A power plant feedwater flow fluctuation processing method and related equipment

Abstract

The present application relates to the technical field of power plant SIS system, in particular to a power plant feedwater flow fluctuation processing method and related equipment, the method comprises collecting power plant feedwater flow original data and preprocessing, obtaining preprocessed flow time series data set; based on the preprocessed flow time series data set and real-time unit working condition parameters, the size of the sliding window is dynamically determined; based on the size of the sliding window, the preprocessed flow time series data set is calculated and corrected in time series median, the corrected final time series median is obtained, and it is output for power plant unit performance calculation. The present application solves the technical problems that the existing power plant feedwater flow fluctuation processing method has poor adaptability, limited filtering effect, and leads to inaccurate SIS system performance calculation.

Description

Technical Field

[0001] This invention relates to the field of power plant SIS system technology, specifically to a method and related equipment for handling power plant feedwater flow fluctuations. Background Technology

[0002] During the operation of power plant units, feedwater flow rate is a core parameter affecting the calculation of unit operating performance. Its data stability directly determines the accuracy of key performance assessments such as turbine heat consumption calculation and unit performance analysis. Accurate feedwater flow rate data provides crucial basis for unit operation optimization and energy consumption control; conversely, inaccurate data can lead to performance assessment deviations and affect the formulation of unit operation strategies. However, in actual operation, feedwater flow rate is easily affected by multiple factors, resulting in drastic fluctuations. Specifically, this manifests in two ways: First, changes in unit operating conditions can trigger sudden changes in feedwater demand, leading to flow fluctuations. For example, during rapid load increases, feedwater demand increases sharply with the load growth, causing instantaneous changes in pressure and flow rate within the feedwater pipeline, directly resulting in feedwater flow instability. Second, non-operating-condition factors such as pressure pulsations within the feedwater pipeline, electromagnetic interference from surrounding electrical equipment affecting the flow sensor, and sensor measurement errors can cause instantaneous peaks or troughs in the collected flow data, forming "false fluctuations" that do not match the actual feedwater conditions. For instance, when the sensor is in a strong electromagnetic environment, it may output abnormally high or low values, significantly deviating from the true feedwater flow rate.

[0003] Currently, the methods for handling feedwater flow fluctuations in the industry are mainly divided into two categories, but both have obvious limitations: one is the fixed-window mean filtering method, which smooths the data by calculating the average flow rate within a fixed time window. This method not only has poor filtering effect on instantaneous spike fluctuations, but the fixed time window cannot adapt to the dynamic changes in unit operating conditions. When operating conditions are rapidly adjusted, the fixed window cannot reflect the actual trend of flow changes in a timely manner, which can easily lead to lag or over-filtering problems in the smoothed data, making it impossible to accurately represent the current feedwater flow status and thus affecting the accuracy of performance calculations. The other method is to perform simple moving average processing on the flow data. This method also has the defects of poor handling of instantaneous fluctuations and inability to adapt to changes in operating conditions, making it difficult to meet the flow data processing needs under varying operating conditions of the unit.

[0004] For example, in the actual operation of a 300MW thermal power unit, during the load change process from 240MW to 180MW, the feedwater flow data was processed using the mean filtering method over a 60-second statistical period. The results showed that the flow fluctuation was still 15%–20%. The turbine heat rate calculation based on this fluctuation data had significant errors, failing to provide reliable guidance for optimized unit operation and potentially even misleading operational adjustment strategies. In summary, existing feedwater flow processing methods cannot simultaneously address operating condition adaptability, fluctuation filtering accuracy, and response timeliness, making it difficult to meet the high-precision requirements of power plant SIS (System-on-System) feedwater flow data. Therefore, there is an urgent need for a power plant feedwater flow processing method that can adapt to dynamic changes in unit operating conditions, accurately filter real and spurious fluctuations, and respond rapidly. This would improve the accuracy of SIS system performance calculations and provide support for efficient unit operation and energy consumption optimization. Summary of the Invention

[0005] In order to overcome the defects of the prior art, the present invention aims to provide a method and related equipment for handling power plant feedwater flow fluctuations, so as to solve the technical problems of poor adaptability and limited filtration effect of the existing power plant feedwater flow fluctuation handling methods, which leads to inaccurate SIS system performance calculation.

[0006] This invention is achieved through the following technical solution: In a first aspect, the present invention provides a method for handling fluctuations in power plant feedwater flow, comprising: Raw data of power plant feedwater flow rate was collected and preprocessed to obtain a preprocessed flow time series dataset. Based on the preprocessed flow time series dataset and real-time unit operating parameters, the size of the sliding window is dynamically determined. Based on the sliding window size, the median and trend correction are performed on the preprocessed flow time series dataset to obtain the corrected final median, which is then output for power plant unit performance calculation.

[0007] Preferably, the process of collecting and preprocessing raw data on power plant feedwater flow includes: Raw water flow data is collected at 1-second intervals through the power plant's SIS system data interface to form a raw flow time series dataset. Outliers in the original traffic time series dataset are identified using the quartile method. The first quartile Q1, the third quartile Q3, and the interquartile range IQR = Q3 - Q1 are calculated. Data that exceed the range [Q1 - 1.5 × IQR, Q3 + 1.5 × IQR] are identified as outliers. Outliers are replaced by the mean of the two valid data points before and after them. If there are ≤3 consecutive missing data points, the valid data from the previous time step is used to fill the gaps. If there are >3 consecutive missing data points, the three valid data points before and after the missing data point are used to fill the gaps using linear interpolation, resulting in the preprocessed traffic time series dataset.

[0008] Furthermore, when replacing outliers, if there are fewer than two valid data points before and after the outlier, the average of all available valid data points is used for replacement; when filling missing values, if the last three valid data points of the missing segment cannot be obtained, the first three valid data points of the missing segment are used for linear extrapolation filling; if the first three valid data points of the missing segment cannot be obtained, the last three valid data points of the missing segment are used for linear extrapolation filling.

[0009] Preferably, the specific process of dynamically determining the sliding window size based on the preprocessed flow time-series dataset and real-time unit operating parameters includes: The unit load is collected as real-time operating parameters through the power plant's SIS system at intervals of 1 second. The unit operating conditions are divided into stable operating conditions, variable load operating conditions, and low load / start-stop operating conditions. The stable operating condition is when the unit load is 80% to 100% of the rated load, the variable load condition is when the unit load is 50% to 80% of the rated load, and the low load / start-stop condition is when the unit load is less than 50% of the rated load or is in the process of starting or stopping the unit. Under stable operating conditions, the sliding window size is set to N=10; under variable load conditions, N=6; and under low load / start-stop conditions, N=4. The current operating condition is compared with the interval threshold in real time, and the window switching response time is ≤1s to obtain the dynamic sliding window size.

[0010] The preferred method for calculating the time series median is as follows: Based on the sliding window size, the preprocessed traffic time series dataset is traversed by sliding, with the window moving backward by 1 data point each time. The N data points in each window are sorted in ascending order. If N is odd, the data in the middle position after sorting is taken as the basic time series median. If N is even, the average of the two data points in the middle position after sorting is taken as the basic time series median.

[0011] The preferred trend correction process is as follows: Take the median of all basic time series within the last 5 minutes, and assign time weights to each data point. The weight coefficient for data within the last 1 minute is 1.0, the weight coefficient for data within 1 to 2 minutes is 0.8, the weight coefficient for data within 2 to 3 minutes is 0.6, and the weight coefficient for data within 3 to 5 minutes is 0.4. With time as the independent variable and the median of the basic time series as the dependent variable, a weighted linear regression model Q=kt+b is established, where k is the trend slope and b is the intercept, and the trend term is obtained by solving the model. The base time series median for each window is superimposed with the trend term at the corresponding time point to obtain the corrected final time series median.

[0012] Preferably, the output and performance calculations specifically include: The corrected final time series median is input into the power plant's SIS system for unit thermal efficiency calculation, turbine heat consumption calculation, and energy consumption analysis. At the same time, the effect is verified by calculating the difference between the maximum and minimum values ​​of the smoothed flow data within 5 minutes. If the fluctuation range is greater than 5% of the rated flow for 5 consecutive minutes, an alarm signal is triggered. If the flow data and steam parameter changes are consistent with the normal operation logic of the unit and the fluctuation of the performance calculation results is within a reasonable range, the processing effect is judged to be up to standard.

[0013] Secondly, the present invention also provides a power plant feedwater flow fluctuation processing system, comprising: The data preprocessing module is used to collect raw data of power plant feedwater flow and preprocess it to obtain a preprocessed flow time series dataset. The sliding window determination module is used to dynamically determine the size of the sliding window based on the preprocessed flow time series dataset and real-time unit operating parameters; The performance calculation module is used to calculate the time-series median and correct the trend of the preprocessed flow time-series dataset based on the sliding window size, obtain the corrected final time-series median, and output it for power plant unit performance calculation.

[0014] Thirdly, the present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the power plant feedwater flow fluctuation processing method described above.

[0015] Fourthly, the present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the power plant feedwater flow fluctuation processing method described above.

[0016] Compared with the prior art, the present invention has the following beneficial technical effects: This invention provides a method for processing power plant feedwater flow fluctuations. It collects raw feedwater flow data from the power plant and performs targeted preprocessing. First, it uses the quartile method to accurately identify and replace instantaneous peaks, troughs, and other outliers. Then, it flexibly uses forward filling or linear interpolation to fill in missing data based on the duration of the missing data, effectively eliminating false fluctuations and data breaks in the raw data. This lays a high-quality data foundation for subsequent fluctuation processing and avoids interference from abnormal or missing data in subsequent calculations. Simultaneously, it dynamically adjusts the sliding window size based on real-time unit load conditions. Under stable conditions, a larger window is used to filter minor disturbances; under varying load conditions, a medium window is used to balance smoothing effect and response speed; and under low load / start-stop conditions, a smaller window is used to avoid data lag. This effectively solves the problem of poor adaptability of existing fixed-window mean filtering or simple moving average methods, achieving real-time matching of window parameters and flow fluctuation characteristics under different operating conditions. The window switching response time is ≤1s, ensuring rapid adaptation to changes in operating conditions. Furthermore, it accurately filters instantaneous fluctuations by calculating the median of the sliding window time series. Compared to mean filtering, which is susceptible to extreme values, median filtering is more resistant to outliers and can significantly reduce the interference of instantaneous peaks / troughs on data smoothing. Combined with weighted linear regression to extract long-term trends and correct the median, it avoids underestimating or overestimating long-term slow trends due to reliance on local medians, ensuring that the processed data is both stable and closely reflects the actual variation of feedwater flow. Finally, the corrected final time-series median is output to the SIS system, providing stable and accurate flow data input for performance evaluations such as unit thermal efficiency calculation and turbine heat rate calculation. This effectively overcomes the performance calculation deviations in the SIS system caused by the limited filtering effect and poor adaptability of traditional methods, significantly improving the accuracy and reliability of performance calculation results. Simultaneously, by monitoring the flow fluctuation amplitude and logical matching with other parameters in real time to verify the effect, it can promptly warn of sensor failures or abnormal operating conditions, further ensuring the scientific nature of unit operation performance analysis and optimization strategy formulation, and providing reliable data support for efficient power plant unit operation and energy consumption control. Attached Figure Description

[0017] Figure 1 This is a flowchart of the power plant feedwater flow fluctuation handling method in an embodiment of the present invention; Figure 2 This is a framework diagram of the power plant feedwater flow fluctuation handling method in an embodiment of the present invention; Figure 3 This is a schematic diagram of the power plant feedwater flow fluctuation treatment system in an embodiment of the present invention; In the diagram: 1. Data preprocessing module; 2. Sliding window determination module; 3. Performance calculation module. Detailed Implementation

[0018] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0019] The present invention will now be described in further detail with reference to the accompanying drawings: The purpose of this invention is to provide a method and related equipment for handling power plant feedwater flow fluctuations, so as to solve the technical problems of poor adaptability and limited filtration effect of existing power plant feedwater flow fluctuation handling methods, which leads to inaccurate SIS system performance calculations.

[0020] See Figure 1 and Figure 2 In one embodiment of the present invention, a method for handling power plant feedwater flow fluctuations is provided, comprising: Step 1: Collect raw data of power plant feedwater flow and preprocess it to obtain a preprocessed flow time series dataset; Specifically, the process of collecting and preprocessing raw data on power plant feedwater flow includes: Raw water flow data is collected at 1-second intervals through the power plant's SIS system data interface to form a raw flow time series dataset. Outliers in the original traffic time series dataset are identified using the quartile method. The first quartile Q1, the third quartile Q3, and the interquartile range IQR = Q3 - Q1 are calculated. Data that exceed the range [Q1 - 1.5 × IQR, Q3 + 1.5 × IQR] are identified as outliers (instantaneous peak data and trough data). Outliers are replaced by the mean of the two valid data points before and after them.

[0021] For example, if the 10th data point is an outlier, and the 8th, 9th, 11th, and 12th data points are valid, then the 10th data point is replaced with the value of (8th + 9th + 11th + 12th) / 4.

[0022] If there are ≤3 consecutive missing data points (corresponding to a missing time ≤3s), the valid data from the previous time step is used to fill the gaps. If there are >3 consecutive missing data points (corresponding to a missing time >3s), the first 3 valid data points of the missing segment (if the missing segment starts from the 20th point, then the 17th, 18th, and 19th points are used) and the last 3 valid data points of the missing segment (if the missing segment ends at the 30th point, then the 31st, 32nd, and 33rd points are used) are used for linear interpolation to fill the gaps, thus obtaining the preprocessed traffic time series dataset.

[0023] In the outlier replacement process, if there are fewer than two valid data points before and after the outlier data point, the average of all available valid data points is used for replacement. In the missing value filling process, if the last three valid data points of the missing segment cannot be obtained, the first three valid data points of the missing segment are used for linear extrapolation filling. If the first three valid data points of the missing segment cannot be obtained, the last three valid data points of the missing segment are used for linear extrapolation filling.

[0024] Step 2: Based on the preprocessed flow time series dataset and real-time unit operating parameters, dynamically determine the sliding window size; Specifically, the specific process includes: The unit load is collected as real-time operating parameters through the power plant's SIS system at intervals of 1 second. The unit operating conditions are divided into stable operating conditions, variable load operating conditions, and low load / start-stop operating conditions. The stable operating condition is when the unit load is 80% to 100% of the rated load (e.g., 240 to 300MW for a 300MW unit). The variable load operating condition is when the unit load is 50% to 80% of the rated load (e.g., 150 to 240MW for a 300MW unit). The low load / start-stop operating condition is when the unit load is less than 50% of the rated load or when the unit is in the process of starting or stopping. Under stable operating conditions, the sliding window size is set to N=10 (corresponding to 10 seconds of traffic data coverage), under variable load conditions, N=6, and under low load / start-stop conditions, N=4 (corresponding to 4 seconds of traffic data coverage). The current operating conditions are compared with the interval threshold in real time, and the window switching response time is ≤1 second to obtain the dynamic sliding window size.

[0025] Step 3: Calculate the time-series median and correct the trend of the preprocessed flow time-series dataset based on the sliding window size to obtain the corrected final time-series median, and output it for power plant unit performance calculation.

[0026] Specifically, the process of calculating the time series median is as follows: Based on the sliding window size, the preprocessed traffic time series dataset is traversed by sliding, with the window moving backward by 1 data point each time. The N data points in each window are sorted in ascending order. If N is odd, the data in the middle position after sorting is taken as the basic time series median. If N is even, the average of the two data points in the middle position after sorting is taken as the basic time series median.

[0027] In this embodiment, based on the window size N determined in step 2, a sliding window traversal is performed on the preprocessed traffic data (the window moves backward by one data point each time). The N data points within each window are sorted in ascending order: 1) If N is odd (e.g., N=9), take the 5th data point after sorting as the base time series median; 2) If N is even (e.g., N=10), take the average of the 5th and 6th data after sorting as the basic time series median.

[0028] For example, if the preprocessed traffic data in a certain window is [85,88,86,92,87,89,88,90,87,88] (unit: t / h), and after sorting it is [85,86,87,87,88,88,88,89,90,92], and N=10 (even number), then the median of the basic time series = (88+88) / 2 = 88t / h.

[0029] Specifically, the process of trend correction is as follows: Take the median of all basic time series within the last 5 minutes, and assign time weights to each data point. The weight coefficient for data within the last 1 minute is 1.0, the weight coefficient for data within 1 to 2 minutes is 0.8, the weight coefficient for data within 2 to 3 minutes is 0.6, and the weight coefficient for data within 3 to 5 minutes is 0.4. With time as the independent variable and the median of the basic time series as the dependent variable, a weighted linear regression model Q=kt+b is established, where k is the trend slope and b is the intercept, and the trend term is obtained by solving the model. The base time series median for each window is superimposed with the trend term at the corresponding time point to obtain the corrected final time series median.

[0030] In this embodiment, the trend term is extracted using weighted linear regression: 1) Take the median of all basic time series within the last 5 minutes (300 data points in 5 minutes), and assign a time weight to each data point: the weight coefficient of the 60 data points within the last 1 minute is 1.0, the weight coefficient of the 60 data points within 1 to 2 minutes is 0.8, the weight coefficient of the 60 data points within 2 to 3 minutes is 0.6, and the weight coefficient of the 120 data points within 3 to 5 minutes is 0.4. The weight decreases linearly with time. 2) Using time as the independent variable (t, unit: s) and the median of the basic time series as the dependent variable (Q), establish a weighted linear regression model Q=kt+b (k is the trend slope and b is the intercept), and solve for the trend term; 3) Overlay the base time series median of each window with the trend term at the corresponding time to obtain the corrected final time series median. For example: if the base median of a certain window is 88t / h and the trend term at the corresponding time is +0.2t / h, then the final median = 88 + 0.2 = 88.2t / h.

[0031] Specifically, the output and performance calculations include: The corrected final time-series median is input into the power plant's SIS system for unit thermal efficiency calculation, turbine heat consumption calculation, and energy consumption analysis. For example, when calculating turbine heat consumption, the smoothed feedwater flow rate data is combined with steam parameters and other data, and the calculation is performed according to the turbine heat consumption calculation formula to obtain a more accurate turbine heat consumption value. Simultaneously, the effect is verified by calculating the difference between the maximum and minimum values ​​of the smoothed flow rate data within 5 minutes. If the fluctuation range is greater than 5% of the rated flow rate for 5 consecutive minutes (e.g., when the rated flow rate is 1000 t / h, the fluctuation range is greater than 50 t / h), an alarm signal is triggered. If the changes in flow rate data and steam parameters conform to the normal operating logic of the unit and the fluctuation of the performance calculation results is within a reasonable range, the processing effect is deemed satisfactory.

[0032] Taking a thermal power unit as an example, the specific implementation process of the processing method of this invention is as follows: Data collection The SIS system interface collects feedwater flow rate at 1-second intervals. The raw flow rate data exhibits instantaneous spikes at a load of 240MW (steady-state condition). For example, at a certain moment, due to slight electromagnetic interference to the sensor, the collected flow rate data suddenly rises to a value far exceeding the normal range, forming a noticeable spike.

[0033] Data preprocessing Outliers are removed using the quartile method. For detected spikes, the mean of the valid data before and after them is used for replacement. If consecutive missing data exist, forward padding or linear interpolation is used to fill the gaps, depending on the number of missing data points, to ensure data integrity and continuity. For example, if two consecutive data points are missing, data from the previous valid time point is used for padding; if four consecutive data points are missing, linear interpolation is used to fill the gaps using three valid data points before and after each missing segment.

[0034] Dynamic window confirmation The power plant's SIS system collects operating parameters such as unit load, steam pressure, and feedwater temperature in real time. When the unit load is stable between 240 and 300 MW (80% to 100% of rated load), steam pressure fluctuation is ≤ ±0.2 MPa, and feedwater temperature fluctuation is ≤ ±3℃, it is considered a stable operating condition, and the sliding window size N=10 is set. If the unit load begins to decrease and enters the variable load range, the data processing unit detects the change in operating conditions in real time and switches the window size to N=6 within 1 second.

[0035] Time series median calculation and trend correction For the preprocessed traffic data, a sliding window is used for traversal according to a defined window size N. Within each window, the data is sorted and the baseline time-series median is calculated. For example, within a window with N=10, the average of the 5th and 6th data points after sorting is taken as the baseline time-series median. Then, a weighted linear regression method is used to extract the trend term of the baseline time-series median over the past 5 minutes. The trend term is then superimposed on the baseline time-series median to obtain the corrected final time-series median. For example, if the baseline median for a certain window is 90 t / h, and the corresponding trend term is +0.3 t / h, then the final median = 90 + 0.3 = 90.3 t / h.

[0036] Performance calculation data output and effect verification: The smoothed final time-series median is used as an input parameter for the SIS system, employed in performance calculations such as turbine heat rate calculation. The monitoring and alarm module collects real-time data on the fluctuation range of the smoothed flow rate and changes in other parameters related to performance calculations. If the flow rate fluctuation is ≤5% of the rated flow rate within 5 consecutive minutes, other parameter changes conform to the normal operating logic of the unit, and the performance calculation results are stable, the treatment effect is deemed satisfactory, and the system is operating stably. Actual operation verification shows that this method effectively reduces feedwater flow fluctuations, improves the accuracy of SIS system performance calculations, and provides a more reliable basis for optimized unit operation.

[0037] In summary, this embodiment also provides a method for processing power plant feedwater flow fluctuations. By collecting raw power plant feedwater flow data and performing targeted preprocessing, it first uses the quartile method to accurately identify and replace instantaneous peaks, troughs, and other outliers. Then, based on the duration of missing data, it flexibly employs forward filling or linear interpolation to fill in the gaps, effectively eliminating false fluctuations and data breaks in the raw data. This lays a high-quality data foundation for subsequent fluctuation processing and avoids interference from abnormal or missing data in subsequent calculations. Simultaneously, it dynamically adjusts the sliding window size based on real-time unit load conditions. Under stable conditions, a larger window filters minor disturbances; under varying load conditions, a medium window balances smoothing effects and response speed; and under low load / start-stop conditions, a smaller window avoids data lag. This effectively solves the problem of poor adaptability of existing fixed-window mean filtering or simple moving average methods, achieving real-time matching of window parameters and flow fluctuation characteristics under different operating conditions. The window switching response time is ≤1s, ensuring rapid adaptation to changes in operating conditions. Furthermore, it accurately filters out instantaneous fluctuations by calculating the median of the sliding window time series. Compared to mean filtering, which is susceptible to extreme values, the median is more resistant to outliers, significantly reducing the interference of instantaneous peaks / troughs on data smoothing. Combined with weighted linear regression to extract long-term trends and correct the median, it avoids underestimating or overestimating long-term slow trends due to reliance on local medians, ensuring that the processed data is both stable and closely reflects the actual variation of feedwater flow. Finally, the corrected final time-series median is output to the SIS system, providing stable and accurate flow data input for performance evaluations such as unit thermal efficiency calculation and turbine heat consumption calculation. This effectively overcomes the performance calculation deviations in the SIS system caused by the limited filtering effect and poor adaptability of traditional methods, significantly improving the accuracy and reliability of performance calculation results. Simultaneously, by monitoring the flow fluctuation amplitude and logical matching with other parameters in real time to verify the effect, it can promptly warn of sensor failures or abnormal operating conditions, further ensuring the scientific nature of unit operation performance analysis and optimization strategy formulation, and providing reliable data support for efficient power plant unit operation and energy consumption control.

[0038] Example 2 according to Figure 3 As shown, this embodiment also provides a power plant feedwater flow fluctuation processing system, including: Data preprocessing module 1 is used to collect raw data of power plant feedwater flow and preprocess it to obtain a preprocessed flow time series dataset. The sliding window determination module 2 is used to dynamically determine the size of the sliding window based on the preprocessed flow time series dataset and real-time unit operating parameters; The performance calculation module 3 is used to calculate the time-series median and correct the trend of the preprocessed flow time-series dataset based on the sliding window size, to obtain the corrected final time-series median, and output it for power plant unit performance calculation.

[0039] Example 3 The present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, such as a power plant feedwater flow fluctuation processing program.

[0040] When the processor executes the computer program, it implements the above-mentioned method for handling power plant feedwater flow fluctuations, for example: Raw data of power plant feedwater flow rate was collected and preprocessed to obtain a preprocessed flow time series dataset. Based on the preprocessed flow time series dataset and real-time unit operating parameters, the size of the sliding window is dynamically determined. Based on the sliding window size, the median and trend correction are performed on the preprocessed flow time series dataset to obtain the corrected final median, which is then output for power plant unit performance calculation.

[0041] Alternatively, when the processor executes the computer program, it implements the functions of each module in the above system, for example: Data preprocessing module 1 is used to collect raw data of power plant feedwater flow and preprocess it to obtain a preprocessed flow time series dataset. The sliding window determination module 2 is used to dynamically determine the size of the sliding window based on the preprocessed flow time series dataset and real-time unit operating parameters; The performance calculation module 3 is used to calculate the time-series median and correct the trend of the preprocessed flow time-series dataset based on the sliding window size, to obtain the corrected final time-series median, and output it for power plant unit performance calculation.

[0042] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the mobile terminal.

[0043] For example, the computer program can be divided into a data preprocessing module 1, a sliding window determination module 2, and a performance calculation module 3; The specific functions of each module are as follows: Data preprocessing module 1 is used to collect raw data of power plant feedwater flow and preprocess it to obtain a preprocessed flow time series dataset. The sliding window determination module 2 is used to dynamically determine the size of the sliding window based on the preprocessed flow time series dataset and real-time unit operating parameters; The performance calculation module 3 is used to calculate the time-series median and correct the trend of the preprocessed flow time-series dataset based on the sliding window size, to obtain the corrected final time-series median, and output it for power plant unit performance calculation.

[0044] The mobile terminal can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The mobile terminal may include, but is not limited to, a processor and memory.

[0045] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the mobile terminal, connecting various parts of the mobile terminal via various interfaces and lines.

[0046] The memory can be used to store the computer program and / or module. The processor implements various functions of the mobile terminal by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0047] The memory may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function (such as sound playback, image playback, etc.); the data storage area may store data created based on the use of the mobile phone (such as audio data, phonebook, etc.). Furthermore, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disks, RAM, plug-in hard disks, SmartMediaCards (SMC), Secure Digital (SD) cards, FlashCards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0048] Example 4 The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for handling fluctuations in power plant feedwater flow.

[0049] If the modules / units integrated in the mobile terminal are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0050] Based on this understanding, all or part of the processes in the above method can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the above-described aggregated reinforcement learning resource scheduling method. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate form.

[0051] The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0052] It should be noted that the content contained in the computer-readable medium may be appropriately added to or subtracted from the content as required by the legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium may not include electrical carrier signals and telecommunication signals.

[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for handling fluctuations in power plant feedwater flow, characterized in that, include: Raw data of power plant feedwater flow rate was collected and preprocessed to obtain a preprocessed flow time series dataset. Based on the preprocessed flow time series dataset and real-time unit operating parameters, the size of the sliding window is dynamically determined. Based on the sliding window size, the median and trend correction are performed on the preprocessed flow time series dataset to obtain the corrected final median, which is then output for power plant unit performance calculation.

2. The method for handling power plant feedwater flow fluctuations according to claim 1, characterized in that, The specific process of collecting and preprocessing the raw data of power plant feedwater flow includes: Raw water flow data is collected at 1-second intervals through the power plant's SIS system data interface to form a raw flow time series dataset. Outliers in the original traffic time series dataset are identified using the quartile method. The first quartile Q1, the third quartile Q3, and the interquartile range IQR = Q3 - Q1 are calculated. Data that exceed the range [Q1 - 1.5 × IQR, Q3 + 1.5 × IQR] are identified as outliers. Outliers are replaced by the mean of the two valid data points before and after them. If there are ≤3 consecutive missing data points, the valid data from the previous time step is used to fill the gaps. If there are >3 consecutive missing data points, the three valid data points before and after the missing data point are used to fill the gaps using linear interpolation, resulting in the preprocessed traffic time series dataset.

3. The method for handling power plant feedwater flow fluctuations according to claim 2, characterized in that, When replacing outliers, if there are fewer than two valid data points before and after the outlier, the average of all available valid data points is used for replacement. When filling missing values, if the last three valid data points of the missing segment cannot be obtained, the first three valid data points of the missing segment are used for linear extrapolation filling. If the first three valid data points of the missing segment cannot be obtained, the last three valid data points of the missing segment are used for linear extrapolation filling.

4. The method for handling power plant feedwater flow fluctuations according to claim 1, characterized in that, The specific process of dynamically determining the sliding window size based on the preprocessed flow time-series dataset and real-time unit operating parameters includes: The unit load is collected as real-time operating parameters through the power plant's SIS system at intervals of 1 second. The unit operating conditions are divided into stable operating conditions, variable load operating conditions, and low load / start-stop operating conditions. The stable operating condition is when the unit load is 80% to 100% of the rated load, the variable load condition is when the unit load is 50% to 80% of the rated load, and the low load / start-stop condition is when the unit load is less than 50% of the rated load or is in the process of starting or stopping the unit. Under stable operating conditions, the sliding window size is set to N=10; under variable load conditions, N=6; and under low load / start-stop conditions, N=4. The current operating condition is compared with the interval threshold in real time, and the window switching response time is ≤1s to obtain the dynamic sliding window size.

5. The method for handling power plant feedwater flow fluctuations according to claim 1, characterized in that, The specific process for calculating the time series median is as follows: Based on the sliding window size, the preprocessed traffic time series dataset is traversed by sliding, with the window moving backward by 1 data point each time. The N data points in each window are sorted in ascending order. If N is odd, the data in the middle position after sorting is taken as the basic time series median. If N is even, the average of the two data points in the middle position after sorting is taken as the basic time series median.

6. The method for handling power plant feedwater flow fluctuations according to claim 1, characterized in that, The specific process of trend correction is as follows: Take the median of all basic time series within the last 5 minutes, and assign time weights to each data point. The weight coefficient for data within the last 1 minute is 1.0, the weight coefficient for data within 1 to 2 minutes is 0.8, the weight coefficient for data within 2 to 3 minutes is 0.6, and the weight coefficient for data within 3 to 5 minutes is 0.

4. With time as the independent variable and the median of the basic time series as the dependent variable, a weighted linear regression model Q=kt+b is established, where k is the trend slope and b is the intercept, and the trend term is obtained by solving the model. The base time series median for each window is superimposed with the trend term at the corresponding time point to obtain the corrected final time series median.

7. The method for handling power plant feedwater flow fluctuations according to claim 1, characterized in that, The output and performance calculations specifically include: The corrected final time series median is input into the power plant's SIS system for unit thermal efficiency calculation, turbine heat consumption calculation, and energy consumption analysis. At the same time, the effect is verified by calculating the difference between the maximum and minimum values ​​of the smoothed flow data within 5 minutes. If the fluctuation range is greater than 5% of the rated flow for 5 consecutive minutes, an alarm signal is triggered. If the flow data and steam parameter changes are consistent with the normal operation logic of the unit and the fluctuation of the performance calculation results is within a reasonable range, the processing effect is judged to be up to standard.

8. A power plant feedwater flow fluctuation handling system, characterized in that, include: The data preprocessing module is used to collect raw data of power plant feedwater flow and preprocess it to obtain a preprocessed flow time series dataset. The sliding window determination module is used to dynamically determine the size of the sliding window based on the preprocessed flow time series dataset and real-time unit operating parameters; The performance calculation module is used to calculate the time-series median and correct the trend of the preprocessed flow time-series dataset based on the sliding window size, obtain the corrected final time-series median, and output it for power plant unit performance calculation.

9. A mobile terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the power plant feedwater flow fluctuation processing method as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the power plant feedwater flow fluctuation processing method as described in any one of claims 1-7.

Images