Dynamic compensation control method for variable load rate of supercritical unit

Abstract

The invention relates to the technical field of control regulation, in particular to a supercritical unit variable load rate dynamic compensation control method, which comprises the following steps: acquiring unit operation data, performing credibility restoration, and outputting a credible parameter set; a time sequence mapping table is constructed based on the credible data, and boiler heat storage lag time, steam turbine control valve delay and a steam temperature interstage coupling coefficient are quantified; correcting the boiler water supply feedforward instruction by using the heat storage lag time to generate a compensated water supply instruction; according to the time sequence mapping table and the main steam pressure change rate, a multi-genetic algorithm is adopted to optimize a control valve constraint rule, and a dynamically constrained control valve action instruction is generated; an inter-stage coupling coefficient is called to generate a swing angle compensation amount, and water spraying temperature reduction disturbance is blocked; safety locking parameters are monitored to trigger buffering intervention in a grading mode, and compensation instructions are integrated to implement cooperative correction. The problem of out-of-step response of multiple execution units under the variable load working condition is solved, and the unit adjusting rate and the control stability are remarkably optimized.

Description

Technical Field

[0001] The present invention relates to the field of control and regulation technology, and in particular to a method for dynamic compensation control of variable load rate of a supercritical unit. Background Art

[0002] A supercritical unit is a thermal power generation technology that operates the working fluid water in the boiler at a high pressure and high temperature state above the critical point, converting it into a supercritical fluid, thereby eliminating the vapor-liquid phase change process. Based on the second law of thermodynamics, this fluid characteristic enhances the heat transfer and flow characteristics of the working fluid, improves the overall efficiency of the Rankine cycle, and directly improves the thermal energy conversion efficiency. Therefore, at the same output power, fossil fuel consumption is reduced, and by optimizing combustion control, pollutant generation is effectively suppressed, showing significant advantages in reducing carbon emissions and improving energy sustainability.

[0003] In the dynamic compensation control of supercritical units with variable load rates, there is a core technical pain point of mismatched dynamic responses of multiple subsystems. Especially when the load command changes rapidly, there is a significant time lag and coupling interference between the coordination command output by the variable load rate dynamic programming algorithm and the dynamic response characteristics of the actual execution units (such as boiler feedforward compensation, steam turbine throttle pre-opening, and multi-stage steam temperature compensation). This makes it difficult for the entire control chain to maintain dynamic balance. Specifically, it manifests itself in instability of working fluid energy matching and a decrease in the dynamic compensation accuracy of key parameters (steam temperature and stress) during load changes. For example, if there is no adaptive matching mechanism for the dynamic response time difference between the feedwater flow advance control based on the transfer function model on the boiler side and the dynamic throttle pre-opening action based on fuzzy control on the steam turbine side, the release of boiler heat storage will lag behind the increase in turbine power demand during the load increase phase, which will cause an unexpected drop in main steam pressure, thereby triggering the safety protection system's locking intervention on the variable load rate, limiting the unit's regulation performance. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the present invention provides a supercritical unit variable load rate dynamic compensation control method to solve the problem of dynamic response sequence loss of multiple execution units of boilers and steam turbines in variable load rate regulation.

[0005] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows:

[0006] The present invention provides a method for dynamic compensation control of a supercritical unit with variable load rate, comprising:

[0007] Step 1: Acquire unit operation data, which includes direct parameters, derived parameters, and external instructions, perform credibility detection and anomaly repair on the unit operation data, and output credible operation data;

[0008] Step 2: Based on the trusted operation data, a multi-execution unit response timing mapping table is constructed to quantify the boiler heat storage lag time, the turbine valve delay, and the steam temperature inter-stage coupling coefficient;

[0009] Step 3, using the heat storage lag time in the timing mapping table to correct the boiler feed water feed forward instruction to generate a compensated feed water instruction;

[0010] Step 4: Based on the timing mapping table and the main steam pressure change rate, a genetic algorithm is used to optimize the turbine valve regulating change rate constraint rule to generate a dynamically constrained valve regulating action instruction;

[0011] Step 5: calling the inter-stage coupling coefficient, generating a burner swing angle compensation amount to block the water spray temperature reduction disturbance, and outputting a swing angle compensation instruction;

[0012] Step 6: monitor the safety locking parameters, and trigger the buffer intervention mechanism in a graded manner, which is integrated into the compensated water supply instruction, the dynamically constrained valve adjustment action instruction, and the swing angle compensation instruction.

[0013] Furthermore, in the variable load rate dynamic compensation control method for a supercritical unit according to the present invention, step 1 comprises:

[0014] The main steam pressure dual-point data is collected through triple redundant transmitters, and the steam temperature data is collected through dual thermocouples as basic input parameters;

[0015] Based on the basic input parameters, derived parameters are calculated, wherein the derived parameters include generating a boiler heat storage release rate based on an integrated difference between feed water and steam flow rates, and generating a turbine valve regulating time lag based on a valve instruction and feedback timestamp difference;

[0016] Perform sliding window standard deviation detection on pressure or temperature parameters, and trigger a conflict alarm when the deviation between boiler and steam turbine pressure measurement points exceeds the threshold;

[0017] In response to the conflict warning or detection anomaly, forward linear prediction interpolation or thermodynamic model virtual parameter replacement is used for the abnormal data to output a credible parameter set.

[0018] Furthermore, in the method for dynamic compensation control of variable load rate of a supercritical unit according to the present invention, step 2 comprises:

[0019] Based on the load change rate, the negative exponential function is fitted piecewise, and the boiler heat storage lag time is calibrated according to the boiler heat storage release rate in the credible parameter set.

[0020] Divide the turbine valve delay level into grades according to the discrete intervals of pressure change rate, and associate the turbine valve delay data in the credible parameter set;

[0021] Extract the history data part of the credible parameter set of the history action sequence of the water injection valve and the burner swing angle, and calculate the water injection amount and the swing angle disturbance intensity coefficient;

[0022] After each load change process is completed, the mapping parameters of the timing mapping table are updated according to the actual response value and the estimated deviation.

[0023] Further, the supercritical unit variable load rate dynamic compensation control method comprises the following steps:

[0024] Obtain a target load value, and generate a reference feedwater change amount through a load and feedwater transfer function;

[0025] Read the heat storage lag time, and calculate a time lag compensation amount;

[0026] Superimpose the time lag compensation amount on the reference feedwater change amount to obtain an initial compensation instruction;

[0027] Dynamically adjust the compensation weight according to the load change direction, and perform gain processing on the initial compensation instruction;

[0028] When the drum water level exceeds a threshold value, disable the compensation superposition and gain processing, and output the reference feedwater change amount as a compensated feedwater instruction; otherwise, output the initial compensation instruction after gain processing as the compensated feedwater instruction.

[0029] Further, the supercritical unit variable load rate dynamic compensation control method comprises the following steps:

[0030] Obtain a turbine governing valve delay parameter from the timing mapping table, and monitor the main steam pressure change rate;

[0031] The first genetic algorithm optimizes a governing valve change rate reference curve every 30 seconds according to the turbine governing valve delay parameter and the main steam pressure change rate, minimizes the pressure fluctuation and the thermal stress increase rate, and outputs the reference curve;

[0032] The second genetic algorithm generates a dynamic damping coefficient every 5 seconds by combining the reference curve and the thermal stress state, and restricts the actual governing valve change rate according to the dynamic damping coefficient;

[0033] The third genetic algorithm analyzes the pressure and governing valve action history record to evolve a fuzzy rule library after the variable load ends, and updates the constraint rule set.

[0034] Further, the supercritical unit variable load rate dynamic compensation control method further comprises the following steps:

[0035] The second genetic algorithm performs adaptive adjustment on the dynamic damping coefficient according to the boiler heat storage lag time and the pressure change rate;

[0036] When the thermal stress exceeds the limit value, the second genetic algorithm forces to reduce the upper limit of the damping coefficient;

[0037] When the water level deviation is detected, the second genetic algorithm triggers the coefficient downshift;

[0038] The third genetic algorithm automatically rolls back to the historical version controller when the rule base evolution fails.

[0039] Further, the supercritical unit variable load rate dynamic compensation control method provided by the present application, the step 5 comprises:

[0040] Obtain the water injection action parameter and the combustion zone wall temperature data from the credible operation data, and calculate the combustion zone wall temperature gradient change amount;

[0041] Extract the inter-stage coupling coefficient from the time sequence mapping table, and correct the gain weight according to the real-time load interval;

[0042] Multiply the wall temperature gradient change amount and the corrected gain weight to generate a negative feedback swing angle compensation amount;

[0043] Convert the negative feedback swing angle compensation amount into a swing angle compensation instruction;

[0044] When the main steam temperature change rate exceeds the preset safety threshold, the swing angle compensation instruction is temporarily suspended.

[0045] Further, the supercritical unit variable load rate dynamic compensation control method provided by the present application further comprises:

[0046] Real-time acquisition of the action rate parameter in the dynamic constraint valve action instruction and the thermal stress monitoring value;

[0047] When the action rate parameter exceeds the preset rate threshold, perform a swing angle compensation operation to compress the change rate of the swing angle compensation instruction in proportion;

[0048] When the thermal stress monitoring value exceeds the limit value to trigger the valve constraint, the valve constraint operation sends a disable signal to the swing angle compensation operation, and synchronously disables the gain weight of the swing angle compensation instruction.

[0049] Further, the supercritical unit variable load rate dynamic compensation control method provided by the present application, the step 6 comprises:

[0050] When the water level deviation exceeds the first safety threshold, immediately lock the variable load instruction transmission channel;

[0051] When the superheat degree is lower than the preset lower limit or the thermal stress exceeds the preset upper limit, activate a time-adjustable buffer window;

[0052] In the buffer window period, the compensation feedwater instruction, the dynamic constraint valve action instruction and the swing angle compensation instruction are forced to be parameter corrected.

[0053] Real-time monitoring of overheating degree and thermal stress recovery state;

[0054] If the overheating degree is still lower than the safe lower limit or the thermal stress is still higher than the safe upper limit at the end of the buffer period, the load rate is reduced to the preset safe value.

[0055] Further, the supercritical unit variable load rate dynamic compensation control method of the present application further comprises:

[0056] Extract the lockout event data from the hierarchical buffer intervention process, and use the lockout event data to update the time sequence mapping table;

[0057] Based on the real-time load change rate, the preset mapping relationship is queried to dynamically adjust the duration of the buffer window;

[0058] When the multi-source check alarm is activated, the duration of the buffer window is extended by a preset proportion;

[0059] After the time sequence mapping table is updated, the buffer window duration calculation is retriggered.

[0060] The present application has the following advantages:

[0061] The present application dynamically responds to the time sequence mapping table quantization key parameter delay characteristics through multiple execution units, pre-corrects the phase of the boiler feedwater feedforward instruction by the heat storage lag time at the initial variable load stage, dynamically optimizes the valve action reference curve by synchronously combining the steam turbine valve delay parameter, aligns the boiler energy release and steam turbine power demand response sequence from the root, and suppresses the risk of unexpected drop of main steam pressure; The three-stage compensation mechanism realizes process synchronization optimization: the feedwater time lag compensation improves the steam-feedwater flow matching accuracy, the dynamic damping coefficient constrains the valve action rate to suppress overshoot, and the inter-stage coupling coefficient driven swing angle compensation blocks the chain interference of water spray desuperheating on combustion stability, so that the multiple execution units maintain dynamic balance throughout the variable load stage; The closed-loop feedback mechanism updates the time sequence mapping table parameters by collecting the actual response deviation through the buffer intervention process, adjusts the buffer window duration under self-adaptive working conditions, continuously matches the compensation reference and real-time load change characteristics, and comprehensively improves the variable load rate regulation range and key parameter control stability of the unit. BRIEF DESCRIPTION OF DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, for those skilled in the art, other drawings can also be obtained without creative labor based on the drawings.

[0063] Figure 1 The flowchart of the supercritical unit variable load rate dynamic compensation control method provided by the present application is shown in the figure. DETAILED DESCRIPTION

[0064] In order to make the objects, technical solutions and advantages of the present application clearer, the following will clearly and completely describe the technical solutions of the present application with reference to the specific embodiments of the present application and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application. The technical solutions provided by the embodiments of the present application are described in detail below with reference to the drawings. In order to better understand the objects of the present application, the present application is further described in detail below.

[0065] Referring to Figure 1 The present application provides a supercritical unit variable load rate dynamic compensation control method, comprising:

[0066] Step 1, obtaining unit operation data, the unit operation data including direct parameters, derived parameters and external instructions, performing credibility detection and abnormality repair on the unit operation data, and outputting credible operation data;

[0067] Step 2, based on the credible operation data, constructing a multi-execution unit response time sequence mapping table, quantifying the boiler heat storage lag time, the steam turbine valve delay and the steam temperature inter-stage coupling coefficient;

[0068] Step 3, using the heat storage lag time in the time sequence mapping table to correct the boiler feedwater feedforward instruction, and generating a compensated feedwater instruction;

[0069] Step 4, according to the time sequence mapping table and the main steam pressure change rate, using a genetic algorithm to optimize the steam turbine valve change rate constraint rule, and generating a dynamically constrained valve action instruction;

[0070] Step 5, calling the inter-stage coupling coefficient to generate a burner swing angle compensation amount to block water injection temperature reduction disturbance, and outputting a swing angle compensation instruction;

[0071] Step 6, monitoring safety locking parameters, and triggering a hierarchical buffer intervention mechanism integrated in the compensated feedwater instruction, the dynamically constrained valve action instruction and the swing angle compensation instruction.

[0072] The acquisition of unit operating data in step 1 is the input to the control system. Direct parameters such as main steam pressure and temperature are collected through a redundant sensor network, and feedwater flow is obtained in conjunction with devices such as mass flow meters. Derived parameters are calculated from direct parameters: the integrated difference between feedwater and steam flow reflects the boiler's heat storage release rate in real time, and the difference between valve command and feedback timestamps dynamically calibrates the turbine valve throttle lag. Data credibility management adopts a three-tier architecture: sliding window standard deviation detects transient anomalies, cross-measurement point deviation comparison identifies multi-source conflicts, and communication layer CRC check ensures transmission integrity. The anomaly repair mechanism performs forward interpolation to handle transient faults, calls on thermodynamic model virtual parameters to compensate for persistent deviations, and ultimately outputs a full-dimensional data stream with credibility labels, providing a verified input source for subsequent control.

[0073] The timing mapping table constructed based on trusted operating data in step 2 is the core hub for coordinating multiple execution units. The boiler heat storage lag time is calibrated using a piecewise negative exponential function, and a prediction model is established based on the dynamic characteristics of the heat storage release rate with the load change rate. The steam turbine valve delay level division is achieved by discretizing the pressure change rate interval, and the measured valve lag data is associated with real-time calibration. The steam temperature inter-stage coupling coefficient is calculated by analyzing the disturbance intensity of the historical action sequence of the water spray valve and the burner swing angle, and quantifying the impact of the change in water spray volume on combustion stability. After each load change cycle, the deviation between the actual response time and the estimated value is fed back to the mapping table, and the parameters are updated using the sliding weighting method to enable the timing benchmark to continuously evolve to adapt to changes in the unit's operating status.

[0074] The correction of the boiler feed-water feedforward compensation in step 3 is achieved through two-layer control. The target load value is converted into a reference feed-water change through the load-feed-water transfer function; at the same time, the heat storage lag time in the timing mapping table is read, and the time lag compensation amount is calculated and superimposed on the reference instruction to form the initial compensation instruction. The compensation weight is dynamically adjusted according to the direction of load increase and decrease: the compensation intensity is strengthened in the load increase phase, and the compensation effect is weakened in the load reduction phase. A boiler water level safety protection mechanism is added. When the water level exceeds the threshold, it automatically switches to the non-compensation mode to output the reference instruction to avoid water level instability caused by compensation superposition. A limiter is set at the output end to prevent the instruction from changing suddenly beyond the range of the actuator.

[0075] The optimization of the turbine valve control constraint rules in step 4 uses a collaborative architecture of three genetic algorithms. The first genetic algorithm runs in a 30-second cycle, integrating the valve control delay parameters and the main steam pressure change rate to generate a baseline curve that minimizes pressure fluctuations and thermal stress growth rates. The second genetic algorithm is refreshed every 5 seconds, and a dynamic damping coefficient is applied to the baseline curve based on the current thermal stress state to generate the actual valve control change rate instruction after the constraint. The third genetic algorithm is started after the load change is completed, analyzes the pressure-valve control action records during the actual control process, evolves the fuzzy rule base, and updates the global constraint strategy. The data sharing bus between algorithms transmits thermal stress warnings and water level deviation status in real time, triggering coefficient downshifting or rule rollback protection.

[0076] The burner swing angle compensation in step 5 maintains the steam temperature stable by blocking the water spray cooling disturbance. The water spray action parameters and the combustion zone wall temperature data are extracted from the reliable operation data, and the wall temperature gradient change is calculated to characterize the disturbance intensity. The inter-stage coupling coefficient in the timing mapping table is called, and the gain weight is corrected according to the unit load rate: the gain sensitivity is reduced in the low load range, and the compensation strength is increased in the high load range. The wall temperature change is multiplied by the correction gain to generate a negative feedback compensation amount, which is output as a swing angle control command through the servo converter. When the main steam temperature change rate exceeds the safety threshold, the command output is suspended and the steam temperature protection interlock is activated to prevent the compensation action from exacerbating parameter oscillation.

[0077] Step 6: Security Integration and Dynamic Protection

[0078] The hierarchical buffer intervention mechanism forms a closed-loop safety protection. The first-level interlock responds to the highest priority risk and immediately cuts off the load command transmission channel when the water level exceeds the threshold. The second-level buffer handles secondary risks: when the superheat is insufficient or the thermal stress exceeds the limit, the time-adjustable buffer window is activated, and the water supply command, the valve action command, and the swing angle compensation command are forced to enable the coordinated correction parameters. The window length is dynamically adjusted according to the load change rate. When the load changes drastically, the window is shortened to improve the response speed. The parameter recovery status is verified by a three-out-two logic. If the standard is not met at the end of the buffer period, the load rate is reduced to the preset safety value. At the same time, the event data is fed back to the timing mapping table update module to drive the system self-learning.

[0079] Specifically, the variable load rate dynamic compensation control method for a supercritical unit according to the present invention comprises the following steps:

[0080] The main steam pressure dual-point data is collected through triple redundant transmitters, and the steam temperature data is collected through dual thermocouples as basic input parameters;

[0081] Based on the basic input parameters, derived parameters are calculated, wherein the derived parameters include generating a boiler heat storage release rate based on an integrated difference between feed water and steam flow rates, and generating a turbine valve regulating time lag based on a valve instruction and feedback timestamp difference;

[0082] Perform sliding window standard deviation detection on pressure or temperature parameters, and trigger a conflict alarm when the deviation between boiler and steam turbine pressure measurement points exceeds the threshold;

[0083] In response to the conflict warning or detection anomaly, forward linear prediction interpolation or thermodynamic model virtual parameter replacement is used for the abnormal data to output a credible parameter set.

[0084] The unit's operating data collection utilizes a triple physical redundancy architecture to enhance data reliability. Three high-precision transmitters are deployed at the boiler outlet and turbine inlet, respectively, for main steam pressure. The output signals from each set of transmitters are filtered for valid values ​​using 2oo3 voting logic to eliminate single-point drift errors. Main steam and reheat steam temperature monitoring utilizes dual armored thermocouples, each connected to an independent isolated transmitter channel to prevent common-mode errors caused by signal cross-interference. This ensures that basic input parameters are validated as physical quantities verified at the physical layer.

[0085] The derived parameter calculation engine is built into the real-time processing unit of the distributed control system. The boiler's stored heat release rate is determined by synchronously collecting feedwater and steam flow rates using a high-precision electromagnetic flowmeter. A 100ms periodic integral difference calculation is performed on these two signals to dynamically quantify the boiler's stored energy release intensity. The turbine valve throttle lag parameter quantification module uses a high-precision clock to record the time difference between the issuance of the valve control command and the arrival of the valve position feedback, accurately reflecting the mechanical delay characteristics of the actuator. These derived parameters provide a dynamic benchmark for multi-system coordinated control.

[0086] The data credibility verification mechanism utilizes a two-tier cascaded detection strategy. The first-tier detection analyzes the standard deviation of the raw pressure and temperature data streams using a one-second sliding window. Fluctuations exceeding 2% of the full scale for three consecutive sampling points indicate a transient credibility degradation. The second-tier detection compares the data from the boiler and turbine pressure measurement points in real time. If the deviation exceeds the 0.5 MPa threshold and persists for more than three seconds, a cross-domain data conflict alarm is triggered and a backup communication channel is activated. The anomaly flag signal is then transmitted simultaneously to the repair processing unit.

[0087] The abnormal data processing unit executes differentiated repair logic based on the tag type. Data points with instantaneous credibility degradation trigger the forward linear predictor, which constructs a linear regression equation based on the first three valid sampling points to generate a replacement value. Cross-domain conflict alarms trigger the thermodynamic equilibrium model reconstruction module, which outputs virtual pressure parameters based on the current fuel input, main steam flow, and the law of conservation of energy. The repaired data is reconstructed by aligning the timestamp with the original data stream, and the output data set with a credibility indicator is used for subsequent control modules.

[0088] Specifically, in the variable load rate dynamic compensation control method for a supercritical unit according to the present invention, step 2 includes:

[0089] Based on the load change rate, the negative exponential function is fitted piecewise, and the boiler heat storage lag time is calibrated according to the boiler heat storage release rate in the credible parameter set.

[0090] Divide the turbine valve delay level into discrete intervals according to the pressure change rate, and associate the turbine valve delay data in the credible parameter set;

[0091] Extract the historical data of the credible parameter set of the historical action sequence of the water spray valve and the burner swing angle, and calculate the water spray volume and the swing angle disturbance intensity coefficient;

[0092] After each load change process is completed, the mapping parameters of the timing mapping table are updated according to the actual response value and the estimated deviation.

[0093] The unit's operating data collection utilizes a triple physical redundancy architecture to enhance data reliability. Three high-precision transmitters are deployed at the boiler outlet and turbine inlet, respectively, for main steam pressure. The output signals from each set of transmitters are filtered for valid values ​​using 2oo3 voting logic to eliminate single-point drift errors. Main steam and reheat steam temperature monitoring utilizes dual armored thermocouples, each connected to an independent isolated transmitter channel to prevent common-mode errors caused by signal cross-interference. This ensures that basic input parameters are validated as physical quantities verified at the physical layer.

[0094] The derived parameter calculation engine is built into the real-time processing unit of the distributed control system. The boiler's stored heat release rate is determined by synchronously collecting feedwater and steam flow rates using a high-precision electromagnetic flowmeter. A 100ms periodic integral difference calculation is performed on these two signals to dynamically quantify the boiler's stored energy release intensity. The turbine valve throttle lag parameter quantification module uses a high-precision clock to record the time difference between the issuance of the valve control command and the arrival of the valve position feedback, accurately reflecting the mechanical delay characteristics of the actuator. These derived parameters provide a dynamic benchmark for multi-system coordinated control.

[0095] The data credibility verification mechanism utilizes a two-tier cascaded detection strategy. The first-tier detection analyzes the standard deviation of the raw pressure and temperature data streams using a one-second sliding window. Fluctuations exceeding 2% of the full scale for three consecutive sampling points indicate a transient credibility degradation. The second-tier detection compares the data from the boiler and turbine pressure measurement points in real time. If the deviation exceeds the 0.5 MPa threshold and persists for more than three seconds, a cross-domain data conflict alarm is triggered and a backup communication channel is activated. The anomaly flag signal is then transmitted simultaneously to the repair processing unit.

[0096] The abnormal data processing unit executes differentiated repair logic based on the tag type. Data points with instantaneous credibility degradation trigger the forward linear predictor, which constructs a linear regression equation based on the first three valid sampling points to generate a replacement value. Cross-domain conflict alarms trigger the thermodynamic equilibrium model reconstruction module, which outputs virtual pressure parameters based on the current fuel input, main steam flow, and the law of conservation of energy. The repaired data is reconstructed by aligning the timestamp with the original data stream, and the output data set with a credibility indicator is used for subsequent control modules.

[0097] Specifically, in the variable load rate dynamic compensation control method for a supercritical unit according to the present invention, step 3 includes:

[0098] Obtain the target load value and generate the reference water supply change through the load and water supply transfer function;

[0099] Reading the heat storage lag time and calculating the time lag compensation amount;

[0100] Superimposing the time lag compensation amount on the reference water supply change amount to obtain an initial compensation instruction;

[0101] Dynamically adjust the compensation weight according to the load change direction and perform gain processing on the initial compensation instruction;

[0102] When the drum water level exceeds the threshold, the compensation superposition and gain processing are disabled, and the reference feed water change is output as the compensated feed water instruction; otherwise, the initial compensation instruction after gain processing is output as the compensated feed water instruction.

[0103] The boiler's thermal storage lag time is calibrated using a dynamic load response model. The load change rate is divided into segments based on their direction and magnitude. During the load increase phase, a negative exponential decay function is used, with the decay coefficient inversely proportional to the load change rate. During the load decrease phase, a linear correction model is used. The segmented function coefficients are regressed and calibrated using historical data on boiler thermal storage release rates from a trusted parameter set. This allows for real-time correlation between the boiler's thermal storage release intensity and the variable load rate. The calibrated thermal storage lag time accurately reflects the boiler's thermal inertia dynamics.

[0104] The turbine valve delay level architecture establishes a pressure change response spectrum. The main steam pressure rate of change is divided into three discrete intervals: the low-pressure rate of change interval (<0.2 MPa / s) corresponds to the long delay level, the medium-pressure rate of change interval (0.2-0.5 MPa / s) corresponds to the medium delay level, and the high-pressure rate of change interval (≥0.5 MPa / s) corresponds to the short delay level. The thresholds for each interval are adaptively calibrated based on the turbine valve delay data from the trusted parameter set, achieving a precise match between the dynamic pressure response characteristics and the mechanical delay characteristics.

[0105] The quantification of inter-stage steam temperature disturbance coupling is based on historical action sequence analysis. The system extracts historical operation records of the water spray control valve opening and burner swing angle stored in the trusted parameter set and calculates the disturbance intensity ratio of the water spray change ΔW_p to the swing angle offset Δθ. The disturbance intensity coefficient is calculated using an integral accumulation method over the variable load cycle to eliminate the influence of transient fluctuations. This coefficient represents the degree to which the water spray desuperheating operation interferes with combustion stability.

[0106] A deviation-driven mechanism is established for updating the parameters in the time series mapping table. After each variable load cycle, the actual duration of the boiler's thermal storage release, the actual response time of the turbine throttle valve, and the actual intensity of the water spray-angle disturbance are collected. The absolute deviation rate between the actual value and the mapping table's estimated value is calculated. If the deviation rate exceeds 15%, the parameter update program is activated. A sliding weighted average method (with a historical weight of 0.6 and a new data weight of 0.4) is used to refresh the thermal storage lag time model parameters, throttle valve delay level threshold, and disturbance intensity coefficient in the mapping table, ensuring that the time series benchmark continuously approaches the unit's true dynamic characteristics.

[0107] Specifically, in the variable load rate dynamic compensation control method for a supercritical unit according to the present invention, step 4 includes:

[0108] Obtaining the turbine valve regulating delay parameter from the timing mapping table and monitoring the main steam pressure change rate at the same time;

[0109] The first genetic algorithm optimizes the valve change rate reference curve every 30 seconds based on the turbine valve delay parameter and the main steam pressure change rate, aiming to minimize pressure fluctuations and thermal stress growth rate, and outputs the reference curve;

[0110] The second genetic algorithm generates a dynamic damping coefficient every 5 seconds by combining the reference curve and the thermal stress state, and constrains the actual change rate of the throttle valve accordingly;

[0111] After the load change is completed, the third genetic algorithm analyzes the pressure and valve action history records to evolve the fuzzy rule base and update the constraint rule set.

[0112] The target load value is acquired in real time through the grid dispatch system interface and converted into a load change rate parameter by the command parsing unit. The load-feedwater transfer function uses a preset function structure. Based on the unit's design characteristics, a mathematical relationship is established between the load change and the boiler feedwater flow rate. The output reference feedwater change serves as the feedforward control reference value. This reference value is constrained by the actuator limiter module to prevent it from exceeding the feedwater pump's regulation range.

[0113] The time-lag compensation calculation module reads the thermal storage lag time parameter generated in step 2 and establishes a compensation model based on the dynamic characteristics of the boiler's energy storage release. The compensation value is calculated using a negative exponential function: ΔW_c = α × ΔW_base × e^(-β × T_b), where α is the system characteristic coefficient and β is the dynamic attenuation factor. A bidirectional limiter is placed at the compensation value output to prevent the compensation value amplitude from exceeding the regulation capability of the field equipment.

[0114] The compensation superposition control unit uses an adder to combine the baseline command and the compensation amount to generate an initial compensation command. This initial command is transmitted to the gain processing module, which configures a variable weighting factor based on the load change direction. During the load increase phase, the gain factor γ is set to greater than 1 (γ = 1.2 - 0.02 × dL / dt) to increase the compensation strength; during the load decrease phase, γ = 0.8 is used to reduce the compensation effect. Gain processing is implemented using a digital multiplier to avoid analog circuit drift errors.

[0115] The safety protection mechanism features an independent monitoring channel. Drum water level parameters are verified using a two-out-of-three voting logic. If a preset safety threshold is exceeded, the override control unit is activated: the compensation overlay path is forcibly disconnected, the gain processing module is bypassed, and the original baseline feedwater change is directly passed to the output. In the non-override state, the initial compensation command after gain processing is generated through the output buffer to generate the final compensated feedwater command. Control command output utilizes an industrial bus protocol to ensure timely actuator response.

[0116] Specifically, the variable load rate dynamic compensation control method for a supercritical unit according to the present invention further includes:

[0117] The second genetic algorithm performs adaptive adjustment on the dynamic damping coefficient according to the boiler heat storage lag time and the pressure change rate;

[0118] When the thermal stress exceeds the limit, the second genetic algorithm is forced to reduce the upper limit of the damping coefficient;

[0119] When water level deviation is detected, the second genetic algorithm triggers coefficient downshifting;

[0120] When the third genetic algorithm fails to evolve the rule base, it automatically rolls back to the historical version controller.

[0121] The second genetic algorithm implements an adaptive adjustment mechanism for the dynamic damping coefficient. The input sources are the boiler heat storage lag time and the real-time main steam pressure change rate quantified in step 2. The optimal damping coefficient is calculated using a multi-objective optimization function. With pressure fluctuation suppression and unit stress control as the core objectives, the coefficient is dynamically updated based on the boiler-turbine energy balance constraints. This adjustment process is performed every 5 seconds, and the output damping coefficient is applied to the main throttle rate constraint loop.

[0122] The thermal stress safety protection subsystem establishes forced coefficient reduction logic. It receives real-time monitoring of the high-pressure rotor's thermal stress. When the stress exceeds a preset upper threshold, the second genetic algorithm activates a protection protocol: the upper limit of the dynamic damping coefficient is locked to 80% of the original setting. This forced reduction continues until the stress falls back to a safe range. This reduction is achieved by modifying the optimization function's constraint boundaries to avoid sudden changes in instructions caused by hard switching.

[0123] Water level deviation triggers a coefficient gradient downshift mechanism. The deviation between the drum water level and the setpoint is monitored, and the water level protection priority strategy is activated when the deviation exceeds the deadband threshold. A second genetic algorithm lowers the currently operating damping coefficient. The downshift step size is positively correlated with the water level deviation, with each 10mm increase in deviation corresponding to a 0.1 coefficient downshift. The downshift result is applied in real time to the throttle constraint command generation circuit to accelerate the restoration of water level equilibrium.

[0124] The third genetic algorithm configures a path for handling rule base evolution anomalies. After each fuzzy rule base evolution, the control stability indicators of the new rules are verified. If the pressure overshoot rate exceeds 15% or the thermal stress growth rate exceeds the warning value during simulation testing, the evolution is considered a failure. At this time, the rule rollback procedure is automatically triggered, extracting the valid rule set from the historical version library for the last three cycles, overwriting the abnormal rule base, and reactivating the controller output.

[0125] Specifically, in the variable load rate dynamic compensation control method for a supercritical unit according to the present invention, step 5 includes:

[0126] Obtain water spray volume action parameters and combustion zone wall temperature data from reliable operation data, and calculate the temperature gradient change of the combustion zone wall;

[0127] Extracting the inter-stage coupling coefficient from the timing mapping table and correcting the gain weight according to the real-time load interval;

[0128] Multiplying the wall temperature gradient change by the corrected gain weight to generate a negative feedback swing angle compensation amount;

[0129] converting the negative feedback swing angle compensation amount into a swing angle compensation instruction;

[0130] When the main steam temperature change rate exceeds a preset safety threshold, the output of the swing angle compensation instruction is suspended.

[0131] Water injection rate parameters are acquired in real time through the boiler feedwater control system, which simultaneously receives temperature data from a multi-point armored thermocouple array located in the burner area. The calculation of the combustion zone wall temperature gradient utilizes a time-domain processing strategy: when the unit load factor is below 75%, a sliding mean filter combined with a gradient difference algorithm is applied to extract low-frequency variation trends. Above 75%, a fast Fourier transform is used to isolate high-frequency disturbance components. This process eliminates measurement noise and accurately characterizes the intensity of thermal disturbances caused by water injection cooling.

[0132] The interstage coupling coefficient uses the quantized parameters stored in the timing mapping table in step 2 to adjust the gain weights segmented by the current load level. The load condition is divided into three characteristic ranges: the low-load range (30%-50% of rated load) uses a correction factor of 0.6 to reduce compensation sensitivity; the medium-load range (50%-80%) uses a factor of 1.0; and the high-load range (80%-100%) increases the factor to 1.3 to enhance response. The weight correction module is implemented using a digital multiplier to prevent the impact of temperature drift on analog circuit accuracy.

[0133] Negative feedback swing angle compensation is generated using a vector dot product. The wall temperature gradient change array is dot-multiplied with the corrected gain weight matrix to output the compensation value array. A rate limiter is implemented at the compensation output to constrain the compensation change rate to within 2% / s to prevent actuator overshoot and combustion instability. The compensation amplitude is also protected by the burner's mechanical travel limits. If the calculated result exceeds the physical travel range, it is automatically scaled to a feasible range.

[0134] The swing angle compensation command conversion utilizes a two-stage processing model. The digital compensation value is input into a D / A converter to generate an analog voltage signal, which is then simultaneously converted to a 4-20mA current drive signal via a servo amplifier. Before the current signal is transmitted to the burner's swing angle actuator, an electrical isolation barrier is added to eliminate ground loop interference. A timeout monitoring mechanism is implemented in the command transmission channel, automatically switching to a redundant channel if the signal delay exceeds 500ms.

[0135] The main steam temperature safety interlock system deploys three independent temperature measurement points. A two-out-of-three logic vote is performed on the temperature change rate at these three points. When the preset safety threshold is exceeded, a protective action is immediately triggered: a lockout command is sent to the swing angle compensation output module, freezing the current command output. The lockout state persists until the temperature change rate falls back into the safe zone and remains stable for three seconds, after which it automatically resets and restores the command transmission channel. The interlock signal also activates feedforward compensation in the steam temperature control system, assisting in rapid temperature stabilization.

[0136] Specifically, the variable load rate dynamic compensation control method for a supercritical unit according to the present invention further includes:

[0137] Real-time acquisition of the action rate parameters and thermal stress monitoring values ​​in the dynamic constraint valve action instructions;

[0138] When the action rate parameter exceeds a preset rate threshold, performing a swing angle compensation operation to proportionally compress the rate of change of the swing angle compensation instruction;

[0139] When the thermal stress monitoring value exceeds the limit and triggers the valve constraint, the valve constraint operation sends a disable signal to the swing angle compensation operation, and synchronously disables the gain weight of the swing angle compensation instruction.

[0140] The throttle valve action rate parameter is obtained in real time from the dynamically constrained throttle valve action command data packet. This parameter represents the instantaneous rate of change of the turbine throttle valve opening. Thermal stress monitoring values ​​are collected by an array of embedded sensors in the high-pressure rotor and calculated using a temperature-stress conversion model to output a digital signal. These two parameters are synchronously transmitted to the collaborative decision-making unit via the control bus, with a transmission delay of less than 50 milliseconds to ensure timeliness.

[0141] The dynamic compression mechanism for swing angle compensation utilizes linear proportional regulation. When the rate parameter exceeds a preset rate threshold (e.g., 5% / s), the decision unit calculates the ratio of the current swing angle compensation command rate of change to the rate limit threshold. The upper limit of the rate of change is compressed by the same amount as the rate limit. The compression operation is dynamic within the execution cycle and is automatically released when the rate parameter falls back below the threshold.

[0142] A hard-interlocked logic is established for the disable signal triggered by the throttle constraint. When the throttle constraint protection is triggered by an over-limit thermal stress monitoring value, the constraint operation generates a 32-bit disable instruction code. This instruction code is broadcast over the real-time data bus and carries a timestamp and information about the stress level exceeded. A signal decoder is configured on the receiving end of the swing angle compensation operation. Upon recognizing a valid disable code, it performs three actions: freezing the gain weight regulator, maintaining the current gain coefficient constant, and blocking external weight correction requests. The disable state persists until the thermal stress returns to the safe zone and a constraint release signal is received.

[0143] Specifically, in the variable load rate dynamic compensation control method for a supercritical unit according to the present invention, step 6 includes:

[0144] When the water level deviation exceeds the first safety threshold, the load variable instruction transmission channel is immediately blocked;

[0145] When the superheat falls below the preset lower limit or the thermal stress exceeds the preset upper limit, a buffer window with adjustable time is activated;

[0146] During the buffer window period, the post-compensation water supply instruction, the dynamically constrained valve adjustment action instruction and the swing angle compensation instruction are forcibly enabled to perform parameter correction;

[0147] Real-time monitoring of superheat and thermal stress recovery status;

[0148] If the superheat is still below the safety lower limit or the thermal stress is still above the safety upper limit at the end of the buffer period, the load rate will be reduced to the preset safety value.

[0149] Water level deviation monitoring utilizes three channels of redundant measurement signals. When any water level measurement exceeds the first safety threshold (typically set at ±50mm), a first-level lockout protection is immediately triggered: a hardware-level interrupt signal is sent to the load command transmission bus, physically disconnecting the command transmission path. This lockout state persists until the water level returns to the safe zone for three seconds, at which point the system automatically resets the connection channel and restores command flow. This mechanism blocks load adjustments under unsafe conditions.

[0150] Superheat and thermal stress safety monitoring utilizes dual-factor triggering logic. An adjustable time buffer window is activated when the superheat monitoring value falls below a preset lower limit (e.g., 15°C superheat) or when the high-pressure rotor thermal stress exceeds a preset upper limit (e.g., 80% of allowable stress). The initial duration of the window is dynamically set based on the current load change rate: 1 second for load changes ≤ 3% / min, and 0.5 seconds for changes > 3% / min. The window activation signal simultaneously initiates the cross-system coordinated command readiness sequence.

[0151] During the buffer window, three types of compensation commands are forcibly activated. The compensated feedwater command skips the normal control cycle and is directly output to the boiler feedwater valve. The dynamically constrained throttle action command applies a double control gain to accelerate response. The swing angle compensation command removes rate-of-change limiting. These three types of commands work together to maximize correction intensity, with their output taking precedence over the normal control loop, directly driving the actuator to restore critical parameters.

[0152] Recovery status monitoring utilizes a millisecond-level refresh mechanism. The superheat recovery indicator is required to return to a preset lower limit of +5°C (typically 20°C overheat). The thermal stress recovery target is to fall back to a preset upper limit of -10% (typically 70% of allowable stress). Real-time parameters are collected every 100 milliseconds. Recovery is considered successful if three consecutive sampling points meet the criteria, terminating the buffer window early.

[0153] If the target is not met at the end of the buffer period, load rate control will be implemented. If either the superheat or thermal stress indicator fails to meet the safety recovery criteria at the end of the predetermined buffer period, the control unit will send a load reduction command to the load regulation system. The target load rate is directly reduced to a preset safety value (typically 1% / min). The load reduction process continues until all safety parameters enter the steady-state safety zone. During the load reduction command, load increase requests are blocked to prevent deterioration of operating conditions.

[0154] Specifically, the variable load rate dynamic compensation control method for a supercritical unit according to the present invention further includes:

[0155] extracting blocking event data from the hierarchical buffer intervention process, and using the blocking event data to update the timing mapping table;

[0156] Querying a preset mapping relationship based on the real-time load change rate, and dynamically adjusting the duration of the buffer window;

[0157] When the multi-source verification alarm is activated, the length of the buffer window is extended according to a preset ratio;

[0158] After the timing mapping table is updated, the calculation of the buffer window duration is retriggered.

[0159] The lockout event data extraction mechanism is deployed in the hierarchical buffer intervention execution unit. When a water level lockout or buffer window activation event is triggered, the event type, trigger time, key parameter deviation, and recovery time are recorded in real time to generate a structured event log. The event log is filtered through a data cleaning module to filter invalid records and extract a dataset of valid lockout features. This dataset is aligned with historical operating data using timestamps and then input into the time series mapping table update engine to drive parameter optimization.

[0160] Dynamic adjustment of the buffer window duration is based on a load change rate response model. A preset mapping relationship between load change rate and optimal buffer duration is established: a load change rate ≤ 2% / min corresponds to a base duration of 1 second. Each 1% / min increase in load change compresses the duration by 0.1 seconds, up to a maximum of 0.5 seconds. This adjustment is performed before each buffer window activation. The current load change rate is queried in real time and the target duration is calculated using interpolation. The calculated duration is written to the buffer controller configuration register and takes effect immediately during the intervention period.

[0161] The multi-source verification alarm response mechanism configures a priority handling strategy. When a pressure measurement point conflict alarm or temperature data verification anomaly is activated, an extension factor is applied based on the alarm level: a level 1 alarm (single measurement point failure) extends the window by 20%, and a level 2 alarm (multiple measurement point conflicts) extends the window by 50%. This extension is applied after the base duration calculation to generate the final effective window duration. Once the alarm is resolved, the base calculation rules are automatically restored, eliminating the effect of the extension factor.

[0162] Updates to the timing mapping table trigger window recalculation. When the mapping table parameters are updated, a version change event is generated, and the event signal is transmitted to the buffer control unit. The control unit immediately initiates the window duration recalculation process: it reloads the load-duration relationship parameters from the latest mapping table and updates the window duration baseline value based on the real-time load change rate. The new baseline value only applies to subsequent intervention cycles; the currently executing buffer window maintains its original duration, preventing control instability caused by configuration changes during operation.

[0163] This invention achieves multi-executor response sequence alignment by constructing a dynamic timing mapping table. Based on trusted operating data, the boiler's thermal storage lag, turbine throttle delay, and steam temperature interstage coupling coefficient are quantified to generate a multi-executor response timing mapping table. This mapping table serves as a coordination benchmark, enabling boiler-side feedwater feedforward compensation to pre-correct the command phase based on the thermal storage lag. Furthermore, turbine throttle constraints are combined with throttle delay parameters to dynamically optimize action timing, fundamentally eliminating response sequence misalignment caused by command transmission lag.

[0164] A three-level dynamic compensation mechanism achieves real-time synchronization of the response process. The three types of compensation instructions generated in steps 3, 4, and 5, respectively, work in synergy: boiler feedwater lag compensation eliminates energy supply delays, turbine throttle dynamic damping constraints suppress overshoot, and swing angle compensation blocks the propagation of water spray cooling disturbances. In the hierarchical buffer intervention of step 6, the three types of instructions are forcibly activated for joint correction. When insufficient superheat or excessive thermal stress is detected, the cross-system coordination of feedwater, throttle, and swing angle compensation is activated through the buffer window, correcting deviations in the execution unit's actions in real time.

[0165] A closed-loop feedback system maintains long-term synchronization accuracy. After each load-variable cycle, the timing mapping table parameters are updated based on the actual response value and the estimated deviation. Lockout event data is fed back to the mapping table update module to continuously optimize the coordination benchmark. The window duration is recalculated based on the load change rate to adapt to new operating conditions. This mechanism enables the coordination benchmark to evolve dynamically, compensating for the impact of long-term changes in equipment characteristics on synchronization accuracy.

Claims

1. A method for dynamic compensation control of variable load rate of a supercritical unit, characterized in that: include: Step 1: Acquire unit operation data, which includes direct parameters, derived parameters, and external instructions, perform credibility detection and anomaly repair on the unit operation data, and output credible operation data; Step 2: Based on the trusted operation data, a multi-execution unit response timing mapping table is constructed to quantify the boiler heat storage lag time, the turbine valve delay, and the steam temperature inter-stage coupling coefficient; Step 3, using the heat storage lag time in the timing mapping table to correct the boiler feed water feed forward instruction to generate a compensated feed water instruction; Step 4: Based on the timing mapping table and the main steam pressure change rate, a genetic algorithm is used to optimize the turbine valve regulating change rate constraint rule to generate a dynamically constrained valve regulating action instruction; Step 5: calling the inter-stage coupling coefficient, generating a burner swing angle compensation amount to block the water spray temperature reduction disturbance, and outputting a swing angle compensation instruction; Step 6: monitor the safety locking parameters, and integrate the hierarchical triggering buffer intervention mechanism into the compensated water supply instruction, the dynamically constrained valve adjustment action instruction, and the swing angle compensation instruction.

2. The method for dynamic compensation control of variable load rate of a supercritical unit according to claim 1, characterized in that: The step 1 comprises: The main steam pressure dual-point data is collected through triple redundant transmitters, and the steam temperature data is collected through dual thermocouples as basic input parameters; Based on the basic input parameters, derived parameters are calculated, wherein the derived parameters include generating a boiler heat storage release rate based on an integrated difference between feed water and steam flow rates, and generating a turbine valve regulating time lag based on a valve instruction and feedback timestamp difference; Perform sliding window standard deviation detection on pressure or temperature parameters, and trigger a conflict alarm when the deviation between boiler and steam turbine pressure measurement points exceeds the threshold; In response to the conflict warning or detection anomaly, forward linear prediction interpolation or thermodynamic model virtual parameter replacement is used for the abnormal data to output a credible parameter set.

3. The supercritical unit variable load rate dynamic compensation control method according to claim 2, characterized in that: The step 2 includes: Based on the load change rate, the negative exponential function is fitted piecewise, and the boiler heat storage lag time is calibrated according to the boiler heat storage release rate in the credible parameter set. Divide the turbine valve delay level into discrete intervals according to the pressure change rate, and associate the turbine valve delay data in the credible parameter set; Extract the historical data of the credible parameter set of the historical action sequence of the water spray valve and the burner swing angle, and calculate the water spray volume and the swing angle disturbance intensity coefficient; After each load change process is completed, the mapping parameters of the timing mapping table are updated according to the actual response value and the estimated deviation.

4. The method for dynamic compensation control of variable load rate of a supercritical unit according to claim 3, characterized in that: The step 3 includes: Obtain the target load value and generate the reference water supply change through the load and water supply transfer function; Reading the heat storage lag time and calculating the time lag compensation amount; Superimposing the time lag compensation amount on the reference water supply change amount to obtain an initial compensation instruction; Dynamically adjust the compensation weight according to the load change direction and perform gain processing on the initial compensation instruction; When the drum water level exceeds the threshold, the compensation superposition and gain processing are disabled, and the reference feed water change is output as the compensated feed water instruction; otherwise, the initial compensation instruction after gain processing is output as the compensated feed water instruction.

5. The method for dynamic compensation control of variable load rate of a supercritical unit according to claim 4, characterized in that: The step 4 comprises: Obtaining the turbine valve regulating delay parameter from the timing mapping table and monitoring the main steam pressure change rate at the same time; The first genetic algorithm optimizes the valve change rate reference curve every 30 seconds based on the turbine valve delay parameter and the main steam pressure change rate, aiming to minimize pressure fluctuations and thermal stress growth rate, and outputs the reference curve; The second genetic algorithm generates a dynamic damping coefficient every 5 seconds by combining the reference curve and the thermal stress state, and constrains the actual change rate of the throttle valve accordingly; After the load change is completed, the third genetic algorithm analyzes the pressure and valve action history records to evolve the fuzzy rule base and update the constraint rule set.

6. The method for dynamic compensation control of variable load rate of a supercritical unit according to claim 5, characterized in that: Also includes: The second genetic algorithm performs adaptive adjustment on the dynamic damping coefficient according to the boiler heat storage lag time and the pressure change rate; When the thermal stress exceeds the limit, the second genetic algorithm is forced to reduce the upper limit of the damping coefficient; When water level deviation is detected, the second genetic algorithm triggers coefficient downshifting; When the third genetic algorithm fails to evolve the rule base, it automatically rolls back to the historical version controller.

7. The method for dynamic compensation control of variable load rate of a supercritical unit according to claim 6, characterized in that: The step 5 comprises: Obtain water spray volume action parameters and combustion zone wall temperature data from reliable operation data, and calculate the temperature gradient change of the combustion zone wall; Extracting the inter-stage coupling coefficient from the timing mapping table and correcting the gain weight according to the real-time load interval; Multiplying the wall temperature gradient change by the corrected gain weight to generate a negative feedback swing angle compensation amount; converting the negative feedback swing angle compensation amount into a swing angle compensation instruction; When the main steam temperature change rate exceeds a preset safety threshold, the output of the swing angle compensation instruction is suspended.

8. The method for dynamic compensation control of variable load rate of a supercritical unit according to claim 7, characterized in that: Also includes: Real-time acquisition of the action rate parameters and thermal stress monitoring values ​​in the dynamic constraint valve action instructions; When the action rate parameter exceeds a preset rate threshold, performing a swing angle compensation operation to proportionally compress the rate of change of the swing angle compensation instruction; When the thermal stress monitoring value exceeds the limit and triggers the valve constraint, the valve constraint operation sends a disable signal to the swing angle compensation operation, and synchronously disables the gain weight of the swing angle compensation instruction.

9. The method for dynamic compensation control of variable load rate of a supercritical unit according to claim 8, characterized in that: The step 6 comprises: When the water level deviation exceeds the first safety threshold, the load variable instruction transmission channel is immediately blocked; When the superheat falls below the preset lower limit or the thermal stress exceeds the preset upper limit, a buffer window with adjustable time is activated; During the buffer window period, the post-compensation water supply instruction, the dynamically constrained valve adjustment action instruction and the swing angle compensation instruction are forcibly enabled to perform parameter correction; Real-time monitoring of superheat and thermal stress recovery status; If the superheat is still below the safety lower limit or the thermal stress is still above the safety upper limit at the end of the buffer period, the load rate will be reduced to the preset safety value.

10. The method for dynamic compensation control of variable load rate of a supercritical unit according to claim 9, characterized in that: Also includes: extracting blocking event data from the hierarchical buffer intervention process, and using the blocking event data to update the timing mapping table; Querying a preset mapping relationship based on the real-time load change rate, and dynamically adjusting the duration of the buffer window; When the multi-source verification alarm is activated, the length of the buffer window is extended according to a preset ratio; After the timing mapping table is updated, the calculation of the buffer window duration is retriggered.

Images