A method for monitoring abnormal temperature parameters at the outlet of a horizontal continuous casting mold

By analyzing the monotonicity and change trend of the outlet temperature of the horizontal continuous casting crystallizer, combining the cooling coefficient, and calculating the risk weight, the problem of inaccurate monitoring of temperature parameters caused by relying on empirical thresholds in the existing technology is solved, and more accurate temperature parameters are achieved Abnormal detection prevents the occurrence of steel leakage of the crystallizer.

CN118527611BActive Publication Date: 2024-12-06HANDAN HENGGONG METALLURGICAL MACHINERY CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411008705.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-26
Publication Date
2024-12-06
Estimated Expiration
2044-07-26

AI Technical Summary

Technical Problem

In the prior art, abnormal detection of outlet temperature parameters of horizontal continuous casting crystallizers depends on experience thresholds, resulting in inaccurate monitoring results and ineffective prevention of steel leakage accidents of crystallizers.

Method used

By collecting crystallizer data in local time periods, analyzing the monotonicity and change trend of outlet temperature, calculating the temperature fluctuation index and cooling coefficient, and performing abnormal detection in combination with risk weights to generate more accurate abnormal monitoring results for temperature parameters.

Benefits of technology

It improves the accuracy of temperature parameter abnormal monitoring, reduces the occurrence of crystallizer steel leakage accidents, and ensures the stability and safety of continuous casting production.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118527611B_ABST
    Figure CN118527611B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of data anomaly detection, and proposes a method for monitoring the abnormality of the outlet temperature parameter of a horizontal continuous casting crystallizer, comprising: collecting crystallizer data at different collection times within a local time period; marking a target collection time and judging whether it is a suspected abnormal collection time, if the target collection time is not a suspected abnormal collection time, assigning a risk weight to the target collection time; if the target collection time is a suspected abnormal collection time, obtaining a temperature fluctuation index and a cooling coefficient at the suspected abnormal collection time, and then obtaining a risk weight at the suspected abnormal collection time; obtaining a risk-weighted distance between two different collection times, performing anomaly detection on the collection time according to the crystallizer data corresponding to the collection time and the risk-weighted distance, and obtaining an abnormal monitoring result of the outlet temperature parameter of a horizontal continuous casting crystallizer. The present invention aims to solve the problem that the abnormal monitoring of temperature parameters is inaccurate due to the abnormal monitoring of temperature parameters based only on empirical thresholds.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data anomaly detection, and in particular to a method for monitoring anomaly of outlet temperature parameters of a horizontal continuous casting crystallizer. Background Art

[0002] Compared with traditional casting methods, horizontal continuous casting can significantly improve production efficiency, reduce energy consumption, and optimize material utilization. It is an important means for steel companies to achieve high-quality and low-cost production goals. In the process of using a crystallizer for continuous casting, it is necessary to monitor the abnormal temperature parameters of the horizontal continuous casting crystallizer outlet, especially the crystallizer leakage accident, to prevent the serious accident of crystallizer leakage, which will cause the interruption of the continuous casting production process, damage the crystallizer, and cause huge losses.

[0003] Generally, in order to monitor the steel leakage accident of the crystallizer, sensors are usually arranged on the crystallizer and thresholds are set according to experience. When the sensor data exceeds the empirical threshold, it is considered that the horizontal continuous casting crystallizer outlet temperature parameters are abnormal, and there is a possibility of a steel leakage accident in the crystallizer. However, the abnormal detection results of the horizontal continuous casting crystallizer outlet temperature parameters are greatly affected by the value of the empirical threshold. When the value of the empirical threshold is unreasonable, the abnormal monitoring results of the temperature parameters are often inaccurate. There is a problem that the abnormal monitoring of the temperature parameters is inaccurate when the abnormal monitoring of the temperature parameters is performed only based on the empirical threshold. Summary of the invention

[0004] The present invention provides a method for monitoring abnormal temperature parameters at the outlet of a horizontal continuous casting crystallizer, so as to solve the problem that the abnormal temperature parameter monitoring is inaccurate when the abnormal temperature parameter monitoring is performed only based on the empirical threshold. The technical scheme adopted is as follows:

[0005] An embodiment of the present invention provides a method for monitoring abnormal temperature parameters at the outlet of a horizontal continuous casting crystallizer, the method comprising the following steps:

[0006] Collecting crystallizer data at different collection times within a local time period, the crystallizer data including cooling water inlet flow rate, cooling temperature difference and outlet temperatures at a preset number of horizontal continuous casting crystallizer outlet positions at each collection time;

[0007] Any collection time is recorded as the target collection time, and according to the monotonicity of the outlet temperature of the horizontal continuous casting crystallizer outlet position at all collection times, it is judged whether the target collection time is a suspected abnormal collection time. If the target collection time is not a suspected abnormal collection time, the risk weight of the target collection time is assigned;

[0008] If the target collection time is a suspected abnormal collection time, the temperature fluctuation index of the suspected abnormal collection time is obtained according to the difference between the outlet temperature change trend of the horizontal continuous casting crystallizer outlet position at all collection times and the outlet temperature at the suspected abnormal collection time; the cooling coefficient of the suspected abnormal collection time is obtained according to the cooling water inlet flow rate and cooling temperature difference between the suspected abnormal collection time and the earliest collection time, and the risk weight of the suspected abnormal collection time is obtained in combination with the temperature fluctuation index of the suspected abnormal collection time;

[0009] According to the crystallizer data and risk weights corresponding to two different collection moments, the risk-weighted distances at the two different collection moments are obtained, and the collection moments are detected for anomalies according to the crystallizer data and risk-weighted distances corresponding to the collection moments, so as to obtain the abnormal monitoring results of the outlet temperature parameters of the horizontal continuous casting crystallizer.

[0010] Further, the specific method of determining whether the target collection time is a suspected abnormal collection time includes:

[0011] According to the monotonicity of the outlet temperature of the same horizontal continuous casting mold outlet position, the derivatives of a preset number of horizontal continuous casting mold outlet positions at the target acquisition time are obtained;

[0012] When the maximum value of all derivatives of the target acquisition time at the exit position is less than or equal to 0, the target acquisition time is not a suspected abnormal acquisition time;

[0013] When the maximum value of all derivatives of the target acquisition moment at the exit position is greater than 0, the target acquisition moment is a suspected abnormal acquisition moment.

[0014] Furthermore, the temperature fluctuation index at the time of suspected abnormality collection is obtained by:

[0015] The ratio of the derivative of the outlet position corresponding to the first outlet temperature at the suspected abnormality collection moment to the maximum value of all derivatives at the outlet position at the suspected abnormality collection moment is recorded as the monotonic change characteristic value at the suspected abnormality collection moment;

[0016] According to the difference between the outlet temperatures at the suspected abnormal collection time, the temperature difference change characteristic value at the suspected abnormal collection time is obtained, and combined with the monotonic change characteristic value at the suspected abnormal collection time, the temperature fluctuation index at the suspected abnormal collection time is obtained.

[0017] Further, according to the difference between the outlet temperatures at the suspected abnormal collection time, the temperature difference change characteristic value at the suspected abnormal collection time is obtained, and the temperature fluctuation index at the suspected abnormal collection time is obtained by combining the monotonic change characteristic value at the suspected abnormal collection time. The specific method of obtaining is:

[0018] The maximum value of a preset number of outlet temperatures at the suspected abnormality collection moment is recorded as the first outlet temperature at the suspected abnormality collection moment, and all the outlet temperatures at the suspected abnormality collection moments are arranged in sequence starting from the first outlet temperature in the clockwise direction of the horizontal continuous casting crystallizer outlet position to obtain the outlet temperature sequence at the suspected abnormality collection moment;

[0019] Any outlet temperature except the first outlet temperature in the outlet temperature sequence at the suspected abnormality collection time is recorded as the target outlet temperature, the difference between the target outlet temperature and the first outlet temperature in the temperature sequence is recorded as the outlet temperature difference of the target outlet temperature, and the outlet temperature differences of the outlet temperatures are arranged according to the order of the outlet temperatures in the outlet temperature sequence to obtain the outlet temperature difference sequence at the suspected abnormality collection time;

[0020] The temperature difference change characteristic value at the suspected abnormality collection time is obtained according to the outlet temperature difference sequence at the suspected abnormality collection time.

[0021] Furthermore, the temperature fluctuation index at the suspected abnormality collection time is obtained by combining the monotonic change characteristic value at the suspected abnormality collection time. The specific method of obtaining the temperature fluctuation index at the suspected abnormality collection time is as follows:

[0022] The difference between the second outlet temperature difference and the first outlet temperature difference in the outlet temperature difference sequence at the suspected abnormality collection time is recorded as the first difference at the suspected abnormality collection time, the difference between the second outlet temperature difference and the third outlet temperature difference in the outlet temperature difference sequence at the suspected abnormality collection time is recorded as the second difference at the suspected abnormality collection time, and the product of the first difference and the second difference at the suspected abnormality collection time is recorded as the temperature difference change characteristic value at the suspected abnormality collection time.

[0023] Furthermore, the temperature fluctuation index at the time of suspected abnormality collection is obtained by:

[0024] The product of the temperature difference change characteristic value and the monotonic change characteristic value at the suspected abnormal collection time is recorded as the temperature fluctuation index at the suspected abnormal collection time.

[0025] Furthermore, the specific method for obtaining the cooling coefficient at the time of suspected abnormal collection is as follows:

[0026] The ratio of the cooling water inlet flow rate and the cooling temperature difference at the suspected abnormal collection time is recorded as the cooling efficiency at the suspected abnormal collection time, the ratio of the cooling water inlet flow rate and the cooling temperature difference at the earliest collection time is recorded as the standard cooling efficiency, and the ratio of the cooling efficiency at the suspected abnormal collection time to the standard cooling efficiency is recorded as the cooling coefficient at the suspected abnormal collection time.

[0027] Furthermore, the risk weight of the suspected abnormal collection moment is obtained by:

[0028] The ratio of the temperature fluctuation index to the cooling coefficient at the suspected abnormal collection moment is recorded as the risk weight at the suspected abnormal collection moment.

[0029] Further, the risk-weighted distances at two different collection moments are obtained according to the characteristic values ​​and risk weights corresponding to the two different collection moments, including the specific method of:

[0030] The sum of the Euclidean distances between all corresponding eigenvalues ​​at two different collection moments is recorded as the metric distance at two different collection moments, and the product of the sum of the risk weights at two different collection moments and the metric distance at two different collection moments is recorded as the risk-weighted distance at two different collection moments.

[0031] Further, the abnormality detection of the acquisition time is performed according to the crystallizer data corresponding to the acquisition time and the risk-weighted distance to obtain the abnormal monitoring result of the horizontal continuous casting crystallizer outlet temperature parameter, including the specific method of:

[0032] Arrange all the mold data corresponding to each possible abnormal acquisition moment in the local time period into a row in the same order, and arrange the row data corresponding to all possible abnormal acquisition moments in the local time period from top to bottom according to the order of the acquisition moments to obtain the feature matrix of the local time period;

[0033] Taking the risk-weighted distances at different collection moments as the distances at different collection moments, perform anomaly detection on the feature matrix of the local time period to obtain the anomaly probability value of any collection moment in the local time period;

[0034] The collection time when the abnormal probability value is greater than the preset segmentation threshold is recorded as the collection time when the horizontal continuous casting crystallizer outlet temperature parameter becomes abnormal.

[0035] The beneficial effects of the present invention are:

[0036] The present invention first analyzes the distribution of the changing trend of the outlet temperature at the outlet position of the horizontal continuous casting crystallizer according to the characteristic that the accident of steel leakage in the crystallizer will definitely lead to an upward trend in the outlet temperature, and judges whether the target collection time is a suspected abnormal collection time. If the target collection time is not a suspected abnormal collection time, the risk weight of the target collection time is assigned. If the target collection time is a suspected abnormal collection time, the crystallizer steel leakage accident may or may not occur at the suspected abnormal collection time, so the suspected abnormal collection time is continued to be analyzed; according to the changing trend of the outlet temperature at the outlet position of the horizontal continuous casting crystallizer, and the characteristics that the steel leakage of the crystallizer, i.e., the strength is insufficient or too thin in one direction of the billet shell, and the leakage of molten steel occurs at the position of insufficient strength or too thinness, the difference between the outlet temperatures at the suspected abnormal collection time is analyzed to obtain the temperature fluctuation index at the suspected abnormal collection time; according to the characteristics that the smaller the cooling temperature difference of the cooling water and the larger the water inlet flow rate, the faster the water circulation speed of the cooling device, the better the cooling efficiency of the cooling water, the thicker the thickness of the acquired billet shell and the higher the hardness, the suspected abnormal collection time is analyzed. The cooling water inlet flow rate and cooling temperature difference between the normal collection time and the earliest collection time are used to obtain the cooling coefficient at the suspected abnormal collection time. Combined with the temperature fluctuation index at the suspected abnormal collection time, the risk weight of the suspected abnormal collection time is obtained. The risk weight of the suspected abnormal collection time is an evaluation of the possibility of a crystallizer steel leakage accident at the suspected abnormal collection time. The greater the risk weight of the suspected abnormal collection time, the greater the possibility of a crystallizer steel leakage accident at the suspected abnormal collection time. Finally, according to the characteristic values ​​and risk weights corresponding to the two different collection times, the risk-weighted distances of the two different collection times are obtained, and the collection times are detected according to the crystallizer data and risk-weighted distance corresponding to the collection times, and the abnormal monitoring results of the horizontal continuous casting crystallizer outlet temperature parameters are obtained, that is, according to the changing trend of the characteristic values ​​corresponding to the collection times and the possibility evaluation of the crystallizer steel leakage accident, the possibility of abnormality in the horizontal continuous casting crystallizer outlet temperature parameters is obtained, and temperature parameter abnormality monitoring is realized to solve the problem of inaccurate temperature parameter abnormality monitoring caused by only monitoring temperature parameters based on empirical thresholds. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0038] Figure 1 A schematic flow chart of a method for monitoring abnormal temperature parameters at the outlet of a horizontal continuous casting crystallizer provided by one embodiment of the present invention;

[0039] Figure 2 A schematic diagram of the installation position of an optical fiber temperature sensor provided by an embodiment of the present invention;

[0040] Figure 3 A schematic diagram of outlet temperature difference characteristics provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] See also Figure 1 , which shows a flow chart of a method for monitoring abnormal temperature parameters at the outlet of a horizontal continuous casting crystallizer provided by an embodiment of the present invention, the method comprising the following steps:

[0043] Step S001, collecting crystallizer data at different collection times within a local time period, the crystallizer data including the cooling water inlet flow rate, cooling temperature difference and outlet temperatures of a preset number of horizontal continuous casting crystallizer outlet positions at each collection time.

[0044] A preset number of optical fiber temperature sensors are installed at equal distances at the outlet of the horizontal continuous casting crystallizer, and the outlet temperature of the horizontal continuous casting crystallizer at different times is collected using the optical fiber temperature sensors. Optical fiber temperature sensors are installed at the water inlet and outlet of the cooling device, and the water temperature at the water inlet and outlet of the cooling device is collected using the optical fiber temperature sensors. A turbine flowmeter is installed at the water inlet of the cooling device, and the turbine flowmeter is used to collect the cooling water inlet flow.

[0045] Among them, this embodiment installs 4 fiber optic temperature sensors at the outlet of the horizontal continuous casting crystallizer, that is, one fiber optic temperature sensor is installed every 90 degrees. The collection time interval of the outlet temperature, cooling water inlet flow, water temperature at the water inlet position and water temperature at the water outlet position is 10 seconds, and the number of data collected of the same type is 300. As other implementation methods, the implementer can decide the preset number of fiber optic temperature sensors, the collection time interval and the number of data collected of the same type according to actual conditions, and this application does not impose any special restrictions.

[0046] The time periods corresponding to 300 different moments are recorded as the local time periods of the last data collection moment. The present application performs abnormal monitoring of the horizontal continuous casting crystallizer outlet temperature parameters based on the local time periods.

[0047] Schematic diagram of the installation position of the optical fiber temperature sensor Figure 2 In the schematic diagram of the installation position of the optical fiber temperature sensor, 1, 2, 3, and 4 are all optical fiber temperature sensors, and 5 is a cross section at the exit of the horizontal continuous casting crystallizer.

[0048] The difference between the water temperature at the water outlet and the water temperature at the water inlet collected at the same collection time is recorded as the cooling temperature difference at the same collection time.

[0049] The cooling water inlet flow rate, cooling temperature difference and outlet temperatures of a preset number of horizontal continuous casting crystallizer outlet positions at the same collection moment are all recorded as the crystallizer data at the same collection moment.

[0050] At this point, the crystallizer data at different collection times are obtained.

[0051] Step S002: record any collection moment as the target collection moment, and judge whether the target collection moment is a suspected abnormal collection moment according to the monotonicity of the outlet temperature at the outlet position of the horizontal continuous casting crystallizer at all collection moments. If the target collection moment is not a suspected abnormal collection moment, assign a risk weight to the target collection moment.

[0052] Any collection time is recorded as the target collection time. This embodiment is described by taking any collection time as an example.

[0053] According to the change trend of the outlet temperature at the outlet position of the continuous casting crystallizer at the same level at the target collection time and the previous collection time, the derivatives of a preset number of target collection times are obtained.

[0054] For each horizontal continuous casting crystallizer outlet position, all outlet temperatures collected at the same horizontal continuous casting crystallizer outlet position are used as the ordinate, and the order of the outlet temperature collection time is used as the abscissa. All outlet temperatures collected at the same horizontal continuous casting crystallizer outlet position are curve fitted to obtain the temperature curve at the horizontal continuous casting crystallizer outlet position. The temperature curve at the horizontal continuous casting crystallizer outlet position is derived at the position corresponding to the target collection time to obtain the derivative at the horizontal continuous casting crystallizer outlet position at the target collection time.

[0055] Among them, using polynomial fitting technology to perform curve fitting is a well-known technology and will not be repeated here. This embodiment uses polynomial fitting technology to perform curve fitting. As other implementation methods, on the basis of achieving the purpose of curve fitting, the implementer can use other existing methods such as least squares method to fit the curve, and this application does not make any special restrictions.

[0056] It should be noted that each optical fiber temperature sensor at the outlet of the horizontal continuous casting crystallizer has a corresponding temperature curve and a corresponding derivative at the outlet of the horizontal continuous casting crystallizer at the target acquisition time, and a preset number of derivatives at the target acquisition time are obtained in total.

[0057] According to the data distribution of all derivatives of the target collection time at the exit position, the risk weight of the target collection time in the normal state is assigned to obtain the monotonic change characteristic value of the target collection time that may have abnormalities, specifically:

[0058] When the maximum value of all derivatives at the outlet position at the target collection time is less than or equal to 0, the outlet temperatures collected at all outlet positions at the target collection time show a downward trend. Combined with the fact that the accident of crystallizer steel leakage will definitely lead to the characteristic of an upward trend in the outlet temperature, therefore, when the maximum value of all derivatives at the outlet position at the target collection time is less than or equal to 0, the target collection time will definitely not be the time when the crystallizer steel leakage accident may occur, that is, the target collection time is in a normal state, and the target collection time is recorded as the target collection time in the normal state.

[0059] The suspected abnormal acquisition moment whose maximum value among all derivatives of the exit position is less than or equal to 0 is marked as the target acquisition moment of the normal state, and the risk weight of the target acquisition moment of the normal state is assigned to the first preset parameter. The first preset parameter is a preset parameter, and the value of the first preset parameter in this embodiment is 1.

[0060] The risk weight of the target collection moment in the normal state can be obtained in the same way, that is, each collection moment in the normal state within the local time period has a corresponding risk weight.

[0061] Step S003: If the target collection time is a suspected abnormal collection time, obtain the temperature fluctuation index of the suspected abnormal collection time according to the difference between the outlet temperature change trend of the horizontal continuous casting crystallizer outlet position at all collection times and the outlet temperature at the suspected abnormal collection time; obtain the cooling coefficient of the suspected abnormal collection time according to the cooling water inlet flow rate and cooling temperature difference between the suspected abnormal collection time and the earliest collection time, and obtain the risk weight of the suspected abnormal collection time in combination with the temperature fluctuation index of the suspected abnormal collection time.

[0062] In order to facilitate subsequent understanding and analysis, the suspected abnormal acquisition moment when the maximum value of all derivatives of the outlet position is greater than 0 is recorded as the target acquisition moment where an abnormality may exist, and the target acquisition moment where an abnormality may exist is recorded as the suspected abnormal acquisition moment. It can be understood that at the suspected abnormal acquisition moment, a crystallizer leak accident may occur, or it may not occur. Therefore, further analysis is required.

[0063] When the maximum value of all derivatives at the exit position at the suspected abnormal collection moment is greater than 0, the monotonic change characteristic value of the suspected abnormal collection moment is obtained according to the data distribution of all derivatives at the exit position at the suspected abnormal collection moment.

[0064]

[0065] In the formula, is the monotonic change characteristic value at the suspected abnormal acquisition moment; It is the maximum value of all derivatives at the exit position at the time of suspected abnormal acquisition; It is the derivative of the outlet position corresponding to the first outlet temperature at the suspected abnormal collection moment.

[0066] However, when the maximum value of all derivatives at the outlet position at the suspected abnormal acquisition time is greater than 0, a mold leak may have occurred, or it may be that the maximum value of the derivative is greater than 0 due to the fluctuation of the outlet temperature. Compare with the number 0 and calculate the monotonic change characteristic value of the suspected abnormal collection moment that may have an abnormality.

[0067] when When it is greater than 0, the larger the derivative of the outlet position corresponding to the first outlet temperature at the suspected abnormal collection moment is relative to the maximum value of all derivatives, the larger the monotonic change characteristic value at the suspected abnormal collection moment is. At this time, the more obvious the monotonic change characteristics of the outlet temperature difference are. At the suspected abnormal collection moment, the more obvious the temperature rise change at the outlet position is, the greater the possibility of uneven shell thickness, and the greater the possibility of an impending crystallizer leakage accident.

[0068] At this point, the risk weight of the target acquisition time in the normal state and the monotonic change characteristic value of the suspected abnormal acquisition time are obtained.

[0069] Under normal production conditions of the continuous casting process, the thickness of the billet shell produced from the crystallizer outlet is uniform and the strength is high, and such a billet shell has the ability to insulate. However, if the billet shell strength is insufficient during the production process, the molten steel will leak out from the gap between the crystallizer and the billet shell, leading to a serious accident of crystallizer steel leakage. Insufficient billet shell strength is often caused by uneven billet shell thickness. Therefore, crystallizer steel leakage means that the billet shell is insufficient or too thin in one direction, and molten steel leaks at the position where the strength is insufficient or too thin.

[0070] When the continuous casting process is in normal production, the thickness of the billet shell is uniform and the strength is high. Even if there may be unsolidified molten steel inside the billet shell, the temperature around the billet shell remains consistent due to the excellent heat insulation ability of the billet shell, and the outlet temperature at different positions of the horizontal continuous casting crystallizer outlet is also consistent. When the thickness of the billet shell is uneven, the smaller the thickness of the billet shell, the worse the heat insulation ability. Therefore, the temperature around the billet shell is inconsistent, and the outlet temperature at different positions of the horizontal continuous casting crystallizer outlet is also inconsistent. The closer to the thinnest position of the billet shell, the higher the outlet temperature.

[0071] According to the characteristic that when the thickness of the slab shell is uneven, the outlet temperature is higher closer to the thinnest position of the slab shell, and according to the difference between the outlet temperatures of a preset number of horizontal continuous casting crystallizer outlet positions at the suspected abnormality collection moment, the temperature difference coefficient at the suspected abnormality collection moment is obtained.

[0072] Specifically, in order to evaluate the outlet temperature difference caused by the uneven thickness of the billet shell, firstly, the temperature difference change characteristic value at the suspected abnormality collection time is obtained according to the difference between the outlet temperatures at the suspected abnormality collection time.

[0073] The maximum value of a preset number of outlet temperatures at the suspected abnormality collection moment is recorded as the first outlet temperature at the suspected abnormality collection moment. All outlet temperatures at the suspected abnormality collection moments are arranged in sequence starting from the first outlet temperature in the clockwise direction of the horizontal continuous casting crystallizer outlet position to obtain the outlet temperature sequence at the suspected abnormality collection moment.

[0074] Any outlet temperature other than the first outlet temperature in the outlet temperature sequence at the suspected abnormality collection moment is recorded as the target outlet temperature. This embodiment takes any outlet temperature other than the first outlet temperature in the outlet temperature sequence at the suspected abnormality collection moment as an example for explanation. The difference between the target outlet temperature and the first outlet temperature in the temperature sequence is recorded as the outlet temperature difference of the target outlet temperature.

[0075] The outlet temperature difference of any outlet temperature except the first outlet temperature in the outlet temperature sequence at the suspected abnormal collection moment can be obtained in the same way, that is, for any outlet temperature except the first outlet temperature in the outlet temperature sequence at the suspected abnormal collection moment, there is a corresponding outlet temperature difference.

[0076] The outlet temperature difference of each outlet temperature except the first outlet temperature in the outlet temperature sequence at the suspected abnormality collection moment is arranged according to the order of the outlet temperatures in the outlet temperature sequence to obtain the outlet temperature difference sequence at the suspected abnormality collection moment.

[0077] When the thickness of the billet shell is uneven, the outlet temperature of the horizontal continuous casting crystallizer outlet position will be different. The outlet position of the first outlet temperature is the position closest to the thinnest billet shell thickness among the preset number of horizontal continuous casting crystallizer outlet positions. Since the crystallizer leaks steel when the billet shell is insufficient or too thin in one direction, and the closer to the thinnest billet shell thickness position, the lower the outlet temperature, in the outlet temperature difference sequence at the suspected abnormal collection time, the outlet temperature difference shows the characteristics of increasing first and then decreasing.

[0078] In this embodiment, four optical fiber temperature sensors are installed at the outlet of the horizontal continuous casting crystallizer. The characteristic diagram of the outlet temperature difference is as follows: Figure 3As shown. In the schematic diagram of the outlet temperature difference characteristic, 1, 2, 3, and 4 are all optical fiber temperature sensors, 5 is the cross section of the horizontal continuous casting crystallizer outlet position, and 6 is the position where the thickness of the billet shell is the thinnest, wherein the position of the optical fiber temperature sensor is the collection position of the outlet temperature of the horizontal continuous casting crystallizer outlet position. Therefore, the outlet temperature of the horizontal continuous casting crystallizer outlet position collected at 1 is the highest, which is the first outlet temperature at the suspected abnormal collection moment. The order of the outlet temperatures in the outlet temperature sequence at the suspected abnormal collection moment corresponds to the order of collection positions 1, 2, 3, and 4. Since the distances between positions 1, 2, 3, and 4 and the position with the thinnest billet shell thickness are in the order of 1, 4, 2, and 3 from far to near, the order of the temperatures in the outlet temperature collection positions from high to low is 1, 4, 2, and 3. Therefore, in the outlet temperature difference sequence at the suspected abnormal collection moment, the outlet temperature difference shows the characteristic of first increasing and then decreasing.

[0079] In order to facilitate understanding that the outlet temperature difference contained in the outlet temperature difference sequence shows the characteristic of increasing first and then decreasing according to the arrangement order in the outlet temperature difference sequence, the outlet temperature of the four horizontal continuous casting crystallizer outlet positions in the outlet temperature difference characteristic diagram is used as an example to illustrate. For example, when the temperature of the molten steel in the shell is 1500℃, since the order of the temperature from high to low in the outlet temperature collection position is 1, 4, 2, and 3, the outlet temperatures of positions 1, 4, 2, and 3 in the outlet temperature difference characteristic diagram are assumed to be 500℃, 400℃, 300℃, and 200℃, respectively, and the data in the outlet temperature sequence are 500, 300, 200, and 400, respectively. Therefore, the data in the outlet temperature difference sequence at the suspected abnormal collection time are 200, 300, and 100, respectively, showing the characteristic of increasing first and then decreasing. Among them, the values ​​of the outlet temperature at each position and the molten steel temperature in the shell are only used as examples to help understand that the outlet temperature difference shows the characteristic of increasing first and then decreasing.

[0080] According to the difference between the outlet temperatures in the outlet temperature difference sequence at the suspected abnormality collection moment, the temperature difference change characteristic value at the suspected abnormality collection moment is obtained. The temperature difference change characteristic value at the suspected abnormality collection moment is positively correlated with the difference between the outlet temperatures in the outlet temperature difference sequence at the suspected abnormality collection moment.

[0081] Preferably, as an embodiment of the present application, the difference between the second outlet temperature difference and the first outlet temperature difference in the outlet temperature difference sequence at the suspected abnormality collection moment is recorded as the first difference at the suspected abnormality collection moment, the difference between the second outlet temperature difference and the third outlet temperature difference in the outlet temperature difference sequence at the suspected abnormality collection moment is recorded as the second difference at the suspected abnormality collection moment, and the product of the first difference and the second difference at the suspected abnormality collection moment is recorded as the temperature difference change characteristic value at the suspected abnormality collection moment.

[0082] Preferably, as another embodiment of the present application, the information entropy of all outlet temperatures in the outlet temperature difference sequence at the suspected abnormal collection moment is recorded as the temperature difference change characteristic value at the suspected abnormal collection moment. On the basis of achieving the purpose of evaluating the difference between the outlet temperatures in the outlet temperature difference sequence, the implementer may use other prior art methods such as approximate entropy, variance, etc. to evaluate the difference between the outlet temperatures in the outlet temperature difference sequence, and the present application does not impose any special restrictions.

[0083] According to the temperature difference change characteristic value and the monotonic change characteristic value at the suspected abnormality collection time, the temperature fluctuation index at the suspected abnormality collection time is obtained. The temperature fluctuation index at the suspected abnormality collection time is positively correlated with the temperature difference change characteristic value and the monotonic change characteristic value at the suspected abnormality collection time.

[0084] Preferably, as an embodiment of the present application, the product of the temperature difference change characteristic value and the monotonic change characteristic value at the suspected abnormality collection time is recorded as the temperature fluctuation index at the suspected abnormality collection time.

[0085] When the temperature difference change characteristic value and the monotonic change characteristic value at the suspected abnormal collection moment are larger, the temperature fluctuation index at the suspected abnormal collection moment is larger. At this time, the difference between the outlet temperatures at the suspected abnormal collection moment is more significant, and the monotonic change characteristics of the outlet temperature difference are more obvious. At the suspected abnormal collection moment, the more obvious the temperature rise change at the outlet position is, the greater the possibility of uneven shell thickness, and the greater the possibility of an impending crystallizer leakage accident.

[0086] At this point, the temperature fluctuation index of the suspected abnormal collection moment is obtained. The temperature fluctuation index of any collection moment can be obtained in the same way, that is, each collection moment that may have an abnormality in the local time period has a corresponding temperature fluctuation index.

[0087] The cooling device of the horizontal continuous casting crystallizer generally adopts a water-cooled design, that is, cooling water is used for cooling. The smaller the cooling temperature difference of the cooling water and the larger the water inlet flow rate, the faster the water circulation speed of the cooling device and the better the cooling efficiency of the cooling water. At this time, the obtained billet shell is thicker and harder.

[0088] The cooling coefficient at the suspected abnormal collection time is obtained according to the cooling water inlet flow rate and the cooling temperature difference between the suspected abnormal collection time and the earliest collection time.

[0089] Preferably, as an embodiment of the present application, the ratio of the cooling water inlet flow rate and the cooling temperature difference at the suspected abnormal collection time is recorded as the cooling efficiency at the suspected abnormal collection time, the ratio of the cooling water inlet flow rate and the cooling temperature difference at the earliest collection time is recorded as the standard cooling efficiency, and the ratio of the cooling efficiency at the suspected abnormal collection time to the standard cooling efficiency is recorded as the cooling coefficient at the suspected abnormal collection time.

[0090] Specifically, the calculation formula of the cooling coefficient at the time of suspected abnormal collection is:

[0091]

[0092] In the formula, is the cooling coefficient at the time of suspected abnormal collection; The cooling efficiency at the time of suspected abnormal collection; This is standard cooling performance.

[0093] The smaller the cooling efficiency at the time of suspected abnormal collection is relative to the standard cooling efficiency, the smaller the cooling coefficient at the time of suspected abnormal collection is. At this time, the cooling effect of the cooling water at the time of suspected abnormal collection is worse, that is, the greater the possibility of a crystallizer leakage accident at the time of suspected abnormal collection.

[0094] At this point, the cooling coefficient at the time of suspected abnormal collection is obtained.

[0095] According to the temperature fluctuation index and cooling coefficient at the suspected abnormal collection moment, the risk weight of the suspected abnormal collection moment is obtained.

[0096]

[0097] In the formula, The risk weight for the suspected abnormal collection moment; The temperature fluctuation index at the time of suspected abnormal collection; is the cooling coefficient at the time of suspected abnormal collection.

[0098] The greater the temperature fluctuation index at the suspected abnormal collection moment and the smaller the cooling coefficient, the greater the risk weight at the suspected abnormal collection moment. At this time, the cooling effect of the cooling water at the suspected abnormal collection moment is worse, the temperature rise change at the outlet position is more obvious, and the possibility of uneven shell thickness is greater, that is, the possibility of a crystallizer leakage accident at the suspected abnormal collection moment is greater.

[0099] At this point, the risk weight of the suspected abnormal collection moment is obtained. The risk weight of any suspected abnormal collection moment can be obtained in the same way, that is, each collection moment that may have an abnormality in the local time period has a corresponding risk weight.

[0100] Step S004: Obtain the risk-weighted distances between the two different collection moments according to the crystallizer data and risk weights corresponding to the two different collection moments, perform anomaly detection on the collection moments according to the crystallizer data and risk-weighted distances corresponding to the collection moments, and obtain the abnormal monitoring results of the horizontal continuous casting crystallizer outlet temperature parameters.

[0101] For any collection moment in a local time period, there are corresponding preset numbers of outlet temperatures, cooling water inlet flow rates and cooling temperature differences at the outlet positions of the horizontal continuous casting crystallizer. These data are all crystallizer data at the collection moment, that is, each collection moment has 6 corresponding crystallizer data.

[0102] According to the crystallizer data and risk weights corresponding to the two different collection moments, the risk-weighted distances at the two different collection moments are obtained.

[0103] The sum of the Euclidean distances between all the same type of crystallizer data corresponding to two different acquisition moments is recorded as the metric distance at the two different acquisition moments, and the product of the sum of the risk weights at the two different acquisition moments and the metric distance at the two different acquisition moments is recorded as the risk-weighted distance at the two different acquisition moments. The calculation formula of the risk-weighted distance at the two different acquisition moments is:

[0104]

[0105] in, For collection time and The risk-weighted distance For collection time risk weights; For collection time risk weights; For collection time and The metric distance.

[0106] All the mold data corresponding to each possible abnormal acquisition moment in the local time period are arranged into a row in the same order, and the row data corresponding to all possible abnormal acquisition moments in the local time period are arranged from top to bottom according to the sequence of the acquisition moments to obtain the feature matrix of the local time period.

[0107] Taking the risk-weighted distances at different collection moments as the distances at different collection moments, the SOS anomaly detection algorithm is used to perform anomaly detection on the feature matrix of the local time period to obtain the anomaly probability value of any collection moment in the local time period.

[0108] The OTSU maximum inter-class variance method is used for the abnormal probability values ​​of all collection moments to obtain the segmentation threshold. The collection moment with an abnormal probability value greater than the segmentation threshold is recorded as an abnormal moment. The abnormal moment is the collection moment when the horizontal continuous casting crystallizer outlet temperature parameter is abnormal. At the abnormal moment, there is a possibility of steel leakage in the continuous casting production process. It is necessary to increase the cold input of the horizontal continuous casting crystallizer cooling device, improve the solidification degree of the billet shell, and enhance the thickness and hardness of the billet shell to avoid the occurrence of steel leakage accidents.

[0109] Among them, this embodiment selects the OTSU maximum inter-class variance method to obtain the segmentation threshold according to the abnormal probability value. On the basis of achieving the purpose of obtaining the segmentation threshold according to the abnormal probability value, the implementer can adopt other existing methods such as K-means, region growing method, etc. to obtain the segmentation threshold according to the abnormal probability value, and this application does not make any special restrictions.

[0110] At this point, the abnormal monitoring results of the horizontal continuous casting crystallizer outlet temperature parameters are obtained.

[0111] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for monitoring abnormal temperature parameters at the outlet of a horizontal continuous casting crystallizer, characterized in that: The method comprises the following steps: Collecting crystallizer data at different collection times within a local time period, the crystallizer data including cooling water inlet flow rate, cooling temperature difference and outlet temperatures at a preset number of horizontal continuous casting crystallizer outlet positions at each collection time; Any collection time is recorded as the target collection time, and according to the monotonicity of the outlet temperature of the horizontal continuous casting crystallizer outlet position at all collection times, it is judged whether the target collection time is a suspected abnormal collection time. If the target collection time is not a suspected abnormal collection time, the risk weight of the target collection time is assigned; If the target collection time is a suspected abnormal collection time, the temperature fluctuation index of the suspected abnormal collection time is obtained according to the difference between the outlet temperature change trend of the horizontal continuous casting crystallizer outlet position at all collection times and the outlet temperature at the suspected abnormal collection time; the cooling coefficient of the suspected abnormal collection time is obtained according to the cooling water inlet flow rate and cooling temperature difference between the suspected abnormal collection time and the earliest collection time, and the risk weight of the suspected abnormal collection time is obtained in combination with the temperature fluctuation index of the suspected abnormal collection time; According to the mold data and risk weights corresponding to the two different collection moments, the risk-weighted distances at the two different collection moments are obtained, and the collection moments are detected for abnormality according to the mold data and risk-weighted distances corresponding to the collection moments, so as to obtain the abnormal monitoring results of the horizontal continuous casting mold outlet temperature parameters; The specific method of determining whether the target acquisition time is a suspected abnormal acquisition time includes: According to the monotonicity of the outlet temperature of the same horizontal continuous casting mold outlet position, the derivatives of a preset number of horizontal continuous casting mold outlet positions at the target acquisition time are obtained; When the maximum value of all derivatives of the target acquisition time at the exit position is less than or equal to 0, the target acquisition time is not a suspected abnormal acquisition time; When the maximum value of all derivatives of the target acquisition moment at the exit position is greater than 0, the target acquisition moment is a suspected abnormal acquisition moment; The specific method for obtaining the temperature fluctuation index at the suspected abnormal collection time is: The ratio of the derivative of the outlet position corresponding to the first outlet temperature at the suspected abnormality collection moment to the maximum value of all derivatives at the outlet position at the suspected abnormality collection moment is recorded as the monotonic change characteristic value at the suspected abnormality collection moment; According to the difference between the outlet temperatures at the suspected abnormal collection time, the temperature difference change characteristic value at the suspected abnormal collection time is obtained, and combined with the monotonic change characteristic value at the suspected abnormal collection time, the temperature fluctuation index at the suspected abnormal collection time is obtained; The specific method for obtaining the temperature fluctuation index at the suspected abnormal collection time is: The product of the temperature difference change characteristic value and the monotonic change characteristic value at the suspected abnormal collection time is recorded as the temperature fluctuation index at the suspected abnormal collection time; The specific method for obtaining the cooling coefficient at the suspected abnormal collection time is: The ratio of the cooling water inlet flow rate and the cooling temperature difference at the suspected abnormality collection time is recorded as the cooling efficiency at the suspected abnormality collection time, the ratio of the cooling water inlet flow rate and the cooling temperature difference at the earliest collection time is recorded as the standard cooling efficiency, and the ratio of the cooling efficiency at the suspected abnormality collection time to the standard cooling efficiency is recorded as the cooling coefficient at the suspected abnormality collection time; The specific method for obtaining the risk weight of the suspected abnormal collection moment is as follows: The ratio of the temperature fluctuation index to the cooling coefficient at the suspected abnormal collection moment is recorded as the risk weight at the suspected abnormal collection moment; The specific method of obtaining the risk-weighted distances at two different collection times according to the crystallizer data and risk weights corresponding to the two different collection times is as follows: The sum of the Euclidean distances between all the same type of crystallizer data corresponding to two different acquisition moments is recorded as the metric distance of the two different acquisition moments, and the product of the sum of the risk weights of the two different acquisition moments and the metric distance of the two different acquisition moments is recorded as the risk weighted distance of the two different acquisition moments; The method of performing abnormal detection on the acquisition time according to the mold data corresponding to the acquisition time and the risk-weighted distance to obtain the abnormal monitoring result of the horizontal continuous casting mold outlet temperature parameter includes: Arrange all the mold data corresponding to each possible abnormal acquisition moment in the local time period into a row in the same order, and arrange the row data corresponding to all possible abnormal acquisition moments in the local time period from top to bottom according to the order of the acquisition moments to obtain the feature matrix of the local time period; Taking the risk-weighted distances at different collection moments as the distances at different collection moments, perform anomaly detection on the feature matrix of the local time period to obtain the anomaly probability value of any collection moment in the local time period; The collection time when the abnormal probability value is greater than the preset segmentation threshold is recorded as the collection time when the horizontal continuous casting crystallizer outlet temperature parameter becomes abnormal.

2. The method for monitoring abnormal temperature parameters at the outlet of a horizontal continuous casting mold according to claim 1, characterized in that: According to the difference between the outlet temperatures at the suspected abnormal collection time, the temperature difference change characteristic value at the suspected abnormal collection time is obtained, and the temperature fluctuation index at the suspected abnormal collection time is obtained by combining the monotonic change characteristic value at the suspected abnormal collection time. The specific method of obtaining is: The maximum value of a preset number of outlet temperatures at the suspected abnormality collection moment is recorded as the first outlet temperature at the suspected abnormality collection moment, and all the outlet temperatures at the suspected abnormality collection moments are arranged in sequence starting from the first outlet temperature in the clockwise direction of the horizontal continuous casting crystallizer outlet position to obtain the outlet temperature sequence at the suspected abnormality collection moment; Any outlet temperature except the first outlet temperature in the outlet temperature sequence at the suspected abnormality collection moment is recorded as the target outlet temperature, the difference between the target outlet temperature and the first outlet temperature in the temperature sequence is recorded as the outlet temperature difference of the target outlet temperature, and the outlet temperature differences of the outlet temperatures are arranged in the order of the outlet temperatures in the outlet temperature sequence to obtain the outlet temperature difference sequence at the suspected abnormality collection moment; The temperature difference change characteristic value at the suspected abnormality collection time is obtained according to the outlet temperature difference sequence at the suspected abnormality collection time.

3. The method for monitoring abnormal temperature parameters at the outlet of a horizontal continuous casting crystallizer according to claim 2, characterized in that: The temperature difference change characteristic value at the suspected abnormality collection time is obtained according to the outlet temperature difference sequence at the suspected abnormality collection time. The specific method of obtaining the temperature difference change characteristic value at the suspected abnormality collection time is: The difference between the second outlet temperature difference and the first outlet temperature difference in the outlet temperature difference sequence at the suspected abnormality collection time is recorded as the first difference at the suspected abnormality collection time, the difference between the second outlet temperature difference and the third outlet temperature difference in the outlet temperature difference sequence at the suspected abnormality collection time is recorded as the second difference at the suspected abnormality collection time, and the product of the first difference and the second difference at the suspected abnormality collection time is recorded as the temperature difference change characteristic value at the suspected abnormality collection time.

Citation Information

Patent Citations

  • Process of continuous casting with detection of possibility of break out

    CA1328925C

  • Method for preventing crystallizer of continuous casting machine from sticking breakout

    CN103878335A

  • Crystallizer water control method based on casting blank temperature simulation

    CN116100003A

  • Crystallizer cooling system for horizontal continuous casting

    CN217393692U

  • Detecting method of abnormal casting condition inside mold in continuous casting

    JP2001239353A