Multi-target prediction method and system based on multi-output regression module

Abstract

The invention relates to the technical field of power plants, and discloses a multi-target prediction method and system based on a multi-output regression module, and the method comprises the following steps: obtaining equipment related parameter data of a steam turbine, a boiler and a generator of a power plant, numbering the data, and storing the data as a data set; calculating an equipment operation index # imgabs0 # and an equipment load index # imgabs1 # according to the data set after the serial number is acquired; the power generation state index # imgabs4 # is evaluated according to the equipment operation index # imgabs2 # and the equipment load index # imgabs3 #; and making a corresponding response according to the evaluation result of the power generation state index # imgabs5 #. The working state of the equipment is comprehensively evaluated through multiple pieces of data, abnormal conditions are found in time, misjudgment and missed judgment caused by a single value are avoided, the equipment is accurately recognized and diagnosed, and potential downtime and loss are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power plants, and specifically to a method and system for multi-objective prediction based on a multi-output regression module. Background Art

[0002] Power plants are important infrastructure for modern social and economic development, providing a stable power supply for industrial production, commercial activities, and residential life, ensuring the normal operation of the economic society. The widespread application of electricity has promoted the development of various industries, contributing to economic growth and social progress. The construction and development of different types of power plants help to diversify the energy structure, reduce dependence on traditional fossil fuels, and improve the security and stability of energy supply. At the same time, the development of new energy power plants is of great significance for addressing climate change and protecting the environment. The construction and development of power plants have promoted innovation and progress in related technologies, such as the research and development of large generator sets, the application of smart grid technology, and the improvement of new energy power generation technology. These technological innovations not only improve the power generation efficiency and reliability of power plants but also provide technical support for the development of other fields. As an indispensable core of energy supply in modern society, the development process of power plants has witnessed technological progress and social changes. From early direct current power generation to the coexistence of various types of power plants today, it not only meets the growing power demand but also plays a crucial role in promoting economic development, optimizing the energy structure, and driving technological innovation. With the continuous emergence of new energy technologies, power plants will continue to lead the transformation in the energy field and contribute to a sustainable future.

[0003] During the daily operation of power plants, to ensure the stable operation of equipment, the background operation station needs to monitor and compare various indicators in real time. Currently, the regression method for target prediction problems usually refers to predicting the value of a single target variable based on the input features. However, in the actual context of power plants, simply predicting a single target variable is often insufficient to comprehensively reflect the operating state of equipment, easily ignoring the mutual influence between variables, leading to misjudgments of potential equipment failures or performance degradation, and being unable to accurately identify and diagnose, resulting in equipment damage. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a method for multi-objective prediction based on a multi-output regression module, which has the advantages of comprehensively evaluating the working state of equipment through multiple data, timely detecting abnormal situations, avoiding misjudgments and omissions caused by a single value, accurately identifying and diagnosing equipment, reducing potential downtime and losses, dynamically adjusting the equipment operation mode according to the equipment load index, not only ensuring the efficient operation of equipment under different load conditions but also contributing to energy conservation and emission reduction.

[0005] To achieve the above object, the present invention provides the following technical solutions: First aspect, a method for multi-objective prediction based on a multi-output regression module, comprising the following steps: Step 1: Obtain the equipment-related parameter data of the power plant steam turbine, boiler, and generator, and number and save them as a data set; Step 2: Calculate the equipment operation index and the equipment load index based on the numbered data set obtained; Step 3: Evaluate the power generation status index according to the equipment operation index and the equipment load index; Step 4: Make corresponding responses according to the evaluation result of the power generation status index

[0006] Preferably, the equipment-related parameter data includes the operation data, equipment load data, and equipment response data of the equipment. Numbering and saving the operation data, equipment load data, and equipment response data of the equipment as a data set specifically includes: the equipment operation data set

[0007] Preferably, the equipment operation data set includes the operation data of the steam turbine, boiler, and generator and the current operation environment, and its expression is: In the expression, successively represent the operation data of the steam turbine, boiler, and generator and the current operation environment data, and the superscript represents the corresponding specific operation parameters. The operation parameters of the steam turbine are the power, speed, pressure, temperature, and flow rate of the steam turbine. The operation parameters of the boiler are the pressure, temperature, pollutant emission amount, combustion amount, and water consumption of the boiler. The operation parameters of the generator are the power, voltage, current, temperature, and frequency of the generator. The operation environment data is the environmental temperature, environmental humidity, and air dust content.

[0008] Preferably, the equipment load data set includes the load data of the steam turbine, boiler, and generator, and its expression is: In the expression, successively represent the load data of the steam turbine, boiler, and generator, and the superscript represents the corresponding specific load parameters. The load data of the steam turbine is the electrical load and the thermal load. The load data of the boiler is the evaporation amount and the heat supply amount. The load data of the generator is the load and the electrical load.

[0009] Preferably, the equipment response data set​ Including the response data of the steam turbine, boiler, and generator, and its expression is: , representing the response data of the steam turbine, boiler, and generator in sequence, and the superscript representing the specific response parameters of the corresponding equipment. The specific response parameters of the response data of the steam turbine, boiler, and generator include response time, adjustment accuracy, and stability during the adjustment period.

[0010] Preferably: The calculation formula of the equipment operation index is:

[0011] In the calculation formula, , , respectively represent the real-time operation data of the steam turbine, boiler, and generator, , , respectively represent the standard operation data of the steam turbine, boiler, and generator, represents the real-time operation environment data, represents the standard operation environment data, represents the weight of the environment data, represents the weight of the operation data of the steam turbine, boiler, and generator.

[0012] Preferably: The calculation formula of the equipment load index is:

[0013] In the calculation formula, respectively represent the real-time response data of the steam turbine, boiler, and generator, respectively represent the standard response data of the steam turbine, boiler, and generator, represents the impact of the response factor on the load, represents the weight of the response factor, , , respectively represent the real-time load data of the steam turbine, boiler, and generator, , , respectively represent the standard load data of the steam turbine, boiler, and generator, represents the weight of the equipment load.

[0014] Preferably: The calculation formula of the power generation status index is:

[0015] In the calculation formula, represents the number of faults of the device, represents the total operating duration of the device, represents the fault frequency of the device.

[0016] In a second aspect, the present invention provides a system for multi-objective prediction based on a multi-output regression module, which is characterized in that it includes a parameter acquisition module, an operation index calculation module, an evaluation module, and a prediction response module; The parameter acquisition module acquires the device-related parameter data of the power plant steam turbine, boiler, and generator, and numbers and saves them as a data set; The operation index calculation module calculates the device operation index and the device load index ; The evaluation module evaluates the power generation status index based on the device operation index and the device load index ; The prediction response module makes corresponding responses according to the evaluation result of the power generation status index .

[0017] Preferably, the device load data set includes the load data of the steam turbine, boiler, and generator, and its expression is: , in the expression, successively represent the load data of the steam turbine, boiler, and generator, and the superscript represents the corresponding specific load parameter. The load data of the steam turbine are the electrical load and the thermal load, the load data of the boiler are the evaporation capacity and the heat supply amount, and the load data of the generator are the load and the electrical load.

[0018] Preferably, the device response data set includes the response data of the steam turbine, boiler, and generator, and its expression is: , successively represent the response data of the steam turbine, boiler, and generator, and the superscript represents the specific response parameter of the corresponding device. The specific response parameters of the response data of the steam turbine, boiler, and generator include the response time, the adjustment accuracy, and the stability during the adjustment period.

[0019] Preferably, the calculation formula of the device operation index is:

[0020] In the calculation formula, , , respectively represent the real-time operation data of the steam turbine, boiler, and generator, , , respectively represent the standard operation data of the steam turbine, boiler, and generator, represents the real-time operation environment data, represents the standard operation environment data, represents the weight of the environment data, represents the weights of the operation data of the steam turbine, boiler, and generator.

[0021] Preferably, the calculation formula of the equipment load index is:

[0022] In the calculation formula, respectively represent the real-time response data of the steam turbine, boiler, and generator, respectively represent the standard response data of the steam turbine, boiler, and generator, represents the impact of the response factor on the load, represents the weight of the response factor, , , respectively represent the real-time load data of the steam turbine, boiler, and generator, , , respectively represent the standard load data of the steam turbine, boiler, and generator, represents the weight of the equipment load.

[0023] Preferably, the calculation formula of the power generation status index is:

[0024] In the calculation formula, represents the number of equipment failures, represents the total operation duration of the equipment, represents the failure frequency of the equipment.

[0025] Preferably, when the equipment operation index exceeds the maximum threshold of equipment operation, and the power generation status index exceeds the maximum threshold of the power generation status index, it represents that the current equipment has a failure, and a failure signal is sent to the maintenance personnel for maintenance.

[0026] Preferably, when the equipment load index is lower than the low threshold of equipment load, it represents that the current is low-load operation, and the energy-saving mode is started. When the equipment load index When it is higher than the high threshold of the device load, it represents that the current is in high-load operation. Optimize the load distribution to improve the device efficiency.

[0027] Compared with the prior art, the present invention provides a method for multi-objective prediction based on a multi-output regression module, which has the following beneficial effects: The present invention provides a method for multi-objective prediction based on a multi-output regression module. Save multiple device-related parameter data as a device operation data set and a device load data set and a device response data set , which can monitor the operation status, load condition and response to control instructions of the device in real time, provide a large amount of data basis for the evaluation and judgment of the device, realize a comprehensive evaluation of the working state of the device, timely discover abnormal situations, comprehensively evaluate the performance of the device, avoid misjudgment and missed judgment caused by a single value, and accurately identify and diagnose the device.

[0028] The present invention takes into account the number of device failures and the total operating duration, and adjusts the influence of the power generation state index through the failure frequency, which helps to more accurately reflect the actual operating condition of the device. Especially in the case of frequent failures, it can more sensitively capture the changes in the device state. A simple and effective fault detection mechanism based on threshold judgment can help to timely discover and handle device problems, reduce potential downtime and losses, and dynamically adjust the device operation mode according to the device load index, which not only ensures the efficient operation of the device under different load conditions, but also helps to save energy and reduce emissions. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a flowchart of the method for multi-objective prediction based on a multi-output regression module in an embodiment of the present invention.

[0030] Figure 2 It is a flowchart of the system for multi-objective prediction based on a multi-output regression module in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0031] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0032] Please refer to Figure 1 , the method for multi-objective prediction based on a multi-output regression module includes the following steps: Step 1: Obtain the equipment-related parameter data of the power plant's steam turbine, boiler, and generator, and number and save them as a data set. The equipment-related parameter data includes the equipment's operation data, equipment load data, and equipment response data. Numbering and saving the equipment's operation data, equipment load data, and equipment response data as a data set specifically includes: the equipment operation data set , the equipment load data set , and the equipment response data set ; The equipment operation data set includes the operation data of the steam turbine, boiler, and generator and the current operating environment. Its expression is: , in the expression, , represent the operation data of the steam turbine, boiler, and generator and the current operating environment data in sequence. The superscript represents the corresponding specific operating parameters. The operating parameters of the steam turbine are the power, speed, pressure, temperature, and flow rate of the steam turbine. The operating parameters of the boiler are the pressure, temperature, pollutant emissions, combustion amount, and water consumption of the boiler. The operating parameters of the generator are the power, voltage, current, temperature, and frequency of the generator. The operating environment data is the ambient temperature, ambient humidity, and air dust content; The equipment operation data set records the real-time operation data of the steam turbine, boiler, and generator and the current operating environment, including various operating parameters of the equipment, can monitor the operating status of the equipment in real time, discover abnormal situations in a timely manner, and provide a rich data basis for the evaluation and calculation of the equipment operation index; The equipment load data set includes the load data of the steam turbine, boiler, and generator. Its expression is: , in the expression, represent the load data of the steam turbine, boiler, and generator in sequence. The superscript represents the corresponding specific load parameters. The load data of the steam turbine are the electrical load and thermal load. The load data of the boiler are the evaporation capacity and heat supply. The load data of the generator are the load and electrical load; The equipment load data set contains the load data of the steam turbine, boiler, and generator, can understand the overloaded or underloaded conditions of the equipment in real time, provides a large amount of data basis for the calculation of the equipment load index, and is convenient for analyzing and evaluating the load of the equipment; The equipment response data set includes the response data of the steam turbine, boiler, and generator. Its expression is: , represent the response data of the steam turbine, boiler, and generator in sequence. The superscript Represent the specific response parameters of the corresponding equipment. The specific response parameters of the response data of steam turbines, boilers, and generators include response time, regulation accuracy, and stability during the regulation period; The equipment response data set records the response of steam turbines, boilers, and generators to control commands, including parameters such as response time, regulation accuracy, and stability during the regulation period, evaluates the control performance of the equipment, ensures its rapid and accurate response to operation commands, discovers problems in the control system in advance, improves the reliability of the system, and provides a data basis for the evaluation of the equipment; Step 2: Calculate the equipment operation index based on the numbered data set obtained and the equipment load index ; The equipment operation index is calculated as follows:

[0033] In the calculation formula, , , respectively represent the real-time operation data of the steam turbine, boiler, and generator, , , respectively represent the standard operation data of the steam turbine, boiler, and generator, represents the real-time operation environment data, represents the standard operation environment data, represents the weight of the environment data, represents the weights of the operation data of the steam turbine, boiler, and generator, , respectively represent the ratios between the real-time data and the standard values of the environment, steam turbine, boiler, and generator, represents the impact of the environment on the equipment operation; By combining the ratios of the real-time operation data and the standard operation data of the steam turbine, boiler, and generator, as well as the ratios of the real-time operation environment data and the standard operation environment data, and through the weight coefficients, the importance of the environment data and the equipment operation data in the calculation can be adjusted, which can more comprehensively reflect the operation status of the equipment and help to more accurately evaluate the power generation status index; The equipment load index is calculated as follows:

[0034] In the calculation formula, respectively represent the real-time response data of the steam turbine, boiler, and generator, respectively represent the standard response data of the steam turbine, boiler, and generator, Represents the ratio between the total real-time response data of the steam turbine, boiler, and generator and the standard value, and represents the impact of the response factor on the load in the formula. Represents the weight of the response factor. , , Respectively represent the real-time load data of the steam turbine, boiler, and generator. , , Respectively represent the standard load data of the steam turbine, boiler, and generator. Represents the weight of the equipment load. Represents the ratio between the real-time load data and the standard load data of the steam turbine, boiler, and generator. Comprehensively considers the ratio of the real-time response data to the standard response data of the steam turbine, boiler, and generator, as well as the ratio of the real-time load data to the standard load data. By adjusting the importance of the response factor and the equipment load in the calculation through the weight coefficient, it can more comprehensively reflect the load status of the equipment and help to more accurately evaluate the power generation status index. Step three: According to the equipment operation index , the equipment load index Evaluate the power generation status index . The power generation status index The calculation formula is:

[0035] In the calculation formula, Represents the number of equipment failures. Represents the total operation duration of the equipment. Represents the failure frequency of the equipment, which takes into account the number of equipment failures and the total operation duration. Adjusting the impact of the power generation status index through the failure frequency helps to more accurately reflect the actual operation status of the equipment. Especially in the case of frequent failures, it can more sensitively capture the changes in the equipment status. Step four: According to the evaluation result of the power generation status index , when the equipment operation index exceeds the maximum threshold of the equipment operation, and the power generation status index exceeds the maximum threshold of the power generation status index, it means that the current equipment has a failure. Send a failure signal to the maintenance personnel for maintenance. When the equipment load index is lower than the low threshold of the equipment load, it means that the current is low-load operation. Start the energy-saving mode. When the equipment load index is higher than the high threshold of the equipment load, it means that the current is high-load operation. Optimize the load distribution to improve the equipment efficiency. A simple and effective fault detection mechanism based on threshold judgment can help detect and handle equipment problems in a timely manner, reduce potential downtime and losses, and dynamically adjust the equipment operation mode according to the equipment load index, which not only ensures the efficient operation of the equipment under different load conditions but also helps with energy conservation and emission reduction.

[0036] In a specific embodiment of the present invention, as Figure 2 shown, a system for multi-objective prediction based on a multi-output regression module is provided, including a parameter acquisition module, an operation index calculation module, an evaluation module, and a prediction response module; The parameter acquisition module acquires the equipment-related parameter data of the steam turbine, boiler, and generator in the power plant, and numbers and saves them as a data set; The operation index calculation module calculates the equipment operation index , the equipment load index ; The evaluation module evaluates the power generation status index according to the equipment operation index and the equipment load index ; The prediction response module makes corresponding responses according to the evaluation result of the power generation status index .

[0037] In another embodiment of the present invention, a system for multi-objective prediction based on a multi-output regression module is provided, including a parameter acquisition module, an operation index calculation module, an evaluation module, and a prediction response module; The parameter acquisition module acquires the equipment-related parameter data of the steam turbine, boiler, and generator in the power plant, and numbers and saves them as a data set; The operation index calculation module calculates the equipment operation index , the equipment load index ; The evaluation module evaluates the power generation status index according to the equipment operation index and the equipment load index ; The prediction response module makes corresponding responses according to the evaluation result of the power generation status index .

[0038] Preferably, the equipment load data set includes the load data of the steam turbine, boiler, and generator, and its expression is: , in the expression, represent the load data of the steam turbine, boiler, and generator in sequence, and the superscript Represent the corresponding specific load parameters. The load data of the steam turbine are electric load and heat load. The load data of the boiler are evaporation capacity and heat supply. The load data of the generator are load and electric load.

[0039] Preferably, the equipment response data set includes the response data of the steam turbine, boiler, and generator, and its expression is: , represent the response data of the steam turbine, boiler, and generator in sequence. The superscript represents the specific response parameters of the corresponding equipment. The specific response parameters of the response data of the steam turbine, boiler, and generator include response time, regulation accuracy, and stability during the regulation period.

[0040] Preferably, the calculation formula of the equipment operation index is:

[0041] In the calculation formula, , , represent the real-time operation data of the steam turbine, boiler, and generator respectively. , , represent the standard operation data of the steam turbine, boiler, and generator respectively. represents the real-time operation environment data. represents the standard operation environment data. represents the weight of the environment data. represents the weights of the operation data of the steam turbine, boiler, and generator.

[0042] Preferably, the calculation formula of the equipment load index is:

[0043] In the calculation formula, represent the real-time response data of the steam turbine, boiler, and generator respectively. represent the standard response data of the steam turbine, boiler, and generator respectively. represents the impact of the response factor on the load. represents the weight of the response factor. , , represent the real-time load data of the steam turbine, boiler, and generator respectively. , , represent the standard load data of the steam turbine, boiler, and generator respectively. represents the weight of the equipment load.

[0044] Preferably, the power generation status index is calculated as follows:

[0045] In the calculation formula, represents the number of faults of the device, represents the total operating duration of the device, represents the fault frequency of the device.

[0046] Preferably, when the device operation index exceeds the maximum threshold of device operation, and the power generation status index exceeds the maximum threshold of the power generation status index, it means that the current device has a fault, and a fault signal is sent to the maintenance personnel for maintenance.

[0047] Preferably, when the device load index is lower than the low threshold of device load, it means that the current is low-load operation, and the energy-saving mode is started. When the device load index is higher than the high threshold of device load, it means that the current is high-load operation, and the load distribution is optimized to improve the device efficiency.

[0048] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for multi-target prediction based on a multi-output regression module, characterized in that: The following steps are involved: Obtain equipment-related parameter data of power plant steam turbines, boilers, and generators, and save them as data sets by numbering; Calculate the equipment operation index based on the data set after obtaining the number , Equipment Load Index ; According to the equipment operation index , Equipment Load Index Power Generation Status Index Conduct assessments; According to the power generation status index Evaluate the results and respond accordingly.

2. The method for multi-target prediction based on a multi-output regression module according to claim 1, characterized in that: The equipment-related parameter data includes equipment operation data, equipment load data, and equipment response data. The equipment operation data, equipment load data, and equipment response data are numbered and saved as a data set, specifically: equipment operation data set , Equipment Load Dataset , Device Response Dataset .

3. The method for multi-target prediction based on a multi-output regression module according to claim 2, characterized in that: The equipment operation data set Including the operating data of the steam turbine, boiler, generator and the current operating environment, the expression is: , in the expression, , which represent the operating data of the steam turbine, boiler, generator and the current operating environment data respectively. Represents the corresponding specific operating parameters, the operating parameters of the steam turbine are the power, speed, pressure, temperature, and flow of the steam turbine; the operating parameters of the boiler are the pressure, temperature, pollutant emissions, combustion volume, and water consumption of the boiler; the operating parameters of the generator are the power, voltage, current, temperature, and frequency of the generator; the operating environment data are the ambient temperature, ambient humidity, and air dust content.

4. The method for multi-target prediction based on a multi-output regression module according to claim 3, characterized in that: The equipment load data set Including the load data of steam turbine, boiler and generator, the expression is: , in the expression, Represents the load data of steam turbine, boiler and generator respectively. Represents the corresponding specific load parameters. The load data of the steam turbine are electrical load and thermal load. The load data of the boiler are evaporation capacity and heating capacity. The load data of the generator are load and electrical load.

5. The method for multi-target prediction based on a multi-output regression module according to claim 4, characterized in that: The device response data set Including the response data of steam turbine, boiler and generator, the expression is: , Represents the response data of steam turbine, boiler and generator respectively. Represents the specific response parameters of the corresponding equipment. The specific response parameters of the response data of the steam turbine, boiler, and generator include response time, regulation accuracy, and stability during the regulation period.

6. The method for multi-target prediction based on a multi-output regression module according to claim 5, characterized in that: The equipment operation index The calculation formula is: In the calculation formula, , , Represents the real-time operating data of steam turbine, boiler and generator respectively. , , Represents the standard operating data of steam turbine, boiler and generator respectively. Represents real-time operating environment data, Represents standard operating environment data, Represents the weight of environmental data, Represents the weight of the operating data of the steam turbine, boiler, and generator.

7. The method for multi-target prediction based on a multi-output regression module according to claim 6, characterized in that: The equipment load index The calculation formula is: In the calculation formula, Represent the real-time response data of steam turbine, boiler and generator respectively. Represent the standard response data of steam turbine, boiler and generator respectively. Represents the impact of the response factor on the load, represents the weight of the response factor, , , Represents the real-time load data of steam turbine, boiler and generator respectively. , , Represents the standard load data of steam turbine, boiler and generator respectively. Represents the weight of the equipment load.

8. The method for multi-target prediction based on a multi-output regression module according to claim 7, characterized in that: The power generation status index The calculation formula is: In the calculation formula, Represents the number of failures of the device. Represents the total operating time of the device. Represents the failure frequency of the device.

9. The method for multi-target prediction based on a multi-output regression module according to claim 8, characterized in that: When the device is running index Exceeding the maximum threshold of equipment operation, the power generation status index When the power generation status index exceeds the maximum threshold, it means that the current equipment is faulty, and a fault signal is sent to the maintenance personnel for maintenance. When the equipment load index If the load index is lower than the low threshold, it means that the current load is low and the energy saving mode is activated. When the load is higher than the high threshold of the equipment load, it means that the current operation is at high load. The load distribution should be optimized to improve the equipment efficiency.

10. A system for multi-target prediction based on a multi-output regression module, characterized in that: It includes parameter acquisition module, operation index calculation module, evaluation module and prediction response module; The parameter acquisition module obtains the equipment-related parameter data of the power plant's steam turbine, boiler, and generator, and numbers and saves them as data sets; The operation index calculation module calculates the equipment operation index based on the data set after obtaining the number , Equipment Load Index ; Evaluation module, based on equipment operation index , Equipment Load Index Power Generation Status Index Conduct assessments; Prediction response module, based on power generation status index Evaluate the results and respond accordingly.

Images