Method for detecting sticking of a strand in a slab caster

Abstract

The present application relates to a kind of slab crystallizer cast slab sticking detection method, belong to metallurgical continuous casting method technical field.The technical scheme of the present application is: install laser range finder and obtain the height variation of crystallizer mould powder slag face, obtain the characteristic spectrum of the height variation of slag face with time by fast discrete fourier transform, compare it with the frequency of crystallizer vibration, directly judge whether the cast slab sticking in crystallizer occurs.The beneficial effects of the present application are: without crystallizer modification, its installation, maintenance is convenient, using direct detection means to judge the occurrence of cast slab sticking, its prediction result is accurate and reliable, cast slab sticking phenomenon can be detected immediately after occurrence, reduce the risk of missing steel caused by late alarm, can reliably, accurately, quickly carry out crystallizer sticking and missing steel prediction.

Description

Technical Field

[0001] This invention relates to a method for detecting slab adhesion in a slab crystallizer, belonging to the technical field of metallurgical continuous casting methods. Background Technology

[0002] In the continuous casting production process, ensuring safe and stable production and guaranteeing the quality of the cast billets are the most important tasks in continuous casting research. Steel leakage is the most dangerous production accident, and steel leakage due to billet adhesion accounts for more than 80% of such accidents.

[0003] Beginning in the late 1970s, various types of leak detection sensors and leak prediction technologies were developed, among which the most typical are the friction resistance method, the heat flux sensor method, and the thermocouple method. The friction resistance method operates under harsh conditions and is complex to install; more than a dozen factors can affect its detection results. Because its accuracy is only about 60%, it is currently only suitable as an auxiliary method for other prediction techniques. The heat flux sensor method can only reflect the macroscopic and overall state of the crystallizer, and cannot provide information such as the local heat transfer conditions within the crystallizer; therefore, its prediction accuracy is also low.

[0004] Thermocouple temperature measurement is a relatively widely used method, but crystallizer leakage prediction systems using this method require modification of the crystallizer copper plates and the installation of numerous thermocouples, with complex wiring requirements. During operation, it faces numerous adverse environmental factors such as high crystallizer temperatures, high dust and humidity levels, and electromagnetic interference, resulting in high thermocouple failure rates and poor alarm accuracy. Furthermore, the thermocouple temperature measurement method requires the billet bonding point to pass through the thermocouple before triggering the logic judgment, often leading to delayed alarm times and a risk of untimely alarms. Summary of the Invention

[0005] The purpose of this invention is to provide a method for detecting slab adhesion in a slab crystallizer. This method requires no modification to the crystallizer, is easy to install and maintain, and uses direct detection to determine the occurrence of slab adhesion. The prediction results are accurate and reliable, and the method can detect slab adhesion immediately after it occurs, reducing the risk of steel leakage due to delayed alarms. It can reliably, accurately, and quickly predict slab adhesion and steel leakage in the crystallizer, effectively solving the aforementioned problems in the background technology.

[0006] The technical solution of the present invention is: a method for detecting slab adhesion in a slab crystallizer, comprising the following steps:

[0007] S1. Install a laser rangefinder at the bottom of the tundish, making it parallel to the wide copper plate of the crystallizer and matching the measurement position with the inner wall of the wide copper plate.

[0008] S2. After the continuous casting machine starts casting, an automatic slag feeder is used to add protective slag to the crystallizer.

[0009] S3. Acquire the measurement signal from the laser rangefinder into the PLC;

[0010] S4. Collect and store the measurement signals from the laser rangefinder and the frequency and amplitude signals of the crystallizer vibration from the PLC on the computer.

[0011] S5. Calculate the number of analysis points S, S=round(1000×K·T-1·f-1), where S is the number of analysis points, in units of points; K is the frequency multiple; T is the data acquisition interval time, in units of ms; f is the frequency of crystallizer vibration, in units of Hz; round is the rounding function.

[0012] S6. After processing the laser rangefinder measurement signal, perform fast discrete Fourier transform on each signal.

[0013] S7. The intensity of the frequency points in the spectrum obtained after the fast discrete Fourier transform is sampled and analyzed. The analysis results are compared with the frequency of the crystallizer vibration to determine whether the billet has bonded.

[0014] In step S1, a total of ten laser rangefinders are installed. The laser rangefinders are installed in the following positions: two above the center of the inner and outer arcs of the crystallizer; four 300-400mm and 600-800mm to the left and right of the center of the inner and outer arcs of the crystallizer; and four to the right and right of the center of the inner and outer arcs of the crystallizer. The measuring position of the laser rangefinder is 3 to 5mm away from the inner wall of the wide copper plate.

[0015] In step S4, the computer collects and stores 10 sets of measurement signals from the laser rangefinder and the frequency and amplitude signals of the crystallizer vibration from the PLC every 20ms interval.

[0016] In step S6, the 10 sets of laser rangefinder measurement signals are processed in the following way: select the most recent S consecutive measurement data points, calculate the average of these S measurement data points, and then subtract the average from all S data points to obtain the processed data.

[0017] In step S7, the intensity of the first 10 frequency points in each of the 10 spectra obtained after the fast discrete Fourier transform is analyzed. If the intensity of the K+1 frequency point in any spectrum is greater than 40% of the total intensity of the first 10 frequency points in that spectrum and greater than 75% of the crystallizer amplitude, it is determined that the billet has bonded; otherwise, the billet is normal.

[0018] The beneficial effects of this invention are: no crystallizer modification is required, its installation and maintenance are convenient, it uses direct detection to determine the occurrence of billet sticking, its prediction results are accurate and reliable, and it can be detected immediately after the billet sticking phenomenon occurs, reducing the risk of steel leakage due to late alarm, and it can reliably, accurately and quickly predict the occurrence of steel leakage due to crystallizer sticking. Attached Figure Description

[0019] Figure 1 This is a time-domain diagram of the first group of laser rangefinder measurement signals in Embodiment 1 of the present invention;

[0020] Figure 2 This is a spectrum analysis diagram obtained after the fast discrete Fourier transform of the measurement data of the first group of laser rangefinders in Embodiment 1 of the present invention.

[0021] Figure 3 This is a time-domain diagram of the laser rangefinder measurement signal in the second group of Embodiment 1 of the present invention;

[0022] Figure 4 This is a spectrum analysis diagram obtained after fast discrete Fourier transform of the measurement data of the second group of laser rangefinders in Embodiment 1 of the present invention;

[0023] Figure 5 This is a time-domain diagram of the laser rangefinder measurement signal in Embodiment 2 of the present invention;

[0024] Figure 6 This is a spectrum analysis diagram obtained after the laser rangefinder measurement data in Embodiment 2 of the present invention has undergone fast discrete Fourier transform. Implementation

[0025] To make the purpose, technical solutions, and advantages of the invention's embodiments clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only a small part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the protection scope of the present invention.

[0026] A method for detecting slab adhesion in a slab crystallizer includes the following steps:

[0027] S1. Install a laser rangefinder at the bottom of the tundish, making it parallel to the wide copper plate of the crystallizer and matching the measurement position with the inner wall of the wide copper plate.

[0028] S2. After the continuous casting machine starts casting, an automatic slag feeder is used to add protective slag to the crystallizer.

[0029] S3. Acquire the measurement signal from the laser rangefinder into the PLC;

[0030] S4. Collect and store the measurement signals from the laser rangefinder and the frequency and amplitude signals of the crystallizer vibration from the PLC on the computer.

[0031] S5. Calculate the number of analysis points S, S=round(1000×K·T-1·f-1), where S is the number of analysis points, in units of points; K is the frequency multiple; T is the data acquisition interval time, in units of ms; f is the frequency of crystallizer vibration, in units of Hz; round is the rounding function.

[0032] S6. After processing the laser rangefinder measurement signal, perform fast discrete Fourier transform on each signal.

[0033] S7. The intensity of the frequency points in the spectrum obtained after the fast discrete Fourier transform is sampled and analyzed. The analysis results are compared with the frequency of the crystallizer vibration to determine whether the billet has bonded.

[0034] In step S1, a total of ten laser rangefinders are installed. The laser rangefinders are installed in the following positions: two above the center of the inner and outer arcs of the crystallizer; four 300-400mm and 600-800mm to the left and right of the center of the inner and outer arcs of the crystallizer; and four to the right and right of the center of the inner and outer arcs of the crystallizer. The measuring position of the laser rangefinder is 3 to 5mm away from the inner wall of the wide copper plate.

[0035] In step S4, the computer collects and stores 10 sets of measurement signals from the laser rangefinder and the frequency and amplitude signals of the crystallizer vibration from the PLC every 20ms interval.

[0036] In step S6, the 10 sets of laser rangefinder measurement signals are processed in the following way: select the most recent S consecutive measurement data points, calculate the average of these S measurement data points, and then subtract the average from all S data points to obtain the processed data.

[0037] In step S7, the intensity of the first 10 frequency points in each of the 10 spectra obtained after the fast discrete Fourier transform is analyzed. If the intensity of the K+1 frequency point in any spectrum is greater than 40% of the total intensity of the first 10 frequency points in that spectrum and greater than 75% of the crystallizer amplitude, it is determined that the billet has bonded; otherwise, the billet is normal. Example 1

[0038] This invention is operated according to the following steps:

[0039] (1) The crystallizer of the present invention has a wide copper plate with a width of 1580 mm. Ten laser rangefinders are installed at the bottom of the intermediate ladle, parallel to the wide copper plate of the crystallizer, and the measuring position is 3 to 5 mm away from the inner wall of the wide copper plate. The laser rangefinders are installed at the following positions: two at the center of the inner and outer arcs of the crystallizer, four at the left 300 mm and 600 mm to the left of the center of the inner and outer arcs of the crystallizer, and four at the right 300 mm and 600 mm to the right of the center of the inner and outer arcs of the crystallizer, for a total of 10 laser rangefinders.

[0040] (2) After the continuous casting machine starts casting, an automatic slag feeder is used to add protective slag to the crystallizer.

[0041] (3) Collect the laser rangefinder measurement signal into the PLC.

[0042] (4) The measurement signals of 10 laser rangefinders, the frequency signal and amplitude signal of crystallizer vibration are collected from the PLC at 20ms intervals on the computer and stored. At the current analysis time, the crystallizer vibration frequency is 2.5Hz and the crystallizer amplitude is 4mm.

[0043] (5) Calculate the number of analysis points S based on the crystallizer vibration frequency, S = round(1000 × K·T) -1 ·f -1 Where S is the number of analysis points (in units); K is the frequency multiple; T is the data acquisition interval (in milliseconds); f is the frequency of crystallizer vibration (in Hz); and round is the rounding function. In this embodiment, K is 5, therefore, according to this formula, S = round(1000 × 5 × 1 / 20 × 1 / 2.5) = 100.

[0044] (6) The 10 sets of laser rangefinder measurement signals were processed in the following way: the 100 most recent consecutive measurement data points were selected, the average of these 100 measurement data points was calculated, and the average of these 100 data points was subtracted from the average to obtain the processed data. Then, the processed data were subjected to Fast Discrete Fourier Transform.

[0045] (7) From the 10 spectra obtained after the Fast Discrete Fourier Transform, the intensity of the first 10 frequency points is analyzed. For example Figure 1 This is a graph showing the changes over time of 100 points within the last 2 seconds of the signal measured and processed by the first group of laser rangefinders in this embodiment. Figure 2The spectrum represents the first 10 frequency points obtained after the fast discrete Fourier transform of the data set. The 6th frequency point is the frequency of the crystallizer vibration, which is 2.5 Hz. The intensity of the collected data at this frequency point is 0.21 mm. The crystallizer amplitude is 4 mm, and 75% of this amplitude is 3 mm. Therefore, the intensity at the 6th frequency point is not greater than 75% of the crystallizer amplitude, and it is not considered that the billet is stuck to the copper plate and vibrating together. It is determined that no billet sticking has occurred at this detection point. Figure 3 This is a graph showing the changes over time of 100 points within the last 2 seconds of the signal measured and processed by the second group of laser rangefinders in this embodiment. Figure 4 The spectrum of the first 10 frequency points obtained after fast discrete Fourier transform of the data set is shown below. The intensities of the 10 frequency points are 0 mm, 0.31 mm, 0.30 mm, 0.84 mm, 0.47 mm, 3.58 mm, 0.46 mm, 0.84 mm, 0.45 mm, and 0.04 mm, respectively. The 6th frequency point is the frequency of the crystallizer vibration, which is 2.5 Hz. The intensity of the collected data at this frequency point is 3.58 mm, while the crystallizer amplitude is 4 mm, and 75% of the amplitude is 3 mm. Therefore, the intensity of the 6th frequency point is greater than 75% of the crystallizer amplitude. The total intensity of the first 10 frequency points is 7.29 mm, and 40% of the amplitude is 2.916 mm. The intensity of the 6th frequency point is greater than 40% of the total intensity of the first 10 frequency points. Therefore, it is believed that the billet is stuck to the copper plate and vibrates together, and the vibration at this frequency is dominant. It can be considered that this vibration is not caused by other accidental factors, and therefore, it is determined that billet sticking has occurred. Example 2

[0046] This invention is operated according to the following steps:

[0047] (1) The crystallizer of the present invention has a wide copper plate with a width of 2050mm. Ten laser rangefinders are installed at the bottom of the intermediate ladle, parallel to the wide copper plate of the crystallizer, and the measuring position is 3 to 5mm away from the inner wall of the wide copper plate. The laser rangefinders are installed at the following positions: two at the center of the inner and outer arcs of the crystallizer, four at 400mm and 800mm to the left of the center of the inner and outer arcs of the crystallizer, and four at 400mm and 800mm to the right of the center of the inner and outer arcs of the crystallizer, for a total of 10 laser rangefinders.

[0048] (2) After the continuous casting machine starts casting, an automatic slag feeder is used to add protective slag to the crystallizer.

[0049] (3) Collect the laser rangefinder measurement signal into the PLC.

[0050] (4) The measurement signals of 10 laser rangefinders, the frequency signal and amplitude signal of crystallizer vibration are collected from the PLC at 20ms intervals on the computer and stored. At the current analysis time, the crystallizer vibration frequency is 3Hz and the crystallizer amplitude is 3.5mm.

[0051] (5) Calculate the number of analysis points S based on the crystallizer vibration frequency, S = round(1000 × K·T) -1 ·f -1 Where S is the number of analysis points (in units); K is the frequency multiple; T is the data acquisition interval (in milliseconds); f is the frequency of crystallizer vibration (in Hz); and round is the rounding function. In this embodiment, K is 5, therefore, according to this formula, S = round(1000 × 5 × 1 / 20 × 1 / 3) = 83.

[0052] (6) The measurement signals of the 10 laser rangefinders were processed in the following way: the 83 most recent consecutive measurement data points were selected, the average of these 83 measurement data points was calculated, and the average of these 83 data points was subtracted from the average to obtain the processed data. Then, the processed data were subjected to Fast Discrete Fourier Transform.

[0053] (7) From the 10 spectra obtained after the Fast Discrete Fourier Transform, the intensity of the first 10 frequency points is analyzed. For example Figure 5 This is a graph showing the changes over time of 83 points within the most recent 1.66 seconds of the signal processed by the first group of laser rangefinders in this embodiment. Figure 6 The spectrum of the first 10 frequency points obtained after fast discrete Fourier transform of the data set is shown. The 6th frequency point is the frequency of crystallizer vibration, which is 3Hz, and the intensity of the collected data at this frequency point is 0.44mm. The crystallizer amplitude is 3.5mm, and 75% of it is 2.625mm. Therefore, the intensity at the 6th frequency point is not greater than 75% of the crystallizer amplitude, and it is not considered that the billet is stuck to the copper plate and vibrating together. It is judged that no billet adhesion has occurred at this detection point. Subsequent analysis of the frequency points of the 10 sets of data did not meet the intensity judgment condition. Therefore, it is judged that no billet adhesion occurred in this analysis.

Claims

1. A method for detecting the sticking of a strand in a slab caster, characterized by Includes the following steps: S1. Install a laser rangefinder at the bottom of the tundish, making it parallel to the wide copper plate of the crystallizer and matching the measurement position with the inner wall of the wide copper plate. S2. After the continuous casting machine starts casting, an automatic slag feeder is used to add protective slag to the crystallizer. S3. Acquire the measurement signal from the laser rangefinder into the PLC; S4. Collect and store the measurement signals from the laser rangefinder and the frequency and amplitude signals of the crystallizer vibration from the PLC on the computer. S5, calculate analysis points S, S = round(1000 x K x T -1 ·f -1 ), wherein S is the analysis points, units are pieces; K is the frequency multiplication number; T is the data collection interval time, units are ms; f is the frequency of the crystallizer vibration, units are Hz; round is the rounding function; S6. After processing the laser rangefinder measurement signal, perform fast discrete Fourier transform on each signal. S7. The intensity of the frequency points in the spectrum obtained after the fast discrete Fourier transform is sampled and analyzed. The analysis results are compared with the frequency of the crystallizer vibration to determine whether the billet has bonded.

2. A method of detecting a slabbing crystallizer strand sticking according to claim 1, characterized in that: In step S1, a total of ten laser rangefinders are installed. The laser rangefinders are installed in the following positions: two above the center of the inner and outer arcs of the crystallizer; four 300-400mm and 600-800mm to the left and right of the center of the inner and outer arcs of the crystallizer; and four to the right and right of the center of the inner and outer arcs of the crystallizer. The measuring position of the laser rangefinder is 3 to 5mm away from the inner wall of the wide copper plate.

3. A method of detecting a slabbing crystallizer strand sticking according to claim 1, characterized in that: In step S4, the computer collects and stores 10 sets of measurement signals from the laser rangefinder and the frequency and amplitude signals of the crystallizer vibration from the PLC every 20ms interval.

4. The method for detecting slab adhesion in a slab crystallizer according to claim 3, characterized in that: In step S6, the 10 sets of laser rangefinder measurement signals are processed in the following way: select the most recent S consecutive measurement data points, calculate the average of these S measurement data points, and then subtract the average from all S data points to obtain the processed data.

5. The method for detecting slab adhesion in a slab crystallizer according to claim 3, characterized in that: In step S7, the intensity of the first 10 frequency points in each of the 10 spectra obtained after the fast discrete Fourier transform is analyzed. If the intensity of the K+1 frequency point in any spectrum is greater than 40% of the total intensity of the first 10 frequency points in that spectrum and greater than 75% of the crystallizer amplitude, it is determined that the billet has bonded; otherwise, the billet is normal.

Images