Method and system for controlling safe opening function of high-pressure bypass valve of generator set

Abstract

The invention belongs to the technical field of intelligent control, and provides a safe opening function control method and system for a high-pressure bypass valve of a generator set, and the method comprises the steps: obtaining pressure measurement values outputted by a plurality of pressure controllers at each side of the high-pressure bypass valve of the generator set, and obtaining a downstream steam pressure value; if an abnormal pressure measurement value meeting an overrun condition in the pressure measurement values output by the plurality of pressure controllers reaches a set proportion, a basic valve opening signal is triggered; according to the pressure measurement value and the downstream steam pressure value, determining a pressure anomaly probability by using a pre-established pressure diagnosis model; and according to the triggering state and the pressure abnormal probability of the basic valve opening signal, a safe opening function control instruction is generated and output to a high-pressure bypass valve of the generator set. According to the scheme provided by the invention, the accuracy, reliability and stability of a control link are improved through hierarchical design of multi-parameter acquisition, proportion trigger judgment, intelligent model diagnosis and fusion decision control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control, and particularly relates to a control method and system for safety opening of a high-pressure bypass valve of a generator set. BACKGROUND

[0002] In the field of large-scale thermal power generation, nuclear power and other energy production, as the core equipment to ensure the reliability of power grid power supply, the pressure stability of the steam system of a million kilowatt generator set directly determines the safety of the unit operation. Among them, the high-pressure bypass valve, as the key executive component of steam pressure regulation, mainly functions to quickly discharge excess steam under the conditions of abnormal increase of boiler main steam pressure, turbine load shedding or start-stop, etc., to avoid major equipment failures such as boiler superheater overpressure and turbine blade erosion, and is the core link of the unit safety protection system.

[0003] With the development of generator sets towards high parameters and large capacity, the traditional safety opening function control scheme of high-pressure bypass valve gradually exposes many problems. Specifically, on the one hand, the traditional scheme has a high risk of protection misoperation and is easy to cause unit fluctuations. The existing control logic mostly adopts a single controller triggering mechanism, for example, three pressure controllers are configured on each side of the high-pressure bypass valve. When any pressure controller detects that the pressure is out of limit, the valve opening action is triggered. Such design does not consider the misjudgment problem caused by sensor aging, electromagnetic interference, measurement deviation, etc. of a single pressure controller.

[0004] On the other hand, the control logic lacks dynamic adaptability and has a lagging response. The traditional scheme only relies on the real-time measurement value of the pressure controller to trigger the action. In the scene of turbine load sudden drop, it is easy to cause short-term overpressure, while in the scene of pressure transient fluctuation, unnecessary valve opening action may be triggered due to misjudgment, affecting the system stability.

[0005] Therefore, the traditional safety opening function control scheme of high-pressure bypass valve of a generator set has the technical problems of low control precision, poor reliability and insufficient stability. SUMMARY

[0006] The present application provides a control method and system for safety opening of a high-pressure bypass valve of a generator set, to solve the defects of low control precision, poor reliability and insufficient stability of the traditional safety opening function control scheme of high-pressure bypass valve of a generator set.

[0007] In one aspect, the present application provides a control method for safety opening of a high-pressure bypass valve of a generator set, comprising:

[0008] obtaining pressure measurement values output by a plurality of pressure controllers on each side of the high-pressure bypass valve of the generator set, and obtaining a valve rear steam pressure value;

[0009] If the abnormal pressure measurement value satisfying the over-limit condition in the pressure measurement values output by the plurality of pressure controllers reaches a set proportion, a basic opening valve signal is triggered;

[0010] According to the pressure measurement value and the post-valve steam pressure value, a pressure abnormality probability is determined by using a pre-established pressure diagnosis model;

[0011] According to the triggering state of the basic opening valve signal and the pressure abnormality probability, a safety opening function control instruction is generated and output to the generator set high-pressure bypass valve.

[0012] According to the generator set high-pressure bypass valve safety opening function control method provided by the application, each side of the generator set is provided with three pressure controllers corresponding to the high-pressure bypass valve, and the set proportion is two-thirds.

[0013] According to the generator set high-pressure bypass valve safety opening function control method provided by the application, according to the pressure measurement value and the post-valve steam pressure value, a pressure abnormality probability is determined by using a pre-established pressure diagnosis model, including:

[0014] The pressure measurement value and the post-valve steam pressure value are subjected to data cleaning and noise reduction processing to obtain preprocessed pressure data;

[0015] The preprocessed pressure data is subjected to feature extraction to obtain key feature parameters;

[0016] The key feature parameters are input into the pre-established pressure diagnosis model to obtain the pressure abnormality probability output by the pressure diagnosis model.

[0017] According to the generator set high-pressure bypass valve safety opening function control method provided by the application, the pressure diagnosis model includes:

[0018] A time sequence feature extraction module is configured to extract a time sequence dependent relationship in the key feature parameters, obtain a time sequence feature vector, map the time sequence feature vector into a probability value in a set interval, and output a first prediction probability;

[0019] A multi-dimensional correlation analysis module is configured to extract a nonlinear correlation relationship between multiple parameters in the key feature parameters, obtain an abnormal class score value, map the abnormal class score value into a probability value in a set interval, and output a second prediction probability;

[0020] A fusion decision module is configured to fuse the first prediction probability and the second prediction probability by using a weighted average fusion strategy, and output a pressure abnormality probability.

[0021] According to the generator set high-pressure bypass valve safety opening function control method provided by the application, the time sequence feature extraction module includes:

[0022] The first input layer is used for converting the key feature parameters into a tensor form;

[0023] The first LSTM hidden layer is used for extracting short-term dependencies in the key feature parameters in the tensor form, to obtain time sequence dynamic features;

[0024] The second LSTM hidden layer is used for further extracting long-term dependencies in the key feature parameters according to the short-term dependencies, to obtain time sequence transition features;

[0025] The full connection layer is used for performing nonlinear conversion and dimension compression on the time sequence dynamic features and the time sequence transition features, to obtain a time sequence feature vector;

[0026] The first output layer is used for mapping the time sequence feature vector into a probability value in a set interval, to output a first prediction probability.

[0027] According to the generator set high-pressure bypass valve safety opening function control method provided by the application, the multi-dimensional correlation analysis module comprises:

[0028] The second input layer is used for splicing the static features and the derived features in the key feature parameters and converting them into a two-dimensional matrix form;

[0029] The decision tree integration layer is used for constructing a nonlinear mapping relationship between the features in the key feature parameters in the two-dimensional matrix form and pressure abnormalities through integrated learning of multiple decision trees, and adding up abnormal original scores output by all decision trees to output an abnormal class score value;

[0030] The second output layer is used for mapping the abnormal class score value into a probability value in a set interval, to output a second prediction probability.

[0031] According to the generator set high-pressure bypass valve safety opening function control method provided by the application, a safety opening function control instruction is generated according to the triggering state of the basic opening valve signal and the pressure abnormality probability, and the safety opening function control instruction comprises:

[0032] If the triggering state of the basic opening valve signal is that the signal is triggered, but the pressure abnormality probability is lower than a set probability threshold value, the safety opening function control instruction is to suspend the valve opening action;

[0033] If the triggering state of the basic opening valve signal is that the signal is triggered, and the pressure abnormality probability is greater than or equal to the set probability threshold value, the safety opening function control instruction is to accelerate the valve opening action.

[0034] According to the generator set high-pressure bypass valve safety opening function control method provided by the application, after the safety opening function control instruction is generated and output to the generator set high-pressure bypass valve, the method further comprises:

[0035] If the post-valve steam pressure value does not drop to the safe pressure range within the set time length, the current valve opening speed is increased to the target valve opening speed until the post-valve steam pressure value drops to the safe pressure range.

[0036] According to the application, the method for controlling the safety opening function of the high-pressure bypass valve of the generator set further comprises:

[0037] The valve control log is generated periodically according to a set period;

[0038] The valve control log, the pressure measurement value, the post-valve steam pressure value and the pressure abnormality probability are displayed in sections.

[0039] In another aspect, the application further provides a system for controlling the safety opening function of the high-pressure bypass valve of the generator set, comprising:

[0040] An acquisition module is configured to acquire pressure measurement values output by a plurality of pressure controllers on each side of the high-pressure bypass valve of the generator set and to acquire a post-valve steam pressure value;

[0041] A triggering module is configured to trigger a basic valve opening signal if abnormal pressure measurement values satisfying an over-limit condition among the pressure measurement values output by the plurality of pressure controllers reach a set proportion;

[0042] A diagnosis module is configured to determine a pressure abnormality probability by using a pre-established pressure diagnosis model according to the pressure measurement values and the post-valve steam pressure value;

[0043] A control module is configured to generate and output a safety opening function control instruction to the high-pressure bypass valve of the generator set according to a triggering state of the basic valve opening signal and the pressure abnormality probability.

[0044] The method and system for controlling the safety opening function of the high-pressure bypass valve of the generator set provided by the application effectively solve the core problems of high misoperation risk, response lag and weak diagnosis capability in the traditional control scheme through the layered design of multi-parameter acquisition, proportional triggering judgment, intelligent model diagnosis and fusion decision control, avoid misjudgment caused by single pressure controller failure or measurement deviation by setting the basic triggering condition that abnormal pressure measurement values reach a set proportion, further filter unnecessary actions caused by transient pressure fluctuation by combining the secondary verification of the post-valve steam pressure value, realize accurate judgment and dynamic response of pressure abnormality by introducing the pressure diagnosis model, avoid the response lag problem of the traditional logic, provide the root clues of pressure abnormality for the operation personnel through the linkage analysis of the pressure abnormality probability output by the pressure diagnosis model and the basic triggering signal, reduce the workload of the operation personnel, reduce the damage risk of the equipment caused by long-term abnormal operation, prolong the service life of the high-pressure bypass valve and the associated equipment, and improve the accuracy, reliability and stability of the control link. BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative effort.

[0046] Figure 1 is a flowchart of the safety opening function control method of the high-pressure bypass valve of the generator set provided by the embodiment of the present application;

[0047] Figure 2 is a control circuit diagram of the safety opening control of the A-side high-pressure bypass valve of the million unit provided by the embodiment of the present application;

[0048] Figure 3 is a control circuit diagram of the safety opening control of the B-side high-pressure bypass valve of the million unit provided by the embodiment of the present application;

[0049] Figure 4 is a structural diagram of the safety opening function control system of the high-pressure bypass valve of the generator set provided by the embodiment of the present application. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solutions and advantages of the present application more clear, the technical solutions of the present application will be described clearly and completely in the following with reference to the drawings in the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present application.

[0051] The details of the safety opening function control method and system of the high-pressure bypass valve of the generator set provided by the embodiments of the present application will be described below with reference to the drawings. Figures 1 to 4 As shown in the drawings,

[0052] the safety opening function control method of the high-pressure bypass valve of the generator set provided by the embodiments of the present application mainly includes the following steps: Figure 1

[0053] Step 110: Obtain the pressure measurement values output by the plurality of pressure controllers on each side of the high-pressure bypass valve of the generator set, and obtain the steam pressure value after the valve.

[0054] In the embodiment, the generator set mainly refers to the million unit, and the million unit refers to the generator set with a power of one million kilowatts, which has the characteristics of large scale, high parameters, complex system, etc., and its safe and stable operation is crucial to the power supply reliability of the power grid.

[0055] ​In practical applications, each high-pressure bypass valve on the AB side of the million-unit set corresponds to three pressure controllers, and through the pre-deployment of edge computing gateways, the pressure measurement values output by the pressure controllers can be collected at a sampling frequency of 50 Hz. At the same time, position sensors and pressure feedback modules can be installed on the actuators of the high-pressure bypass valves on the AB side, so as to collect the valve opening and the steam pressure value behind the valve in real time, thereby providing feedback signals for closed-loop control.

[0056] Figure 2 and Figure 3 respectively show the safety opening control circuit wiring state of the high-pressure bypass valves on the AB side of the million-unit set, as Figure 2 shown, the safety opening control circuit of the high-pressure bypass valve on the A side of the million-unit set contains K19.4-14 (pressure control A1-1), K19.5-14 (pressure control A2-1), K19.6-14 (pressure control A3-1), K19.4-24 (pressure control A1-2), K19.5-24 (pressure control A2-2), K19.6-24 (pressure control A3-2) multiple pressure controllers, each of which is responsible for collecting pressure measurement values at the corresponding position, and providing pressure measurement basis for the safety opening control logic of the high-pressure bypass valve on the A side.

[0057] At the same time, according to the pressure control logic provided in the embodiment, different safety valves can be correspondingly driven and configured with safety valve monitoring devices, as Figure 2 shown, composed of Y22, Y26 series safety valves and corresponding K22, K26 series monitoring devices, such as Y22.1-A1 (B valve safety valve 1) matched with K22.2-A1 (B valve safety valve monitoring 1), used to perform safety opening actions on the corresponding high-pressure bypass valve and monitor the valve state in real time.

[0058] Similarly, as Figure 3 shown, the safety opening control circuit of the high-pressure bypass valve on the B side of the million-unit set contains K23.4-14 (pressure control B1-1), K23.5-14 (pressure control B2-1), K23.6-14 (pressure control B3-1), K23.4-24 (pressure control B1-2), K23.5-24 (pressure control B2-2), K23.6-24 (pressure control B3-2) multiple pressure controllers, each of which is responsible for collecting pressure measurement values at the corresponding position, and providing pressure measurement basis for the safety opening control logic of the high-pressure bypass valve on the B side.

[0059] At the same time, Figure 3 shown, the control and monitoring hardware scheme composed of Y21, Y25 series safety valves and corresponding K21, K25 series monitoring devices, such as Y21.1-A1 (A valve safety valve 1) matched with K21.2-A1 (A valve safety valve monitoring 1), responsible for the safety opening and state monitoring of the corresponding high-pressure bypass valve.

[0060] Step 120: If the abnormal pressure measurement satisfying the over-limit condition reaches a set proportion in the pressure measurements output by the plurality of pressure controllers, a basic opening valve signal is triggered.

[0061] In this embodiment, since three pressure controllers are arranged on each side of the generator set high-pressure bypass valve, if at least two pressure measurements output by the pressure controllers on each side satisfy the over-limit condition, the basic opening valve signal can be triggered, that is, the set proportion can be set to two-thirds.

[0062] Step 130: According to the pressure measurement and the post-valve steam pressure value, a pressure anomaly probability is determined by using a pre-established pressure diagnosis model.

[0063] In this embodiment, the pressure diagnosis model adopts a machine learning model, which can realize pressure anomaly prediction and then output the pressure anomaly probability, thereby providing model data basis for subsequent comprehensive control decision.

[0064] Step 140: According to the triggering state of the basic opening valve signal and the pressure anomaly probability, a safety opening function control instruction is generated and output to the generator set high-pressure bypass valve.

[0065] In this embodiment, by comprehensively considering the triggering state of the basic opening valve signal and the pressure anomaly probability, a safety opening function control instruction more in line with actual needs can be output.

[0066] In an embodiment, according to the pressure measurement and the post-valve steam pressure value, a pressure anomaly probability is determined by using a pre-established pressure diagnosis model, specifically including:

[0067] First, the pressure measurement and the post-valve steam pressure value are subjected to data cleaning and noise reduction processing to obtain preprocessed pressure data.

[0068] In actual application, a sliding average filtering method can be used to process the pressure original signal containing the pressure measurement and the post-valve steam pressure value to remove high-frequency interference noise; abnormal data such as sudden value caused by sensor instantaneous failure can be identified and removed by 3σ criterion. Missing data can be completed by linear interpolation to ensure data continuity; subsequently, the collected analog signals can be converted into 0-10V standardized digital signals for model calculation and logical judgment.

[0069] Then, the preprocessed pressure data is subjected to feature extraction to obtain key feature parameters.

[0070] In this embodiment, time domain features such as pressure mean value, variance, peak value, rising rate, and steady-state duration can be extracted, and correlation features such as the difference between the post-valve steam pressure value and the pressure measurement and the correction coefficient of the environmental temperature on the pressure measurement can be extracted.

[0071] Finally, the key feature parameters are input into the pressure diagnosis model established in advance to obtain a pressure anomaly probability output by the pressure diagnosis model.

[0072] In an embodiment, the pressure diagnosis model specifically comprises:

[0073] a time sequence feature extraction module, configured to extract a time sequence dependency relationship in the key feature parameters, obtain a time sequence feature vector, map the time sequence feature vector into a probability value in a set interval, and output a first prediction probability.

[0074] In this embodiment, the time sequence feature extraction module specifically comprises:

[0075] a first input layer, configured to convert the key feature parameters into a tensor form.

[0076] It can be understood that the first input layer can perform standardization processing on the preprocessed time sequence feature data and convert the time sequence feature data into a three-dimensional tensor format that can be processed by the time sequence feature extraction module, so as to eliminate the influence of dimension differences on model prediction.

[0077] a first LSTM hidden layer, configured to extract a short-term dependency relationship in the key feature parameters in the tensor form, and obtain a time sequence dynamic feature.

[0078] In actual application, the first LSTM hidden layer can preliminarily extract a short-term dependency relationship in the time sequence data, such as a short-term fluctuation rule of pressure and a change correlation between adjacent sampling points, and filter redundant information through a forgetting gate to retain the time sequence dynamic feature.

[0079] a second LSTM hidden layer, configured to further extract a long-term dependency relationship in the key feature parameters according to the short-term dependency relationship, and obtain a time sequence transition feature.

[0080] It can be understood that the second LSTM hidden layer can further mine a long-term dependency relationship in the time sequence feature, such as a continuous rising or falling trend of pressure and a transition trend from a steady state to an abnormal state, obtain the time sequence transition feature, and randomly shield part of neurons to improve the generalization ability of the model.

[0081] a fully connected layer, configured to perform nonlinear conversion and dimension compression on the time sequence dynamic feature and the time sequence transition feature, and obtain a time sequence feature vector.

[0082] In this embodiment, the fully connected layer can perform nonlinear conversion and dimension compression on the time sequence dynamic feature and the time sequence transition feature, strengthen the weight of key trend features (such as a pressure mutation rate and a super-threshold duration), and inhibit parameter overfitting through L2 regularization, and finally output the time sequence feature vector.

[0083] a first output layer, configured to map the time sequence feature vector into a probability value in a set interval, and output a first prediction probability.

[0084] In this embodiment, the first output layer can map the compressed time sequence feature vector into a probability value in the interval [0, 1] to output a first prediction probability. The closer the first prediction probability is to 1, the higher the possibility of determining the pressure abnormality based on the time trend.

[0085] The multi-dimensional correlation analysis module is configured to extract a nonlinear correlation between multiple parameters in the key feature parameters, obtain an abnormal class score value, and map the abnormal class score value into a probability value in a set interval to output a second prediction probability.

[0086] In this embodiment, the multi-dimensional correlation analysis module specifically includes:

[0087] The second input layer is configured to splice the static features and the derived features in the key feature parameters and convert them into a two-dimensional matrix form.

[0088] In actual applications, the second input layer can splice the static features and the derived features and convert them into a two-dimensional matrix format that can be processed by XGBoost. The feature values can be mapped to the interval [0, 1] through maximum and minimum normalization.

[0089] The decision tree integration layer is configured to construct a nonlinear mapping relationship between the features in the key feature parameters and the pressure abnormality through ensemble learning of multiple decision trees, and accumulate the abnormal original scores output by all the decision trees to output an abnormal class score value.

[0090] It can be understood that the decision tree integration layer can construct a nonlinear mapping relationship between the features and the pressure abnormality through ensemble learning of multiple CART (Classification and Regression Tree) trees. Each decision tree is based on a different feature subset for splitting, can capture local feature correlations, and can avoid overfitting through regularization to control the complexity of the decision tree.

[0091] The second output layer is configured to map the abnormal class score value into a probability value in a set interval to output a second prediction probability.

[0092] In this embodiment, the second output layer can map the abnormal class score value output by the decision tree integration layer into a probability value in the interval [0, 1], i.e., a second prediction probability. The closer the second prediction probability is to 1, the higher the possibility of determining the pressure abnormality based on the multi-parameter correlation.

[0093] The fusion decision module is configured to fuse the first prediction probability and the second prediction probability through a weighted average fusion strategy to output a pressure abnormality probability.

[0094] In this embodiment, the fusion decision layer adopts a weighted average fusion strategy, which can fuse the first prediction probability based on the LSTM output and the second prediction probability based on the XGBoost output to generate the final pressure abnormality probability, thereby solving the one-sidedness problem of a single model.

[0095] In an embodiment, according to the trigger state of the basic opening signal and the pressure abnormality probability, a safety opening function control instruction is generated, specifically including:

[0096] In one case, if the trigger state of the basic opening signal is that the signal is triggered, but the pressure abnormality probability is lower than the set probability threshold, the safety opening function control instruction is to suspend the opening action.

[0097] In this embodiment, the set probability threshold can be 0.7. If the pressure abnormality probability is lower than 0.7, it means that the model predicts that the pressure is normal, and at this time, the model prediction result deviates from the trigger result based on the pressure measured value output by the pressure controller, so the opening action needs to be suspended and then act after confirmation in the actual field.

[0098] In another case, if the trigger state of the basic opening signal is that the signal is triggered, and the pressure abnormality probability is greater than or equal to the set probability threshold, the safety opening function control instruction is to accelerate the opening action.

[0099] It can be understood that if the pressure abnormality probability is greater than or equal to 0.7, it means that the model predicts that the pressure is abnormal, and at this time, the model prediction result is consistent with the trigger result based on the pressure measured value output by the pressure controller, so as to ensure that the pressure can be adjusted to the normal range in time, the opening action can be accelerated.

[0100] In practical application, the accelerated opening action can be to increase the opening degree speed, or to increase the number of controlled valves, for example, the previous basic opening signal only controls two high-pressure bypass valves on the same side to open, and the subsequent accelerated opening action trigger link can control two high-pressure bypass valves on the other side to open synchronously.

[0101] In an embodiment, after generating and outputting the safety opening function control instruction to the generator high-pressure bypass valve, the above method can further include:

[0102] If the valve steam pressure value does not drop to the safe pressure range within the set time, the current opening speed is increased to the target opening speed until the valve steam pressure value drops to the safe pressure range.

[0103] It can be understood that the set time length can be 5 seconds, that is, if the valve behind the steam pressure value does not decrease to the safe pressure range within 5 seconds, it means that the pressure is not effectively controlled in time, at this time the opening valve speed can be further increased, such as from 5% / s to 8% / s, until the valve behind the steam pressure value decreases to the safe pressure range, the control strategy can further improve the response speed and control efficiency of the high pressure bypass valve safety opening function.

[0104] In an embodiment, the above-mentioned generator set high pressure bypass valve safety opening function control method can further include:

[0105] First, generate a valve control log periodically according to a set period.

[0106] In actual application, the automatic process of timing triggering, data automatic collection, structured packaging and storage archiving can be used to realize the periodic generation of the valve control log, and the specific steps are as follows:

[0107] First, configure a timing task in the unit DCS system or edge computing platform, set the log generation period, and specifically recommend 5 minutes / time, which can be dynamically shortened to 1 minute / time under critical conditions. When the timing task is triggered, the system automatically collects the core data in the period from the pressure controller database, valve control execution unit and pressure diagnosis model output module.

[0108] Second, package the collected core data in the structured format of log period, basic information, pressure data, valve control data, model data and abnormal record, wherein the basic information includes unit number, current load and operating condition, and the abnormal record marks whether the basic opening valve signal is triggered, whether there is controller data abnormality and model alarm information.

[0109] Third, the system automatically stores the structured log in PDF and CSV formats to the local log server and cloud backup database, and at the same time, establishes directory classification archive according to year, month and day, supports quick retrieval through unit number, log period and abnormal type, etc. Key words, ensure the integrity and traceability of log data, and provide data support for subsequent operation and maintenance analysis and model optimization.

[0110] Then, the valve control log, pressure measurement value, valve behind the steam pressure value and pressure abnormal probability are displayed in zones.

[0111] In this embodiment, based on the visualization platform of the DCS system, the four-zone linkage and dynamic interaction design are used to realize the zoned display of the valve control log, pressure measurement value, valve behind the steam pressure value and pressure abnormal probability, and the specific scheme is as follows:

[0112] In one aspect, the visual interface is divided into four functional partitions, each partition is independently displayed and supports data linkage.

[0113] The first partition is a valve control log display area, which adopts a table form to arrange the latest 100 log records in reverse chronological order, and the table columns include log period, unit operating condition, valve action state, and abnormality marker. Clicking on a log can expand to view the detailed data in the period, such as opening degree change curve and over-limit record, and filtering by abnormal log or normal log is supported.

[0114] The second partition is a pressure measurement value real-time monitoring area, which is displayed in a combination of line chart and digital panel. The left panel displays the current pressure measurement values of three pressure controllers on each side in real time, which can be distinguished by different colors, with over-limit values marked in red. The right line chart dynamically draws the measurement value change trend of each controller in the last hour, with over-limit threshold lines marked in the chart, facilitating intuitive observation of pressure fluctuation rules.

[0115] The third partition is a valve post steam pressure display area, which adopts a combination of column chart and dynamic threshold bar. The column chart displays the average value of valve post steam pressure in the last 20 log periods, and the dynamic threshold bar displays the current safe pressure range in real time. When the pressure exceeds the range, the threshold bar automatically changes color to alarm, and the action state of the high bypass valve in the period is also displayed to assist in judging the valve control effect.

[0116] The fourth partition is a pressure abnormality probability display area, which adopts a combination of instrument panel and trend chart. The instrument panel displays the abnormality probability output by the current model in real time. The larger the probability value, the closer the instrument panel pointer to the abnormal area, and the color gradually changes from green to red. The trend chart draws the change curve of the abnormality probability in the last hour, with a probability threshold line marked. When the probability exceeds the threshold, it automatically flashes to prompt.

[0117] In addition, each partition supports data linkage operation, such as clicking on the line of a controller in the pressure measurement value monitoring area to automatically filter the logs containing over-limit records of the controller in the valve control log area, and clicking on the peak point in the abnormality probability display area to display the valve post steam pressure value and high pressure bypass valve state at the corresponding time, realizing multi-dimensional data correlation analysis, helping operation personnel quickly grasp the equipment operating state, and improving monitoring efficiency and abnormality judgment accuracy.

[0118] Based on the same overall inventive concept, the present application also protects a generator set high pressure bypass valve safety opening function control system. The generator set high pressure bypass valve safety opening function control system provided by the present application is described below, and the generator set high pressure bypass valve safety opening function control system described below can be mutually corresponding and referred to the generator set high pressure bypass valve safety opening function control method described above.

[0119] For example, Figure 4As shown, the generator set high-pressure bypass valve safety opening function control system provided by the embodiment of the present application specifically comprises:

[0120] The acquisition module 210 is configured to acquire pressure measurement values output by a plurality of pressure controllers on each side of the generator set high-pressure bypass valve, and acquire a post-valve steam pressure value.

[0121] The triggering module 220 is configured to trigger a basic opening valve signal if abnormal pressure measurement values satisfying an over-limit condition among the pressure measurement values output by the plurality of pressure controllers reach a set proportion.

[0122] The diagnosis module 230 is configured to determine a pressure abnormality probability by using a pre-established pressure diagnosis model according to the pressure measurement values and the post-valve steam pressure value.

[0123] The control module 240 is configured to generate and output a safety opening function control instruction to the generator set high-pressure bypass valve according to a triggering state of the basic opening valve signal and the pressure abnormality probability.

[0124] As to the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and thus will not be described in detail here.

[0125] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method of controlling a safety opening function of a high pressure bypass valve of a genset, characterized by, The method comprises the following steps: acquiring pressure measurement values output by multiple pressure controllers on each side of a high-pressure bypass valve of a generator set, and acquiring a post-valve steam pressure value; if abnormal pressure measurement values satisfying an over-limit condition among the pressure measurement values output by the multiple pressure controllers reach a set proportion, triggering a basic opening valve signal; determining a pressure abnormality probability by using a pre-established pressure diagnosis model according to the pressure measurement values and the post-valve steam pressure value; generating and outputting a safety opening function control instruction to the high-pressure bypass valve of the generator set according to a triggering state of the basic opening valve signal and the pressure abnormality probability.

2. The control method of the safety opening function of the high pressure bypass valve of the generator set according to claim 1, characterized by, Each side of the high-pressure bypass valve of the generator set is provided with three pressure controllers, and the set proportion is two-thirds.

3. The control method of the safety opening function of the high pressure bypass valve of the genset according to claim 1, characterized by, The method for determining the pressure abnormality probability according to the pressure measurement values and the post-valve steam pressure value by using the pre-established pressure diagnosis model comprises the following steps: performing data cleaning and noise reduction processing on the pressure measurement values and the post-valve steam pressure value to obtain preprocessed pressure data; extracting key feature parameters from the preprocessed pressure data; inputting the key feature parameters into the pre-established pressure diagnosis model to obtain a pressure abnormality probability output by the pressure diagnosis model.

4. The control method of the safety opening function of the high pressure bypass valve of the genset according to claim 3, characterized by, The pressure diagnosis model comprises: a time sequence feature extraction module, configured to extract a time sequence dependent relationship in the key feature parameters, obtain a time sequence feature vector, map the time sequence feature vector into a probability value in a set interval, and output a first prediction probability; a multi-dimensional correlation analysis module, configured to extract a nonlinear correlation relationship among multiple parameters in the key feature parameters, obtain an abnormal class score value, and map the abnormal class score value into a probability value in a set interval, and output a second prediction probability; a fusion decision module, configured to fuse the first prediction probability and the second prediction probability by using a weighted average fusion strategy, and output a pressure abnormality probability.

5. The control method of the safety opening function of the high pressure bypass valve of the genset according to claim 4, characterized by, The time sequence feature extraction module comprises: a first input layer, configured to convert the key feature parameters into a tensor form; a first LSTM hidden layer, configured to extract a short-term dependent relationship in the key feature parameters in the tensor form, and obtain a time sequence dynamic feature; a second LSTM hidden layer, configured to further extract a long-term dependent relationship in the key feature parameters according to the short-term dependent relationship, and obtain a time sequence transition feature; a full connection layer, configured to perform nonlinear conversion and dimension compression on the time sequence dynamic feature and the time sequence transition feature, and obtain a time sequence feature vector; a first output layer, configured to map the time sequence feature vector into a probability value in a set interval, and output a first prediction probability.

6. The control method of the safety opening function of the high pressure bypass valve of the genset according to claim 4, characterized by, The multi-dimensional correlation analysis module comprises: a second input layer, configured to splice static features and derived features in the key feature parameters, and convert the key feature parameters into a two-dimensional matrix form; a decision tree integration layer, configured to construct a nonlinear mapping relationship between features and pressure abnormalities in the key feature parameters in the two-dimensional matrix form by using ensemble learning of multiple decision trees, and accumulate abnormal original scores output by all decision trees to obtain an abnormal class score value; a second output layer, configured to map the abnormal class score value into a probability value in a set interval, and output a second prediction probability.

7. The control method of the generator set high pressure bypass valve safety open function according to claim 1, characterized by, According to the triggering state of the basic opening valve signal and the pressure abnormality probability, a safety opening function control instruction is generated, including: If the triggering state of the basic opening valve signal is that the signal is triggered, and the pressure abnormality probability is greater than or equal to the set probability threshold, the safety opening function control instruction is to accelerate the opening valve action. After generating and outputting the safety opening function control instruction to the generator set high pressure bypass valve, the method further includes:

8. The control method of the generator set high pressure bypass valve safety open function according to claim 1, characterized by, If the post-valve steam pressure value does not drop to the safe pressure range within the set time, the current opening valve speed is increased to the target opening valve speed until the post-valve steam pressure value drops to the safe pressure range. The method further includes:

9. The control method of the generator set high pressure bypass valve safety open function according to claim 1, characterized by, Generating a valve control log periodically according to a set period; The valve control log, the pressure measurement value, the post-valve steam pressure value, and the pressure abnormality probability are displayed in zones. Including:

10. A control system for a safety open function of a high pressure bypass valve of a genset, the control system comprising: An acquisition module is configured to acquire pressure measurement values output by a plurality of pressure controllers on each side of a generator set high pressure bypass valve, and to acquire a post-valve steam pressure value; A triggering module is configured to trigger a basic opening valve signal if abnormal pressure measurement values that meet over-limit conditions in the pressure measurement values output by the plurality of pressure controllers reach a set proportion; A diagnosis module is configured to determine a pressure abnormality probability by using a pre-established pressure diagnosis model according to the pressure measurement values and the post-valve steam pressure value; A control module is configured to generate and output a safety opening function control instruction to the generator set high pressure bypass valve according to the triggering state of the basic opening valve signal and the pressure abnormality probability. ​

Images