A method for eliminating steam from a water return pipe of an oxygen lance body

Abstract

This invention belongs to the field of metallurgical engineering technology. It provides a method for eliminating steam in the return water pipeline of an oxygen lance, comprising: collecting pressure rise rate data and corresponding regulating valve control parameters from historical production processes of the oxygen lance; screening qualified production records where no secondary entrainment of high-speed gas flow occurred; grouping the screened qualified production records using a clustering method based on regulating valve control parameters to construct several rate intervals; attempting self-recovery without interrupting operation; and immediately implementing tiered responses and decisive intervention until the risk is eliminated for any risks indicating humidity or water droplet signals. This differentiated correction mechanism avoids excessive intervention that could affect production efficiency and can quickly cut off the steam source in emergencies, fundamentally preventing safety accidents caused by secondary entrainment leading to explosions. Furthermore, it allows the use of inexpensive saturated steam instead of expensive nitrogen, achieving the dual benefits of safety and energy saving.

Description

Technical Field

[0001] This invention belongs to the field of metallurgical engineering technology, specifically a method for eliminating steam in the return water pipeline of an oxygen lance. Background Technology

[0002] In the process of steelmaking in converters or electric arc furnaces, the oxygen lance is the core equipment for supplying oxygen to the molten pool. To prevent backflow of high-temperature flue gas, protect the oxygen lance body, and achieve cooling and sealing, steam is usually introduced into the oxygen lance body and its return water pipe. However, in actual production, the oxygen lance return water pipe often carries "wet steam" in which saturated steam and condensate coexist. If the large amount of liquid water droplets carried in it are sprayed out with the steam and fall into the molten steel pool, they will instantly vaporize and expand in volume, causing violent splashing or even explosion, which seriously threatens personal and equipment safety. Therefore, multi-stage steam-water separation devices (such as baffles, wire mesh, cyclone separators, etc.) are commonly used to separate and discharge the condensate in the wet steam to obtain relatively dry steam for sealing.

[0003] However, existing steam-water separation methods have serious defects under variable load conditions. Steelmaking production has a typical cyclical rhythm: the oxygen lance needs to be frequently inserted and lifted, causing the steam flow rate in the return water pipe to surge from zero to 3-5 times the normal value in a very short time, or to drop from the normal value to zero. This violent flow fluctuation will cause the airflow velocity inside the separator to instantly exceed the design limit, thereby re-rolling up the condensate that has been separated and collected at the bottom of the tank and carrying it out of the tank, forming a "secondary entrainment" phenomenon. Secondary entrainment causes the separator to completely fail at the moment of start-up and shutdown, and water-laden steam is still sprayed towards the oxygen lance seal. The aforementioned safety hazards have not been eliminated. To address this problem, existing technologies mostly use the method of adding flow-limiting orifice plates to the air inlet pipe to passively limit the maximum flow rate. However, the orifice plate diameter needs to be repeatedly calculated based on empirical formulas, and when the smelting rhythm or steam source pressure changes, the orifice plate cannot be adaptively adjusted. It often requires frequent shutdowns to replace orifice plates of different specifications, which is cumbersome and has a slow response, making it difficult to fundamentally solve the secondary entrainment problem under variable load conditions.

[0004] Therefore, the present invention provides a method for eliminating steam in the return water pipe of the oxygen lance body. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0006] The technical solution adopted by this invention to solve its technical problem is: a method for eliminating steam in the return water pipe of an oxygen lance body, comprising:

[0007] Data on the pressure rise rate during the historical production process of the oxygen lance and the corresponding control parameters of the regulating valve were collected. Qualified production records that did not experience secondary entrainment of high-speed airflow were screened out. The selected qualified production records were grouped using the clustering method of regulating valve control parameters to construct several rate intervals.

[0008] When the oxygen lance insertion signal is obtained, the steam pressure at the inlet or outlet of the separation device is detected in real time, the current actual pressure rise rate is calculated, and it is compared with the rate range to retrieve the set of regulating valve control parameters that match and have the strongest historical elimination effect.

[0009] During the oxygen lance insertion process, the control valve is opened based on the retrieved control parameters. The actual pressure rise rate and precursor characteristics of secondary entrainment are continuously monitored during the control process. When the actual pressure rise rate deviates from the expected value or precursor characteristics of secondary entrainment appear, the opening speed of the control valve is adjusted in real time until the pressure rise rate falls back to the safe range and the secondary entrainment phenomenon is eliminated.

[0010] The beneficial effects of this invention are as follows:

[0011] This invention transforms discrete steelmaking production experience into a structured and reusable system by constructing a historical learning and mapping library. It abandons the practice of dividing the data at equal intervals according to the rate of pressure rise, and instead uses the control parameters of the regulating valve as the cluster center to inversely summarize the rate interval. Through mechanisms such as selective shrinking of overlapping regions and merging of similar samples with low sample size, it eliminates the matching ambiguity between parameters and intervals. The mapping library established in this way ensures that each set of control parameters forms a unique, safe, and historically verified optimal association with its applicable rate interval, providing an unambiguous and highly robust decision-making basis for subsequent real-time adaptive control.

[0012] This invention achieves millisecond-level response for condition perception and optimal decision-making through real-time matching and safe extrapolation. At the moment of oxygen lance insertion, it detects the pressure rise rate in real time and quickly retrieves the control parameters with the strongest historical elimination effect from the mapping library. For boundary values ​​and out-of-range conditions, it adopts priority selection and safe extrapolation strategies (extending the opening time) to ensure safe operation under extreme conditions and record data for library updates. This makes each valve control equivalent to calling the most successful operation in hundreds of furnace runs in history, eliminating the dependence on manual experience or fixed flow restrictor plates.

[0013] This invention balances production rhythm and safety through adaptive correction and tiered response. For potential risks such as excessive pressure rate, it prioritizes slowing down and other mitigation strategies, attempting self-recovery without interrupting operation. For risks that have already occurred, such as humidity or water droplet signals, it immediately implements tiered response, decisively intervening until the risk is eliminated. This differentiated correction mechanism avoids excessive intervention that could affect production efficiency and can quickly cut off the steam source in emergencies, fundamentally preventing secondary entrainment and explosions. At the same time, it can use inexpensive saturated steam to replace expensive nitrogen, achieving the dual benefits of safety and energy saving. Attached Figure Description

[0014] The invention will now be further described with reference to the accompanying drawings.

[0015] Figure 1 This is a flowchart illustrating the steps of an embodiment of the present invention;

[0016] Figure 2 This is a partial flowchart of step one of the embodiments of the present invention. Detailed Implementation

[0017] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0018] Example

[0019] Please see Figure 1 and Figure 2 As shown in the embodiment of the present invention, a method for eliminating steam in the return water pipe of an oxygen lance body includes the following steps:

[0020] Step 1: Collect data on the pressure rise rate during the historical production process of the oxygen lance and the corresponding control parameters of the regulating valve. Select qualified production records that have not experienced secondary entrainment of high-speed airflow. Use the clustering method of regulating valve control parameters to group the selected qualified production records and construct several rate intervals.

[0021] The specific process of step one is as follows: First, collect the historical records of oxygen lance production furnaces. The historical records are retrieved from the database of the steelmaking workshop's automated control system.

[0022] Specifically, each historical record includes two types of information: one is the steam pressure rise rate data measured by the pressure sensor installed at the inlet or outlet of the steam-water separator at the moment of oxygen lance insertion, characterized by the pressure increase per second; the other is the control parameters of the regulating valve corresponding to that furnace, specifically including the start time of valve opening from the fully closed position, the valve opening command at each moment during the opening process, and the total time taken from fully closed to fully open.

[0023] Based on the historical records obtained above, qualified production records that did not experience secondary entrainment of high-speed airflow were selected.

[0024] It should be noted that the process for determining whether secondary entrainment has occurred is as follows:

[0025] Check the records of the humidity sensor or water droplet detection device installed on the outlet pipe of the separation unit within ten seconds after the oxygen lance is inserted. If the detected humidity value is always lower than the preset threshold and there is no sudden increase in water droplet signal (i.e. the water droplet signal is not triggered), it is determined that no secondary entrainment has occurred in this batch, and the corresponding production record is retained as a qualified production record. Conversely, if the humidity value instantly exceeds the threshold or the water droplet signal is triggered, it is determined that secondary entrainment has occurred, and the corresponding production record is discarded.

[0026] The selected qualified production records are grouped using a clustering method based on control parameters. The process is as follows:

[0027] Statistically analyze the different control parameters of the regulating valves that appear in all qualified production records, and group those parameters with completely consistent opening speed curves or equal total opening time into the same group, using each group of parameters as a cluster center;

[0028] All qualified records are assigned to the corresponding cluster centers according to the control parameters they actually use, forming several control parameter groups;

[0029] For example, all furnaces using the "five-second uniform speed start" parameter are grouped into one group, all furnaces using the "three-second fast start plus two-second slow start" parameter are grouped into another group, and so on;

[0030] For each control parameter group, extract the pressure rise rate values ​​corresponding to all furnaces within the group, find the minimum and maximum values ​​of the pressure rise rate values, and thus obtain the pressure rise rate range.

[0031] Each control parameter group corresponds to a continuous rate range, the boundaries of which are determined by the extreme values ​​of the actual data. Since the rate ranges of different control parameter groups may overlap—for example, the rate range of the "five-second constant speed opening" group is 0.2 to 0.5 MPa per second, while the range of the "three-second fast opening plus two-second slow opening" group is 0.4 to 0.7 MPa per second—the two ranges overlap between 0.4 and 0.5 MPa. For the overlapping area, further decision-making is required.

[0032] Within the overlapping area, the control parameter with better elimination effect is selected as the recommended parameter. The comparison of elimination effect is based on the average value of steam outlet humidity or the frequency of occurrence of secondary entrainment alarm in the furnace records of each group within the overlapping area.

[0033] When comparing the elimination effects of two or more sets of control parameters, the set with the lower average humidity or the lower alarm frequency is considered to have a better effect. If both indicators exist simultaneously, the alarm frequency should be given priority (because alarms directly correspond to safety hazards).

[0034] Parameters with better performance retain their original range, while parameters with poorer performance have their range boundaries shrunk, that is, overlapping parts are removed, and the end point or start point of their range is adjusted to the edge of the overlapping area.

[0035] It should be noted that if the number of furnaces in a certain control parameter group does not reach the target number, for example, less than five furnaces, the reliability of that group of parameters is low and it is not appropriate to directly treat it as an independent rate range. In this case, the furnaces in that group should be merged into the control parameter group that is most similar to it.

[0036] The similarity criteria are that the difference in the total opening time of the two sets of parameters is less than 10 percent and the shape of the opening curves is similar. After merging, the pressure rise rate range of the merged group is recalculated, and the best-performing set of parameters in the merged group is used as the representative parameter.

[0037] Based on the above clustering method of control parameters, several rate intervals are constructed, and each rate interval uniquely corresponds to a set of control valve parameters. The boundary of each rate interval is determined by the extreme value of the pressure rise rate in which the set of parameters was successfully applied in actual historical data. Each set of parameters within the interval is the optimal choice with verified elimination effect within that set.

[0038] Step one involves collecting historical production data and filtering out qualified production records without secondary entrainment. Using a clustering method based on control valve parameters, a mapping library between pressure rise rate and control parameters was successfully constructed, achieving at least the following results:

[0039] Instead of dividing intervals according to the pressure rise rate, this approach adopts a reverse construction method with the control parameters of the regulating valve as the cluster center. First, different opening speed curves or total opening times actually used in historical successful cases are used as natural grouping criteria. Then, the pressure rise rate range corresponding to each group of parameters is summarized in reverse. Through mechanisms such as optimal shrinkage of overlapping areas and similar merging of low sample groups, the matching ambiguity of the same parameter corresponding to multiple intervals or a single interval corresponding to multiple parameters is eliminated. The mapping library established in this way not only transforms discrete production experience into a structured and reusable knowledge base, but also ensures that each group of control parameters forms a unique, safe, and historically verified optimal association with its applicable rate range, providing an unambiguous and highly robust decision-making basis for subsequent real-time adaptive control.

[0040] Step 2: When the oxygen lance insertion signal is obtained, the steam pressure at the inlet or outlet of the separation device is detected in real time, the current actual pressure rise rate is calculated, and compared with the rate range. The set of regulating valve control parameters that match and have the strongest historical elimination effect is retrieved.

[0041] Among them, the best elimination effect means that under this set of control parameters, the probability of secondary entrainment is the lowest or the steam dryness is the highest.

[0042] The specific process of step two is as follows: when the oxygen lance insertion signal is obtained (when the oxygen lance insertion command is issued), the pressure rise rate is detected in real time.

[0043] Specifically, the pressure sensor installed at the inlet or outlet of the steam-water separator starts to continuously collect steam pressure values. The collection frequency is set to record once every 0.1 seconds. From the moment the oxygen lance insertion signal is triggered, the pressure data is continuously recorded, and the current actual pressure rise rate is calculated using a sliding time window method. The duration of the sliding time window is set to two seconds based on historical experience. That is, every 0.1 seconds, the pressure change value within the most recent two seconds is taken, the total pressure increase is calculated and divided by two seconds to obtain the pressure rise rate in units of pressure increase per second.

[0044] To avoid misjudgment caused by instantaneous pressure fluctuations, the collected raw pressure data is smoothed. For every five consecutive pressure values ​​collected, the maximum and minimum values ​​are removed, and the arithmetic mean of the remaining three values ​​is taken. The average value is used as the effective pressure value at that moment, and the pressure rise rate is calculated based on the smoothed effective pressure value.

[0045] After obtaining the current actual rate of pressure increase, compare it one by one with all rate intervals in the mapping library constructed in step one above.

[0046] The comparison principle is to find the rate range to which the actual pressure rise rate value belongs, that is, the value is greater than or equal to the minimum boundary value of a certain range and less than or equal to the maximum boundary value of the range. Since each rate range has been guaranteed to uniquely correspond to a set of regulating valve control parameters, after the comparison is completed, the set of regulating valve control parameters associated with that range is directly retrieved. This set of parameters is the parameter with the strongest elimination effect in that rate range in historical production, specifically manifested as the lowest frequency of secondary entrainment alarms or the lowest average steam outlet humidity.

[0047] If, during the comparison process, it is found that the current pressure rise rate value falls exactly on the boundary of two adjacent rate intervals (i.e., equal to the maximum value of one interval and equal to the minimum value of another interval), the control parameter set with the lower alarm frequency is prioritized. If the alarm frequencies of the two intervals are the same, the control parameter set with the smaller average steam outlet humidity is selected.

[0048] If the calculated pressure rise rate exceeds the range of all rate intervals in the mapping library due to extremely low or high upstream steam pressure (i.e., below the minimum interval boundary value or above the maximum interval boundary value), the parameters in the mapping library cannot be directly called. Instead, a safe extrapolation strategy is adopted: If the pressure rise rate is lower than the minimum value in the mapping library, the control parameter corresponding to the minimum pressure rise rate interval is called, but the total valve opening time is extended by 30% to open the valve more smoothly; if the pressure rise rate is higher than the maximum value in the mapping library, the control parameter corresponding to the minimum pressure rise rate interval (i.e., the most conservative parameter) is called, and the total opening time is extended by 50% to ensure that secondary entrainment will not occur even under the most dangerous operating conditions. At the same time, this out-of-range event is recorded as supplementary data for subsequent step one when updating the mapping library.

[0049] After retrieving the matching control parameters, wait for the valve control command to be issued during the oxygen lance insertion process.

[0050] Step 2 involves real-time sensing of operating conditions, rapid matching of historical optimal parameters, and output of reliable control parameters. This allows for automatic adaptation of the optimal valve opening strategy the moment the oxygen lance is inserted, reducing secondary entrainment caused by uncontrolled pressure rise rate and ensuring the safety and stability of the steam elimination process.

[0051] Through a two-step collaborative process of historical learning in step one and real-time matching in step two, fundamental improvements have been achieved, moving from passive flow limiting to proactive prediction, from fixed adjustment to adaptive optimization, and from experience-based dependence to data-driven approaches. This has enhanced the safety, economy, and intelligence of the steelmaking process.

[0052] Step 3: During the oxygen lance insertion process, the regulating valve is opened based on the retrieved regulating valve control parameters. During the control process, the actual pressure rise rate and the precursor characteristics of secondary entrainment are continuously monitored. When the actual pressure rise rate is detected to deviate or the precursor characteristics of secondary entrainment appear, the opening speed of the regulating valve is adjusted in real time until the pressure rise rate falls back to the safe range and the secondary entrainment phenomenon is eliminated.

[0053] The specific process of step three is as follows: when the oxygen lance begins actual insertion, the valve opening control is activated according to the retrieved control parameters;

[0054] Specifically, the control parameters include the valve's opening speed curve or total opening time from fully closed to fully open, based on which an opening command is sent to the actuator of the regulating valve, causing the valve to open gradually at a preset rate.

[0055] For example, if the parameter is "five seconds of uniform opening", the valve opening will be increased uniformly within five seconds, increasing by 20% per second; if the parameter is "three seconds of fast opening plus two seconds of slow opening", the opening will increase by about 33% per second in the first three seconds and by about 17% per second in the last two seconds.

[0056] During the opening of the regulating valve, continuous real-time monitoring and dynamic correction are performed. The process is as follows:

[0057] The monitoring content includes:

[0058] The actual pressure rise rate is monitored in the same way as in step two above. If the actual pressure rise rate is detected to exceed 20 percent of the expected value, the real-time correction mechanism is triggered.

[0059] The specific operation process of the real-time correction mechanism is as follows:

[0060] Reduce the valve opening speed to 50% of the current speed and maintain it for 0.5 seconds. For example, if the valve is currently increasing its opening speed at 10% per second, then reduce it to 5% per second. During this 0.5 seconds, continue to monitor the pressure rise rate, humidity value, and water droplet signal.

[0061] After 0.5 seconds, the actual pressure rise rate is re-detected and compared with the expected value. If the pressure rise rate has fallen back to within 120% of the expected value (i.e., not exceeding 20%), the risk is determined to be eliminated. The valve is opened at the decelerated opening speed until the valve is fully open and the original speed is no longer restored.

[0062] If the pressure rise rate still does not decrease after 0.5 seconds and still exceeds 20% of the expected value, but the humidity value and water droplet signal are still normal (i.e., there are still no precursory features), it is considered that simply relying on deceleration cannot effectively suppress the pressure rise rate. At this time, switch to "deceleration hold" mode: keep the valve opening at the current value and do not continue to increase it. Observe for one second. During this second, check the pressure rise rate, humidity value and water droplet signal every 0.2 seconds.

[0063] During the one-second observation period of deceleration holding, if the rate of pressure rise gradually decreases and falls back to within 120% of the expected value, the deceleration holding mode is exited, and the valve is restarted at the current speed (i.e. the decelerated speed) until it is fully open.

[0064] If, during the deceleration and holding process, the humidity value begins to rise abnormally (e.g., the increase exceeds 30% of the baseline value within two consecutive sampling cycles) or the water droplet detection device emits a water droplet signal, it indicates that the potential risk has evolved into an actual precursor to secondary entrainment. At this point, the deceleration and holding strategy should be abandoned immediately, and a graded response procedure should be switched to address the precursory characteristics (i.e., deceleration and holding, minor pullback, emergency shutdown, etc.).

[0065] The precursor characteristics of secondary entrainment are mainly obtained by humidity sensors or water droplet detection devices installed on the outlet pipe of the separation device, mainly the instantaneous rate of change of humidity value and the triggering of water droplet signals.

[0066] Among them, the precursor features of secondary entrainment include, but are not limited to:

[0067] First, the instantaneous change rate of the humidity value exceeds the normal fluctuation range. Specifically, under normal steam drying conditions, the outlet humidity value is usually stable near a low reference value, and its change rate per second does not exceed five percent. When the humidity sensor detects that the increase in humidity value exceeds thirty percent of the reference value within two consecutive sampling cycles (each sampling cycle is 0.1 seconds) and shows a continuous upward trend, it is determined to be a precursor characteristic of secondary entrainment.

[0068] Second, if the water droplet detection device shows intermittent and discontinuous water droplet signals, it should remain silent in the absence of secondary entrainment. If water droplet signals are detected to appear occasionally, such as once or twice per second, and each signal lasts for no more than 0.1 seconds, it indicates that a small number of tiny water droplets have been carried to the outlet by the airflow. This is an early warning that secondary entrainment is about to occur.

[0069] Third, there is an abnormal correlation between humidity value and pressure rise rate. When the actual pressure rise rate exceeds the upper limit of the current rate range in the mapping library, if the outlet humidity value jumps by more than 15 percent in the following second, it constitutes a combined precursor feature. This situation indicates that the high-speed airflow has begun to pick up the water at the bottom of the tank, but has not yet formed a continuous water entrainment.

[0070] Fourth, if the instantaneous fluctuation of the pressure difference between the inlet and outlet of the separation device exceeds 20% of the steady-state value, this pressure difference data can be obtained by pressure sensors installed at the inlet and outlet of the separation device. Under normal operating conditions, the pressure difference changes slowly with the valve opening. If the pressure difference fluctuates violently within 0.5 seconds (increasing or decreasing by more than 20% of the steady-state value), it indicates that the airflow inside the tank is turbulent, which is a hydrodynamic precursor to the impending secondary entrainment.

[0071] If any one of the above four conditions is met, it is determined that there is a risk of secondary smuggling and a real-time correction mechanism is triggered.

[0072] The specific operation process of the real-time correction mechanism is as follows:

[0073] First, stop increasing the valve opening according to the original speed curve, and instead maintain the valve opening at the current value for 0.5 seconds. The purpose of this is to temporarily stabilize the airflow speed inside the tank, giving the water droplets that have been rolled up a chance to settle back to the bottom. After 0.5 seconds, re-detect the pressure rise rate and humidity signal.

[0074] If the signs of secondary entrainment disappear, the valve will continue to open at a slower rate than originally planned. Specifically, the remaining opening time will be doubled. For example, if the original plan was to open the valve fully in two seconds, the remaining opening will be gradually opened over four seconds.

[0075] If the precursors to secondary entrainment do not disappear, implement the following tiered response measures:

[0076] Level 1, deceleration and maintenance: maintain the valve opening at the current value and do not increase it. Observe for one second. During this second, check the humidity signal and water droplet detection status every 0.2 seconds. If the precursor characteristics gradually weaken (for example, the rate of humidity increase drops from 30% to below 15%), switch to the gentle opening mode, extend the remaining opening time to three times the original remaining time, and continue to open the valve.

[0077] The second level is a slight pullback. If the precursory characteristics do not weaken after a one-second deceleration period, or even if the humidity value continues to rise or the water droplet signal becomes more frequent, a slight pullback operation is performed: the valve is gradually closed at a rate of 5% per second until the opening drops to 80% of the current value. During the pullback, the humidity signal is continuously monitored. Once the humidity value begins to decrease, the pullback is stopped immediately, and the valve is reopened at the slowest safe opening speed (the parameter with the longest total opening time in the mapping library).

[0078] Level 3, Emergency Shutdown: If the humidity value continues to rise and exceeds twice the alarm threshold during the pullback process, or if the water droplet detection device emits continuous water droplet signals (more than five times per second), it is determined that secondary entrainment has seriously occurred. At this time, an emergency shutdown command is immediately issued to completely close the regulating valve within 0.5 seconds to cut off the steam source. At the same time, an alarm signal is sent to the steelmaking automation control system to remind the operator that there is an abnormality in the current oxygen lance sealing steam. After the valve is closed, wait for three seconds until the airflow in the tank is completely still and all the water droplets that have been rolled up have settled back to the bottom. Then, reopen the valve at the most conservative opening speed (the total time of the slowest parameter in the mapping library is extended by 50%) to restore the steam supply.

[0079] In the above-mentioned graded response measures, after each level is implemented, the status of the precursor features must be reassessed, and a decision must be made based on the assessment results to upgrade to the next level or downgrade back to normal control. The entire correction process will be fully recorded, including the trigger time, the type of precursor feature, the measures taken, and the final effect, which will serve as the basis for optimization when updating the mapping library in the subsequent step one.

[0080] After the correction is completed, the regulating valve is fully open or has reached a stable opening based on the correction results. The oxygen lance sealing steam flow is stable, and the risk of secondary entrainment has been eliminated. Continue to maintain low-frequency monitoring of the outlet humidity and pressure rise rate of the separation unit (e.g., once every two seconds) until the oxygen lance lift signal appears.

[0081] It should be noted that the reason why different correction methods are used for the two triggering conditions is that the risk levels and physical nature they represent are fundamentally different: the pressure rise rate exceeding the standard is a "potential risk" that has not yet caused actual water entrainment. At this time, although the airflow speed is fast, it has not yet entrained the condensate at the bottom, and there is still room for buffering. Therefore, the priority is to adopt a slowing down and other mitigation strategies, and try to recover on its own without affecting the valve opening progress as much as possible. On the other hand, the appearance of secondary entrainment precursors such as humidity jump or water droplet signals is a "risk that has occurred". It proves that the water at the bottom of the tank has been entrained by the airflow and has begun to be transported to the outlet. If decisive intervention is not made, it will quickly develop into a serious water entrainment accident. Therefore, more proactive and mandatory graded response measures such as suspension, reversal or even emergency shutdown must be implemented immediately. The former is a "preventive" early warning adjustment, while the latter is an emergency correction of "already occurred".

[0082] Step three involves real-time monitoring of the pressure rise rate and precursor characteristics of secondary entrainment, implementing graded corrections during the valve opening process. For "potential risks" where only the pressure rate exceeds the limit, mitigation strategies such as deceleration and deceleration maintenance are employed to ensure self-recovery without affecting the production rhythm. For "occurring risks" where humidity spikes or water droplet signals appear, mandatory interventions such as pause maintenance, minor correction, or even emergency closure are immediately implemented to ensure that secondary entrainment is contained at its inception or localized stage. The entire process forms a safety closed loop from early warning to correction. After correction, monitoring is switched to low frequency until the oxygen lance is lifted. At the same time, event data is fully recorded for updating the mapping library in step one. Thus, while ensuring the safety of steam elimination, adaptive optimization of the control strategy and smooth connection of the production rhythm are achieved.

[0083] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for eliminating steam in the return water pipe of an oxygen lance body, characterized in that: include: Data on the pressure rise rate during the historical production process of the oxygen lance and the corresponding control parameters of the regulating valve were collected. Qualified production records that did not experience secondary entrainment of high-speed airflow were screened out. The selected qualified production records were grouped using the clustering method of regulating valve control parameters to construct several rate intervals. When the oxygen lance insertion signal is obtained, the steam pressure at the inlet or outlet of the separation device is detected in real time, the current actual pressure rise rate is calculated, and it is compared with the rate range to retrieve the set of regulating valve control parameters that match and have the strongest historical elimination effect. During the oxygen lance insertion process, the control valve is opened based on the retrieved control parameters. The actual pressure rise rate and precursor characteristics of secondary entrainment are continuously monitored during the control process. When the actual pressure rise rate deviates or precursor characteristics of secondary entrainment appear, the opening speed of the control valve is adjusted in real time until the pressure rise rate falls back to a safe range and the secondary entrainment phenomenon is eliminated.

2. The method for eliminating steam in the return water pipe of an oxygen lance body according to claim 1, characterized in that: The process for selecting qualified production records that did not experience secondary entrainment by high-speed airflow is as follows: Check the humidity sensor or water droplet detection device installed on the outlet pipe of the separation unit and record the data within ten seconds after the oxygen lance is inserted. If the detected humidity value is always lower than the preset threshold and the water droplet signal is not triggered, it is determined that no secondary entrainment has occurred and it is considered a qualified production record.

3. The method for eliminating steam in the return water pipe of an oxygen lance body according to claim 1, characterized in that: The process of constructing several rate ranges is as follows: Statistically analyze the different control parameters of the regulating valves that appear in all qualified production records, and group the parameters with completely consistent opening speed curves or equal total opening time into the same group, using each group of parameters as a cluster center; All qualified records are assigned to the corresponding cluster centers according to the control parameters they actually use, forming several control parameter groups; For each control parameter group, extract the minimum and maximum values ​​of the pressure rise rate for all furnaces within the group to obtain the corresponding pressure rise rate range. When the rate ranges of different control parameter groups overlap, the control parameter with good elimination effect is selected as the recommended parameter in the overlapping area. The parameter with good effect retains the original range, while the parameter with poor effect shrinks the range boundary to eliminate the overlapping part. Several rate ranges are constructed using the above method, and each rate range uniquely corresponds to a set of control parameters for the regulating valve.

4. The steam elimination method for the oxygen lance body return water pipeline according to claim 3, characterized in that: The process of selecting control parameters with good elimination effect is as follows: When comparing the elimination effects of two or more sets of control parameters, the set with the lower average humidity or the lower alarm frequency is considered to have a better effect. If both indicators exist simultaneously, the alarm frequency should be given priority.

5. The method for eliminating steam in the return water pipe of an oxygen lance body according to claim 1, characterized in that: When calculating the current actual pressure rise rate, a sliding time window method is used, with the window duration set to two seconds, and the calculation is performed once every 0.1 seconds. The collected raw pressure data is smoothed, and after collecting five consecutive pressure values, the maximum and minimum values ​​are removed, and the arithmetic mean of the remaining three values ​​is taken as the effective pressure value.

6. The method for eliminating steam in the return water pipe of an oxygen lance body according to claim 1, characterized in that: The process of retrieving the set of control valve parameters that best matches and has the strongest historical elimination effect is as follows: Find the rate range to which the actual pressure rise rate belongs, and directly retrieve the control parameters of the associated regulating valve. If the current pressure rise rate falls on the boundary of two adjacent rate intervals during the comparison process, the control parameters with the lower alarm frequency will be selected first. If the alarm frequencies of the two intervals are the same, the control parameters with the lower average steam outlet humidity will be selected. If the calculated actual pressure rise rate exceeds all rate ranges, a safe extrapolation strategy is adopted: if the pressure rise rate is lower than the minimum value in the mapping library, the control parameter corresponding to the minimum pressure rise rate range is called, and the total valve opening time is extended by 30%; if the pressure rise rate is higher than the maximum value in the mapping library, the control parameter corresponding to the minimum pressure rise rate range is called, and the total opening time is extended by 50%.

7. The method for eliminating steam in the return water pipe of an oxygen lance body according to claim 1, characterized in that: When the actual pressure rise rate is detected to exceed 20% of the expected value but no warning signs appear, reduce the valve opening speed to 50% of the current speed and maintain it for 0.5 seconds. If the pressure rise rate falls back to within 120% of the expected value, maintain the decelerated speed and continue opening. If it still does not fall back and there are still no warning signs, switch to deceleration holding mode and maintain the valve opening at the current value for one second. If warning signs appear during the deceleration holding process, switch to graded response measures.

8. The method for eliminating steam in the return water pipe of an oxygen lance body according to claim 1, characterized in that: The precursory features of secondary entrainment include: The instantaneous rate of change of humidity exceeds the normal fluctuation range; the water droplet detection device shows intermittent and discontinuous water droplet signals; the correlation between humidity and pressure rise rate is abnormal; the instantaneous fluctuation amplitude of the pressure difference between the inlet and outlet of the separation device exceeds 20% of the steady-state value.

9. The method for eliminating steam in the return water pipe of an oxygen lance body according to claim 1, characterized in that: When signs of secondary entrainment appear, stop increasing the valve opening according to the original speed curve, and instead maintain the valve opening at the current value for 0.5 seconds. After 0.5 seconds, re-detect the pressure rise rate and humidity signal. If the signs of secondary entrainment disappear, the valve will continue to open at a slower rate than originally planned, doubling the original remaining opening time.

10. A method for eliminating steam in the return water pipe of an oxygen lance body according to claim 9, characterized in that: If the signs of secondary entrainment do not disappear, implement tiered response measures: Level 1, Deceleration Hold: Maintain the valve opening at the current value and observe for one second. If the precursory characteristics gradually weaken, switch to the gradual opening mode and extend the remaining opening time to three times the original remaining time. Level 2, slight pullback: If the precursory characteristics do not weaken after the deceleration is maintained, the valve is gradually closed at a rate of 5% per second until the opening drops to 80% of the current value. Once the humidity value begins to decrease, the pullback stops and the valve is reopened at the slowest safe opening speed. Level 3, Emergency Shutdown: If the humidity value exceeds twice the alarm threshold or the water droplet signal appears more than five times per second during the callback process, the regulating valve will be completely shut off within 0.5 seconds, an alarm will be issued, and the valve will be reopened at the most conservative opening speed after waiting for three seconds.

Images