Data-driven intelligent control data transmission system for steam energy-saving desuperheaters

Abstract

This invention discloses a data-driven intelligent control data transmission system for steam energy-saving desuperheaters, relating to the technical field of steam energy-saving desuperheaters. The system includes a control structure extraction module, an intelligent control event analysis module, a comparative abnormal control event analysis module, an abnormal response model construction module, a control structure differentiation module, a control priority ranking module, and a real-time control data response module. By analyzing the relationship between the sampling frequency and data update frequency of each data transmission path during abnormal responses in historical events using control variables, the system determines the data impact relationship under different control structures and provides early warning for potential data asynchrony. This prevents data from being transmitted to the next node's control system before anomalies are confirmed, thereby reducing the reduction in desuperheater regulation function caused by data asynchrony at the source, and improving the system's control accuracy and stability.

Description

Technical Field

[0001] This invention relates to the field of data control technology, specifically a data-driven intelligent control data transmission system for steam energy-saving desuperheaters. Background Technology

[0002] In industrial production, steam, as an important energy medium, is widely used in many fields such as power generation, chemical industry, pharmaceuticals, and food processing. However, in practical applications, the temperature and pressure of steam often need to be precisely adjusted according to different process requirements. A steam desuperheater is a device used to reduce high-temperature steam to a suitable temperature, which is crucial for ensuring the safety and stability of the production process and improving energy efficiency. With the continuous development of industrial automation technology, sensor technology, computer technology, and intelligent control algorithms, the introduction of intelligent control technology can achieve precise control, optimized operation, remote monitoring, and fault diagnosis of steam desuperheaters, thereby improving their performance and reliability, reducing operating costs, and meeting the requirements of modern industrial production for high efficiency, energy saving, and safety. However, some problems still exist alongside intelligent control, such as data synchronization issues caused by data transmission. Because the desuperheater contains multiple sensors and control systems, data transmission requires a certain logical sequence. Asynchronous data can make it difficult for the control system to accurately coordinate the work of various components. For example, if temperature monitoring data updates slowly, while flow regulation and water spray control adjust according to the latest flow data, it may lead to a mismatch between water spray control and actual temperature changes, affecting the control accuracy and stability of the system. Summary of the Invention

[0003] The purpose of this invention is to provide a data-driven intelligent control and data transmission system for steam energy-saving desuperheaters, in order to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a data-driven intelligent control data transmission system for steam energy-saving desuperheaters, the system comprising a steam energy-saving desuperheater control structure extraction module, an intelligent control event analysis module, a comparison abnormal control event analysis module, an abnormal response model construction module, a control structure differentiation module, a control priority sorting module, and a real-time control data response module; The steam energy-saving desuperheater control structure extraction module is used to extract the control structure based on the historical control data transmission path of the steam energy-saving desuperheater; the control structure includes steady-state control structure and special control structure; The intelligent control event analysis module is used to obtain the corresponding data update rate in the control event record control data transmission path, in order to analyze the normal update rate range under the steady-state control structure. The abnormal control event analysis module is used to use abnormal control events recorded by the control structure that is the same as the steady-state control structure as a control group, and analyzes and outputs the abnormal update rate range. The abnormal response model construction module is used to construct the corresponding abnormal response model under the steady-state control structure based on the abnormal update rate range output by the control abnormal control event analysis module. The control structure differentiation module is used to distinguish the degree of difference between the abnormal control events corresponding to special control structures and the data update rates recorded by steady-state control structures. The regulation priority ranking module is used to construct an abnormal response model for special regulation structures with differences greater than a threshold, and output the priority of the regulation data transmission path corresponding to the special regulation structure. The real-time control data response module is used to acquire the real-time control structure, predict abnormal transmissions before the control data transmission path response based on priority, and issue early warning responses based on the prediction results.

[0005] Furthermore, the steam energy-saving desuperheater control structure extraction module forms a control data transmission path between adjacent signal transmission ends and receiving ends, and generates all independent control data transmission paths under a control event according to the control sequence. The steady-state control structure refers to a temperature sensor monitoring the steam temperature and feeding the temperature signal back to the control system through the first control data transmission path. The control system compares the actual temperature with the set temperature. If the actual temperature is higher than the set temperature, the control system will issue a command to the water spray valve through the second control data transmission path to spray water for cooling. At the same time as water spraying, the flow sensor monitors the steam flow rate and sends it to the control system through the third control data transmission path. The control system then transmits a signal to the water spray valve through the fourth control data transmission path to adjust the opening of the valve, so that the water spray volume matches the steam flow rate. A special control structure refers to a control structure that differs from the control structure in which the data transmission order of the first, second, third, and fourth control data transmission paths differs from that of the steady-state control structure.

[0006] Furthermore, the intelligent control event analysis module includes a normal control event filtering unit, a data update rate analysis unit, and a normal update rate range output unit; The normal control event screening unit is used to screen the control events corresponding to the desuperheater when the steam temperature is reduced to below the threshold after intelligent control and the control time is less than or equal to the preset time threshold as normal control events. The data update rate analysis unit is used to obtain the data update rate of each control data transmission path record in normal control events that are recorded as steady-state control structures. The data update rate refers to the number of times the required data is updated per unit time in the corresponding control data transmission path. The normal update rate range output unit is used to extract the minimum and maximum data update rates R1min and R1max of each control data transmission path to form the normal update rate range A of each control data transmission path under the steady-state control structure, where A = [R1min, R1max].

[0007] Furthermore, the abnormal control event analysis module includes an abnormal factor determination unit and an abnormal update rate range output unit; The abnormal factor determination unit is used to mark the response nodes of the control event corresponding to the steady-state control structure. The response node refers to the intelligent device that collects or responds to data. The response node at both ends of the control data transmission path is extracted as a monitoring item; all monitoring items in the control event corresponding to the steady-state control structure are generated. The monitoring items corresponding to the location where regulation stops or anomalies occur in the regulation event are marked as core anomaly monitoring items. At the same time, the monitoring items of the related paths affected by the core anomaly monitoring items are marked as related anomaly monitoring items. An event set is formed by extracting anomaly regulation events from the control group that contain only one core anomaly monitoring item and no more than two related anomaly monitoring items. The related anomaly monitoring items must meet the requirement that the degree of anomaly is less than 50% of that of the core anomaly monitoring items. The degree of anomaly is measured by the percentage of the data update rate deviating from the normal range. Divide the event set into units based on the same core anomaly monitoring item and the same type of anomaly control events; calculate the update rate deviation Q of each unit event set in the same steady-state control structure, Q={[∑(G1-G0)} 2 ] / m} 1 / 2 G1 represents the absolute difference between the update rate of abnormal control event records and the maximum value of the normal update rate interval A, R1max, in the unit event set, i.e., G1 = |abnormal update rate - R1max|, G0 represents the average of all differences in the unit event set, and m represents the total number of abnormal control events recorded in the unit event set. Abnormal control events where the difference between the update rate deviation Q and the rate difference threshold is greater than the rate deviation threshold are selected as factor abnormal control events. The determination of abnormal control events is to control the factors affecting the occurrence of abnormal control events by changes in data update frequency, rather than other variables; the calculation of deviation reflects the overall dispersion of the data, and the smaller the deviation, the closer the differences within the set are; The abnormal update rate interval output unit is used to extract the minimum and maximum data update rates R2min and R2max of the corresponding control data transmission path in the abnormal control event based on the abnormal control event, and form the abnormal update rate interval B, B=[R2min,R2max].

[0008] Furthermore, the anomaly response model construction module includes a sampling frequency extraction unit, a data pair matching unit, and a model output unit; The sampling frequency extraction unit is used to extract the sampling frequency F1 of the core abnormal monitoring item of each factor abnormal regulation event record in each unit event set; The data pair matching unit is used to match the sampling frequency F1 output by the sampling frequency extraction unit with the data update sampling frequency R2 of the corresponding factor abnormal regulation event record, using the same factor abnormal regulation event as the matching condition to generate a data pair D, D=(F1,R2); where F1 is the independent variable of the data pair and R2 is the dependent variable of the data pair. The model output unit is used to traverse all factor abnormal control events within a unit event set and obtain corresponding data pairs, constructing the functional relationship Y corresponding to each unit event set. R =a×X F +ε, substitute the data pair D, calculate the influence coefficient a and error term ε, and output the abnormal response model Y corresponding to each unit event set. R .

[0009] The sampling frequency is analyzed because different data sampling devices may have different sampling frequencies before different data transmission links. The sampling frequency affects the data update rate to a certain extent. By analyzing the relationship between the sampling frequency and the data update rate from the selected abnormal control events, the influence relationship of each data transmission path can be effectively established. Thus, before the monitored items enter the data transmission path, the sampling data can be used for analysis and prediction to reduce the possibility of control abnormalities caused by data delays in steam energy-saving desuperheaters.

[0010] Furthermore, the regulatory structure differentiation module includes a special regulatory structure classification unit and a difference discrimination calculation unit; The special control structure classification unit is used to classify the abnormal control events corresponding to the special control structure separately according to the classification method of each unit event set in the steady-state control structure; The difference discrimination calculation unit is used to extract the data update rate of each type of abnormal control event record under the special control structure, and calculate the difference between the data update rate and the maximum data update rate of the abnormal control event record under the corresponding steady-state control structure of the same type. The number P of the difference greater than the difference threshold is marked, and the difference discrimination L corresponding to each type of abnormal control event under the special control structure is calculated, L=P / N, where N represents the total number of records of each type of abnormal control event.

[0011] Furthermore, the regulation priority ranking module includes a special response model construction unit and a priority analysis unit; The special response model construction unit is used when the difference discrimination degree is greater than or equal to the difference discrimination degree threshold. A difference discrimination degree greater than the threshold indicates that although the data control transmission path of the anomaly is the same in the special control structure, the data update frequency that causes the anomaly is very different. Therefore, the model needs to be re-analyzed and reconstructed for this type. The transmission signal is sent to the anomaly response model construction module to obtain the data pairs of each anomaly control event under the special control structure and to construct the special response model of the special control structure. The priority analysis unit is used to analyze special control structures where there are simultaneous response data control transmission paths, specifically: Obtain the influence coefficients and path importance weights of the corresponding response model records for the simultaneous response data regulation transmission path. Sort the products of the influence coefficients and path importance weights in descending order to generate a priority sequence for the corresponding data regulation transmission path.

[0012] The larger the influence coefficient, the greater the impact on the data update frequency. The order of the coefficients is used as the priority order of the data control transmission path corresponding to the model.

[0013] Furthermore, the real-time control data response module compares the acquired real-time control structure with the historical steady-state control structure and special control structures. When the sampling frequency for each monitoring item is obtained, the abnormal response model of the corresponding data control path is used to obtain the real-time model output value. If the real-time model output value belongs to the corresponding abnormal update rate interval B, an early warning is issued when the data control transmission path does not respond to adjacent nodes; if the real-time model output value does not belong to the corresponding abnormal update rate interval B, monitoring continues. When the real-time control structure records the sampling frequency in multiple different types of data control transmission paths, it analyzes the output value of the abnormal response model for each type of path according to priority; it issues an early warning when the abnormal update rate range is met, and continues monitoring if it is not met.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention analyzes the control structure in a steam energy-saving desuperheater, distinguishes the transmission paths of different data signals, defines the macroscopic transmission mode through microscopic analysis, and thus realizes the visualization and analyzability of data. 2. This invention also analyzes the relationship between the sampling frequency and data update frequency of each data transmission path in historical events by controlling variables, in order to determine the data influence relationship under different control structures, and realize the prediction and early warning when data may be out of sync, so as to prevent the transmission anomaly from being confirmed only when the control system of the next node responds. This reduces the reduction of the decelerator's adjustment function caused by data asynchrony from the source, thereby improving the control accuracy and stability of the system. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the data transmission system for the intelligent control of a data-driven steam energy-saving desuperheater according to the present invention. Detailed Implementation

[0016] Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0017] Example: Figure 1 As shown, the present invention provides a data-driven intelligent control data transmission system for steam energy-saving desuperheaters. The system includes a steam energy-saving desuperheater control structure extraction module, an intelligent control event analysis module, a comparison abnormal control event analysis module, an abnormal response model construction module, a control structure differentiation module, a control priority sorting module, and a real-time control data response module. The steam energy-saving desuperheater control structure extraction module is used to extract the control structure based on the historical control data transmission path of the steam energy-saving desuperheater; the control structure includes steady-state control structure and special control structure; The intelligent control event analysis module is used to obtain the corresponding data update rate in the control event record control data transmission path, in order to analyze the normal update rate range under the steady-state control structure. The abnormal control event analysis module is used to use abnormal control events recorded by the control structure that is the same as the steady-state control structure as a control group, and analyzes and outputs the abnormal update rate range. The abnormal response model construction module is used to construct the corresponding abnormal response model under the steady-state control structure based on the abnormal update rate range output by the control abnormal control event analysis module. The control structure differentiation module is used to distinguish the degree of difference between the abnormal control events corresponding to special control structures and the data update rates recorded by steady-state control structures. The regulation priority ranking module is used to construct an abnormal response model for special regulation structures with differences greater than a threshold, and output the priority of the regulation data transmission path corresponding to the special regulation structure. The real-time control data response module is used to acquire the real-time control structure, predict abnormal transmissions before the control data transmission path response based on priority, and issue early warning responses based on the prediction results.

[0018] The steam energy-saving desuperheater control structure extraction module forms a control data transmission path by connecting adjacent signal transmission ends and receiving ends, and generates all independent control data transmission paths under a control event according to the control sequence. Adjacent means logically adjacent, that is, the path boundary is determined based on the causal relationship of signal transmission. When the output signal of device A is directly used as the input signal of device B, A and B constitute adjacent signal transmission ends / receiving ends, forming an independent control data transmission path. As shown in the embodiment: In the steady-state control structure, the first control data transmission path (temperature sensor → control system) and the second control data transmission path (control system → spray valve) are independent paths because the output signal of the temperature sensor is directly input to the control system (logically adjacent), and the output signal of the control system is directly input to the spray valve (logically adjacent), and there is a clear causal relationship boundary between the two. Similarly, the third path (flow sensor → control system) and the fourth path (control system → spray valve) are also independent paths.

[0019] The steady-state control structure refers to a temperature sensor monitoring the steam temperature and feeding the temperature signal back to the control system through the first control data transmission path. The control system compares the actual temperature with the set temperature. If the actual temperature is higher than the set temperature, the control system will issue a command to the water spray valve through the second control data transmission path to spray water for cooling. At the same time as water spraying, the flow sensor monitors the steam flow rate and sends it to the control system through the third control data transmission path. The control system then transmits a signal to the water spray valve through the fourth control data transmission path to adjust the opening of the valve, so that the water spray volume matches the steam flow rate. A special control structure refers to a control structure whose data transmission order differs from that of the first, second, third, and fourth control data transmission paths in the steady-state control structure. The data transmission order is determined based on the 'signal triggering logic priority,' specifically categorized into 'temperature-priority triggering,' 'flow-priority triggering,' and 'synchronous triggering.' The steady-state control structure is categorized as 'temperature-priority triggering,' meaning the temperature sensor signal is transmitted to the control system first, followed by the flow sensor signal. If 'flow-priority triggering' or 'synchronous triggering' occurs, it is classified as a special control structure. Synchronous triggering limits the simultaneous transmission of temperature and flow signals within ±0.5 seconds.

[0020] As shown in the example: When the steam load changes rapidly, the flow sensor quickly detects the change in steam flow and transmits the signal to the control system. The control system first makes an initial adjustment to the opening of the spray valve based on the change in flow to adapt to the new steam flow. At the same time, the temperature sensor continuously monitors the steam temperature. Since load changes may cause fluctuations in steam temperature, the control system further and more precisely adjusts the opening of the spray valve based on the real-time results of temperature monitoring. If the steam load changes rapidly, the flow signal is transmitted before the temperature signal (time difference > 0.5s), which is a special control structure of 'flow priority triggering'. If the two are transmitted synchronously (time difference ≤ 0.5s), it is also classified as a special control structure, because this transmission logic is fundamentally different from the 'temperature priority' of the steady-state structure.

[0021] The path sequence recorded in the above cases is inconsistent with the data transmission path of the steady-state control structure.

[0022] The intelligent control event analysis module includes a normal control event filtering unit, a data update rate analysis unit, and a normal update rate range output unit; The normal control event screening unit is used to screen control events that occur when the desuperheater reduces the steam temperature to below the threshold after intelligent control and the control time is less than or equal to the preset time threshold. The preset time threshold is determined based on historical control data under the rated operating conditions of the desuperheater. The 95th percentile of the control time of 'steam temperature dropping from above the threshold to the target value' in the steady-state control structure within the past 12 months is taken as the preset time threshold. The data update rate analysis unit is used to obtain the data update rate of each control data transmission path record in normal control events that are recorded as steady-state control structures. The data update rate refers to the number of times the required data is updated per unit time in the corresponding control data transmission path. The normal update rate range output unit is used to extract the minimum and maximum data update rates R1min and R1max of each control data transmission path to form the normal update rate range A of each control data transmission path under the steady-state control structure, where A = [R1min, R1max].

[0023] The abnormal control event analysis module includes an abnormal factor identification unit and an abnormal update rate range output unit. The abnormal factor determination unit is used to mark the response nodes of the control event corresponding to the steady-state control structure. The response node refers to the intelligent device that collects or responds to data, such as the temperature sensor, flow sensor and control system in the energy-saving de-cooler. The response node at both ends of the control data transmission path is extracted as a monitoring item. All monitoring items corresponding to the control event of the steady-state control structure are generated. The monitoring item corresponding to the location where control stops or an anomaly occurs in the control event is marked as the core anomaly monitoring item, and the monitoring items of the associated path affected by the core anomaly monitoring item are marked as associated anomaly monitoring items. An event set is formed by extracting anomaly control events from the control group that contain only 1 core anomaly monitoring item and no more than 2 associated anomaly monitoring items. The associated anomaly monitoring items must meet the requirement that the degree of anomaly is less than 50% of that of the core anomaly monitoring item. The degree of anomaly is measured by the percentage of the data update rate deviating from the normal range. As shown in the example, when the second control data transmission path (control system → water spray valve) is abnormal, the first path (temperature sensor → control system) will have a continuously high temperature due to the water spray valve not operating, and the data update rate will be abnormal (20% deviation from the normal range). At this time, the second path is the core anomaly monitoring item (60% deviation from the normal range), and the first path is the associated anomaly monitoring item. This event meets the screening condition of '1 core + 1 associated' and can be included in the event set. Divide the event set into units based on the same core anomaly monitoring item and the same type of anomaly control events; calculate the update rate deviation Q of each unit event set in the same steady-state control structure, Q={[∑(G1-G0)} 2 ] / m} 1 / 2 G1 represents the absolute difference between the update rate of abnormal control event records and the maximum value of the normal update rate interval A, R1max, in the unit event set, i.e., G1 = |abnormal update rate - R1max|, G0 represents the average of all differences in the unit event set, and m represents the total number of abnormal control events recorded in the unit event set. Abnormal control events where the difference between the update rate deviation Q and the deviation threshold Q0 is less than or equal to the deviation threshold Q0 are selected as factor abnormal control events; the deviation threshold Q0 is calculated by using the standard deviation σ of the 3σ principle to calculate the standard deviation σ of the update rate deviation in the unit event set, and Q0 = 3σ is taken; the rate difference threshold is set based on the maximum value R1max of the normal update rate interval A, and the rate difference threshold is taken as 0.3 × R1max; The determination of abnormal control events is to control the factors affecting the occurrence of abnormal control events by changes in data update frequency, rather than other variables; the calculation of deviation reflects the overall dispersion of the data, and the smaller the deviation, the closer the differences within the set are; The abnormal update rate interval output unit is used to extract the minimum and maximum data update rates R2min and R2max of the corresponding control data transmission path in the abnormal control event based on the abnormal control event, and form the abnormal update rate interval B, B=[R2min,R2max].

[0024] The abnormal response model construction module includes a sampling frequency extraction unit, a data pair matching unit, and a model output unit; The sampling frequency extraction unit is used to extract the sampling frequency F1 of the core abnormal monitoring item of each factor abnormal regulation event record in each unit event set; The data pair matching unit is used to match the sampling frequency F1 output by the sampling frequency extraction unit with the data update sampling frequency R2 of the corresponding factor abnormal regulation event record, using the same factor abnormal regulation event as the matching condition to generate a data pair D, D=(F1,R2); where F1 is the independent variable of the data pair and R2 is the dependent variable of the data pair. The model output unit is used to traverse all factor abnormal control events within a unit event set and obtain corresponding data pairs, constructing the functional relationship Y corresponding to each unit event set. R =a×X F +ε, substitute the data pair D, calculate the influence coefficient a and error term ε, and output the abnormal response model Y corresponding to each unit event set. R The linear relationship between F1 and R2 can be verified through controlled variable experiments: Under the conditions of fixed data volume (512KB / packet) and fixed transmission bandwidth (100Mbps), R2 data corresponding to different sampling frequencies F1 (1Hz-10Hz) are collected. Pearson correlation analysis yields a correlation coefficient r=0.92 (P<0.01), indicating a significant linear correlation between the two. Therefore, a linear function is used to construct the model. The error term ε in the model includes compensation factors for data volume and transmission bandwidth, specifically ε=k1×(data volume / standard data volume)+k2×(standard bandwidth / transmission bandwidth), where k1 and k2 are weighting coefficients (k1=0.15, k2=0.2 obtained by gradient descent fitting). The standard data volume is set to 512KB / packet, and the standard bandwidth is set to 100Mbps to correct the influence of non-sampling frequency factors on R2. The sampling frequency is analyzed because different data sampling devices may have different sampling frequencies before different data transmission links. The sampling frequency affects the data update rate to a certain extent. By analyzing the relationship between the sampling frequency and the data update rate from the selected abnormal control events, the influence relationship of each data transmission path can be effectively established. Thus, before the monitored items enter the data transmission path, the sampling data can be used for analysis and prediction to reduce the possibility of control abnormalities caused by data delays in steam energy-saving desuperheaters.

[0025] The regulatory structure differentiation module includes a special regulatory structure classification unit and a difference discrimination calculation unit; The special control structure classification unit is used to classify the abnormal control events corresponding to the special control structure separately according to the classification method of each unit event set in the steady-state control structure; The difference discrimination calculation unit is used to extract the data update rate of each type of abnormal control event record under the special control structure, and calculate the difference between the data update rate and the maximum data update rate of the abnormal control events recorded under the corresponding steady-state control structure of the same type. The number P of differences greater than the difference threshold is marked, and the difference discrimination L corresponding to each type of abnormal control event under the special control structure is calculated, L = P / N, where N represents the total number of records for each type of abnormal control event. The difference threshold is set based on the historical normal operation data of each control data transmission path under the "steady-state control structure," and the average data update rate of the path during 72 consecutive hours without abnormalities within the past 3 months is statistically analyzed. This average value is used as the "baseline update rate" for the corresponding path. Combined with the operating characteristics of the steam desuperheater, the difference threshold is calculated using the "3σ principle." That is, the standard deviation σ corresponding to the baseline update rate is first calculated, and the difference threshold is set to 3σ.

[0026] The regulation priority ranking module includes a special response model construction unit and a priority analysis unit; The special response model construction unit is used when the difference discrimination degree is greater than or equal to the difference discrimination degree threshold. A difference discrimination degree greater than the threshold indicates that although the data control transmission path of the anomaly occurs in the special control structure, the data update frequency that causes the anomaly differs greatly. Therefore, this type needs to be re-analyzed and the model reconstructed. The transmission signal is sent to the anomaly response model construction module to obtain the data pairs of each anomaly control event under the special control structure and to construct the special response model of the special control structure. The difference discrimination degree threshold is set based on the historical rate difference data of the associated path under the steady-state control structure. The average value μ and the standard deviation σ' of the difference are calculated, and the difference discrimination degree threshold is taken as μ + 2σ'.

[0027] When the difference discrimination is less than the difference discrimination threshold, the model response mode of the special control structure is the same as the abnormal response model of the corresponding steady-state control structure of the same type. The priority analysis unit is used to analyze special control structures where there are simultaneous response data control transmission paths, specifically: The impact coefficients and path importance weights of the corresponding response model records for the simultaneous response data regulation transmission paths are obtained. The products of the impact coefficients and path importance weights are then sorted in descending order to generate a priority sequence for the corresponding data regulation transmission paths. The path importance weights are calculated using Fault Model and Impact Analysis (FMEA). Specifically, this includes the following: For each control data transmission path, a quantitative score is assigned based on three dimensions: "Severity (S)," "Occurrence Probability (O)," and "Detectability (D)" (1-10 points, with higher scores corresponding to higher risks). Specific scoring can be based on the following: S: 10 points: Abnormal path directly caused the steam temperature to exceed the limit by more than 50°C, resulting in equipment shutdown; 8 points: Abnormal path caused temperature deviation of 20℃-50℃, requiring manual intervention; 5 points: Abnormal path caused temperature deviation of 10℃-20℃, which the system can automatically correct; 2 points: The path anomaly only caused data fluctuations of <10℃, which did not affect temperature control; O: 10 points: ≥3 abnormalities per day; 8 points: 1-2 abnormal occurrences per day; 5 points: 3-5 abnormalities per week; 2 points: Average number of abnormalities per month ≤ 2; 1 point: No abnormalities during the quarter; D: 10 points: The anomaly can only be detected after more than 10 minutes from its occurrence; 8 points: Detected 5-10 minutes after the anomaly occurred; 5 points: Detected 1-5 minutes after the anomaly occurred; 2 points: Detected within 1 minute of the anomaly occurring; 1 point: Early warnings can be given before an anomaly occurs; For the first path: temperature sensor → control system; Severity (S): 8 points; This path anomaly, such as a delay in temperature data transmission, will cause the control system to be unable to obtain the real-time temperature, directly causing the water spray valve to adjust lag, which in turn will lead to a temperature deviation of 20℃-50℃, requiring manual intervention. Probability of occurrence (O): 5 points. Based on the statistical data of the past 6 months, anomalies caused by signal interference on this path occur an average of 3 times per week; Detectability (D): 5 points. The existing real-time model can detect the anomaly within 3 minutes of its occurrence by detecting a "data update rate lower than the normal range"; Calculate the Risk Priority Number (RPN): RPN = S × O × D = 8 × 5 × 5 = 200; For the second path: Control system → Spray valve; Severity (S): 5 points. An anomaly in this path (such as a delay in command transmission) will cause the water spray valve to lag, but the temperature sensor can still provide real-time temperature feedback. The control system can correct this by sending additional commands, and the temperature deviation is only 10℃-20℃.

[0028] Probability of occurrence (O): 2 points. Statistics from the past 6 months show that anomalies caused by communication interface failures occur an average of once per month on this path.

[0029] Detectability (D): 2 points. The system can detect an anomaly within 30 seconds of its occurrence via the "missing valve action feedback signal".

[0030] Calculate RPN: RPN = 5 × 2 × 2 = 20.

[0031] For the third path: flow sensor → control system; Severity (S): 5 points. An abnormal path (such as incorrect flow data transmission) will cause a mismatch between the water spray volume and the steam flow rate, and the temperature will exceed the range by 10℃-20℃. However, the temperature sensor can report the deviation, and the system can adjust automatically.

[0032] Probability of occurrence (O): 5 points. Statistics from the past 6 months show an average of 2 occurrences per week (due to sampling deviation of the flow sensor).

[0033] Detectability (D): 5 points. The real-time model can detect anomalies within 4 minutes of their occurrence via "abnormal correlation between flow and temperature data".

[0034] Calculate RPN: RPN = 5 × 5 × 5 = 125; For the fourth path: Control system → Spray valve; Severity (S): 2 points. This path is only used for "fine-tuning the water spray volume". In case of an anomaly, the control system can achieve coarse adjustment through the second path. Temperature fluctuations <10℃ will not affect the core control.

[0035] Probability of occurrence (O): 1 point. No abnormal records in the past 6 months.

[0036] Detectability (D): 2 points. The system can detect the change in flow rate within 1 minute by indicating that "no change in flow rate after fine-tuning".

[0037] Calculate RPN: RPN = 2 × 1 × 2 = 4.

[0038] The total RPN for all paths is calculated as follows: 200 (first path) + 20 (second path) + 125 (third path) + 4 (fourth path) = 349.

[0039] Obtain the importance weights of each path: First path weight = 200 / 349 ≈ 0.573 (rounded to three decimal places); The weight of the second path is approximately 20 / 349. The weight of the third path = 125 / 349 ≈ 0.358; The weight of the fourth path is approximately 0.011 (4 / 349).

[0040] Assuming that the impact coefficient of the first path is 1.2 and the impact coefficient of the second path is 1.8, calculated using the anomaly response model; Priority score = Influence coefficient × Path importance weight: Score for the first path = 1.2 × 0.573 ≈ 0.688; The score for the second path is approximately 1.8 × 0.057 ≈ 0.103.

[0041] Priority ranking: First path (0.688) > Second path (0.103), which is consistent with the actual risk assessment logic that "abnormal temperature data transmission paths pose a greater risk" and avoids the lag in critical path early warning.

[0042] The real-time control data response module compares the acquired real-time control structure with the historical steady-state control structure and special control structures. For each monitoring item, the sampling frequency is used to obtain the abnormal response model of the corresponding data control path. Substituting the specific real-time sampling frequency yields the real-time model output value. If the real-time model output value belongs to the corresponding abnormal update rate interval B, an early warning is issued when the data control transmission path does not respond to adjacent nodes. If the real-time model output value does not belong to the corresponding abnormal update rate interval B, monitoring continues. Based on newly added abnormal event data, R2min and R2max are recalculated, and an early warning extension interval is set to B ± 0.2 × (R2max - R2min). When the real-time model output value falls into the early warning extension interval, a new abnormality type investigation process is triggered. Manual verification confirms whether it is a new abnormality; if confirmed, interval B is updated.

[0043] When the real-time control structure records the sampling frequency in multiple different types of data control transmission paths, it analyzes the output value of the abnormal response model for each type of path according to priority; it issues an early warning when the abnormal update rate range is met, and continues monitoring if it is not met.

[0044] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A data-driven intelligent control and data transmission system for steam energy-saving desuperheaters, characterized in that: The system includes a steam energy-saving desuperheater control structure extraction module, an intelligent control event analysis module, a comparison abnormal control event analysis module, an abnormal response model construction module, a control structure differentiation module, a control priority ranking module, and a real-time control data response module. The steam energy-saving desuperheater control structure extraction module is used to extract the control structure based on the historical control data transmission path of the steam energy-saving desuperheater; the control structure includes a steady-state control structure and a special control structure; The intelligent control event analysis module is used to obtain the corresponding data update rate in the control event record control data transmission path, so as to analyze the normal update rate range under the steady-state control structure. The control abnormal regulation event analysis module is used to take the abnormal regulation events recorded by the regulation structure that is the same as the steady-state regulation structure as the control group, and analyze and output the abnormal update rate range. The abnormal response model construction module is used to construct the corresponding abnormal response model under the steady-state control structure based on the abnormal update rate range output by the abnormal control event analysis module. The control structure differentiation module is used to distinguish the degree of difference between the abnormal control events corresponding to special control structures and the data update rate recorded by steady-state control structures. The regulation priority sorting module is used to construct an abnormal response model for special regulation structures with differences greater than a threshold, and output the regulation data transmission path priority corresponding to the special regulation structure. The real-time control data response module is used to acquire the real-time control structure, predict abnormal transmissions before the control data transmission path response based on priority, and issue an early warning response based on the prediction results.

2. The data-driven intelligent control and data transmission system for steam energy-saving desuperheaters according to claim 1, characterized in that: The steam energy-saving desuperheater control structure extraction module forms a control data transmission path by connecting adjacent signal transmission ends and receiving ends, and generates all independent control data transmission paths under a control event according to the control sequence. The steady-state control structure refers to a temperature sensor monitoring the steam temperature and feeding the temperature signal back to the control system through a first control data transmission path. The control system compares the actual temperature with the set temperature. If the actual temperature is higher than the set temperature, the control system issues a command to the water spray valve through a second control data transmission path to spray water for cooling. At the same time as water spraying, a flow sensor monitors the steam flow rate and sends it to the control system through a third control data transmission path. The control system then transmits a signal to the water spray valve through a fourth control data transmission path to adjust the opening of the valve based on the steam flow rate, so that the water spray volume matches the steam flow rate. The special control structure refers to the control structure that corresponds to a situation where the data transmission order of the first control data transmission path, the second control data transmission path, the third control data transmission path, and the fourth control data transmission path in the steady-state control structure is different.

3. The data-driven intelligent control and data transmission system for steam energy-saving desuperheaters according to claim 1, characterized in that: The intelligent control event analysis module includes a normal control event filtering unit, a data update rate analysis unit, and a normal update rate range output unit. The normal control event screening unit is used to screen the control event corresponding to the situation where the steam temperature of the desuperheater is reduced to below the threshold after intelligent control and the control time is less than or equal to the preset time threshold as a normal control event. The data update rate analysis unit is used to obtain the data update rate of each control data transmission path record for events recorded as steady-state control structures in normal control events. The data update rate refers to the number of times the data to be transmitted is updated per unit time in the corresponding control data transmission path. The normal update rate interval output unit is used to extract the minimum and maximum data update rates R1min and R1max of each control data transmission path to form the normal update rate interval A of each control data transmission path under the steady-state control structure, where A = [R1min, R1max].

4. The data-driven intelligent control and data transmission system for steam energy-saving desuperheaters according to claim 3, characterized in that: The control abnormal regulation event analysis module includes an abnormal factor determination unit and an abnormal update rate interval output unit; The abnormal factor determination unit is used to mark the response nodes of the control event corresponding to the steady-state control structure. The response node refers to the intelligent device that collects data or response data. The response node linked by the control data transmission path is extracted as a monitoring item. Generate all monitoring items in the control event corresponding to the steady-state control structure; The monitoring items corresponding to the location where regulation stops or anomalies occur in the regulation event are marked as core anomaly monitoring items. At the same time, the monitoring items of the related paths affected by the core anomaly monitoring items are marked as related anomaly monitoring items. An event set is formed by extracting anomaly regulation events from the control group that contain only one core anomaly monitoring item and no more than two related anomaly monitoring items. The related anomaly monitoring items must meet the requirement that the degree of anomaly is less than 50% of that of the core anomaly monitoring items. The degree of anomaly is measured by the percentage of the data update rate deviating from the normal range. The event set is divided into a single event set based on abnormal control events that share the same core abnormality monitoring item and belong to the same type. Calculate the update rate deviation Q of each unit event set in the same steady-state control structure, Q={[∑(G1-G0)} 2 ] / m} 1 / 2 G1 represents the absolute difference between the update rate of abnormal control event records and the maximum value of the normal update rate interval A, R1max, in the unit event set, i.e., G1 = |abnormal update rate - R1max|, G0 represents the average of all differences in the unit event set, and m represents the total number of abnormal control events recorded in the unit event set. Abnormal control events where the difference between the update rate deviation Q and the rate difference threshold is greater than the rate deviation threshold are selected as factor abnormal control events. The abnormal update rate interval output unit is used to extract the minimum and maximum data update rates R2min and R2max of the corresponding control data transmission path recorded in the abnormal control event based on the abnormal control event, and form the abnormal update rate interval B, B=[R2min,R2max].

5. The data-driven intelligent control and data transmission system for steam energy-saving desuperheaters according to claim 4, characterized in that: The abnormal response model construction module includes a sampling frequency extraction unit, a data pair matching unit, and a model output unit; The sampling frequency extraction unit is used to extract the sampling frequency F1 of the core abnormal monitoring item of each factor abnormal regulation event record in each unit event set; The data pair matching unit is used to match the sampling frequency F1 output by the sampling frequency extraction unit with the data update sampling frequency R2 of the corresponding factor abnormal regulation event record, and use the same factor abnormal regulation event as the matching condition to generate a data pair D, D=(F1,R2); where F1 is the independent variable of the data pair and R2 is the dependent variable of the data pair. The model output unit is used to traverse all factor abnormal control events within a unit event set and obtain corresponding data pairs, constructing the functional relationship Y corresponding to each unit event set. R =a×X F +ε, substitute the data pair D, calculate the influence coefficient a and error term ε, and output the abnormal response model Y corresponding to each unit event set. R .

6. The data-driven intelligent control and data transmission system for steam energy-saving desuperheaters according to claim 4, characterized in that: The regulatory structure differentiation module includes a special regulatory structure classification unit and a difference differentiation calculation unit; The special control structure classification unit is used to classify the abnormal control events corresponding to the special control structure separately according to the classification method of each unit event set in the steady-state control structure; The difference discrimination calculation unit is used to extract the data update rate of each type of abnormal control event record under the special control structure, and calculate the difference between the data update rate and the maximum data update rate of the abnormal control event record under the corresponding steady-state control structure of the same type. The number P of the difference values ​​greater than the difference threshold is marked, and the difference discrimination L corresponding to each type of abnormal control event under the special control structure is calculated, where L = P / N, and N represents the total number of records of each type of abnormal control event.

7. The data-driven intelligent control and data transmission system for steam energy-saving desuperheaters according to claim 5, characterized in that: The regulation priority ranking module includes a special response model construction unit and a priority analysis unit; The special response model construction unit is used to transmit a signal to the abnormal response model construction module to obtain data pairs of each abnormal control event under the special control structure when the difference discrimination degree is greater than or equal to the difference discrimination degree threshold, and to construct a special response model for the special control structure. The priority analysis unit is used to analyze special control structures where there are simultaneous response data control transmission paths, specifically: Obtain the influence coefficients and path importance weights of the corresponding response model records for the simultaneous response data regulation transmission path. Sort the products of the influence coefficients and path importance weights in descending order to generate a priority sequence for the corresponding data regulation transmission path.

8. The data-driven intelligent control and data transmission system for steam energy-saving desuperheaters according to claim 5, characterized in that: The real-time control data response module compares the acquired real-time control structure with the historical steady-state control structure and special control structure. When the sampling frequency is acquired in each monitoring item, the abnormal response model of the corresponding data control path is responded to. The real-time model output value is obtained by substituting the specific real-time sampling frequency. If the real-time model output value belongs to the corresponding abnormal update rate interval B, an early warning is issued when the data control transmission path does not respond to the adjacent node. If the real-time model output value does not belong to the corresponding abnormal update rate interval B, then continue monitoring; When the real-time control structure records the sampling frequency in multiple different types of data control transmission paths, it analyzes the output value of the abnormal response model for each type of path according to priority; it issues an early warning when the abnormal update rate range is met, and continues monitoring if it is not met.

Images