AVC adjustment rate prediction model construction method, prediction method and device

Abstract

The present invention provides a prediction model construction method, prediction method and device for AVC regulation rate, including: obtaining the first historical operating condition data of a thermal power unit, and establishing a fixed parameter information table according to the characteristics of the thermal power unit; regularly checking the first historical operating condition data according to the first fixed parameter, deleting the operating condition data that exceeds the limit, and obtaining the second historical operating condition data; calculating the AVC regulation rate of each operating condition based on the bus voltage at the first moment and the bus voltage at the second moment, combined with the time difference; performing K-means clustering on the second historical operating condition data and the corresponding AVC regulation rate, training a Gaussian mixture model according to the clustering results, and obtaining the AVC regulation rate prediction model of the thermal power unit. Compared with the prior art, the prediction model of the AVC regulation rate is constructed through historical operating conditions, and the online prediction of the AVC regulation rate based on the real-time operating conditions of the thermal power unit is realized, which improves the convenience and real-time performance.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a prediction model construction method, a prediction method, a prediction model construction device and a prediction device for an AVC regulation rate. Background Art

[0002] Voltage is an important indicator reflecting the quality of power grid electricity, and is of great significance to ensuring the safety, stability and economic operation of the grid. With the continuous increase in the scale of the power grid, the grid structure is becoming increasingly complex, and the normal voltage level of the power grid depends to a large extent on the voltage reactive control of the power plant or substation. As the stabilizer and important reactive source of the power system, the regulation and control of the bus voltage of the thermal power unit plays an important role in the stability of the power grid. The regulation of the bus voltage at the outlet of the thermal power unit is mainly achieved through the automatic voltage control (AVC) technology. At present, the existing technology detects the regulation rate of the plant-level AVC of the thermal power unit through a dynamic closed-loop test under regular and stable load operating conditions, but the operating conditions change in real time, and the test interval is generally long, and the test results are difficult to reflect the regulation performance of the AVC under the current operating conditions. Therefore, it is actually impossible to make predictions based on the current operating conditions of the power grid, and it is difficult to provide effective data support for the refined management of the power grid. In addition, the test requires close cooperation between the power dispatching department, power plant, and test institute, and requires a certain amount of manpower and material resources, which is easy to introduce disturbances to the grid voltage and affect the measurement results. Summary of the invention

[0003] The present invention provides a prediction model construction method, a prediction method, a prediction model construction device and a prediction device for an AVC regulation rate. By constructing a prediction model of the AVC regulation rate based on historical operating conditions, the technical problem that the prior art cannot make predictions based on the current operating conditions of the power grid is solved.

[0004] In order to solve the above technical problems, an embodiment of the present invention provides a method for constructing a prediction model of an AVC adjustment rate, comprising:

[0005] Acquire first historical operating condition data of the thermal power unit, and establish a fixed value parameter information table according to the characteristics of the thermal power unit; wherein the fixed value parameter information table includes a plurality of first fixed value parameters; each of the first fixed value parameters includes an upper limit value and a lower limit value;

[0006] According to the upper limit and lower limit of each of the first fixed value parameters, the first historical operating condition data is regularly checked, the operating condition data exceeding the limit is deleted, and the second historical operating condition data is obtained; wherein the second historical operating condition data includes the actual bus voltage value at the first moment of each operating condition and the actual bus voltage value at the second moment of each operating condition; the first moment is the moment when the bus voltage adjustment instruction issued by the AVC master station is received; the second moment is the moment when the bus voltage adjustment instruction issued by the AVC master station is received;

[0007] Calculate the AVC regulation rate of each operating condition in the second historical operating condition data based on the actual value of the bus voltage at the first moment and the actual value of the bus voltage at the second moment, combined with the time difference between the first moment and the second moment;

[0008] Based on the preset operating condition division conditions, K-means clustering is performed on the second historical operating condition data and the AVC regulation rate of each operating condition in the second historical operating condition data, and a Gaussian mixture model is trained according to the clustering results until convergence to obtain an AVC regulation rate prediction model of the thermal power unit; wherein the AVC regulation rate prediction model is used to predict the AVC regulation rate of the thermal power unit to be tested based on the real-time operating conditions of the thermal power unit to be tested.

[0009] As a preferred solution, the determination condition for the end of the AVC adjustment is specifically:

[0010] When the actual bus voltage value is higher than the first preset value or lower than the second preset value for 120 seconds, it is judged that the condition for ending the AVC regulation is met; wherein, the first preset value is the bus voltage target value corresponding to the bus voltage regulation instruction plus the dead zone value; the second preset value is the bus voltage target value minus the dead zone value; the moment when the AVC regulation ends is 150 seconds before it is judged that the judgment condition is met.

[0011] As a preferred solution, based on the preset working condition division condition, K-means clustering is performed on the second historical working condition data and the AVC adjustment rate of each working condition in the second historical working condition data, specifically:

[0012] Taking the AVC bus voltage set value, bus voltage, generator active power, generator reactive power, machine terminal voltage, fixed current, rotor current, plant 6kV bus voltage and AVC input status as the operating condition division conditions, K-means clustering is performed on the second historical operating condition data and the AVC regulation rate of each operating condition in the second historical operating condition data, and the number of clusters with the smallest sum of squared errors is taken as the optimal number of clusters to obtain the clustering results; wherein, the initial value of k for K-means clustering is 10 to 99.

[0013] As a preferred solution, the first historical operating condition data is regularly checked according to the upper limit and lower limit of each of the first fixed value parameters, and the operating condition data exceeding the limit is deleted to obtain the second historical operating condition data, specifically:

[0014] Each historical operating condition in the first historical operating condition data includes a plurality of operating condition parameters, and the operating condition parameters correspond one-to-one to the fixed value parameters;

[0015] The first historical operating condition data is checked regularly, and the operating condition data in which any operating condition parameter is greater than the upper limit value of the corresponding fixed value parameter or less than the lower limit value of the corresponding fixed value parameter is deleted to obtain the second historical operating condition data.

[0016] As a preferred scheme, before the regular verification of the first historical operating condition data, it also includes: performing upper and lower limit checks on the several first constant value parameters based on empirical values, adjusting the upper and lower limit values ​​of each of the first constant value parameters, and obtaining the first constant value parameters that have passed the upper and lower limit checks.

[0017] As a preferred solution, before the regular verification of the first historical operating condition data, after the upper and lower limit verification of the first fixed value parameters according to the empirical values, it also includes: eliminating the historical operating condition data that has not been put into AVC to obtain the first historical operating condition data that has been initially screened.

[0018] As a preferred solution, before eliminating the historical operating condition data that has not been put into AVC, the method further includes:

[0019] Recording the time period in the first historical operating condition data that meets the alarm condition; wherein the alarm condition is: any operating condition parameter is greater than the maximum value of the historical operation or less than the minimum value of the historical operation; or any operating condition parameter is greater than the maximum value of the empirical value or less than the minimum value of the empirical value;

[0020] After the historical operating condition data without AVC is eliminated, the following is also included:

[0021] The time period that meets the alarm condition is identified, and the corresponding operating condition data is recorded.

[0022] Accordingly, an embodiment of the present invention further provides a method for predicting an AVC adjustment rate, comprising:

[0023] Obtain the current operating condition data of the thermal power unit to be tested;

[0024] Inputting the operating condition data into the AVC regulation rate prediction model constructed by the AVC regulation rate prediction model construction method;

[0025] The classification of the current operating condition is determined by the similarity principle, and the AVC regulation rate of the thermal power unit to be tested is obtained according to the classification.

[0026] Accordingly, the embodiment of the present invention further provides a device for constructing a prediction model of an AVC adjustment rate, an information table establishment module, a verification module, a calculation module and a model construction module, wherein:

[0027] The information table establishment module is used to obtain the first historical operating condition data of the thermal power unit, and establish a fixed value parameter information table according to the characteristics of the thermal power unit; wherein the fixed value parameter information table includes a plurality of first fixed value parameters; each of the first fixed value parameters includes an upper limit value and a lower limit value;

[0028] The verification module is used to regularly verify the first historical operating condition data according to the upper limit value and the lower limit value of each first fixed value parameter, delete the operating condition data that exceeds the limit, and obtain the second historical operating condition data; wherein the second historical operating condition data includes the actual value of the bus voltage at the first moment of each operating condition and the actual value of the bus voltage at the second moment of each operating condition; the first moment is the moment of receiving the bus voltage adjustment instruction issued by the AVC master station; the second moment is the moment of receiving the bus voltage adjustment instruction issued by the AVC master station;

[0029] The calculation module is used to calculate the AVC regulation rate of each operating condition in the second historical operating condition data based on the actual value of the bus voltage at the first moment and the actual value of the bus voltage at the second moment, combined with the time difference between the first moment and the second moment;

[0030] The model building module is used to perform K-means clustering on the second historical operating condition data and the AVC regulation rate of each operating condition in the second historical operating condition data based on preset operating condition division conditions, and train a Gaussian mixture model according to the clustering results until convergence to obtain the AVC regulation rate prediction model of the thermal power unit; wherein the AVC regulation rate prediction model is used to predict the AVC regulation rate of the thermal power unit to be tested based on the real-time operating conditions of the thermal power unit to be tested.

[0031] Accordingly, an embodiment of the present invention further provides an AVC adjustment rate prediction device, comprising: an acquisition module, an input module and a prediction module; wherein:

[0032] The acquisition module is used to acquire the current operating condition data of the thermal power unit to be tested;

[0033] The input module is used to input the operating condition data into the AVC regulation rate prediction model constructed by the AVC regulation rate prediction model construction method;

[0034] The prediction module is used to determine the classification of the current operating conditions through the similarity principle, and obtain the AVC regulation rate of the thermal power unit to be tested according to the classification.

[0035] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0036] The embodiment of the present invention provides a prediction model construction method, a prediction method, a prediction model construction device and a prediction device for an AVC regulation rate, wherein the prediction model construction method comprises: obtaining first historical operating condition data of a thermal power unit, and establishing a fixed value parameter information table according to the characteristics of the thermal power unit; wherein the fixed value parameter information table comprises a plurality of first fixed value parameters; each of the first fixed value parameters comprises an upper limit value and a lower limit value; according to the upper limit value and the lower limit value of each of the first fixed value parameters, the first historical operating condition data is regularly checked, the operating condition data exceeding the limit is deleted, and the second historical operating condition data is obtained; wherein the second historical operating condition data comprises the actual value of the bus voltage at the first moment of each operating condition and the actual value of the bus voltage at the second moment of each operating condition; the first moment is the time when the bus voltage is received from the AVC master station. The moment of the voltage regulation instruction; the second moment is the moment of receiving the bus voltage regulation instruction issued by the AVC master station; based on the actual value of the bus voltage at the first moment and the actual value of the bus voltage at the second moment, combined with the time difference between the first moment and the second moment, the AVC regulation rate of each operating condition in the second historical operating condition data is calculated; based on the preset operating condition division conditions, the second historical operating condition data and the AVC regulation rate of each operating condition in the second historical operating condition data are subjected to K-means clustering, and a Gaussian mixture model is trained according to the clustering results until convergence to obtain the AVC regulation rate prediction model of the thermal power unit; wherein the AVC regulation rate prediction model is used to predict the AVC regulation rate of the thermal power unit to be tested based on the real-time operating conditions of the thermal power unit to be tested. Compared with the existing technology, the prediction model of AVC regulation rate is constructed through historical operating conditions, which realizes the online prediction of AVC regulation rate based on the real-time operating conditions of thermal power units, improving convenience and real-time performance; there is no need to obtain different data from power dispatching departments, power plants, test institutes and other units, saving manpower and material costs; the process of model construction and prediction will not introduce disturbances to the grid voltage, improving the accuracy of prediction; K-means clustering is performed on each operating condition based on preset division conditions, which realizes online AVC regulation rate prediction for different types of operating conditions. Compared with the existing technology that can only measure the AVC regulation rate for specific operating conditions, it has universal applicability and versatility for different operating conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 : A flow chart of an embodiment of a method for constructing a prediction model for AVC regulation rate provided by the present invention.

[0038] Figure 2 : A schematic diagram of a flow chart of an embodiment of the method for predicting the AVC adjustment rate provided by the present invention.

[0039] Figure 3 : A schematic structural diagram of an embodiment of a device for constructing a prediction model for AVC regulation rate provided by the present invention.

[0040] Figure 4 : A schematic diagram of the structure of an embodiment of the AVC adjustment rate prediction device provided by the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0042] Embodiment one:

[0043] Please refer to Figure 1 , Figure 1 A method for constructing a prediction model of an AVC adjustment rate provided in an embodiment of the present invention includes steps S101 to S104, wherein:

[0044] Step S101, obtaining the first historical operating condition data of the thermal power unit, and establishing a constant parameter information table according to the characteristics of the thermal power unit; wherein the constant parameter information table includes a plurality of first constant parameters; each of the first constant parameters includes an upper limit value and a lower limit value.

[0045] In this step, the first fixed value parameter includes but is not limited to 500kV / 220kV bus voltage effective value (high, low), lockout value (high, low), generator active power effective value (high, low), generator active power lockout value (high, low), generator reactive power effective value (high, low), generator reactive power lockout curve (high, low), plant 6kV bus voltage effective value (high, low) and lockout value (high, low), etc. These first fixed value parameters all contain a higher value and a lower value, representing a normal operating range of the first fixed value parameter, wherein the higher value is used as the upper limit value of the first fixed value parameter, and the lower value is used as the lower limit value of the first fixed value parameter. The fixed value parameter information table also includes other fixed value parameters such as voltage regulation dead zone and reactive power regulation dead zone, which facilitates the processing of historical operating condition data in subsequent steps. The selection of the first fixed value parameter and the selection of other fixed value parameters in the fixed value parameter information table are selected according to the characteristics of the thermal power unit itself. In the first historical operating condition data, each historical operating condition includes operating condition parameters corresponding to each fixed value parameter in the fixed value parameter information table, and also includes data such as the time period of occurrence of each operating condition.

[0046] Exemplarily, after the first historical operating condition data of the thermal power unit is obtained in step S101 and the fixed parameter information table is established, before step S102, it also includes: performing upper and lower limit checks on the several first fixed parameters according to empirical values, adjusting the upper limit and lower limit of each of the first fixed parameters, and obtaining the first fixed parameters that have been checked for upper and lower limits. Among them: the empirical value for each fixed parameter also includes the upper limit and lower limit of the empirical value. The empirical value can be used to determine whether the formulation of the fixed parameter information table obtained based on the characteristics of the thermal power unit is reasonable, and to adjust unreasonable fixed parameters. Similarly, for fixed parameters other than the first fixed parameters, such as the voltage regulation dead zone and the reactive power regulation dead zone, corresponding checks can also be performed based on empirical values.

[0047] Furthermore, after the upper and lower limit checks, the time periods in the first historical operating condition data that meet the alarm conditions are recorded. Wherein, the alarm conditions are: any operating condition parameter is greater than the maximum value of the historical operation or less than the minimum value of the historical operation; or any operating condition parameter is greater than the maximum value of the empirical value or less than the minimum value of the empirical value. Secondly, the first historical operating condition data is preliminarily screened to eliminate the historical operating condition data that has not been put into AVC, and then the time periods that meet the alarm conditions are identified, and the corresponding operating condition data are recorded. By implementing the embodiments of the present application, by identifying the alarm time periods and eliminating the historical operating condition data that has not been put into AVC, the consumption of data processing resources can be effectively reduced, while reducing invalid data in the historical operating condition data, improving the classification effect and accuracy of the subsequent model construction, and improving the work efficiency of the model to a certain extent.

[0048] Step S102: regularly check the first historical operating condition data according to the upper limit value and the lower limit value of each of the first constant parameters, delete the operating condition data that exceeds the limit, and obtain the second historical operating condition data.

[0049] Among them, the second historical operating condition data includes the actual value of the bus voltage at the first moment of each operating condition and the actual value of the bus voltage at the second moment of each operating condition; the first moment is the moment when the bus voltage adjustment instruction issued by the AVC master station is received; the second moment is the moment when the bus voltage adjustment instruction issued by the AVC master station is received.

[0050] In this embodiment, each historical operating condition in the first historical operating condition data includes a plurality of operating condition parameters, and the operating condition parameters correspond one-to-one to the fixed value parameters.

[0051] Regularly verify the first historical operating condition data (for example, verify it every day, which can be done by randomly checking several thermal power units. As new operating condition data is accumulated every day, operating condition data that does not meet the conditions can be cleaned up in time through daily verification), delete any operating condition data with an operating condition parameter greater than the upper limit value of the corresponding fixed value parameter or less than the lower limit value of the corresponding fixed value parameter, and obtain the second historical operating condition data.

[0052] In this step, when any unit AVC and plant-level AVC are both in operation, the AVC master station issues a bus voltage regulation instruction, and the bus voltage regulation instruction includes a bus voltage target value. After receiving the bus voltage regulation instruction, the actual bus voltage value changes accordingly toward the target value. When the actual bus voltage value is higher than the first preset value or lower than the second preset value for 120 seconds, it is determined that the conditions for the end of AVC regulation are met, and the moment when AVC regulation ends is determined. The setting value of the dead zone can be referenced, for example: when the actual bus voltage value is greater than or equal to 500kV, the dead zone is 0.5kV; when the actual bus voltage value is less than or equal to 220kV, the dead zone is 0.3kV. It can be further determined that the first preset value is the bus voltage target value corresponding to the bus voltage regulation instruction plus the dead zone value, and the second preset value is the bus voltage target value minus the dead zone value. When the actual value of the bus voltage continues to be higher than the first preset value or lower than the second preset value for five consecutive points (since there is 30 seconds between each two points, the four intervals are 120 seconds), the AVC regulation is judged to be ended. The moment when the AVC regulation ends is the point before the start time when the actual value of the bus voltage is higher than the first preset value or lower than the second preset value, that is, the moment corresponding to 30 seconds ago, and also the moment 150 seconds before the fifth point.

[0053] Step S103, based on the actual value of the bus voltage at the first moment and the actual value of the bus voltage at the second moment, combined with the time difference between the first moment and the second moment, calculate the AVC regulation rate of each operating condition in the second historical operating condition data.

[0054] In this embodiment, the first moment and the second moment are known, and the time difference can be obtained. At the same time, combined with the actual value of the bus voltage at the first moment and the actual value of the bus voltage at the second moment, the AVC regulation rate under each operating condition in the second historical operating condition data can be calculated through existing technology.

[0055] Step S104, based on the preset operating condition division conditions, K-means clustering is performed on the second historical operating condition data and the AVC regulation rate of each operating condition in the second historical operating condition data, and a Gaussian mixture model is trained according to the clustering results until convergence to obtain an AVC regulation rate prediction model of the thermal power unit; wherein the AVC regulation rate prediction model is used to predict the AVC regulation rate of the thermal power unit to be tested based on the real-time operating conditions of the thermal power unit to be tested.

[0056] Optionally, the AVC bus voltage set value, bus voltage, generator active power, generator reactive power, terminal voltage, fixed current, rotor current, 6kV bus voltage for plant use and AVC input status are used as the operating condition division conditions, and the AVC regulation rate of each operating condition in the second historical operating condition data and the second historical operating condition data is subjected to K-means clustering, and the number of clusters with the smallest sum of square errors is taken as the optimal number of clusters to obtain the clustering results. Among them, the initial value of k for K-means clustering is 10 to 99. After the analysis of the K-means clustering converges, the Gaussian mixture model is trained according to the K-means clustering results until convergence to obtain the AVC regulation rate prediction model of the thermal power unit.

[0057] Accordingly, the embodiment of the present invention further provides a method for predicting an AVC adjustment rate, comprising steps S1 to S3, wherein:

[0058] Step S1, obtaining the current operating condition data of the thermal power unit to be tested.

[0059] Step S2: inputting the operating condition data into the AVC regulation rate prediction model constructed by the AVC regulation rate prediction model construction method described in steps S101 to S104.

[0060] Step S3, determine the classification of the current working condition by the similarity principle, and obtain the AVC regulation rate of the thermal power unit to be tested according to the classification. The unit AVC is the basis for the plant-level AVC. If there are two units in a plant, if only one unit AVC and the plant-level AVC are put into operation, then steps S1 to S3 predict the regulation rate of one unit. If both the unit AVC and the plant-level AVC are put into operation, the predicted regulation rate is the rate of both units.

[0061] Accordingly, refer to Figure 3 The embodiment of the present invention further provides a device for constructing a prediction model of an AVC adjustment rate, comprising an information table establishment module 101, a verification module 102, a calculation module 103 and a model construction module 104, wherein:

[0062] The information table establishment module 101 is used to obtain the first historical operating condition data of the thermal power unit, and establish a fixed value parameter information table according to the characteristics of the thermal power unit; wherein the fixed value parameter information table includes a plurality of first fixed value parameters; each of the first fixed value parameters includes an upper limit value and a lower limit value.

[0063] The verification module 102 is used to regularly verify the first historical operating condition data according to the upper limit and lower limit of each of the first fixed value parameters, delete the operating condition data that exceeds the limit, and obtain the second historical operating condition data. The second historical operating condition data includes the actual bus voltage value at the first moment of each operating condition and the actual bus voltage value at the second moment of each operating condition; the first moment is the moment when the bus voltage adjustment instruction issued by the AVC master station is received; the second moment is the moment when the bus voltage adjustment instruction issued by the AVC master station is received.

[0064] The calculation module 103 is used to calculate the AVC regulation rate of each operating condition in the second historical operating condition data based on the actual value of the bus voltage at the first moment and the actual value of the bus voltage at the second moment, combined with the time difference between the first moment and the second moment.

[0065] The model building module 104 is used to perform K-means clustering on the second historical operating condition data and the AVC regulation rate of each operating condition in the second historical operating condition data based on preset operating condition division conditions, and train a Gaussian mixture model according to the clustering results until convergence to obtain the AVC regulation rate prediction model of the thermal power unit; wherein the AVC regulation rate prediction model is used to predict the AVC regulation rate of the thermal power unit to be tested based on the real-time operating conditions of the thermal power unit to be tested.

[0066] Accordingly, refer to Figure 4 The embodiment of the present invention further provides an AVC adjustment rate prediction device, comprising: an acquisition module 1, an input module 2 and a prediction module 3; wherein,

[0067] The acquisition module 1 is used to acquire the current operating condition data of the thermal power unit to be tested.

[0068] The input module 2 is used to input the operating condition data into the AVC regulation rate prediction model constructed by the AVC regulation rate prediction model construction method described in steps S101 to S104.

[0069] The prediction module 3 is used to determine the classification of the current operating conditions through the similarity principle, and obtain the AVC regulation rate of the thermal power unit to be tested according to the classification.

[0070] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0071] The embodiment of the present invention provides a prediction model construction method, a prediction method, a prediction model construction device and a prediction device for an AVC regulation rate, wherein the prediction model construction method comprises: obtaining first historical operating condition data of a thermal power unit, and establishing a fixed value parameter information table according to the characteristics of the thermal power unit; wherein the fixed value parameter information table comprises a plurality of first fixed value parameters; each of the first fixed value parameters comprises an upper limit value and a lower limit value; according to the upper limit value and the lower limit value of each of the first fixed value parameters, the first historical operating condition data is regularly checked, the operating condition data exceeding the limit is deleted, and the second historical operating condition data is obtained; wherein the second historical operating condition data comprises the actual value of the bus voltage at the first moment of each operating condition and the actual value of the bus voltage at the second moment of each operating condition; the first moment is the time when the bus voltage is received from the AVC master station. the moment of the bus voltage regulation instruction; the second moment is the moment of receiving the bus voltage regulation instruction issued by the AVC master station; based on the actual value of the bus voltage at the first moment and the actual value of the bus voltage at the second moment, combined with the time difference between the first moment and the second moment, the AVC regulation rate of each operating condition in the second historical operating condition data is calculated; based on the preset operating condition division condition, the second historical operating condition data and the AVC regulation rate of each operating condition in the second historical operating condition data are subjected to K-means clustering, and a Gaussian mixture model is trained according to the clustering results until convergence to obtain the AVC regulation rate prediction model of the thermal power unit; wherein the AVC regulation rate prediction model is used to predict the AVC regulation rate of the thermal power unit to be tested based on the real-time operating condition of the thermal power unit to be tested. Compared with the existing technology, the prediction model of AVC regulation rate is constructed through historical operating conditions, which realizes the online prediction of AVC regulation rate based on the real-time operating conditions of thermal power units, improving convenience and real-time performance; there is no need to obtain different data from power dispatching departments, power plants, test institutes and other units, saving manpower and material costs; the process of model construction and prediction will not introduce disturbances to the grid voltage, improving the accuracy of prediction; K-means clustering is performed on each operating condition based on preset division conditions, which realizes online AVC regulation rate prediction for different types of operating conditions. Compared with the existing technology that can only measure the AVC regulation rate for specific operating conditions, it has universal applicability and versatility for different operating conditions.

[0072] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for constructing a prediction model for AVC regulation rate, characterized in that: include: Acquire first historical operating condition data of the thermal power unit, and establish a fixed value parameter information table according to the characteristics of the thermal power unit; wherein the fixed value parameter information table includes a plurality of first fixed value parameters; each of the first fixed value parameters includes an upper limit value and a lower limit value; According to the upper limit and lower limit of each of the first fixed value parameters, the first historical operating condition data is regularly checked, the operating condition data exceeding the limit is deleted, and the second historical operating condition data is obtained; wherein the second historical operating condition data includes the actual bus voltage value at the first moment of each operating condition and the actual bus voltage value at the second moment of each operating condition; the first moment is the moment when the bus voltage adjustment instruction issued by the AVC master station is received; the second moment is the moment when the bus voltage adjustment instruction issued by the AVC master station is received; Calculate the AVC regulation rate of each operating condition in the second historical operating condition data based on the actual value of the bus voltage at the first moment and the actual value of the bus voltage at the second moment, combined with the time difference between the first moment and the second moment; Based on the preset working condition division conditions, K-means clustering is performed on the second historical working condition data and the AVC regulation rate of each working condition in the second historical working condition data, and a Gaussian mixture model is trained according to the clustering results until convergence, so as to obtain an AVC regulation rate prediction model of the thermal power unit; wherein the AVC regulation rate prediction model is used to predict the AVC regulation rate of the thermal power unit to be tested based on the real-time operating condition of the thermal power unit to be tested; The K-means clustering is performed on the second historical operating condition data and the AVC adjustment rate of each operating condition in the second historical operating condition data based on the preset operating condition division condition, specifically: Taking the AVC bus voltage set value, bus voltage, generator active power, generator reactive power, machine terminal voltage, fixed current, rotor current, plant 6kV bus voltage and AVC input status as the operating condition division conditions, K-means clustering is performed on the second historical operating condition data and the AVC regulation rate of each operating condition in the second historical operating condition data, and the number of clusters with the smallest sum of squared errors is taken as the optimal number of clusters to obtain the clustering results; wherein, the initial value of k for K-means clustering is 10 to 99.

2. A method for constructing a prediction model of an AVC regulation rate according to claim 1, characterized in that: The determination conditions for the end of the AVC adjustment are specifically: When the actual bus voltage value is higher than the first preset value or lower than the second preset value for 120 seconds, it is judged that the condition for ending the AVC regulation is met; wherein, the first preset value is the bus voltage target value corresponding to the bus voltage regulation instruction plus the dead zone value; the second preset value is the bus voltage target value minus the dead zone value; the moment when the AVC regulation ends is 150 seconds before it is judged that the judgment condition is met.

3. The method for constructing a prediction model of an AVC regulation rate according to claim 1, characterized in that: According to the upper limit and lower limit of each of the first fixed value parameters, the first historical operating condition data is regularly checked, and the operating condition data exceeding the limit is deleted to obtain the second historical operating condition data, specifically: Each historical operating condition in the first historical operating condition data includes a plurality of operating condition parameters, and the operating condition parameters correspond one-to-one to the fixed value parameters; The first historical operating condition data is checked regularly, and the operating condition data in which any operating condition parameter is greater than the upper limit value of the corresponding fixed value parameter or less than the lower limit value of the corresponding fixed value parameter is deleted to obtain the second historical operating condition data.

4. A method for constructing a prediction model of an AVC regulation rate as claimed in claim 3, characterized in that: Before the first historical operating condition data is regularly checked, the method further includes: performing upper and lower limit checks on the first constant value parameters according to empirical values, adjusting the upper and lower limit values ​​of each of the first constant value parameters, and obtaining the first constant value parameters that have been checked for upper and lower limit values.

5. A method for constructing a prediction model of an AVC regulation rate as claimed in claim 4, characterized in that: Before the first historical operating condition data is regularly checked, after the upper and lower limit checks are performed on the first fixed value parameters according to the empirical values, the method further includes: eliminating the historical operating condition data that has not been put into AVC to obtain the first historical operating condition data that has been preliminarily screened.

6. A method for constructing a prediction model for AVC regulation rate according to claim 5, characterized in that: Before the removal of the historical operating condition data that has not been put into AVC, it also includes: Recording the time period in the first historical operating condition data that meets the alarm condition; wherein the alarm condition is: any operating condition parameter is greater than the maximum value of the historical operation or less than the minimum value of the historical operation; or any operating condition parameter is greater than the maximum value of the empirical value or less than the minimum value of the empirical value; After the historical operating condition data without AVC is eliminated, the following is also included: The time period that meets the alarm condition is identified, and the corresponding operating condition data is recorded.

7. A method for predicting AVC adjustment rate, characterized in that: include: Obtain the current operating condition data of the thermal power unit to be tested; Inputting the operating condition data into the AVC regulation rate prediction model constructed by the AVC regulation rate prediction model construction method according to any one of claims 1 to 6; The classification of the current operating condition is determined by the similarity principle, and the AVC regulation rate of the thermal power unit to be tested is obtained according to the classification.

8. A prediction model construction device for AVC adjustment rate, characterized in that: It includes information table establishment module, verification module, calculation module and model construction module, among which, The information table establishment module is used to obtain the first historical operating condition data of the thermal power unit, and establish a fixed value parameter information table according to the characteristics of the thermal power unit; wherein the fixed value parameter information table includes a plurality of first fixed value parameters; each of the first fixed value parameters includes an upper limit value and a lower limit value; The verification module is used to regularly verify the first historical operating condition data according to the upper limit value and the lower limit value of each first fixed value parameter, delete the operating condition data that exceeds the limit, and obtain the second historical operating condition data; wherein the second historical operating condition data includes the actual value of the bus voltage at the first moment of each operating condition and the actual value of the bus voltage at the second moment of each operating condition; the first moment is the moment when the bus voltage adjustment instruction issued by the AVC master station is received; the second moment is the moment when the bus voltage adjustment instruction issued by the AVC master station is received; The calculation module is used to calculate the AVC regulation rate of each operating condition in the second historical operating condition data based on the actual value of the bus voltage at the first moment and the actual value of the bus voltage at the second moment, combined with the time difference between the first moment and the second moment; The model building module is used to perform K-means clustering on the second historical operating condition data and the AVC regulation rate of each operating condition in the second historical operating condition data based on a preset operating condition division condition, and train a Gaussian mixture model according to the clustering result until convergence, so as to obtain an AVC regulation rate prediction model of the thermal power unit; wherein the AVC regulation rate prediction model is used to predict the AVC regulation rate of the thermal power unit to be tested based on the real-time operating condition of the thermal power unit to be tested; The model building module performs K-means clustering on the second historical operating condition data and the AVC adjustment rate of each operating condition in the second historical operating condition data based on the preset operating condition division condition, specifically: The model building module uses the AVC bus voltage set value, bus voltage, generator active power, generator reactive power, terminal voltage, fixed current, rotor current, plant 6kV bus voltage and AVC input status as operating condition division conditions, performs K-means clustering on the second historical operating condition data and the AVC regulation rate of each operating condition in the second historical operating condition data, and takes the number of clusters with the smallest sum of squared errors as the optimal number of clusters to obtain clustering results; wherein the initial value of k for K-means clustering is 10 to 99.

9. A device for predicting AVC adjustment rate, characterized in that: include: Acquisition module, input module and prediction module; among them, The acquisition module is used to acquire the current operating condition data of the thermal power unit to be tested; The input module is used to input the operating condition data into the AVC regulation rate prediction model constructed by the AVC regulation rate prediction model construction method according to any one of claims 1 to 6; The prediction module is used to determine the classification of the current operating conditions through the similarity principle, and obtain the AVC regulation rate of the thermal power unit to be tested according to the classification.

Images