An intelligent regulation method for main steam temperature of a thermal power plant based on artificial intelligence

Abstract

The application discloses an intelligent regulation and control method for main steam temperature of a thermal power plant based on artificial intelligence, and relates to the technical field of regulation and control of thermal power plants; the main steam temperature is effectively regulated and controlled to be in an optimal temperature state; the application deploys operation measuring points of a boiler of the thermal power plant, collects a boiler operation data set in real time, sets a dynamic operation change model, extracts current working condition feature data set from the boiler operation data set based on the dynamic operation change model, obtains a future temperature operation trajectory corresponding to a preset prediction period of the current working condition feature data set, and fuses to generate a future temperature fusion trajectory; the future temperature fusion trajectory is analyzed to obtain main future time nodes; an abnormal future temperature operation trajectory of a previous collection time point is obtained according to the main future time nodes; the current working condition feature data set is regulated and controlled according to the abnormal future temperature operation trajectory to obtain a current regulation and control working condition data set; and the optimal main steam temperature is obtained according to the regulation and control working condition data set.

Description

Technical Field

[0001] This invention relates to the field of thermal power plant control technology, specifically to an intelligent control method for the main steam temperature of thermal power plants based on artificial intelligence. Background Technology

[0002] Main steam temperature is not only a core parameter for the operation of boilers in thermal power plants, but also a core parameter for the safe and economical operation of thermal power units. Its control accuracy directly affects the unit efficiency.

[0003] However, current main steam temperature control in thermal power plants largely relies on traditional PID and cascade control, which is difficult to adapt to the boiler's large inertia, large lag, nonlinearity, and strong coupling operating characteristics. Under conditions such as load fluctuations, coal quality changes, and combustion disturbances, existing methods are prone to improper main steam temperature regulation, either too high or too low, failing to maintain the main steam temperature at the optimal temperature, thus affecting unit efficiency.

[0004] Therefore, this invention provides an intelligent control method for the main steam temperature of thermal power plants based on artificial intelligence. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent control method for the main steam temperature of thermal power plants based on artificial intelligence, so as to solve the problems in the background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent control method for main steam temperature in thermal power plants based on artificial intelligence, characterized in that the method includes the following steps: Deploy operational monitoring points for boilers in thermal power plants to collect boiler operation data in real time; Set up a dynamic operation change model, and extract the current features of the boiler operation dataset based on the dynamic operation change model to obtain the current operating condition feature dataset. Obtain the future temperature operation trajectory corresponding to the current operating condition feature dataset within a preset prediction period, and fuse them to generate a future temperature fusion trajectory. Analyze the future temperature fusion trajectory to obtain the main future time nodes; obtain the abnormal future temperature operation trajectory of the previous acquisition time point based on the main future time nodes; adjust the current operating condition feature dataset based on the abnormal future temperature operation trajectory to obtain the current adjustment operating condition dataset. The optimal main steam temperature is obtained based on the control condition dataset.

[0007] Furthermore, the process of deploying operational monitoring points for thermal power plant boilers and collecting boiler operation data in real time includes: Obtain operational measurement points of the boiler in a thermal power plant; set a continuous number, and then collect a corresponding number of operational measurement point datasets at the collection time point according to the continuous number; Obtain the mean value of the corresponding collection time points in the measurement point operation dataset, and record it as boiler operation data. Then, integrate the boiler operation data of each operation measurement point to generate a boiler operation dataset.

[0008] Furthermore, a dynamic operation change model is set up. The process of extracting current features from the boiler operation dataset based on the dynamic operation change model to obtain the current operating condition feature dataset includes: Based on the unit load rate, a load rate level threshold range is set, and the load change impact characteristics of the load rate level range are set. The optimal boiler operation dataset is obtained based on the load change impact characteristics. At the same time, the optimal main steam temperature generated in the optimal boiler operation dataset is obtained. Based on the optimal main steam temperature, the optimal coefficients corresponding to each optimal boiler operation data point in the optimal boiler operation dataset corresponding to the load rate level threshold range are obtained, and then integrated to generate an optimal coefficient set. The load rate change impact characteristics are used as a benchmark to obtain the input boiler operation dataset, and then the actual optimal operating temperature of the input boiler operation dataset is obtained according to the optimal coefficient set; then a dynamic operation change model is constructed. Based on the dynamic operation change model, the features corresponding to the boiler operation dataset are extracted and denoted as the current operating condition feature dataset.

[0009] Furthermore, the process of obtaining the future temperature operation trajectory corresponding to the current operating condition feature dataset within a preset prediction period, and fusing it to generate a future temperature fusion trajectory, includes: Set the adjacent acquisition period for adjacent acquisition time points; where the total duration of the adjacent acquisition period is equal to the time interval between adjacent acquisition time points; Based on the dynamic operation change model, the current operating condition feature dataset corresponding to each collection time point is obtained, and the actual optimal operating temperature generated by the current operating condition feature dataset is sequentially deduced and smoothed according to the adjacent collection cycles to obtain the corresponding future temperature operating trajectory. The various future temperature operation trajectories are spliced ​​and fused according to the collection time points to generate a future temperature fusion trajectory.

[0010] Furthermore, the process of analyzing the future temperature fusion trajectory to obtain key future time nodes includes: The future operating temperature at the corresponding acquisition time point is obtained by acquiring the future temperature fusion trajectory, and compared with the actual optimal operating temperature at the acquisition time point. If the absolute value of the difference between the temperature generated by future operation and the temperature generated by actual optimal operation exceeds a, where a is a non-zero natural number, then the corresponding data collection time point is recorded as the primary future time node. Conversely, the future temperature trajectory corresponding to the time point of data collection is obtained.

[0011] Furthermore, the process of obtaining the abnormal future temperature trajectory of the previous data collection time point based on the major future time point includes: Set the maximum temperature change value, and determine whether the temperature change value of the future operation generated temperature at the next adjacent acquisition time point in the adjacent acquisition cycle in the future temperature operation trajectory exceeds the maximum temperature change value. If so, the corresponding adjacent acquisition time point is recorded as an abnormal adjacent acquisition time point; otherwise, no action is taken. To obtain the future temperature trajectory of the data collection point preceding the main future time node, first determine whether the preceding data collection point is the main future time node. If so, mark the future temperature operation trajectory at the previous data collection time point as an abnormal future temperature operation trajectory; Conversely, it is determined whether the number of abnormal adjacent acquisition time points of the future temperature operation trajectory of the previous acquisition time point exceeds b, where b is a positive integer; if so, the future temperature operation trajectory of the previous acquisition time point is marked as an abnormal future temperature operation trajectory; otherwise, no processing is performed.

[0012] Furthermore, the process of obtaining the current control condition dataset by adjusting the current operating condition feature dataset based on the abnormal future temperature operating trajectory includes: If the abnormal future temperature operation trajectory corresponds to the previous acquisition time point as the main future time node, then the current operating condition feature dataset of the previous acquisition time point is directly obtained, and then the current operating condition feature dataset is adjusted at the previous acquisition time point according to the actual optimal operating temperature to obtain the current adjustment operating condition dataset. If the previous acquisition time point corresponding to the abnormal future temperature trajectory is not a major future time node, then the current operating condition feature dataset is adjusted at the adjacent acquisition time point of the abnormality to obtain the current control operating condition dataset.

[0013] Furthermore, the process of obtaining the optimal main steam temperature based on the control condition dataset includes: The actual optimal operating temperature at the time of data collection is obtained in real time based on the control condition dataset; a conversion coefficient is set, and the optimal main steam temperature is obtained based on the conversion coefficient.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention collects boiler operation data in real time by deploying operational monitoring points at thermal power plant boilers; sets up a dynamic operation change model, and extracts current features from the boiler operation data based on the dynamic operation change model to obtain the current operating condition feature data; obtains the future temperature operation trajectory corresponding to the current operating condition feature data in a preset prediction period, and fuses it to generate a future temperature fusion trajectory; analyzes the future temperature fusion trajectory to obtain the main future time nodes; obtains the abnormal future temperature operation trajectory of the previous collection time point based on the main future time points; regulates the current operating condition feature data based on the abnormal future temperature operation trajectory to obtain the current regulation operating condition data; obtains the optimal main steam temperature based on the regulation operating condition data; and effectively regulates the main steam temperature to the optimal main steam temperature state. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 This is a flowchart of the present invention.

[0017] Figure 2 This is a flowchart illustrating how the present invention obtains a dataset of current operating condition features.

[0018] Figure 3 This is a flowchart for obtaining the future temperature fusion trajectory according to the present invention.

[0019] Figure 4 This is a flowchart illustrating the process of obtaining the operational trajectory of abnormal future temperatures in this invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figure 1 As shown, an intelligent control method for main steam temperature in thermal power plants based on artificial intelligence is described, the method comprising the following steps: Step S1: Deploy operation monitoring points for the boilers in thermal power plants and collect boiler operation data in real time; Step S2: Set up a dynamic operation change model, and extract the current features from the boiler operation dataset based on the dynamic operation change model to obtain the current operating condition feature dataset; Step S3: Obtain the future temperature operation trajectory corresponding to the current working condition feature dataset in the preset prediction period, and fuse them to generate the future temperature fusion trajectory; Step S4: Analyze the future temperature fusion trajectory to obtain the main future time nodes; obtain the abnormal future temperature operation trajectory of the previous acquisition time point based on the main future time nodes; adjust the current operating condition feature dataset based on the abnormal future temperature operation trajectory to obtain the current adjustment operating condition dataset. Step S5: Obtain the optimal main steam temperature based on the control condition dataset.

[0022] Please see Figure 2 As shown, step S1 requires further clarification. The process of deploying operational monitoring points for the thermal power plant boiler and collecting boiler operation data in real time includes: Step S101: Obtain the operation measurement points of the boiler in the thermal power plant; set the continuous number, and then collect the corresponding number of measurement point operation datasets at the collection time point according to the continuous number. Step S102: Obtain the mean value of the collection time points corresponding to the measurement point operation dataset, record it as boiler operation data, and integrate the boiler operation data of each operation measurement point to generate a boiler operation dataset. In the above embodiments, steps S101-S103 require further clarification; the "continuous quantity" indicates that a corresponding number of data points are continuously collected within the collection time; that is, at a single collection time point, the same operating measurement point does not use an instantaneous single value, but instead performs multiple dense samplings according to a set continuous quantity to form a set of measurement point operating datasets, thereby reducing instantaneous interference, measurement jitter, and random noise, and improving the reliability of the original data. The boiler operating data includes, but is not limited to, total coal consumption, total air supply volume, flue gas oxygen content, desuperheating water reference flow rate, and burner tilt angle position, etc.

[0023] like Figure 2 As shown, step S2 requires further refinement. A dynamic operation change model is set up. The process of extracting current features from the boiler operation dataset based on the dynamic operation change model to obtain the current operating condition feature dataset includes: Step S201: Set the load rate level threshold range based on the unit load rate, and set the load change impact characteristics of the load rate level range. Obtain the optimal boiler operation dataset based on the load change impact characteristics; at the same time, obtain the optimal main steam temperature generated in the optimal boiler operation dataset. Step S202: Based on the optimal main steam temperature, obtain the optimal coefficients corresponding to each optimal boiler operation data point in the optimal boiler operation dataset corresponding to the load rate level threshold range, and integrate them to generate an optimal coefficient set. Step S203: Obtain the input boiler operation dataset based on the characteristics of load rate change, and then obtain the actual optimal operating temperature of the input boiler operation dataset according to the optimal coefficient set; then construct a dynamic operation change model.

[0024] Step S204: Extract the features corresponding to the boiler operation dataset based on the input boiler operation dataset corresponding to the dynamic operation change model, and denot it as the current operating condition feature dataset.

[0025] In the above embodiments, steps S201-S104 require further clarification; the unit load rate is used to represent the ratio of the current actual power generation of the thermal power unit to the rated maximum continuous power generation of the unit, which characterizes how much electricity needs to be generated, directly affecting the amount of coal, and thus affecting the main steam temperature. Specifically, the optimal coefficient set is obtained based on the optimal main steam temperature as the optimal output result, serving as the variation coefficient for obtaining the actual optimal operating temperature, rather than directly obtaining the actual optimal main steam temperature; its function is to establish a mapping relationship between the load rate, boiler operating data, and theoretical temperature through the optimal coefficient set, enabling rapid calculation of the ideal main steam temperature under the current operating conditions. Obtaining the optimal coefficients is not a simple linear process. It involves: based on the optimal boiler operation dataset and optimal main steam temperature corresponding to the current load rate threshold range, using nonlinear fitting algorithms (such as BP neural networks, least squares nonlinear regression, etc.) to fit the nonlinear correlation between each set of optimal boiler operation data and the optimal main steam temperature, obtaining the optimal coefficients for each set of optimal boiler operation data, and finally integrating them to generate the optimal coefficient set. This process must fully consider the strong coupling characteristics of the boiler operation process, avoiding the influence of operating condition fluctuations and parameter linkages that linear fitting cannot adapt to, ensuring that the optimal coefficients can accurately reflect the nonlinear mapping law between operation data and temperature. For example, assuming a certain load rate threshold range is 50%–80%. The BMCR (Boiler Management Controller) optimal boiler operation dataset includes a total coal load of 100–140 t / h, a total air supply of 220–300 km³ / h, a flue gas oxygen content of 3.0%–3.8%, a desuperheating water baseline flow rate of 60–90 t / h, and a burner tilt angle of -10°–10°. The corresponding optimal main steam temperature is fixed at 40°C. When obtaining the optimal coefficients, a simple linear proportional calculation (e.g., a fixed temperature increase of X°C for every 10 t / h increase in coal load) is not used. Instead, a nonlinear fitting algorithm is employed, combining multiple sets of historical optimal operating data within this load range (e.g., coefficients of 0.82 for 100 t / h, 0.85 for 110 t / h, and 0.85 for 140 t / h). The coefficient corresponding to the time is 0.91 (not increasing nonlinearly). Nonlinear correlation coefficients between the optimal operating data of total coal volume, total air volume, and flue gas oxygen content and the optimal main steam temperature of 540℃ are fitted respectively. Among them, the correlation coefficient between total air volume and temperature shows a nonlinear change that first increases and then tends to stabilize as the air volume increases, while the correlation coefficient of flue gas oxygen content shows a nonlinear decreasing trend as the oxygen content increases. Finally, all the fitted nonlinear coefficients are integrated to generate the optimal coefficient set corresponding to the load range. Therefore, the actual operating temperature is also obtained by constructing a dynamic operation change model.

[0026] like Figure 3As shown, step S3 needs further refinement. The process of obtaining the future temperature operation trajectory corresponding to the current operating condition feature dataset within a preset prediction period and fusing it to generate the future temperature fusion trajectory includes: Step S301: Set the adjacent acquisition period for adjacent acquisition time points; wherein, the total period duration of adjacent acquisition periods is equal to the time interval between adjacent acquisition time points; Step S302: Based on the dynamic operation change model, obtain the current operating condition feature dataset corresponding to each collection time point, and sequentially deduce and smooth the actual optimal operating temperature of the current operating condition feature dataset according to adjacent collection cycles to obtain the corresponding future temperature operating trajectory. Step S303: The future temperature operation trajectories are spliced ​​and fused according to the acquisition time points to generate a future temperature fusion trajectory.

[0027] In the above embodiments, steps S301-S303 need further clarification. Since the time interval between adjacent acquisition time points is relatively short, the corresponding current operating condition feature dataset can be considered continuously valid within the time interval. Therefore, adjacent acquisition cycles with a total cycle duration equal to the time interval are set within the time interval. Since adjacent acquisition cycles are extreme cycles, the actual optimal operating temperature is a parameter with large inertia and large lag, and will not change in a very short time. The trend of its future temperature operating trajectory has strong continuity. The future temperature operating trajectory of the adjacent acquisition cycle can be smoothly extrapolated from the current operating condition feature dataset and the dynamic operating change model. The process includes: based on each acquisition time point and its corresponding future temperature operating trajectory, they are arranged in the order of acquisition time. The actual optimal operating temperature corresponding to the previous acquisition time point is used as the starting connection value of the trajectory corresponding to the next acquisition time point. The time-series smoothing algorithm is used to eliminate the numerical jumps and connection spikes between adjacent trajectories, so that multiple independent trajectories are continuously connected and smoothly transitioned on the time axis. Finally, they are spliced ​​and merged into a continuous and smooth future temperature fusion trajectory that covers the entire prediction cycle.

[0028] like Figure 4 As shown, step S4 requires further refinement. The analysis of the future temperature fusion trajectory is needed to obtain the main future time nodes. The process of obtaining the abnormal future temperature trajectory of the previous acquisition time point based on the main future time points includes: Step S401: Obtain the future operating temperature corresponding to the acquisition time point of the future temperature fusion trajectory, and compare it with the actual optimal operating temperature at the acquisition time point. Step S402: If the absolute value of the difference between the future operating temperature and the actual optimal operating temperature exceeds a, where a is a non-zero natural number, then the corresponding data collection time point is recorded as the primary future time node. Step S403: Conversely, obtain the future temperature operation trajectory corresponding to the acquisition time point, and set the maximum temperature change value. Based on the maximum temperature change value, determine whether the temperature change amplitude corresponding to the future operation temperature at the adjacent acquisition time point of the adjacent acquisition cycle in the future temperature operation trajectory exceeds the maximum temperature change value. Step S404: If yes, then record the corresponding adjacent acquisition time point as an abnormal adjacent acquisition time point; otherwise, do not perform any processing. Step S405: Obtain the future temperature trajectory of the previous acquisition time point before the main future time node, and first determine whether the previous acquisition time point is the main future time node; Step S406: If so, mark the future temperature operation trajectory of the previous acquisition time point as an abnormal future temperature operation trajectory; Step S407: Conversely, determine whether the number of abnormal adjacent acquisition time points of the future temperature operation trajectory of the previous acquisition time point exceeds b, where b is a positive integer; if so, mark the future temperature operation trajectory of the previous acquisition time point as an abnormal future temperature operation trajectory; otherwise, do not perform any processing.

[0029] In the above embodiments, steps S401-S407 require further clarification. First, the data acquisition time points of the splicing and fusion are analyzed. Taking the future temperature fusion trajectory as the analysis object, the future operating temperature corresponding to each acquisition time point after trajectory splicing is extracted sequentially. The difference between this temperature and the actual optimal operating temperature calculated by the dynamic operating change model at the same acquisition time point is calculated point by point. By setting a threshold, it is determined whether there is a significant temperature deviation at the acquisition time point, thus achieving preliminary screening of the main future time nodes. Furthermore, when a significant temperature deviation occurs at the current moment, the cause has often been formed in the previous acquisition cycle. Therefore, by further screening the future temperature operating trajectory of the previous acquisition time point and tracing back, the risk can be identified as soon as the abnormal trend appears, rather than waiting until the temperature exceeds the standard before responding.

[0030] It should be further explained that the process of obtaining the current control condition dataset by adjusting the current operating condition feature dataset based on the abnormal future temperature trajectory includes: Step S408: If the abnormal future temperature operation trajectory corresponds to the previous acquisition time point as the main future time node, then directly obtain the current operating condition feature dataset of the previous acquisition time point, and then adjust the current operating condition feature dataset at the previous acquisition time point according to the actual optimal operating temperature to obtain the current adjustment operating condition dataset. Step S409: If the previous acquisition time point corresponding to the abnormal future temperature operation trajectory is not the main future time node, then adjust the current operating condition feature dataset at the adjacent acquisition time point of the abnormality to obtain the current control operating condition dataset.

[0031] In the above embodiments, steps S408-S409 require further clarification; specifically, adjusting the current operating condition feature dataset of abnormal adjacent acquisition time points requires prioritizing the determination of the consistency of fluctuation trends among multiple abnormal adjacent acquisition time points (all experiencing sudden temperature increases, all experiencing sudden temperature decreases, or alternating fluctuations), and then performing targeted adjustments. The specific process includes: extracting the current operating condition feature dataset corresponding to the preceding acquisition time point, simultaneously locating all abnormal adjacent acquisition time points, and recording the temperature change amplitude and fluctuation direction (sudden increase / sudden decrease) of each abnormal point; if the fluctuation trends of multiple abnormal adjacent acquisition time points are consistent (e.g., all experiencing sudden temperature increases), it is determined to be a continuous fluctuation caused by a single factor, and the abnormal point with the largest fluctuation amplitude is selected. The corresponding operating condition parameter adjustment amounts are calculated for adjacent acquisition time points to perform a one-time precise adjustment of the current operating condition feature dataset. At the same time, the fluctuation amplitude of other abnormal points is taken into account, and the adjustment amounts are fine-tuned to ensure that the temperature fluctuations of all abnormal points are controlled within the allowable range after adjustment. If the fluctuation trends of multiple abnormal adjacent acquisition time points are inconsistent (such as some sudden increases and some sudden decreases), it is determined that the fluctuation is caused by multiple factors. The small deviations of the operating condition feature data of the current operating condition feature dataset corresponding to each abnormal adjacent acquisition time point are extracted, and the corresponding adjustment amounts are calculated. The current operating condition feature dataset is adjusted step by step using the method of "segmented fine-tuning and gradual convergence" to avoid a single large adjustment that would aggravate the fluctuations. Finally, the trajectory is restored to a stable state.

[0032] Further clarification is needed for step S5, specifically the process of obtaining the optimal main steam temperature based on the control condition dataset, which includes: Step S501: Obtain the actual optimal operating temperature at the time of data collection based on the control condition dataset; set the conversion coefficient and obtain the optimal main steam temperature based on the conversion coefficient.

[0033] In the above embodiments, step S501 needs further clarification. The process of obtaining the conversion coefficient includes: it can be obtained by fitting historical optimal operating data. For example, historical sample pairs are formed by collecting the actual optimal operating temperature and the corresponding stable optimal main steam temperature under different load rate threshold ranges. The conversion coefficient between the two is obtained by fitting through nonlinear regression, least squares fitting, etc., so that the deviation between the converted temperature value and the historical true optimal main steam temperature is minimized. Alternatively, it can be obtained by dynamically correcting the furnace combustion heat release intensity, introducing disturbance parameters such as real-time furnace combustion heat release intensity, flue gas temperature, and flue gas flow rate, and dynamically correcting the basic conversion coefficient according to the current heat exchange intensity, so that the conversion coefficient can adapt to changes in combustion state and improve temperature conversion accuracy. Other methods include on-site test calibration, real-time online adaptive updating, etc., which will not be elaborated here. By comprehensively determining the conversion coefficient through multiple methods, factors such as load changes, combustion disturbances, and differences in heat exchange characteristics can be fully considered, so that the actual optimal operating temperature can be accurately and stably converted into the optimal main steam temperature, providing a reliable target value for subsequent precise control of the main steam temperature.

[0034] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for intelligent control of main steam temperature in thermal power plants based on artificial intelligence, characterized in that, The method includes the following steps: Deploy operational monitoring points for boilers in thermal power plants to collect boiler operation data in real time; Set up a dynamic operation change model, and extract the current features of the boiler operation dataset based on the dynamic operation change model to obtain the current operating condition feature dataset. Obtain the future temperature operation trajectory corresponding to the current operating condition feature dataset within a preset prediction period, and fuse them to generate a future temperature fusion trajectory. Analyze the future temperature fusion trajectory to obtain the main future time nodes; obtain the abnormal future temperature operation trajectory of the previous acquisition time point based on the main future time nodes; adjust the current operating condition feature dataset based on the abnormal future temperature operation trajectory to obtain the current adjustment operating condition dataset. The optimal main steam temperature is obtained based on the control condition dataset.

2. The method for intelligent control of main steam temperature in thermal power plants based on artificial intelligence according to claim 1, characterized in that, The process of deploying operational monitoring points for thermal power plant boilers and collecting boiler operation data in real time includes: Obtain operational measurement points of the boiler in a thermal power plant; set a continuous number, and then collect a corresponding number of operational measurement point datasets at the collection time point according to the continuous number; Obtain the mean value of the corresponding collection time points in the measurement point operation dataset, and record it as boiler operation data. Then, integrate the boiler operation data of each operation measurement point to generate a boiler operation dataset.

3. The method for intelligent control of main steam temperature in thermal power plants based on artificial intelligence according to claim 2, characterized in that, The process of setting up a dynamic operation change model and extracting current features from the boiler operation dataset based on the dynamic operation change model to obtain the current operating condition feature dataset includes: Based on the unit load rate, a load rate level threshold range is set, and the load change impact characteristics of the load rate level range are set. The optimal boiler operation dataset is obtained based on the load change impact characteristics. At the same time, the optimal main steam temperature generated in the optimal boiler operation dataset is obtained. Based on the optimal main steam temperature, the optimal coefficients corresponding to each optimal boiler operation data point in the optimal boiler operation dataset corresponding to the load rate level threshold range are obtained, and then integrated to generate an optimal coefficient set. The load rate change impact characteristics are used as a benchmark to obtain the input boiler operation dataset, and then the actual optimal operating temperature of the input boiler operation dataset is obtained according to the optimal coefficient set; then a dynamic operation change model is constructed. Based on the dynamic operation change model, the features corresponding to the boiler operation dataset are extracted and denoted as the current operating condition feature dataset.

4. The method for intelligent control of main steam temperature in thermal power plants based on artificial intelligence according to claim 3, characterized in that, The process of obtaining the future temperature trajectory of the current operating condition feature dataset within a preset prediction period and fusing it to generate a future temperature fusion trajectory includes: Set the adjacent acquisition period for adjacent acquisition time points; where the total duration of the adjacent acquisition period is equal to the time interval between adjacent acquisition time points; Based on the dynamic operation change model, the current operating condition feature dataset corresponding to each collection time point is obtained, and the actual optimal operating temperature generated by the current operating condition feature dataset is sequentially deduced and smoothed according to the adjacent collection cycles to obtain the corresponding future temperature operating trajectory. The various future temperature operation trajectories are spliced ​​and fused according to the collection time points to generate a future temperature fusion trajectory.

5. The method for intelligent control of main steam temperature in thermal power plants based on artificial intelligence according to claim 4, characterized in that, The process of analyzing future temperature fusion trajectories to obtain key future time points includes: The future operating temperature at the corresponding acquisition time point is obtained by acquiring the future temperature fusion trajectory, and compared with the actual optimal operating temperature at the acquisition time point. If the absolute value of the difference between the temperature generated by future operation and the temperature generated by actual optimal operation exceeds a, where a is a non-zero natural number, then the corresponding data collection time point is recorded as the primary future time node. Conversely, the future temperature trajectory corresponding to the time point of data collection is obtained.

6. The method for intelligent control of main steam temperature in thermal power plants based on artificial intelligence according to claim 5, characterized in that, The process of obtaining the abnormal future temperature trajectory from the previous data collection time point based on the major future time point includes: Set the maximum temperature change value, and determine whether the temperature change value of the future operation generated temperature at the next adjacent acquisition time point in the adjacent acquisition cycle in the future temperature operation trajectory exceeds the maximum temperature change value. If so, the corresponding adjacent acquisition time point is recorded as an abnormal adjacent acquisition time point; otherwise, no action is taken. To obtain the future temperature trajectory of the data collection point preceding the main future time node, first determine whether the preceding data collection point is the main future time node. If so, mark the future temperature operation trajectory at the previous data collection time point as an abnormal future temperature operation trajectory; Conversely, it is determined whether the number of abnormal adjacent acquisition time points of the future temperature operation trajectory of the previous acquisition time point exceeds b, where b is a positive integer; if so, the future temperature operation trajectory of the previous acquisition time point is marked as an abnormal future temperature operation trajectory; otherwise, no processing is performed.

7. The method for intelligent control of main steam temperature in thermal power plants based on artificial intelligence according to claim 6, characterized in that, The process of obtaining the current control condition dataset by adjusting the current operating condition feature dataset based on the abnormal future temperature trajectory includes: If the abnormal future temperature operation trajectory corresponds to the previous acquisition time point as the main future time node, then the current operating condition feature dataset of the previous acquisition time point is directly obtained, and then the current operating condition feature dataset is adjusted at the previous acquisition time point according to the actual optimal operating temperature to obtain the current adjustment operating condition dataset. If the previous acquisition time point corresponding to the abnormal future temperature trajectory is not a major future time node, then the current operating condition feature dataset is adjusted at the adjacent acquisition time point of the abnormality to obtain the current control operating condition dataset.

8. The method for intelligent control of main steam temperature in thermal power plants based on artificial intelligence according to claim 7, characterized in that, The process of obtaining the optimal main steam temperature from the control condition dataset includes: The actual optimal operating temperature at the time of data collection is obtained in real time based on the control condition dataset; a conversion coefficient is set, and the optimal main steam temperature is obtained based on the conversion coefficient.

Images