A real-time optimization method and system for sliding pressure operation of a steam turbine unit

Abstract

The present invention belongs to the technical field of steam turbine unit operation management, and specifically provides a real-time optimization method and system for the sliding pressure operation of a steam turbine unit, wherein the system includes a sensor integration module, a heat rate prediction module, an optimization algorithm module, a dynamic optimization module and a supervision control end; the present invention calculates and predicts the heat rate of the steam turbine unit based on the operation data of the steam turbine unit through the heat rate prediction module, searches for the optimal main steam pressure using the GSA algorithm, and dynamically optimizes the sliding pressure operation of the steam turbine unit based on the optimal main steam pressure in the dynamic optimization module, thereby realizing real-time optimization of the sliding pressure operation of the steam turbine unit, and monitors all sensors on the steam turbine unit through the acquisition monitoring and evaluation module to judge the acquisition operation performance of the sensor integration module, and evaluates and analyzes the parameter maintenance performance of the steam turbine unit when generating an acquisition qualified signal, which is beneficial to ensuring the operation effect of the steam turbine unit.

Description

Technical Field

[0001] The present invention relates to the technical field of steam turbine unit operation management, and in particular to a steam turbine unit sliding pressure operation real-time optimization method and system. Background Art

[0002] A steam turbine unit is a rotary power machine that efficiently converts steam thermal energy into mechanical energy. Known as the heart of a power plant, its core function is to generate power through the expansion of steam within the turbine, driving generators to generate electricity or directly driving mechanical equipment such as pumps, fans, compressors, and ship propellers. It can also utilize exhaust steam or intermediate extraction steam to meet heating needs.

[0003] A Chinese invention patent, publication number CN110765627A, discloses a big data-based intelligent operation optimization system and method for steam turbine units in thermal power plants. This invention utilizes a queue competition algorithm to solve a general intelligent optimization model for steam turbine units, obtaining optimized values for steam turbine unit steam intake, power generation, extraction, and exhaust rates. This system is then regulated to optimize the distribution of thermal power loads within the steam turbine units, achieving energy savings without requiring equipment modification to reduce costs.

[0004] However, in actual application, the above-mentioned technical solution cannot accurately determine the risk status of data collection on the basis of achieving comprehensive monitoring and reasonable optimization of the steam turbine unit. Moreover, when it is judged that the data collection performance is poor, it is difficult to accurately evaluate the parameter maintenance and control performance of the steam turbine unit. This is not conducive to ensuring the safe, stable and energy-saving operation of the steam turbine unit, and increases the difficulty of steam turbine unit operation supervision.

[0005] In view of the above technical defects, a solution is now proposed. Summary of the Invention

[0006] The purpose of the present invention is to provide a real-time optimization method and system for the sliding pressure operation of a steam turbine unit, which solves the problem that the existing technology is unable to accurately judge the risk status of data collection on the basis of achieving comprehensive monitoring and reasonable optimization of the steam turbine unit, and it is difficult to accurately evaluate the parameter maintenance control performance of the steam turbine unit when it is judged that the data collection performance is poor, which is not conducive to ensuring the safe, stable and energy-saving operation of the steam turbine unit, and the problem that the operation supervision of the steam turbine unit is difficult and the level of intelligence is low.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A real-time optimization system for sliding pressure operation of a steam turbine unit includes a sensor integration module, a heat rate prediction module, an optimization algorithm module, a dynamic optimization module, and a supervisory control terminal. The sensor integration module uses several high-precision sensors installed at key locations of the steam turbine unit to collect various operating data of the steam turbine unit in real time, including main steam pressure, temperature, flow rate, speed, back pressure, and extraction volume. The data collection frequency is set to ten times per second, and the collected data is filtered, denoised, and processed for outliers. The processed operating data is then sent to the heat rate prediction module.

[0009] The heat rate prediction module calculates the heat rate of the steam turbine unit based on the operating data of the steam turbine unit through the heat rate calculation formula, and uses the LS-SVM algorithm to predict the heat rate, and sends the predicted heat rate to the optimization algorithm module; the optimization algorithm module uses the GSA algorithm to search for the optimal main steam pressure within the feasible main steam pressure range with the predicted heat rate as the objective function, and sends the optimal main steam pressure to the dynamic optimization module and the supervision and control end; the dynamic optimization module performs dynamic optimization of the sliding pressure operation of the steam turbine unit based on the optimal main steam pressure, and sends the optimization information to the supervision and control end.

[0010] Furthermore, the heat rate calculation formula is as follows:

[0011] ;

[0012] Wherein, HRt represents the predicted heat rate; Fthr represents the main steam flow rate; Hthr represents the main steam enthalpy; Fhrh represents the hot reheat steam flow rate; Hhrh represents the reheat hot section steam enthalpy; Fffw represents the feed water flow rate; Hffw represents the boiler feed water enthalpy; Fcrh represents the cold reheat steam flow rate; Hcrh represents the reheat cold section enthalpy; Frhs represents the reheat desuperheating water flow rate; Hrhs represents the reheat desuperheating water enthalpy; Wc represents the generator power after the secondary pressure drop correction.

[0013] Furthermore, the sensor integration module is communicatively connected to the acquisition, monitoring, and evaluation module. The acquisition, monitoring, and evaluation module monitors all sensors on the steam turbine unit and marks the corresponding sensor as a monitoring object i, where i is a natural number greater than 1. The monitoring object i is marked as an obstacle object or a normal object through acquisition anomaly evaluation and analysis. If an obstacle object exists within a unit time, an acquisition alarm signal is generated.

[0014] If there is no obstacle object within the unit time, the sensor will cooperate with decision analysis to generate a collection alarm signal or a collection qualified signal, and the collection alarm signal or the collection qualified signal will be sent to the supervision and control end. The supervision and control end will issue an early warning when it receives the collection alarm signal.

[0015] Furthermore, the specific analysis process of the collection anomaly assessment analysis is as follows:

[0016] The operation data collected by the monitoring object i is obtained. If the corresponding operation data is not within the normal collection value range of the monitoring object i, the corresponding operation data is defined as error data; the number of collection times of error data corresponding to the monitoring object i in unit time is counted and marked as the error detection value, and the error detection value is numerically compared with the preset error detection threshold. If the error detection value exceeds the preset error detection threshold, the monitoring object i is marked as an obstacle object.

[0017] Furthermore, if the sampling error detection value does not exceed the preset sampling error detection threshold, the time when the monitoring object i performs data collection is obtained and marked as the actual sampling time, and the deviation time of the actual sampling time compared to the corresponding standard collection time is marked as the resistance measurement time; all the resistance measurement time corresponding to the monitoring object i in the unit time are obtained and the average is calculated to obtain the resistance measurement time table value, and the resistance measurement time is numerically compared with the preset resistance measurement time threshold, and the number of resistance measurement time exceeding the preset resistance measurement time threshold in the unit time is marked as the resistance measurement time deviation value;

[0018] The collection error evaluation value is calculated by weighted summing up the collection error detection value, resistance measurement time table value and resistance measurement time deviation value of the monitoring object i, and the collection error evaluation value is numerically compared with the preset collection error evaluation threshold. If the collection error evaluation value exceeds the preset collection error evaluation threshold, the monitoring object i is marked as an obstacle object; if the collection error evaluation value does not exceed the preset collection error evaluation threshold, the monitoring object i is marked as a normal object.

[0019] Furthermore, the specific analysis process of the sensor in conjunction with decision analysis is as follows:

[0020] Obtain the abnormal evaluation values of all sensors, mark the average of all abnormal evaluation values as the initial evaluation value, compare the initial evaluation value with the preset initial evaluation threshold, and generate an alarm signal if the initial evaluation value exceeds the preset threshold.

[0021] If the initial evaluation value does not exceed the preset initial evaluation threshold, the corresponding standard collection time is defined as the target time. If the operating data collected by the monitoring object i at the corresponding target time is erroneous data or the resistance measurement duration exceeds the preset resistance measurement duration threshold, the monitoring object i is assigned the collection judgment symbol TX-1;

[0022] If there is a sensor assigned with the sampling symbol TX-1 at the corresponding target moment, the corresponding target moment will be marked as the alarm moment, and the number of alarm moments per unit time will be marked as the sampling alarm value. The sampling alarm value will be compared with the preset sampling alarm threshold. If the sampling alarm value exceeds the preset sampling alarm threshold, a sampling alarm signal will be generated; if the sampling alarm value does not exceed the preset sampling alarm threshold, a sampling qualified signal will be generated.

[0023] Furthermore, the acquisition monitoring and evaluation module is communicatively connected to the maintenance detection and evaluation module, and the acquisition monitoring and evaluation module sends the acquisition qualified signal to the maintenance detection and evaluation module. When the maintenance detection and evaluation module receives the acquisition qualified signal, it generates a maintenance qualified signal or a maintenance abnormality signal through the turbine unit maintenance performance evaluation analysis, and sends the maintenance qualified signal or the maintenance abnormality signal to the supervision and control end. When the supervision and control end receives the maintenance abnormality signal, it issues an early warning.

[0024] Furthermore, the specific analysis process of the steam turbine unit maintenance performance evaluation and analysis includes:

[0025] Obtain operating parameters that need to be monitored during the operation of the steam turbine unit, mark the corresponding operating parameters as matching parameters k, where k is a natural number greater than 1; obtain real-time data of the matching parameter k, and when the deviation of the real-time data from the corresponding data standard value exceeds the corresponding preset deviation threshold, determine that the matching parameter k is in a maintenance risk state;

[0026] Obtain the number of occurrences of the matching parameter k in the maintenance risk state per unit time and the duration of each occurrence and define them as the maintenance risk frequency value and maintenance risk duration value respectively. Compare the maintenance risk frequency value and maintenance risk duration value with the corresponding preset maintenance risk frequency threshold value and preset maintenance risk duration threshold value respectively. If the maintenance risk frequency value or the maintenance risk duration value exceeds the corresponding preset threshold value, the matching parameter k is marked as a maintenance loss parameter.

[0027] If the maintenance risk frequency value and the maintenance risk time value do not exceed the corresponding preset threshold value, the maximum deviation of the real-time data of the matching parameter k compared with the corresponding data standard value within the unit time is marked as the maintenance deviation amplitude value, and the maintenance deviation amplitude value, maintenance risk frequency value and maintenance risk time value of the matching parameter k are weighted and summed to obtain the dimensional abnormality characteristic value, and the dimensional abnormality characteristic value is numerically compared with the preset dimensional abnormality characteristic threshold. If the dimensional abnormality characteristic value exceeds the preset dimensional abnormality characteristic threshold, the matching parameter k is marked as a maintenance damage parameter; if a maintenance damage parameter exists within the unit time, a maintenance abnormality signal is generated.

[0028] Furthermore, if there is no maintenance-causing parameter within the unit time, the preset maintenance importance weight value of the matching parameter k is retrieved, the ratio of the dimensional difference feature value of the matching parameter k to the corresponding preset dimensional difference feature threshold is marked as the dimensional difference detection value, the dimensional difference detection value of the matching parameter k is multiplied by the corresponding preset maintenance importance weight value, and the product result is marked as the maintenance risk judgment value;

[0029] The maintenance risk judgment values of all operating parameters that need to be monitored are obtained, and all maintenance risk judgment values are summed up to obtain a maintenance risk assessment value. The maintenance risk assessment value is numerically compared with the preset maintenance risk assessment threshold. If the maintenance risk assessment value exceeds the preset maintenance risk assessment threshold, a maintenance abnormality signal is generated; if the maintenance risk assessment value does not exceed the preset maintenance risk assessment threshold, a maintenance qualified signal is generated.

[0030] Furthermore, the present invention also proposes a real-time optimization method for sliding pressure operation of a steam turbine unit, comprising the following steps:

[0031] Step 1: Collect various operating data of the steam turbine unit, and filter, remove noise and process outliers on the collected data;

[0032] Step 2: predicting the heat rate of the steam turbine unit based on the operating data of the steam turbine unit;

[0033] Step 3: Using the predicted heat rate as the objective function, the GSA algorithm is used to search for the optimal main steam pressure;

[0034] Step 4: Dynamically optimize the sliding pressure operation of the steam turbine unit based on the optimal main steam pressure.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] 1. In the present invention, a sensor integration module collects various operating data of the steam turbine unit in real time and processes the data. A heat rate prediction module calculates and predicts the heat rate of the steam turbine unit based on the operating data of the steam turbine unit. An optimization algorithm module uses the predicted heat rate as the objective function within the feasible main steam pressure range and utilizes the GSA algorithm to search for the optimal main steam pressure. A dynamic optimization module dynamically optimizes the sliding pressure operation of the steam turbine unit based on the optimal main steam pressure, thereby significantly improving the efficiency of the steam turbine unit, reducing energy consumption, enhancing adaptability, and improving economy and safety.

[0037] 2. In the present invention, all sensors on the steam turbine unit are monitored by the acquisition monitoring and evaluation module to determine the acquisition operation performance of the sensor integration module. When an acquisition alarm signal is generated, the operation of the steam turbine unit is suspended and the relevant sensors are inspected and repaired. When an acquisition qualified signal is generated, the parameter maintenance performance of the steam turbine unit is evaluated and analyzed by the maintenance detection and evaluation module. When a maintenance abnormality signal is generated, the operation of the steam turbine unit is suspended and the cause is investigated. This is beneficial to ensuring the operation effect of the steam turbine unit, significantly reducing the management difficulty of the steam turbine unit, and has a high level of intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0039] Figure 1 This is a system block diagram of Embodiment 1 of the present invention;

[0040] Figure 2 This is a system block diagram of Embodiment 2 and Embodiment 3 of the present invention;

[0041] Figure 3 This is a flow chart of the method of embodiment 4 of the present invention. DETAILED DESCRIPTION

[0042] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0043] Example 1: Figure 1 As shown, the present invention proposes a real-time optimization system for sliding pressure operation of a steam turbine unit, comprising a sensor integration module, a heat rate prediction module, an optimization algorithm module, a dynamic optimization module and a supervisory control terminal;

[0044] The sensor integration module collects various operating data of the steam turbine unit in real time, including main steam pressure, temperature, flow, speed, back pressure and extraction volume, through several high-precision sensors installed at key parts of the steam turbine unit. Preferably, the data collection frequency is set to ten times per second. The collected data is filtered, denoised and processed for outliers to improve data quality, and the processed operating data is sent to the heat rate prediction module.

[0045] The heat rate prediction module calculates the heat rate of the steam turbine unit based on the operating data of the steam turbine unit and through the heat rate calculation formula, providing a theoretical benchmark for the algorithm and ensuring the physical rationality of the prediction results. The LS-SVM algorithm is used to predict the heat rate and the predicted heat rate is sent to the optimization algorithm module, thereby providing information support for the analysis process of the optimization algorithm module. The heat rate calculation formula is as follows:

[0046] ;

[0047] Wherein, HRt represents the predicted heat rate; Fthr represents the main steam flow rate; Hthr represents the main steam enthalpy; Fhrh represents the hot reheat steam flow rate; Hhrh represents the reheat hot section steam enthalpy; Fffw represents the feed water flow rate; Hffw represents the boiler feed water enthalpy; Fcrh represents the cold reheat steam flow rate; Hcrh represents the reheat cold section enthalpy; Frhs represents the reheat desuperheating water flow rate; Hrhs represents the reheat desuperheating water enthalpy; Wc represents the generator power after the secondary pressure drop correction.

[0048] The optimization algorithm module searches for the optimal main steam pressure within the feasible main steam pressure range with the predicted heat rate as the objective function using the GSA algorithm, and sends the optimal main steam pressure to the dynamic optimization module and the supervisory control end; the dynamic optimization module dynamically optimizes the sliding pressure operation of the steam turbine unit based on the optimal main steam pressure, and sends the optimization information to the supervisory control end.

[0049] The present invention can realize real-time optimization of the sliding pressure operation of the steam turbine unit through the coordinated work between the modules, thereby improving the efficiency of the steam turbine unit, reducing energy consumption, enhancing adaptability, and improving economy and safety.

[0050] Example 2: Figure 2 As shown, the difference between this embodiment and the first embodiment is that the sensor integration module is communicatively connected to the acquisition monitoring and evaluation module. The acquisition monitoring and evaluation module monitors all sensors on the steam turbine unit and marks the corresponding sensor as a monitoring object i, where i is a natural number greater than 1; the monitoring object i is marked as an obstacle object or a normal object through acquisition anomaly evaluation analysis, and if an obstacle object exists within a unit time, an acquisition alarm signal is generated;

[0051] If there is no obstacle within the unit time, the sensor will cooperate with the decision analysis to generate an acquisition alarm signal or an acquisition qualified signal, and the acquisition alarm signal or the acquisition qualified signal will be sent to the supervision and control end. When the supervision and control end receives the acquisition alarm signal, it will issue an early warning to remind the supervisor to suspend the operation of the steam turbine unit and inspect and repair the relevant sensors on the steam turbine unit as needed to ensure the timeliness and accuracy of data acquisition, avoid affecting the monitoring effect of the steam turbine unit, and help ensure the safe, stable and energy-saving operation of the steam turbine unit. The specific analysis process of acquisition anomaly assessment analysis is as follows:

[0052] Obtain the operation data collected by monitoring object i. If the corresponding operation data is not within the normal collection value range of monitoring object i, the corresponding operation data is defined as error data; count the number of error data collected corresponding to monitoring object i in a unit time and mark it as the error detection value; compare the error detection value with the preset error detection threshold; if the error detection value exceeds the preset error detection threshold, it indicates that the collection operation performance of monitoring object i is poor, and monitoring object i is marked as an obstacle object;

[0053] If the sampling error detection value does not exceed the preset sampling error detection threshold, the time when the monitoring object i performs data collection is obtained and marked as the actual sampling time, and the deviation time of the actual sampling time compared to the corresponding standard collection time is marked as the resistance measurement time; all the resistance measurement time corresponding to the monitoring object i in the unit time are obtained and the average is calculated to obtain the resistance measurement time table value, and the resistance measurement time is numerically compared with the preset resistance measurement time threshold, and the number of resistance measurement time exceeding the preset resistance measurement time threshold in the unit time is marked as the resistance measurement time deviation value;

[0054] The collection deviation evaluation value is obtained by weightedly summing the sampling error detection value, the resistance measurement time table value, and the resistance measurement time deviation value of the monitoring object i; that is, the sampling error detection value, the resistance measurement time table value, and the resistance measurement time deviation value are respectively assigned corresponding preset weight coefficients, and the sampling error detection value, the resistance measurement time table value, and the resistance measurement time deviation value are respectively multiplied by the corresponding preset weight coefficients, and the three sets of product results are summed up to obtain the collection deviation evaluation value; and the larger the value of the collection deviation evaluation value, the worse the overall collection operation performance of the monitoring object i is;

[0055] The collection error value is numerically compared with the preset collection error threshold. If the collection error value exceeds the preset collection error threshold, it indicates that the collection operation performance of the monitoring object i is generally poor, and the monitoring object i is marked as an obstacle object; if the collection error value does not exceed the preset collection error threshold, it indicates that the collection operation performance of the monitoring object i is generally good, and the monitoring object i is marked as a normal object.

[0056] Furthermore, the specific analysis process of the sensor coordination decision analysis is as follows: obtaining the abnormal evaluation values of all sensors, marking the average value of all abnormal evaluation values as the coordination preliminary evaluation value, comparing the preliminary evaluation value with the preset coordination preliminary evaluation threshold, and if the preliminary evaluation value exceeds the preset coordination preliminary evaluation threshold, indicating that the operation status of the sensor integration module is poor, a collection alarm signal is generated;

[0057] If the initial evaluation value does not exceed the preset initial evaluation threshold, the corresponding standard collection time is defined as the target time. If the operating data collected by the monitoring object i at the corresponding target time is erroneous data or the resistance measurement duration exceeds the preset resistance measurement duration threshold, the monitoring object i is assigned the collection judgment symbol TX-1;

[0058] If there is a sensor assigned with the sampling symbol TX-1 at the corresponding target moment, the corresponding target moment will be marked as the alarm moment, and the proportion of the number of alarm moments per unit time will be marked as the sampling alarm value. The sampling alarm value will be compared with the preset sampling alarm threshold. If the sampling alarm value exceeds the preset sampling alarm threshold, it indicates that the operating condition of the sensor integrated module is not good, and a sampling alarm signal is generated; if the sampling alarm value does not exceed the preset sampling alarm threshold, it indicates that the operating condition of the sensor integrated module is good, and a sampling qualified signal is generated.

[0059] Example 3: Figure 2 As shown, the difference between this embodiment and the first and second embodiments is that the acquisition monitoring and evaluation module is communicatively connected to the maintenance detection and evaluation module. The acquisition monitoring and evaluation module sends the acquisition qualified signal to the maintenance detection and evaluation module. When the maintenance detection and evaluation module receives the acquisition qualified signal, it performs maintenance performance evaluation and analysis of the steam turbine unit to generate a maintenance qualified signal or a maintenance abnormality signal.

[0060] The qualified maintenance signal or abnormal maintenance signal is sent to the supervisory control end. When the supervisory control end receives the abnormal maintenance signal, it issues an early warning to remind supervisors to suspend the operation of the steam turbine unit and conduct a cause investigation. It also takes reasonable countermeasures as needed to ensure the control effect of the steam turbine unit operating parameters, which is conducive to the continuous stability of various operating parameters and further improves the operating effect of the steam turbine unit. It has a high level of intelligence. The specific analysis process of the steam turbine unit maintenance performance evaluation and analysis is as follows:

[0061] Obtain operating parameters that need to be monitored during the operation of the steam turbine unit, mark the corresponding operating parameters as matching parameters k, where k is a natural number greater than 1; obtain real-time data of the matching parameter k, and when the deviation of the real-time data from the current corresponding data standard value exceeds the corresponding preset deviation threshold, determine that the matching parameter k is in a maintenance risk state;

[0062] Obtain the number of occurrences of the matching parameter k in the maintenance risk state per unit time and the duration of each occurrence and define them as the maintenance risk frequency value and maintenance risk duration value respectively. Compare the maintenance risk frequency value and maintenance risk duration value with the corresponding preset maintenance risk frequency threshold value and preset maintenance risk duration threshold value respectively. If the maintenance risk frequency value or the maintenance risk duration value exceeds the corresponding preset threshold value, indicating that the maintenance performance of the matching parameter k per unit time is poor, the matching parameter k is marked as a maintenance damage parameter.

[0063] If the maintenance risk frequency value and the maintenance risk time value do not exceed the corresponding preset threshold value, the maximum deviation of the real-time data of the matching parameter k in unit time compared with the corresponding data standard value is marked as the maintenance deviation amplitude value;

[0064] The maintenance bias amplitude, maintenance critical frequency and maintenance critical time of the matching parameter k are weighted and summed to obtain the dimensional difference characteristic value; that is, the maintenance bias amplitude, maintenance critical frequency and maintenance critical time are assigned corresponding preset weight coefficients, and the maintenance bias amplitude, maintenance critical frequency and maintenance critical time are multiplied by the corresponding preset weight coefficients respectively, and the three sets of product results are summed to obtain the dimensional difference characteristic value; and the larger the value of the dimensional difference characteristic value, the worse the control condition of the matching parameter k is, and the worse the maintenance performance of the matching parameter k in unit time is.

[0065] The dimensional difference characteristic value is numerically compared with the preset dimensional difference characteristic threshold. If the dimensional difference characteristic value exceeds the preset dimensional difference characteristic threshold, it indicates that the control condition of the matching parameter k is poor, and the maintenance performance of the matching parameter k in unit time is generally poor, then the matching parameter k is marked as a maintenance-damaging parameter; if there is a maintenance-damaging parameter in unit time, it indicates that the parameter maintenance risk of the steam turbine unit is high, and the operation control condition of the steam turbine unit is poor, then a maintenance abnormality signal is generated.

[0066] Furthermore, if there is no maintenance-causing parameter within a unit time, the preset maintenance importance weight value of the matching parameter k is retrieved, wherein the preset maintenance importance weight values are all positive numbers, and the more important the stable maintenance of the corresponding operating parameter is, the larger the value of the preset maintenance importance weight value matched thereto is; the ratio of the dimensional difference feature value of the matching parameter k to the corresponding preset dimensional difference feature threshold is marked as the dimensional difference detection value, the dimensional difference detection value of the matching parameter k is multiplied by the corresponding preset maintenance importance weight value, and the product result is marked as the maintenance risk judgment value;

[0067] The maintenance risk judgment values of all operating parameters that need to be monitored are obtained, and all maintenance risk judgment values are summed up to obtain a maintenance risk assessment value. The maintenance risk assessment value is numerically compared with the preset maintenance risk assessment threshold. If the maintenance risk assessment value exceeds the preset maintenance risk assessment threshold, it indicates that the parameter maintenance risk of the steam turbine unit is generally high, and a maintenance abnormality signal is generated; if the maintenance risk assessment value does not exceed the preset maintenance risk assessment threshold, it indicates that the parameter maintenance risk of the steam turbine unit is generally low, and the operation control condition of the steam turbine unit is better, and a maintenance qualified signal is generated.

[0068] Example 4: Figure 3 As shown, the difference between this embodiment and the first, second and third embodiments is that the present invention proposes a real-time optimization method for sliding pressure operation of a steam turbine unit, comprising the following steps:

[0069] Step 1: Collect various operating data of the steam turbine unit, and filter, remove noise and process outliers on the collected data;

[0070] Step 2: predicting the heat rate of the steam turbine unit based on the operating data of the steam turbine unit;

[0071] Step 3: Using the predicted heat rate as the objective function, the GSA algorithm is used to search for the optimal main steam pressure;

[0072] Step 4: Dynamically optimize the sliding pressure operation of the steam turbine unit based on the optimal main steam pressure.

[0073] The working principle of the present invention is as follows: when in use, various operating data of the steam turbine unit are collected in real time and processed through the sensor integration module; the heat rate prediction module calculates and predicts the heat rate of the steam turbine unit based on the operating data of the steam turbine unit; the optimization algorithm module uses the predicted heat rate as the objective function within the feasible main steam pressure range and uses the GSA algorithm to search for the optimal main steam pressure; the dynamic optimization module dynamically optimizes the sliding pressure operation of the steam turbine unit based on the optimal main steam pressure, thereby achieving real-time optimization of the sliding pressure operation of the steam turbine unit, improving the efficiency of the steam turbine unit, reducing energy consumption, enhancing adaptability, and improving economy and safety; and the acquisition monitoring and evaluation module monitors all sensors on the steam turbine unit to determine the acquisition and operation performance of the sensor integration module. When an acquisition alarm signal is generated, the operation of the steam turbine unit is suspended and the relevant sensors are inspected and repaired. When an acquisition qualified signal is generated, the maintenance detection and evaluation module evaluates and analyzes the parameter maintenance performance of the steam turbine unit. When a maintenance abnormality signal is generated, the operation of the steam turbine unit is suspended and the cause is investigated. This is conducive to ensuring the operating effect of the steam turbine unit and has a high level of intelligence.

[0074] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A real-time optimization system for sliding pressure operation of a steam turbine unit, characterized in that: It includes a sensor integration module, a heat rate prediction module, an optimization algorithm module, a dynamic optimization module, and a supervisory control terminal. The sensor integration module collects various operating data of the steam turbine unit in real time, filters, removes noise, and processes outliers on the collected data, and sends the processed operating data to the heat rate prediction module. The heat rate prediction module calculates the heat rate of the steam turbine unit based on the operating data of the steam turbine unit using the heat rate calculation formula, and uses the LS-SVM algorithm to predict the heat rate and send the predicted heat rate to the optimization algorithm module. The optimization algorithm module uses the GSA algorithm to search for the optimal main steam pressure within the feasible main steam pressure range with the predicted heat rate as the objective function. The dynamic optimization module dynamically optimizes the sliding pressure operation of the steam turbine unit based on the optimal main steam pressure and sends the optimization information to the supervisory control end. The sensor integration module is communicatively connected to the data acquisition, monitoring and evaluation module. The data acquisition, monitoring and evaluation module monitors all sensors on the steam turbine unit and marks the corresponding sensor as a monitoring object i, where i is a natural number greater than 1. The monitored object i is marked as an obstacle object or a normal object through acquisition anomaly assessment and analysis. If an obstacle object exists within a unit time, an acquisition alarm signal is generated; If there is no obstacle within the unit time, the sensor will cooperate with the decision analysis to generate a collection alarm signal or a collection pass signal, and the collection alarm signal or the collection pass signal will be sent to the supervision and control end. When the supervision and control end receives the collection alarm signal, it will issue an early warning; The specific analysis process of collection anomaly assessment analysis is as follows: Obtaining the operating data collected by the monitoring object i, if the corresponding operating data is not within the normal collection value range of the monitoring object i, then the corresponding operating data is defined as error data; Count the number of error data collected for monitoring object i within a unit time and mark it as the error detection value. Compare the error detection value with the preset error detection threshold. If the error detection value exceeds the preset error detection threshold, the monitoring object i is marked as an obstacle object. If the sampling error detection value does not exceed the preset sampling error detection threshold, the time when the monitored object i collects data is obtained and marked as the actual sampling time, and the deviation time of the actual sampling time compared to the corresponding standard sampling time is marked as the resistance measurement time; Obtain all the resistance measurement durations corresponding to the monitoring object i within the unit time and calculate their average, thereby obtaining the resistance measurement time table value, and compare the resistance measurement duration with the preset resistance measurement duration threshold, and mark the number of resistance measurement durations exceeding the preset resistance measurement duration threshold within the unit time as the resistance measurement time outlier value; The acquisition error evaluation value is calculated by weighted summing the acquisition error detection value, the resistance measurement time table value and the resistance measurement time deviation value of the monitoring object i, and the acquisition error evaluation value is compared with the preset acquisition error evaluation threshold. If the acquisition error evaluation value exceeds the preset acquisition error evaluation threshold, the monitoring object i is marked as an obstacle object; if the acquisition error evaluation value does not exceed the preset acquisition error evaluation threshold, the monitoring object i is marked as a normal object; The specific analysis process of sensor-assisted decision analysis is as follows: Obtain the abnormal evaluation values of all sensors, mark the average of all abnormal evaluation values as the initial evaluation value, compare the initial evaluation value with the preset initial evaluation threshold, and generate an alarm signal if the initial evaluation value exceeds the preset threshold. If the initial evaluation value does not exceed the preset initial evaluation threshold, the corresponding standard collection time is defined as the target time. If the operating data collected by the monitoring object i at the corresponding target time is erroneous data or the resistance measurement duration exceeds the preset resistance measurement duration threshold, the monitoring object i is assigned the collection judgment symbol TX-1; If there is a sensor assigned with the sampling symbol TX-1 at the corresponding target moment, the corresponding target moment is marked as the alarm moment, and the ratio of the number of alarm moments per unit time is marked as the sampling alarm value. The sampling alarm value is numerically compared with the preset sampling alarm threshold. If the sampling alarm value exceeds the preset sampling alarm threshold, a sampling alarm signal is generated. If the acquisition alarm value does not exceed the preset acquisition alarm threshold, an acquisition qualified signal is generated.

2. A steam turbine unit sliding pressure operation real-time optimization system according to claim 1, characterized in that: The heat rate calculation formula is as follows: ; In the formula, HRt represents the predicted heat rate; Fthr represents the main steam flow rate; Hthr represents the main steam enthalpy; Fhrh represents the hot reheat steam flow rate; Hhrh represents the reheat hot section steam enthalpy; Fffw represents the feed water flow rate; Hffw represents the boiler feed water enthalpy; Fcrh represents the cold reheat steam flow rate; Hcrh represents the reheat cold section enthalpy; Frhs represents the reheat desuperheating water flow rate; Hrhs represents the reheat desuperheating water enthalpy; Wc represents the generator power after the secondary pressure drop correction.

3. A steam turbine unit sliding pressure operation real-time optimization system according to claim 1, characterized in that: The acquisition monitoring and evaluation module is communicatively connected to the maintenance detection and evaluation module. The acquisition monitoring and evaluation module sends the acquisition qualified signal to the maintenance detection and evaluation module. When the maintenance detection and evaluation module receives the acquisition qualified signal, it performs maintenance performance evaluation and analysis of the turbine unit to generate a maintenance qualified signal or a maintenance abnormality signal, and sends the maintenance qualified signal or the maintenance abnormality signal to the supervision and control end.

4. A steam turbine unit sliding pressure operation real-time optimization system according to claim 3, characterized in that: The specific analysis process of the steam turbine unit maintenance performance evaluation and analysis includes: Obtain operating parameters that need to be monitored during the operation of the steam turbine unit, mark the corresponding operating parameters as matching parameters k, where k is a natural number greater than 1; obtain real-time data of the matching parameter k, and when the deviation of the real-time data from the corresponding data standard value exceeds the corresponding preset deviation threshold, determine that the matching parameter k is in a maintenance risk state; Obtain the number of occurrences of the matching parameter k in the maintenance risk state per unit time and the duration of each occurrence and define them as the maintenance risk frequency value and maintenance risk duration value respectively. Compare the maintenance risk frequency value and maintenance risk duration value with the corresponding preset maintenance risk frequency threshold value and preset maintenance risk duration threshold value respectively. If the maintenance risk frequency value or the maintenance risk duration value exceeds the corresponding preset threshold value, the matching parameter k is marked as a maintenance loss parameter. If the maintenance risk frequency value and the maintenance risk time value do not exceed the corresponding preset threshold value, the maximum deviation of the real-time data of the matching parameter k compared with the corresponding data standard value within the unit time is marked as the maintenance deviation amplitude value, and the maintenance deviation amplitude value, maintenance risk frequency value and maintenance risk time value of the matching parameter k are weighted and summed to obtain the dimensional abnormality characteristic value, and the dimensional abnormality characteristic value is numerically compared with the preset dimensional abnormality characteristic threshold. If the dimensional abnormality characteristic value exceeds the preset dimensional abnormality characteristic threshold, the matching parameter k is marked as a maintenance damage parameter; if a maintenance damage parameter exists within the unit time, a maintenance abnormality signal is generated.

5. A steam turbine unit sliding pressure operation real-time optimization system according to claim 4, characterized in that: If there is no maintenance-causing parameter within the unit time, the preset maintenance importance weight value of the matching parameter k is retrieved, the ratio of the dimensional difference feature value of the matching parameter k to the corresponding preset dimensional difference feature threshold is marked as the dimensional difference detection value, the dimensional difference detection value of the matching parameter k is multiplied by the corresponding preset maintenance importance weight value, and the product result is marked as the maintenance risk judgment value; Obtaining the maintenance risk judgment values of all operating parameters that need to be monitored, and summing up all the maintenance risk judgment values to obtain a maintenance risk assessment value, comparing the maintenance risk assessment value with a preset maintenance risk assessment threshold, and generating a maintenance abnormality signal if the maintenance risk assessment value exceeds the preset maintenance risk assessment threshold; If the maintenance risk assessment value does not exceed the preset maintenance risk assessment threshold, a maintenance pass signal is generated.

Images