A power grid deep peak shaving optimization control system

Abstract

The application relates to the technical field of power grid peak regulation, and discloses a power grid deep peak regulation optimization control system, which is characterized in that: first, high-precision monitoring equipment is arranged and controlled, and multi-dimensional data collection is synchronously carried out; then, the influence range of the SCR flue gas temperature of the furnace, the coal powder NO X influence index, the optimal baffle opening, and the optimal ammonia injection ratio are calculated and quantified based on the collected data; the deep peak regulation low-load working condition of the thermal power generating unit is accurately clustered and divided based on the collected data and the calculated and quantified results; the core parameter benchmark interval of various typical working conditions is calibrated; the working condition change trend of the boiler combustion, the denitration reaction and the whole system linkage is predicted; finally, parameter adjustment values are generated based on the working condition change trend results and the calculated and quantified results, parameter adjustment is carried out, and an early warning signal is sent for abnormal conditions. The system specifically solves the core problems of insufficient boiler stable combustion capability, poor denitration system adaptability and whole system linkage imbalance in the low-load stage.

Description

Technical Field

[0001] This invention relates to the field of power grid peak shaving technology, specifically to a deep power grid peak shaving optimization control system. Background Technology

[0002] Deep grid peak shaving is a core power regulation method adapted to the high proportion of renewable energy grid connection under the new power system. It is an advanced operation mode compared to conventional peak shaving. It refers to a peak shaving method in which the output of peak shaving resources such as thermal power units is reduced to below 50% of the rated capacity (mainstream can reach 30%-40%, and advanced units can reach 20% or even lower) and continuously and stably operated for more than 4 hours during periods of low grid load and high renewable energy generation. This is different from conventional peak shaving, which only involves small adjustments within the 50%-100% rated load range. The core of this method is to free up power generation space for intermittent renewable energy such as wind power and photovoltaics, significantly reduce the curtailment rate of wind and solar power, and at the same time enable coal-fired power to transform from a traditional power source to a regulation and guarantee power source. It allows for rapid ramp-up to replenish energy when renewable energy output drops sharply, providing spinning reserve for the grid and maintaining frequency and voltage stability. Implementing deep peak shaving for the power grid requires flexible upgrades to thermal power units, such as boiler combustion stabilization, thermal-electric decoupling, and auxiliary equipment frequency conversion. This addresses technical challenges such as stable combustion under low load, environmental compliance, and equipment safety. Furthermore, as a paid service category in the power auxiliary service market, it relies on intelligent dispatching to achieve precise matching between new energy output forecasting and peak shaving resources. Currently, it is also gradually integrating resources such as pumped storage, electrochemical energy storage, and demand-side response to build a diversified deep peak shaving system to reduce excessive reliance on coal-fired power.

[0003] During the deep peak shaving and low load operation phase, existing thermal power units have insufficient boiler combustion stability and poor combustion stability under low load conditions, which can easily lead to problems such as flameout and incomplete combustion. This not only affects the safe and continuous operation of the unit, but also makes it difficult for environmental protection systems such as denitrification to adapt to the operating conditions due to the decrease in furnace temperature and abnormal flue gas flow and composition. The reaction conditions and reagent dosage of environmental protection facilities are out of balance with the actual output of the unit, resulting in a significant decrease in denitrification efficiency and causing emissions of pollutants such as nitrogen oxides to exceed the standards. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] To address the shortcomings of existing technologies, this invention provides a power grid deep peak shaving optimization control system. Through the collaborative cooperation of multiple modules, it can achieve system-wide collaborative optimization of deep peak shaving in the power grid, and specifically solve the core problems of insufficient boiler combustion stability during low load periods, poor adaptability of the denitrification system, and system-wide imbalance.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the present invention provides the following technical solution: a power grid deep peak shaving optimization control system, comprising a multi-dimensional data monitoring module, a multi-dimensional data analysis module, an operating condition classification and prediction module, and a regulation command optimization module;

[0008] The multidimensional data monitoring module includes a measuring point deployment unit and a data division unit. The measuring point deployment unit is used to deploy high-precision monitoring equipment in multiple areas of the boiler, SCR denitrification system and power grid, and simultaneously collect multidimensional data. The data division unit is used to store and divide the multidimensional data collected by the measuring point deployment unit, specifically including: boiler combustion data, denitrification reaction data and whole system linkage data.

[0009] The multi-dimensional data analysis module is used to calculate and quantify the impact of furnace SCR flue gas temperature and pulverized coal NO based on the collected data. x Influencing factors include: index, optimal damper opening, and optimal ammonia injection ratio;

[0010] The operating condition classification and prediction module includes an operating condition feature classification unit and an operating condition trend prediction unit. The operating condition feature classification unit is used to accurately cluster and classify the deep peak-shaving and low-load operating conditions of thermal power units based on the collected data and the results of calculation and quantification, and to define the core parameter benchmark ranges for various typical operating conditions. The operating condition trend prediction unit is used to predict the changing trends of boiler combustion, denitrification reaction, and the entire system linkage.

[0011] The adjustment command optimization module includes an instruction adjustment unit and an early warning unit. The instruction adjustment unit is used to generate parameter adjustment values ​​and adjust parameters based on the results of operating condition change trends and the results of calculation and quantification. The early warning unit is used to issue early warning signals for abnormal situations based on the results of operating condition change trends.

[0012] Preferably, the boiler combustion data includes combustion state parameters during the boiler's low-load operation phase;

[0013] The denitrification reaction data includes the operating status of the denitrification system and reaction condition parameters;

[0014] The system-wide linkage data includes parameters reflecting the linkage and matching status between the boiler, denitrification system, power grid dispatch, and auxiliary equipment systems.

[0015] Preferably, the formula for calculating the influence of SCR flue gas temperature in the furnace is: ; in, This represents the extent to which the flue gas temperature is affected by the SCR in the furnace. Represents covariance; This represents the average temperature of the furnace. This represents the average smoke temperature at the SCR inlet. The standard deviation representing the average temperature of the furnace; The standard deviation of the average smoke temperature at the SCR inlet; This represents the degree of temperature drop in the furnace.

[0016] The pulverized coal NO x The formula for calculating the influence index is: ; in, Represents NO of pulverized coal x Impact Index; This represents the current pulverized coal concentration; Represents the baseline pulverized coal concentration; Represents the effect of pulverized coal concentration on NO x The generated correction coefficients.

[0017] Preferably, the formula for calculating the optimal opening degree of the baffle is: ; in, This represents the optimal opening degree of the baffle. This represents the target value for the SCR inlet smoke temperature; This represents the current value of the SCR inlet smoke temperature; This represents the maximum increase in SCR inlet smoke temperature when the damper is fully open. This represents the maximum opening of the baffle. This represents the boiler efficiency correction factor.

[0018] The formula for calculating the optimal ammonia injection ratio is: ; in, This represents the optimal ammonia injection ratio; Represents the target denitrification efficiency; This represents the maximum achievable denitrification efficiency under the current operating conditions; Represents the smoke temperature correction factor; This represents the flue gas flow correction factor.

[0019] Preferably, the operating condition feature classification unit uses the K-Means clustering algorithm to achieve accurate clustering and benchmark parameter calibration for low-load operating conditions, specifically including:

[0020] A1. Retrieve the multidimensional data set collected by the multidimensional data monitoring module and mark it as... , It covers boiler combustion, denitrification reaction, and overall system linkage parameters, primarily reflecting the linkage and matching status of the boiler, denitrification system, power grid dispatch, and auxiliary systems; it retrieves multi-dimensional data analysis modules to calculate and quantify the impact of furnace SCR flue gas temperature and pulverized coal NO. xThe set of indices, including the influencing index, optimal damper opening, and optimal ammonia injection ratio, is tagged as follows: Then, the K-Means algorithm is used to perform clustering operations.

[0021] A2. The clustering results are iteratively converged based on minimizing the sum of squared errors within each cluster, and typical operating conditions are output. and corresponding cluster centers ;

[0022] A3. For each type of typical working condition The sample means were calculated for the core parameters of boiler combustion, denitrification reaction, and overall system linkage. and standard deviation Then, substitute the values ​​into the benchmark interval formula to determine the reasonable fluctuation range of each parameter.

[0023] Preferably, in A1, the clustering operation formula expression is: ; in, This represents the result of clustering operations; Represents the number of clusters;

[0024] In A2, the iterative convergence formula is expressed as follows: ; in, Represents the sum of squared errors within the class; represent A single data sample; Representing the Cluster center for similar working conditions;

[0025] In the A2 section, the typical operating condition is as follows: The formula expression is: ; in, The Euclidean distance from the sample to the cluster center;

[0026] In A2, the corresponding cluster center The formula expression is: ; in, Representing the Number of samples for each type of working condition;

[0027] In A3, the sample mean The formula expression is: ; in, Representing the Core parameters in similar working conditions A single sample value;

[0028] In A3, the standard deviation The formula expression is: ;

[0029] In A3, the formula expression for the reference interval is: ; in, Representing the Core parameters for similar working conditions The baseline range; This represents the parameter fluctuation threshold set according to the 3σ principle.

[0030] Preferably, the working steps of the working condition trend prediction unit are as follows:

[0031] B1. Use the LSTM time series prediction algorithm to calculate the predicted value of the operating condition change;

[0032] B2. Predicted values ​​of operating condition changes Decomposed into boiler combustion prediction values Predicted value of denitrification reaction System-wide linkage prediction value ;

[0033] B3. Compare each predicted value with typical operating conditions. The baseline range Compare and determine the trend type.

[0034] Preferably, in B1, the formula expression for the predicted value of the operating condition change is: ; in, represent Predicted changes in operating conditions at any given time; represent Time before A set of historical features within a time period; Represents the weights of the LSTM model; This represents the bias of the LSTM model; Represents the predicted time point;

[0035] In B2, the formula expressions are as follows: ; ; ; in, This represents the set of core parameters for boiler combustion. Represents the core parameter set for denitration reaction; Represents the core parameter set for the whole system linkage.

[0036] Preferably, in the B3, the determination method is: When , represents a stable trend; When , represents an abnormal trend; Wherein, Represents any one of boiler combustion, denitration reaction, and whole system linkage.

[0037] Preferably, in the instruction adjustment unit, the formula expression of the parameter adjustment value is: ; Wherein, Represents the parameter adjustment value; Represents the working condition trend correction coefficient; Represents the quantization index weight coefficient.

[0038] Compared with the prior art, the present invention provides a power grid deep peak shaving optimization control system, which has the following beneficial effects:

[0039] 1. Through multi-dimensional data monitoring, the present invention targets high-precision monitoring devices for boiler, SCR denitration system and multi-region layout of power grid, and synchronously collects multi-dimensional data, realizing the basic data support for the collaborative optimization of the whole system from the source, and打通 the linkage data link of boiler, denitration, power grid and auxiliary equipment from the data level, avoiding the optimization disconnection problem caused by the isolation of single-system data; Through multi-dimensional data analysis and calculation, the influence amplitude of SCR flue gas temperature in the furnace, the influence index of pulverized coal NO x The influence index, the optimal opening of the baffle, and the optimal ammonia injection ratio are quantified, accurately depicting the coupling relationship between boiler combustion parameters and denitration reaction conditions, and clarifying the influence law of key parameters such as furnace temperature and pulverized coal concentration under low load on the denitration system.

[0040] 2. This invention employs K-Means clustering algorithm to achieve accurate clustering and benchmark parameter calibration for low-load operating conditions through operating condition classification and prediction, providing a quantitative basis for determining operating conditions. Subsequently, LSTM time series prediction algorithm is used to predict the operating condition change trends of the boiler, denitrification system, and the entire system linkage. By comparing the predicted values ​​with the benchmark interval, stable and abnormal trends are identified, achieving forward-looking quantitative prediction and accurate decomposition of operating condition trends, and early detection of precursors to system linkage mismatch. Through adjustment command optimization, precise parameter adjustment values ​​are generated based on the operating condition trend results and quantitative analysis indicators, for the boiler... The system coordinates and regulates the SCR denitrification and auxiliary systems, and issues early warning signals for abnormal trends. It achieves dynamic optimization of boiler combustion parameters under low load to improve stable combustion capability, and simultaneously adapts to the reaction conditions and reagent dosage of the denitrification system. This solves the problem of reduced denitrification efficiency caused by furnace temperature reduction and abnormal flue gas parameters. It also matches the operating status of the boiler and denitrification system with the grid dispatch load and auxiliary system output, avoiding boiler shutdown, incomplete combustion and excessive nitrogen oxide emissions from the perspective of system-wide coordination, and ensuring safe and continuous operation and environmental compliance of the unit during low load phases. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of the system of the present invention. Detailed Implementation

[0042] 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, 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 are within the scope of protection of the present invention. It is worth noting that this application also relates to prior art. Since prior art is well known to those skilled in the art, it will not be described in detail in this application.

[0043] Please see Figure 1 A deep peak-shaving optimization control system for power grids includes a multi-dimensional data monitoring module, a multi-dimensional data analysis module, an operating condition classification and prediction module, and a regulation command optimization module.

[0044] The multi-dimensional data monitoring module includes a measurement point deployment unit and a data division unit. The measurement point deployment unit is used to deploy high-precision monitoring equipment in multiple areas of the boiler, SCR denitrification system and power grid, and simultaneously collect multi-dimensional data. The data division unit is used to store and divide the multi-dimensional data collected by the measurement point deployment unit, specifically including: boiler combustion data, denitrification reaction data and whole system linkage data.

[0045] Boiler combustion data includes combustion status parameters during the low-load operation phase of the boiler;

[0046] Denitrification reaction data includes the operating status of the denitrification system and reaction condition parameters;

[0047] The system-wide linkage data includes parameters reflecting the linkage and matching status between the boiler, denitrification system, power grid dispatch, and auxiliary equipment systems;

[0048] The multi-dimensional data monitoring module provides basic data support for the coordinated optimization of the entire system from the source, and connects the linkage data links of boiler, denitrification, power grid and auxiliary equipment at the data level, avoiding the optimization disconnect caused by the isolation of data from a single system;

[0049] The multi-dimensional data analysis module is used to calculate and quantify the impact of SCR flue gas temperature in the furnace and pulverized coal NO based on the collected data. x The parameters of the boiler combustion and denitrification system under low load conditions are accurately characterized by the influence index, optimal damper opening, and optimal ammonia injection ratio.

[0050] The formula for calculating the impact of SCR flue gas temperature in the furnace is: ; in, This represents the extent to which the flue gas temperature is affected by the SCR in the furnace. Represents covariance; This represents the average temperature of the furnace. This represents the average smoke temperature at the SCR inlet. The standard deviation representing the average temperature of the furnace; The standard deviation of the average smoke temperature at the SCR inlet; The formula represents the degree of decrease in furnace temperature. It quantifies the influence of furnace temperature on SCR inlet flue gas temperature by using covariance, temperature mean and standard deviation, and clarifies the core role of boiler combustion state in denitrification reaction conditions.

[0051] Pulverized coal NO x The formula for calculating the influence index is: ; in, Represents NO of pulverized coal x Impact Index; This represents the current pulverized coal concentration; Represents the baseline pulverized coal concentration; Represents the effect of pulverized coal concentration on NO x The generated correction factor, calculated by applying the coal powder concentration deviation to the NO concentration, quantifies the effect of coal powder concentration on NO. x The effects generated provide a quantitative basis for boiler combustion adjustment and denitrification agent dosage.

[0052] The formula for calculating the optimal opening of the baffle is: ; in, This represents the optimal opening degree of the baffle. This represents the target value for the SCR inlet smoke temperature; This represents the current value of the SCR inlet smoke temperature; This represents the maximum increase in SCR inlet smoke temperature when the damper is fully open. This represents the maximum opening of the baffle. This represents the boiler efficiency correction coefficient. This formula combines the target flue gas temperature, damper opening, and boiler efficiency to calculate the optimal damper opening, thereby achieving the linkage between boiler flue gas temperature regulation and denitrification reaction condition optimization.

[0053] The formula for calculating the optimal ammonia injection ratio is: ; in, This represents the optimal ammonia injection ratio; Represents the target denitrification efficiency; This represents the maximum achievable denitrification efficiency under the current operating conditions; Represents the smoke temperature correction factor; The formula represents the flue gas flow correction coefficient. It is based on the denitrification efficiency target, flue gas temperature and flue gas flow correction to calculate the optimal ammonia injection ratio, so that the denitrification agent addition matches the flue gas parameter changes under low boiler load.

[0054] The multi-dimensional data analysis module calculates the impact of SCR flue gas temperature in the furnace and the NO content of pulverized coal. x The four major indicators of influence index, optimal damper opening, and optimal ammonia injection ratio establish the correlation between boiler combustion, denitrification reaction, and auxiliary equipment regulation from a quantitative perspective, providing calculable and implementable quantitative references for the coordinated optimization of the entire system.

[0055] The operating condition classification and prediction module includes an operating condition feature classification unit and an operating condition trend prediction unit. The operating condition feature classification unit, based on collected data and quantified results, uses the K-Means clustering algorithm to achieve accurate clustering and benchmark parameter calibration for low-load operating conditions. This enables accurate classification of low-load operating conditions and quantitative definition of core parameters, providing a unified benchmark for operating condition determination. Specifically, it includes:

[0056] A1. Retrieve the multidimensional data set collected by the multidimensional data monitoring module and mark it as... , It covers boiler combustion, denitrification reaction, and overall system linkage parameters, primarily reflecting the linkage and matching status of the boiler, denitrification system, power grid dispatch, and auxiliary systems; it retrieves multi-dimensional data analysis modules to calculate and quantify the impact of furnace SCR flue gas temperature and pulverized coal NO. x The set of indices, including the influencing index, optimal damper opening, and optimal ammonia injection ratio, is tagged as follows: Then, the K-Means algorithm is used to perform clustering operations. The clustering operation formula is as follows: ; in, This represents the result of clustering operations; Represents the number of clusters;

[0057] The clustering operation formula integrates full-dimensional operating data and quantitative indicators, so that the division of operating conditions not only fits the actual operating status of the unit, but also takes into account the coupling characteristics of the boiler and denitrification. The division results can accurately reflect the typical low-load operating conditions under different linkage matching states.

[0058] A2. The clustering results are iteratively converged based on minimizing the sum of squared errors within each cluster, and typical operating conditions are output. and corresponding cluster centers ;

[0059] The expression for the iterative convergence formula is: ; in, Represents the sum of squared errors within the class; represent A single data sample; Representing the Cluster centers for typical working conditions are determined by minimizing the sum of squared errors within each cluster, ensuring the accuracy and stability of the clustering results and representing typical working conditions. Its characteristics are more distinct;

[0060] Typical operating conditions The formula expression is: ; in, The Euclidean distance from the sample to the cluster center;

[0061] Corresponding cluster center The formula expression is: ; in, Representing the Number of samples for each type of working condition;

[0062] Accurate classification of samples is achieved by using Euclidean distance, the core feature centers of various typical working conditions are identified, and quantitative references are provided for working condition matching;

[0063] A3. For each type of typical working condition The sample means were calculated for the core parameters of boiler combustion, denitrification reaction, and overall system linkage. and standard deviation Then, the reasonable fluctuation range of each parameter is determined by substituting it into the benchmark interval formula, and a parameter benchmark for the coordinated operation of the whole system under low load conditions is established, so that subsequent trend prediction and anomaly judgment have clear quantitative standards, avoiding adjustment inaccuracies caused by ambiguity in parameter judgment.

[0064] Sample mean The formula expression is: ; in, Representing the Core parameters in similar working conditions A single sample value;

[0065] Standard deviation The formula expression is: ; In A3, the formula expression for the baseline interval is: ; in, Representing the Core parameters for similar working conditions The baseline range; This represents the parameter fluctuation threshold set according to the 3σ principle;

[0066] The operating condition trend prediction unit is used to predict the changing trends of boiler combustion, denitrification reaction, and overall system linkage, achieving forward-looking quantitative prediction and precise breakdown of operating condition trends, and early detection of precursors to system linkage mismatch. The method is as follows:

[0067] B1. The LSTM time series prediction algorithm is used to calculate the predicted value of the operating condition change. The formula expression is as follows: ; in, represent Predicted changes in operating conditions at any given time; represent Time before A set of historical features within a time period; Represents the weights of the LSTM model; This represents the bias of the LSTM model; Representing the predicted time point, time-series forecasting based on historical feature sets can capture the dynamic changes of parameters under low-load conditions and accurately output future data. The predicted operating conditions of the entire system at any given time provide a time window for advance adjustments;

[0068] B2. Predicted values ​​of operating condition changes Decomposed into boiler combustion prediction values Predicted value of denitrification reaction System-wide linkage prediction value The formula expression is: ; ; ; in, This represents the set of core parameters for boiler combustion. This represents the set of core parameters for the denitrification reaction; It represents the core parameter set of the entire system linkage, and accurately breaks down the predicted values ​​of the entire system according to parameter categories. It can predict the individual trends of boiler combustion, denitrification reaction and the linkage of the entire system, and clarify the operational change trend of each system.

[0069] B3. Compare each predicted value with typical operating conditions. The baseline range By comparison, the trend type can be determined when... This represents a stable trend; when This represents an abnormal trend; in, It represents any situation in boiler combustion, denitrification reaction, and overall system linkage, enabling quantitative judgment of stable and abnormal trends, accurately identifying the precursors of abnormal problems, and transforming the optimization of the entire system from passive adjustment to proactive prediction;

[0070] The adjustment command optimization module includes an instruction adjustment unit and an early warning unit. The instruction adjustment unit is used to generate parameter adjustment values ​​and perform parameter adjustment based on the results of operating condition change trends and the results of calculation and quantification.

[0071] The formula for the parameter adjustment value is: ; in, Represents the parameter adjustment value; Representative working condition trend correction coefficient; The formula integrates the deviation of the operating condition trend and the quantitative indicators, so that the generation of the adjustment value not only conforms to the trend change of the current operating condition, but also takes into account the coupling characteristics of the boiler and denitrification. At the same time, through the flexible adjustment of the operating condition trend correction coefficient and the quantitative indicator weight coefficient, it can adapt to the adjustment needs under different linkage matching states.

[0072] The command and control unit can directly guide the parameter adjustment of boiler combustion, SCR denitrification and auxiliary equipment system, realize the coordinated optimization of the three major systems, make the boiler combustion parameter adjustment adapt to the stable combustion requirements, the denitrification system reagent dosing and reaction condition adjustment adapt to the low load conditions of the boiler, and the auxiliary equipment adjustment adapt to the grid dispatch command and the operating requirements of the boiler and denitrification, thus realizing the linkage and matching of the whole system from the control level.

[0073] The early warning unit is used to issue early warning signals for abnormal situations based on the trend of changes in operating conditions. It can promptly remind users when parameters deviate from the reference range, ensuring the safety of the unit's low-load operation and avoiding problems such as boiler shutdown, incomplete combustion, reduced denitrification efficiency, and excessive pollutant emissions. Ultimately, it achieves safe and continuous operation of thermal power units during the deep peak shaving and low-load phase and coordinated optimization of the entire system.

[0074] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A deep peak-shaving optimization control system for power grids, characterized in that, It includes a multi-dimensional data monitoring module, a multi-dimensional data analysis module, a working condition classification and prediction module, and a regulation command optimization module; The multidimensional data monitoring module includes a measuring point deployment unit and a data division unit. The measuring point deployment unit is used to deploy high-precision monitoring equipment in multiple areas of the boiler, SCR denitrification system and power grid, and simultaneously collect multidimensional data. The data division unit is used to store and divide the multidimensional data collected by the measuring point deployment unit, specifically including: boiler combustion data, denitrification reaction data and whole system linkage data. The multi-dimensional data analysis module is used to calculate and quantify the impact of furnace SCR flue gas temperature and pulverized coal NO based on the collected data. x Influencing factors include: index, optimal damper opening, and optimal ammonia injection ratio; The operating condition classification and prediction module includes an operating condition feature classification unit and an operating condition trend prediction unit. The operating condition feature classification unit is used to accurately cluster and classify the deep peak-shaving and low-load operating conditions of thermal power units based on the collected data and the results of calculation and quantification, and to define the core parameter benchmark ranges for various typical operating conditions. The operating condition trend prediction unit is used to predict the changing trends of boiler combustion, denitrification reaction, and the entire system linkage. The adjustment command optimization module includes an instruction adjustment unit and an early warning unit. The instruction adjustment unit is used to generate parameter adjustment values ​​and adjust parameters based on the results of operating condition change trends and the results of calculation and quantification. The early warning unit is used to issue early warning signals for abnormal situations based on the results of operating condition change trends.

2. The power grid deep peak shaving optimization control system according to claim 1, characterized in that, The boiler combustion data includes combustion status parameters during the boiler's low-load operation phase; The denitrification reaction data includes the operating status of the denitrification system and reaction condition parameters; The system-wide linkage data includes parameters reflecting the linkage and matching status between the boiler, denitrification system, power grid dispatch, and auxiliary equipment systems.

3. The power grid deep peak shaving optimization control system according to claim 2, characterized in that, The formula for calculating the influence of SCR flue gas temperature in the furnace is as follows: ； in, This represents the extent to which the flue gas temperature is affected by the SCR in the furnace. Represents covariance; This represents the average temperature of the furnace. This represents the average smoke temperature at the SCR inlet. The standard deviation representing the average temperature of the furnace; The standard deviation of the average smoke temperature at the SCR inlet; This represents the degree of temperature drop in the furnace. The pulverized coal NO x The formula for calculating the influence index is: ； in, Represents NO of pulverized coal x Impact Index; This represents the current pulverized coal concentration; Represents the baseline pulverized coal concentration; Represents the effect of pulverized coal concentration on NO x The generated correction coefficients.

4. The power grid deep peak shaving optimization control system according to claim 2, characterized in that, The formula for calculating the optimal opening degree of the baffle is: ； in, This represents the optimal opening degree of the baffle. This represents the target value for the SCR inlet smoke temperature; This represents the current value of the SCR inlet smoke temperature; This represents the maximum increase in SCR inlet smoke temperature when the damper is fully open. This represents the maximum opening of the baffle. This represents the boiler efficiency correction factor. The formula for calculating the optimal ammonia injection ratio is: ； in, This represents the optimal ammonia injection ratio; Represents the target denitrification efficiency; This represents the maximum achievable denitrification efficiency under the current operating conditions; Represents the smoke temperature correction factor; This represents the flue gas flow correction factor.

5. A power grid deep peak shaving optimization control system according to claim 4, characterized in that, The operating condition feature classification unit uses the K-Means clustering algorithm to achieve accurate clustering and baseline parameter calibration for low-load operating conditions, specifically including: A1. Retrieve the multidimensional data set collected by the multidimensional data monitoring module and mark it as... , It covers boiler combustion, denitrification reaction, and overall system linkage parameters, primarily reflecting the linkage and matching status of the boiler, denitrification system, power grid dispatch, and auxiliary systems; it retrieves multi-dimensional data analysis modules to calculate and quantify the impact of furnace SCR flue gas temperature and pulverized coal NO. x The set of indices, including the influencing index, optimal damper opening, and optimal ammonia injection ratio, is tagged as follows: Then, the K-Means algorithm is used to perform clustering operations. A2. The clustering results are iteratively converged based on minimizing the sum of squared errors within each cluster, and typical operating conditions are output. and corresponding cluster centers ; A3. For each type of typical working condition The sample means were calculated for the core parameters of boiler combustion, denitrification reaction, and overall system linkage. and standard deviation Then, substitute the values ​​into the benchmark interval formula to determine the reasonable fluctuation range of each parameter.

6. A power grid deep peak shaving optimization control system according to claim 5, characterized in that, In A1, the clustering operation formula expression is: ； in, This represents the result of clustering operations; Represents the number of clusters; In A2, the iterative convergence formula is expressed as follows: ； in, Represents the sum of squared errors within the class; represent A single data sample; Representing the Cluster center for similar working conditions; In the A2 section, the typical operating condition is as follows: The formula expression is: ； in, The Euclidean distance from the sample to the cluster center; In A2, the corresponding cluster center The formula expression is: ； in, Representing the Number of samples for each type of working condition; In A3, the sample mean The formula expression is: ； in, Representing the Core parameters in similar working conditions A single sample value; In A3, the standard deviation The formula expression is: ； In A3, the formula expression for the reference interval is: ； in, Representing the Core parameters for similar working conditions The baseline range; This represents the parameter fluctuation threshold set according to the 3σ principle.

7. A power grid deep peak-shaving optimization control system according to claim 6, characterized in that, The working steps of the operating condition trend prediction unit are as follows: B1. Use the LSTM time series prediction algorithm to calculate the predicted value of the operating condition change; B2. Predicted values ​​of operating condition changes Decomposed into boiler combustion prediction values Predicted value of denitrification reaction System-wide linkage prediction value ; B3. Compare each predicted value with typical operating conditions. The baseline range Compare and determine the trend type.

8. A power grid deep peak-shaving optimization control system according to claim 7, characterized in that, In B1, the formula for predicting the change in operating conditions is as follows: ； in, represent Predicted changes in operating conditions at any given time; represent Time before A set of historical features within a time period; Represents the weights of the LSTM model; This represents the bias of the LSTM model; Represents the predicted time point; In B2, the formula expressions are as follows: ； ； ； Among them, represents the set of core parameters for boiler combustion; represents the set of core parameters for denitrification reaction; represents the set of core parameters for the whole system linkage.

9. A power grid deep peak shaving optimization control system according to claim 7, characterized in that, In B3, the determination method is as follows: when This represents a stable trend; when This represents an abnormal trend; in, This represents any situation in boiler combustion, denitrification reaction, or overall system linkage.

10. A power grid deep peak-shaving optimization control system according to claim 9, characterized in that, The formula expression for the parameter adjustment value in the instruction adjustment unit is: ； in, Represents the parameter adjustment value; Representative working condition trend correction coefficient; This represents the weighting coefficient of the quantitative indicator.

Images