneading involving a high-temperature gas reaction or liquid reaction, such as a converter, a degassing furnace, and a coal gasification reactor. A container that carries molten iron, such as a car, hot metal ladle, and molten steel pan.

According to the present invention, there is a temperature difference between the inner wall surface and the outer wall surface, the heat history information in which the temperature changes with time, and the thickness information when the heat history information is obtained is known. When estimating the future temperature of each area in the container and managing the container state, the container wall state of the inner wall of the container wall (for example, the worn state) can be accurately estimated over a wide range, and the container wall state can be managed with high accuracy. In addition, the container wall state can be managed with high accuracy even if the temperature measurement data has only a partial period.
In particular, there is a process that has a high temperature material inside like a molten steel pan, there is a temperature difference between the inner surface and the outer surface, the material temperature changes unsteadily, the thickness of the container wall inner wall changes due to wear etc. The behavior of the future temperature of the container wall state is estimated from a relatively short temperature measurement period, and the container wall state can be managed with high accuracy without being limited to the period in which the temperature measurement data exists.

Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.
(First embodiment)
In this embodiment, a molten steel pan 300 will be described as an example of a container having a temperature difference between the inner wall surface and the outer wall surface. As shown in FIG. 1A, the temperature distribution on the surface of the outer wall (iron skin) 301 of the molten steel pan 300 is measured by the infrared thermography 200. The temperature distribution data measured by the infrared thermography 200 is input to the container wall state management device 100.

As shown in FIG. 1 (b), when the inner wall (refractory material such as wear brick) 302 of the molten steel pan 300 is melted (melting depth d), on the surface of the outer wall 301 of the molten steel pan 300 after receiving steel. , the temperature u _c of the portion corresponding to the melting point, to reach a fast steady state with even a temperature higher than u _a surrounding (see Figure 2).

  Hereinafter, based on the flowchart of FIG. 3, the container wall state management method according to the present embodiment will be described with reference to FIGS. FIG. 3 is a flowchart illustrating an example of a processing procedure of the container wall state management method according to the present embodiment. The container wall state management device 100 estimates and manages the container wall state, specifically, the remaining thickness of the container wall, by performing the processes of steps S101 to S105 in FIG.

First, in step S101, the analysis area set on the outer wall 301 surface of the molten steel pan 300 is divided into a plurality of appropriate regions (see FIG. 4), and each region i (i = 1, 2,..., N). The temperature u _i (t) is measured with the infrared thermography 200 over time. As the analysis area, the entire area measured by the infrared thermography 200 or an area appropriately selected from the entire measured area can be set. FIG. 5 shows an example of the temperature u _i (t) of each region i measured by the infrared thermography 200. The region i can be set as a pixel unit of the detection element, but a plurality of pixels existing in the vicinity can be collectively calculated as the temperature of the region i. For example, 3840 pixels (= 32 vertical pixels × 120 horizontal pixels) can be set as one area.

  In the present embodiment, the slag line portion of the surface of the outer wall 301 of the molten steel pan 300 is set as the analysis area. In the molten steel pan 300, the heat load is larger in the upper layer (slag line portion) than in the lower layer (metal line portion), and the number of repairs and replacements is increased, so that it is particularly required to manage the remaining thickness of the slag line portion. Because. Also, for example, when the material is different between the upper and lower parts of the slag line, or when the refractory replacement frequency is different between the upper and lower parts, as shown in FIG. May be set as

Next, in step S102, an average temperature of the analysis area (average temperature of all the regions i included in the analysis area) u _a (t) is calculated, and the temperature u _i (t) of each region i and the average of the analysis areas are calculated. The difference (u _i (t) −u _a (t)) from the temperature u _a (t) is calculated. As shown in FIG. 5, when the abnormal value 51 exists for reasons such as slag adhesion, the abnormal value 51 may be excluded.

In this specification, the difference _{(u i (t) -u a} (t)) and the average temperature u _a analysis area temperature u _i (t) of each region i (t) is referred to as "thermal contrast". FIG. 6 shows the thermal contrast in a certain region i. The thermal contrast can be obtained by actually measuring (u _i (t) −u _a (t)) in _a part or a plurality of time zones and obtaining _a prediction curve by curve approximation. This prediction curve is obtained in each region i = 1 to N. Details will be described later.

  Next, in step S103, the thermal contrast peak time in each region i (the time when the thermal contrast is maximized) is obtained. As shown in FIG. 7, the peak time of thermal contrast in each region i shifts to an earlier time after receiving steel as the melting depth d increases (in other words, the remaining thickness decreases). To do. That is, as the melting depth d becomes deeper (in other words, as the remaining thickness becomes thinner), the thermal contrast becomes maximum (peak) at an early stage after receiving steel.

  Further, by observing the peak time of the thermal contrast, it is possible to avoid the influence of the empty pan time (initial temperature distribution before receiving steel). As shown in FIG. 8 (a), even if the melting size is the same, the temperature of the outer wall surface of the molten steel pan changes if the empty pan time is different. That is, when only the temperature of the outer wall surface is observed, it is affected by the empty pan time. Also, the temperature peak time is affected by the pan time. On the other hand, when observing the peak time of the thermal contrast, as shown in FIG. 8B, the value of the thermal contrast itself is changed even if the pan time is different, but the peak time of the thermal contrast is almost fluctuated. No (not affected by empty pot time).

  Furthermore, by observing the peak time of the thermal contrast, the influence of the three-dimensional shape of the melting damage can be avoided. As shown in FIG. 9 (a), when the erosion width (the erosion depth d is the same) is different, the temperature of the outer wall surface of the molten steel pan changes. That is, when only the temperature of the outer wall surface is observed, it is affected by the three-dimensional shape of the melting. On the other hand, when observing the peak time of the thermal contrast, as shown in FIG. 9B, the value of the thermal contrast itself changes even if the erosion width (the erosion depth is the same) is different. The peak time of thermal contrast hardly fluctuates (it is hardly affected by the three-dimensional shape of melting).

  Returning to FIG. 3, in step S104, the peak time distribution of the thermal contrast in the analysis area is examined based on the peak time of the thermal contrast in each region i. FIG. 10A shows the peak time distribution of the thermal contrast in the second use (almost new) of the molten steel pan 300. On the other hand, FIG.10 (b) shows the peak time distribution of the thermal contrast in the 17th use (immediately repair) of the molten steel pan 300. FIG. As shown in the figure, as the number of times of use is increased, the dispersion width of the peak time distribution is increased, and it can be seen that the inner wall 302 of the molten steel pan 300 is being melted.

In the present embodiment, as shown in FIG. 11, three statistics of average peak time S _ave , dispersion time S _Σ , and minimum peak time S _min are managed in the peak time distribution of thermal contrast.

  Next, in step S105, based on the thermal contrast peak time distribution in the analysis area, the residual thickness of the container wall is estimated using, for example, inverse problem analysis. For example, as shown in FIG. 12, a calibration curve representing an increase / decrease from the average remaining thickness with respect to an increase / decrease from the average peak time is determined in advance from the result of thickness measurement at the time of repair or empty pan, etc. The residual thickness of the container wall is converted based on the peak time distribution of the thermal contrast by adding the average residual thickness obtained by the thickness calculation by measurement or inverse problem analysis. The method for setting the average remaining thickness by the inverse problem analysis calculation is not particularly limited, but an example thereof will be described later.

As shown in FIG. 7, the peak time of the thermal contrast shifts to an earlier time after receiving the steel as the melting depth d becomes deeper (in other words, as the remaining thickness becomes thinner). The residual thickness distribution of the wall also corresponds to the peak time distribution of the thermal contrast. That is, the average peak time S _ave corresponds to an average residual thickness, minimum peak time S _min is the residual thickness of the container wall corresponding to the minimum residual thickness, area i to be the minimum time S _min at the peak time distribution curve of the thermal contrast By specifying this, it is possible to specify the region i in which the remaining thickness of the container wall is the minimum remaining thickness.

  FIG. 13 is a residual thickness distribution of the container wall obtained using the calibration curve of FIG. Fig.13 (a) shows the residual thickness distribution of the container wall in the 2nd use of the molten steel pan 300 (corresponding | compatible to Fig.10 (a)). On the other hand, FIG.13 (b) shows the residual thickness distribution of the container wall in the 17th use of the molten steel pan 300 (corresponding | compatible to FIG.10 (b)). As the number of times of use increases, the average remaining thickness decreases, the dispersion width of the remaining thickness distribution increases, and the value of the minimum remaining thickness decreases. .

In the container wall state management apparatus 100, the processing of steps S101 to S105 in FIG. 3 is executed every time the molten steel pan 300 is used, that is, every time the steel is received, for example.
Then, the average peak time S _ave , dispersion time S _Σ , minimum peak time S _min in the peak time distribution of thermal contrast obtained each time, and the remaining thickness of the container wall obtained by converting them are made into a database and a history is recorded. to manage. In particular, in the daily inspection of the molten steel pan 300, the region i where the remaining thickness of the container wall is the minimum remaining thickness may be managed with priority.

  As shown in FIG. 14, when the melting loss of the inner wall 302 progresses with repeated use of the molten steel pan 300, the value of the minimum remaining thickness of the remaining thickness distribution S of the container wall decreases. For example, the threshold value Y of the minimum remaining thickness is set in advance in the container wall state management device 100, and when the value of the minimum remaining thickness reaches the threshold value, it is determined that the life of the molten steel pan 300 is reached. That's fine. In this case, the container wall state management device 100 may notify the user that the life of the molten steel pan 300 is reached and prompt the user to repair the molten steel pan 300 or replace a necessary part.

  The residual thickness of the container wall is measured at the time of repair (replacement) of the molten steel pan 300, and the calibration curve shown in FIG. 12 is corrected based on the measured value, thereby improving the estimation accuracy of the residual thickness. May be.

  As described above, since the thermal contrast peak time shifts in accordance with the remaining thickness of the container wall, the remaining thickness of the container wall can be estimated. In addition, by observing the peak time of the thermal contrast, it is possible to avoid the influence of the empty pan time (initial temperature distribution before receiving steel) and the three-dimensional shape of the melting loss. Therefore, based on the two-dimensional temperature measurement data on the outer wall surface of the molten steel pan 300, the state of the container wall, specifically, the remaining thickness of the container wall and the maximum wear amount of the three-dimensional wear state of the container wall are accurately estimated and managed. can do.

In the above embodiment, the remaining thickness of the container wall is estimated as the container wall state based on the peak time distribution of the thermal contrast. However, it is also possible to detect the presence or absence of a so-called “barrel”. Is possible. A metal bar means that a crack occurs in the inner wall of the molten steel pan 300 and the molten steel enters there. As shown in FIGS. 15 (a) and 15 (b), when a bullion is generated, the bulge is inserted in addition to the average peak time _Save in the peak time distribution of the thermal contrast due to the difference in thermal conductivity with other parts. The peak due to.
Therefore, when a peak appears in addition to the average peak time _Save , it can be estimated that a bullion has occurred.

  FIG. 16 shows a hardware configuration example of a computer system that functions as the container wall state management device 100. The container wall state management device 100 includes a CPU 20, an input device 21, a display device 22, and a recording device 23, and each unit is connected via a bus 24. The recording device 23 includes a ROM, a RAM, an HD, and the like, and stores a computer program that controls the operation of the container wall state management device 100 described above. The function or process of the container wall state management apparatus 100 is realized by the CPU 20 executing the computer program. A database is stored in the recording device 23.

(Second Embodiment)
Next, a preferred embodiment of the present invention will be described in which the thermal contrast is measured (u _i (t) −u _a (t)) in part or in a plurality of time zones and _a prediction curve is obtained by curve approximation.
In the present embodiment, in the analysis area divided into a plurality of regions shown in FIG. 4, for each specific region (for example, i = 20), the future behavior of temperature is estimated to manage the container wall state. I am doing so. Hereinafter, an example of the estimation procedure will be described with reference to the flowchart of FIG.

First, in the first step S81, the thermal history is calculated for each thickness of the object to be measured.
As described above, in the present embodiment, the molten steel pan 300 is used as the object to be measured, and the wall state of the molten steel pan 300 is managed. In the molten steel pan 300, a state in which the molten steel is put inside (hereinafter referred to as a steel receiving pan) and an empty state (hereinafter referred to as an empty pan) are alternately repeated. In FIG. 18 (a), the heat history when the cycle of receiving steel pan → empty pan → receiving steel pan → empty pan... Is repeated four times is shown for each remaining thickness of the wear brick. As shown in FIG. 18 (a), in the iron making process, it is known what operation was performed before, so the thermal history of the iron skin temperature is calculated for each thickness of the molten steel pan refractory (estimation). can do. In this case, the example which performs the one-dimensional heat transfer calculation of the surface temperature of the molten steel pan 300 using the time history (received steel pan, empty pan) of the molten steel pan 300 from before 4 times is shown.

Specifically, the ladle 300, as shown in FIG. 22, for example, information of temperature change u _a in the analysis area in the previous operation are known. Therefore, by changing the thickness of the molten steel pan 300 (in this case, wear brick) when the information on the temperature change u _a is obtained, as shown by reference numeral 30 in FIG. Calculate the heat history for. In this example, the remaining thicknesses of the wear bricks are calculated for “30 mm”, “50 mm”, “70 mm”, “90 mm”, “110 mm”, “130 mm”, and “150 mm”, respectively. Specifically, the constituent material of the molten steel pan 300 is expressed by a one-dimensional approximate unsteady heat conduction equation and calculated using a numerical calculation method such as a difference method. That is, the one-dimensional unsteady transmission of the surface temperature of the molten steel pan 300 using the history of the change in the remaining thickness of the ware brick of the molten steel pan and the operation time history (receiving steel pan, empty pan) from the previous four times (when new). It can be obtained by performing thermal calculation.

Next, it progresses to step S82 and calculates thermal contrast (phi) i with respect to the thickness of the molten steel pan 300 known from the last operation.
This is defined as “thermal contrast φi”, which is defined as the difference from the iron skin temperature Ti at each refractory thickness calculated at step S81 on the basis of the iron skin temperature T0 at the previous refractory thickness calculated at step S81. And do it.
φi = Ti−T0 1 ≦ i ≦ N (1 formula)

  The above-described calculation according to (Expression 1) is performed for each of the plurality of refractory thicknesses calculated in step S81. For example, as shown in FIG. 18 (b), the amount of erosion when the remaining thickness of the wear brick is 110 mm is “−80 mm”, “−60 mm”, “−40 mm”, “−20 mm”, “+20 mm”. , “+40 mm”. Thereby, “thermal contrast φ1” to “thermal contrast φ6” are obtained for each refractory thickness. Here, the calculation is performed excluding the melting loss of 0 mm, but of course, the calculation may be performed including the melting loss of 0 mm.

  Next, the process proceeds to step S83, and a thermal contrast change prediction formula “γ (t)” in “region i = 20” is configured. In the present embodiment, an example is shown in which the prediction formula “γ (t)” is configured by the following (formula 2) using the thermal contrast φi as a basis function.

In (Formula 2), the “coefficient wi” is unknown. This unknown “coefficient wi” is determined in step S84. First, in the analysis area of the molten steel pan 300, the temperature u _c (for example, i = 20) of the region of interest is measured for a predetermined time by the infrared thermography 200, and “the measured thermal contrast value d (t) = u _{c is obtained.} (T) -u _a (t) ”is obtained. The time for measuring the temperature is, for example, about 10 minutes.

Next, in step S85, a future temperature prediction curve is determined. This is achieved by solving the equation obtained by substituting the “measured thermal contrast values d (t1) to d (tm)” acquired in step S84 into the following (Equation 3) to obtain “coefficients w _{1 to} w _N. " Then, by substituting the calculated “coefficients w _{1 to} w _N ” into the above-described (Equation 2), a future prediction curve of the actual measured value of thermal contrast can be determined. The future temperature prediction formula in the “region i = 20” focused at this time is “T0 (t) + γ (t)”. The remaining thickness of the wear brick can be determined by multiplying each remaining thickness of the wear brick used in “thermal contrast φ ₁ ” to “thermal contrast φ _N ” by wi / Σwi and performing a weighted average. It is preferable to set m equal to N or larger than N because it is easy to obtain. In particular, “coefficients w1 to wN” in the case where m is equal to N are more preferable because they can be uniquely determined. “Coefficients w _{1 to} w _N ” when m is larger than _N can be obtained by calculation as an approximate solution using the least square method or the like.

  FIG. 19 shows the relationship between a plurality of thermal contrast basis functions φ1 to φ6, which are determined as described above, around the known thickness and the elapsed time (minutes) from the steel receiving. A prediction formula “γ (t)” of the thermal contrast change is constructed in FIG. 19 by superimposing these thermal contrast basis functions φ1 to φ6. In FIG. 19, the vertical axis represents temperature (° C.) and the horizontal axis represents time (minutes).

  FIG. 20 is an enlarged view of the thermal contrast change prediction formula “γ (t)” in FIG. Although FIG. 20 shows an example in which noise removal is performed, noise removal is not necessarily performed. Noise removal may be performed by a commonly used method such as a moving average method.

FIG. 21 shows a prediction error of the future temperature predicted by using the thermal contrast change prediction formula “γ (t)” of the present embodiment and a future temperature by a conventional prediction method, for example, a general cubic spline interpolation method. A prediction result is shown. In the future temperature prediction curve 61 predicted by the conventional prediction method, the error increases extremely after 20 minutes have passed. In contrast, the future temperature prediction curve 62 predicted by the prediction method of the present embodiment has an error of 1 ° C. or less even after 1 hour, and it can be seen that good accuracy can be maintained.
Here, the cubic spline interpolation formula means that when data of a function y _i = y (x _i ) i = 1, 2,..., N is given, from x _j to x _{j + 1.} In this section, interpolation is performed using the following formula.

  As described above, according to the container wall state management method of the present embodiment, the thickness of the ladle 300 is calculated based on the time transition data of a certain period of temperature measured in a predetermined region of the ladle 300. decide. And based on the thickness information of the determined molten steel pan 300 and the thermal history information of the molten steel pan 300, the future behavior of the time transition of the temperature of the molten steel pan 300 is predicted.

  Thereby, it is possible to reliably predict how the future temperature of the molten steel pan 300 changes according to the time transition only by performing the one-dimensional heat transfer calculation. Moreover, the thickness of each area | region in the said molten steel pan 300 can be reliably estimated based on the peak time of a thermal contrast prediction curve.

  In particular, in the present embodiment, the thermal history of the surface temperature is calculated for the known thickness and the thickness before and after the known thickness, and the calculated known thickness is calculated based on the temperature T0 of the known thickness. A difference in temperature Ti between the front and rear thicknesses was calculated using the above (1), and a plurality of thermal contrasts φ1 to φ6 were defined. Then, using these thermal contrasts φ1 to φ6 as basis functions, a prediction formula γ (t) is constructed, and the future temperature of the molten steel pan 300 is predicted. Therefore, the state of the molten steel pan 300 can be predicted with extremely high accuracy.

(Third embodiment)
Next, a third embodiment of the present invention will be described with reference to FIG.
First, in the first step S91, with respect to the molten steel pan 300, the thermal history is calculated for a plurality of thicknesses by changing the thickness of the molten steel pan 300 when the average temperature change u _a of the analysis area is obtained. This calculation is the same as the calculation performed in step S81 of FIG. 23 in the second embodiment described above, and based on the time history and temperature change history of the molten steel pan 300, the molten steel pan known from the previous measurement. The thermal history is calculated for each of a plurality of thicknesses before and after the thickness of 300.

Next, it progresses to step S92 and the temperature of the specific area | region of the molten steel ladle 300, ie, "area | region i = 20, is measured. As a result of this measurement, a “measured value u _c (t)” for a certain time in “region i = 20” is obtained.

Next, the process proceeds to step S93, and a thermal history and temperature change characteristic curve corresponding to the “measured value u _c (t)” measured in step S92 for a certain time is specified. Then, from the specified characteristic curve, the thickness of the molten steel pan 300 in “region i = 20” is specified.

  Next, it progresses to step S94 and future temperature is estimated from the thickness of the ladle 300 in the "area | region i = 20" specified in step S93. For the estimation of the future temperature, for example, a prediction curve can be obtained by curve approximation based on time transition data of a certain period of temperature measured in “region i = 20”.

  As described above, also in the present embodiment, it is possible to reliably predict how the future temperature of each region in the molten steel pan 300 changes according to the time transition only by performing the one-dimensional heat transfer calculation. Moreover, thickness can be reliably estimated about each area | region of the said steel ladle 300 from the weighting coefficient when comprising a prediction curve. Therefore, the container wall state of the molten steel pan 300 can be managed well based on the thickness information estimated as described above.

  The container wall state management device of the present invention may be applied to a system composed of a plurality of devices or an apparatus composed of a single device.

  The object of the present invention can also be achieved by supplying a computer program for realizing the above-described functions to a system or apparatus and executing the computer (CPU or MPU) of the system or apparatus. In this case, the computer program itself Constitutes the present invention.

  As mentioned above, although this invention was demonstrated with various embodiment, this invention is not limited only to these embodiment, A change etc. are possible within the scope of the present invention. For example, in the above embodiment, the number of infrared thermography 200 is not mentioned, but as shown in FIG. 17, three or more infrared thermography 200 are evenly arranged around the molten steel pan 300 and the entire circumference of the molten steel pan 300 is obtained. You may make it manage the molten steel wall state in.

Hereinafter, an example of a method for determining the average remaining thickness by inverse problem analysis calculation will be described. First, the basic concept of the container wall thickness estimation method will be described. FIG. 25 is a diagram showing a part of the container wall, where x = 0 is the position of the inner wall surface of the container. In the figure, the remaining thickness l of the container wall, the temperature f (t) = U _{M of} the high-temperature substance existing in the container, the temperature u (x, t) of the container wall (measured at the temperature measurement point of the outer wall surface). Temperature h (t)) and heat flux calculated based on temperature h (t) (or heat flux measured at a temperature measurement point on the outer wall surface) g (t).

(Formulation)
(Formula 4) and (Formula 5) represent unsteady heat conduction equations. Note that u _t represents ∂u / 、 _t , and u _xx represents ∂ ² u / ∂x ² . In (Expression 4), α is a thermal diffusion coefficient, and u (x, 0) = u ₀ (x) is an initial value of the temperature of the container wall. In this case, the initial value u (x, 0) = u ₀ (x) of the temperature of the container wall is unknown. Further, the temperature U _{M of the} high-temperature substance, the outside air temperature u _a , the thermal diffusion coefficient α, the Stefan Boltzmann coefficient σ of radiant heat transfer, and the thermal conductivity λ of the container wall are positive constants.

  Here, by introducing (Expression 6) and performing Fourier expansion, the temperature u (x, t) of the container wall of (Expression 4) can be obtained as (Expression 7).

  If x = 1, (Equation 8) is obtained.

  However, when performing arithmetic processing by a computer, there is a problem that the calculation of the second term and the third term in particular on the right side of (Equation 8) tends to cause a truncation error. Therefore, a variable v (x, t) is defined as an alternative to the container wall temperature u (x, t), and (Equation 9) is introduced. In (Equation 9), the initial value v (x, 0) = xg (0) + f (0) of the variable is the initial value g (0) of the heat flux of the inner wall surface and the initial value of the temperature of the inner wall surface ( That is, the temperature of the high-temperature substance) f (0) can be known. Therefore, the variable v (x, t) can be directly calculated by the backward difference method, for example.

  Furthermore, when the difference between the temperature u (x, t) of the container wall and the variable v (x, t) is defined as w (x, t), (Expression 7) is obtained.

  Here, by performing Fourier expansion in the same manner as described above, the variable w (x, t) in (Equation 10) is obtained as in (Equation 11). In (Expression 11), as is clear as compared with (Expression 7), it can be a simple expression without the second and third terms.

Further, w (l, t) = u (l, t) −v (l, t). U (l, t) is a known h (t), and v (l, t) can be directly calculated from the equation (9) by the backward difference method. Accordingly, assuming that w (l, t) is known, an approximate value of B _n (l) can be obtained from (Equation 11).

(Inverse problem for obtaining the remaining thickness l)
In the inverse problem, the remaining thickness l of the container wall is unknown, so the variable v (l, t) is unknown, but u (l, t) is given as the measured value h (t). (Expression 12) is obtained from (Expression 11).

As described below, an approximate value of the remaining thickness l of the container wall can be obtained by optimization calculation. That is, the observation time is set as (T _st , T _end ), and T _st <T ₁ <T ₂ <T _end . Then, T ₁ = t ₁ <t ₂ <... <T _M = T ₂ and a uniform lattice. The assumed value l to> 0 of the remaining thickness is a known condition. In addition, in this specification, the notation of l assumes that ~ is attached on l.

Here, the temperature h which is measured at a temperature measurement point of the outer wall surface (t) is known, the variable v (l ^~, t _i) by the backward difference method by providing a ^~ assumed value l (9 type) Desired. As in (Equation 13), MA is an M × (N + 1) matrix, VB is an (N + 1) × 1 vector, Vb is an M × 1 vector, and B ₀ (l _{^{~), B 1 (l ~}} ), ···, B N (l ~) is required.

Then, define a w by actual measurement (l, t) and the calculation by w ^(l ~, t) represents the difference between (14 type). Substituting B _i (l ^to ) into (Equation 11) as t∈ (T ₂ , T _end ) yields p (l ^to , t), so that p (l ^to , t) approaches 0. by selecting the assumed value l ^~ of the remaining thickness, can be determined ^- the assumed value l as an approximation of the remaining thickness l of the container wall.

FIG. 26 is a flowchart for explaining the container wall thickness estimation processing based on the inverse problem analysis described above. When the temperature h (t _i ) measured at the temperature measurement point on the outer wall surface of the container is input (step S111), the heat flux g (t _i ) is calculated based on the temperature h (t _i ) ( Step S112). It should be noted that the heat flux g (t _i ) measured at the temperature measurement point on the outer wall surface may be input, and in that case, calculation of the heat flux is not necessary.

Subsequently, an assumed value l˜ of the remaining thickness is first set (step S113). Then, by providing the assumed value l to, determine the (Expression 6) from the variable v ^(l ~, t _i) (step S114). Then, B ₀ (l ^to ), B ₁ (l ^to ),..., B _N (l ^to ) are obtained by solving MA × VB = Vb in (Equation 15) (step S115), (16 P (l ^˜ , t) in the formula is calculated (step S116). Steps S113 to S116 are repeated until p (l ^to , t) that is less than or equal to the set value ε is obtained (step S117), and an assumed value l ^to when the value is less than or equal to the set value ε is determined as the remaining thickness l of the container wall. (Step S118).

(A) is a figure which shows the state which is measuring the temperature distribution of the outer wall surface of a molten steel pan with an infrared thermography, (b) is a figure which shows the state in the vicinity of a erosion loss. It is a characteristic view which shows the relationship between time and the temperature of the outer wall surface of a molten steel pan. It is a flowchart which shows an example of the process sequence of the management method of the container wall state which concerns on this embodiment. It is a figure which shows the image which divided | segmented the analysis area of the outer wall surface of a molten steel pan into a suitable some area | region. It is a characteristic view which shows an example of the temperature of each area | region measured by the infrared thermography. It is a characteristic view which shows the relationship between the elapsed time from steel receiving, and thermal contrast. It is a characteristic view which shows the relationship between the elapsed time from steel receiving, and thermal contrast. (A) is a characteristic view which shows the relationship between the elapsed time from steel receiving, and a skin temperature, (b) is a characteristic diagram which shows the relationship between the elapsed time from receiving steel and thermal contrast. (A) is a characteristic view which shows the relationship between the elapsed time from steel receiving, and a skin temperature, (b) is a characteristic diagram which shows the relationship between the elapsed time from receiving steel and thermal contrast. (A) is a characteristic diagram which shows the peak time distribution of the thermal contrast in the 2nd use, (b) is a characteristic diagram which shows the peak time distribution of the thermal contrast in the 17th use. It is a characteristic view which shows the peak time distribution of thermal contrast. FIG. 4 is a characteristic diagram showing a calibration curve representing an increase / decrease from an average peak time and an increase / decrease from an average remaining thickness. (A) is a characteristic view which shows the remaining thickness distribution in the 2nd use, (b) is a characteristic diagram which shows the remaining thickness distribution in the 17th use. It is a figure for demonstrating the state from which melting loss progresses and the value of the minimum residual thickness of the residual thickness distribution of a container wall becomes small. (A) is a characteristic diagram which shows the peak time distribution of thermal contrast when there is a bullion insert, and (b) is a characteristic diagram showing the peak time distribution of thermal contrast when there is no bullion insert. It is a figure which shows the hardware structural example of the computer system which functions as a container wall state management apparatus. It is a figure which shows the example of arrangement | positioning of infrared thermography. (A) is a characteristic diagram showing an example of performing one-dimensional heat transfer calculation of the surface temperature of the molten steel pan using the time history of the molten steel pan from the previous four times (receiving steel pan, empty pan), (b ) Is a characteristic diagram showing an example of calculating a plurality of thermal contrasts by changing the thickness of the molten steel pan measured last time. It is a figure which shows the relationship between the several thermal-contrast basis function before and behind known thickness, and the elapsed time from receiving steel. It is the characteristic view which expanded and showed the prediction formula (gamma) (t) of the thermal contrast change in FIG. It is a characteristic view which shows the prediction error of the future temperature estimated using the prediction formula (gamma) (t) of the thermal contrast change of this embodiment, and the future temperature prediction result by the conventional prediction method. It is a characteristic view which shows the temperature of the area | region which is paying attention in this embodiment, and the change of the average temperature of an analysis area. It is a flowchart which shows 2nd Embodiment of the future temperature estimation method of a molten steel pan, and demonstrates an example of the procedure which estimates the behavior of the future temperature of a molten steel pan using the prediction curve of thermal contrast. 3rd Embodiment of the future temperature estimation method of a to-be-measured object is shown, Based on the time transition data of a certain period of the temperature measured in a part of to-be-measured object, curve future approximation of the to-be-measured object's future temperature It is a flowchart explaining an example of the procedure which estimates a behavior. It is a figure showing a part of container wall for demonstrating an example of the method of setting average residual thickness by inverse problem analysis calculation. It is a flowchart for demonstrating the thickness estimation process of a container wall by an inverse problem analysis.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Container wall state management apparatus 200 Infrared thermography 300 Molten steel pan 301 Outer wall 302 Inner wall

Claims (9)

The future temperature of each region in the container having a temperature difference between the inner wall surface and the outer wall surface, the heat history information whose temperature changes with time, and the thickness information when the heat history information is obtained is known Is a container wall state management method for managing the container wall state by estimating
A procedure for measuring the temperature u _i (t) at time t of each region i obtained by dividing the analysis area set on the outer wall surface of the container into N pieces over time by thermography;
A thermal contrast calculation procedure for calculating a thermal contrast d (t) that is a difference between the temperature u _i (t) of each region i and the average temperature u _a (t) of the analysis area;
The peak time of the thermal contrast in the focused area i = c, and the peak time acquisition procedure for obtaining estimates the total measuring whether we overall temperature changes in a predetermined time t1 to tm,
A peak time distribution obtaining procedure for obtaining a thermal contrast peak time distribution in the analysis area based on a thermal contrast peak time in the region of interest i = c obtained in the peak time obtaining procedure;
Before Kipi over click time acquisition procedure,
A thermal history calculation step of increasing or decreasing the thickness of the container in the area i to be measured, and calculating a thermal history for a plurality of thicknesses by a thermal history calculation means;
From the thermal history, the temperature Ti (t) of each region i at time t and the analysis area for the thickness of the object to be measured that is known and the thickness set to a value around that thickness, respectively. A thermal contrast calculation step of calculating a thermal contrast φi (t) that is a difference from the average temperature T0 (t) from “φi (t) = Ti (t) −T0 (t) 1 ≦ i ≦ N”;
Prediction formula construction step for constructing a prediction formula “γ (t)” as shown below using a plurality of thermal contrasts φi (t) calculated in the thermal contrast calculation step as basis functions for the region of interest i = c When,

(Where N is the number of basis functions φi (t) , i is an index thereof, wi is a weighting coefficient of basis functions φi (t) when describing the prediction formula γ (t) using N basis functions, t represents an arbitrary time.)
For the region of interest i = c, the thermal contrast measured value d (t) = u _c (t) −u _a (t) is obtained, and the “coefficient” in the prediction formula “γ (t)” constructed in the prediction formula construction step is calculated. a coefficient determination step for determining wi "using the following equation:

_{(However, φi (t 1), φi} (t 1), φi (t 1) (i = 1,2, ···, N) is t = t _1, t _2, basis functions in · · · t _{m , d (t 1), d} (t 2), ···, d (t m) is t = t ₁ for the focused area i = c, t _2, thermal contrast Found in · · · t _m, m Indicates the number of measured values of thermal contrast.)
Prediction for determining the future prediction curve of the measurement object measured by T0 (t) + γ (t) by substituting the “coefficients w1 to wN” determined in the coefficient determination step into the above-described (Equation 2). A curve determination step ;
A container wall state management method comprising:   The container wall state management method according to claim 1, wherein N ≦ m in the coefficient determination step.   3. The container wall state management method according to claim 2, wherein N = m in the coefficient determination step. The future temperature of each region in the container having a temperature difference between the inner wall surface and the outer wall surface, the heat history information whose temperature changes with time, and the thickness information when the heat history information is obtained is known A container wall state management device for managing the container wall state by estimating
The difference between the temperature u _i (t) at the time t of each region i obtained by dividing the analysis area set on the outer wall surface of the container into N pieces, which is measured over time by thermography, and the average temperature of the analysis area Thermal contrast calculation means for calculating a certain thermal contrast;
The peak time of the thermal contrast in the focused area i = c, and peak time obtaining means for obtaining estimates a temperature change of the entire measurement or al in a predetermined time t1 to tm,
Based on the thermal contrast peak time in the interest region i = c determined by the peak time acquiring unit, and a peak time distribution obtaining means for obtaining a peak time distribution of the thermal contrast in the analysis area,
Before Kipi over click time acquisition means,
A thermal history calculation step of increasing or decreasing the thickness of the container in the area i to be measured, and calculating a thermal history for a plurality of thicknesses by a thermal history calculation means;
From the thermal history, the temperature Ti (t) of each region i at time t and the analysis area for the thickness of the object to be measured that is known and the thickness set to a value around that thickness, respectively. A thermal contrast calculation step of calculating a thermal contrast φi (t) that is a difference from the average temperature T0 (t) from “φi (t) = Ti (t) −T0 (t) 1 ≦ i ≦ N”;
Prediction formula construction step for constructing a prediction formula “γ (t)” as shown below using a plurality of thermal contrasts φi (t) calculated in the thermal contrast calculation step as basis functions for the region of interest i = c When,

(Where N is the number of basis functions φi (t) , i is an index thereof, wi is a weighting coefficient of basis functions φi (t) when describing the prediction formula γ (t) using N basis functions, t represents an arbitrary time.)
For the region of interest i = c, the thermal contrast measured value d (t) = u _c (t) −u _a (t) is obtained, and the “coefficient” in the prediction formula “γ (t)” constructed in the prediction formula construction step is calculated. a coefficient determination step for determining wi "using the following equation:

_{(However, φi (t 1), φi} (t 1), φi (t 1) (i = 1,2, ···, N) is t = t _1, t _2, basis functions in · · · t _{m , d (t 1), d} (t 2), ···, d (t m) is t = t ₁ for the focused area i = c, t _2, thermal contrast Found in · · · t _m, m Indicates the number of measured values of thermal contrast.)
Prediction for determining the future prediction curve of the measurement object measured by T0 (t) + γ (t) by substituting the “coefficients w1 to wN” determined in the coefficient determination step into the above-described (Equation 2). A curve determination step ;
The container wall state management apparatus characterized by performing.   5. The container wall state management device according to claim 4, wherein N ≦ m in the coefficient determination step.   6. The container wall state management apparatus according to claim 5, wherein N = m in the coefficient determination step. The future temperature of each region in the container having a temperature difference between the inner wall surface and the outer wall surface, the heat history information whose temperature changes with time, and the thickness information when the heat history information is obtained is known A computer program for managing the state of the container wall by estimating
The difference between the temperature u _i (t) at the time t of each region i obtained by dividing the analysis area set on the outer wall surface of the container into N pieces, which is measured over time by thermography, and the average temperature of the analysis area Thermal contrast calculation processing for calculating a certain thermal contrast;
The peak time of the thermal contrast in the focused area i = c, and the peak time acquisition process of obtaining estimates the total measuring whether we overall temperature changes in a predetermined time t1 to tm,
Based on the thermal contrast peak time in the region i = c to the focus determined by the peak time acquisition process is performed and a peak time distribution obtaining process of obtaining the peak time distribution of the thermal contrast in the computer in the analysis area,
Before Kipi over click time acquisition process,
A thermal history calculation step of increasing or decreasing the thickness of the container in the area i to be measured, and calculating a thermal history for a plurality of thicknesses by a thermal history calculation means;
From the thermal history, the temperature Ti (t) of each region i at time t and the analysis area for the thickness of the object to be measured that is known and the thickness set to a value around that thickness, respectively. A thermal contrast calculation step of calculating a thermal contrast φi (t) that is a difference from the average temperature T0 (t) from “φi (t) = Ti (t) −T0 (t) 1 ≦ i ≦ N”;
Prediction formula construction step for constructing a prediction formula “γ (t)” as shown below using a plurality of thermal contrasts φi (t) calculated in the thermal contrast calculation step as basis functions for the region of interest i = c When,

(Where N is the number of basis functions φi (t) , i is an index thereof, wi is a weighting coefficient of basis functions φi (t) when describing the prediction formula γ (t) using N basis functions, t represents an arbitrary time.)
For the region of interest i = c, the thermal contrast measured value d (t) = u _c (t) −u _a (t) is obtained, and the “coefficient” in the prediction formula “γ (t)” constructed in the prediction formula construction step is calculated. a coefficient determination step for determining wi "using the following equation:

_{(However, φi (t 1), φi} (t 1), φi (t 1) (i = 1,2, ···, N) is t = t _1, t _2, basis functions in · · · t _{m , d (t 1), d} (t 2), ···, d (t m) is t = t ₁ for the focused area i = c, t _2, thermal contrast Found in · · · t _m, m Indicates the number of measured values of thermal contrast.)
Prediction for determining the future prediction curve of the measurement object measured by T0 (t) + γ (t) by substituting the “coefficients w1 to wN” determined in the coefficient determination step into the above-described (Equation 2). A computer program comprising a curve determination step.   The computer program according to claim 7, wherein N ≦ m in the coefficient determination step.   9. The computer program according to claim 8, wherein N = m in the coefficient determination step.

Images