A mechanism data fusion dynamic simulation method for a conventional island of a photo-thermal power station

Abstract

The application discloses a kind of photothermal power station conventional island mechanism data fusion dynamic simulation method, belong to the technical field of photothermal power station conventional island, including steps S1: establishing conventional island dynamic mechanism model;Step S2: obtain the real-time data or historical data input on conventional island, pre-process data to form measurement value;Step S3: based on dynamic mechanism model and measurement value, execute online state estimation, obtain unmeasurable state;Step S4: execute slowly varying parameter online identification, obtain key output variable;Step S5: calculate the residual error of model output and measurement value, train and update residual error correction module, output correction quantity and give confidence;Step S6: output corrected twin state, performance index and health index.The present application provides a kind of solution to the problems of insufficient accuracy, parameters cannot be traced back, high quality requirements for training data, weak online correction ability for continuous operation in the field under the dynamic working condition of prior art.

Description

Technical Field

[0001] This invention relates to the technical field of conventional islands in solar thermal power plants, and in particular to a dynamic simulation method for the mechanistic data fusion of conventional islands in solar thermal power plants. Background Technology

[0002] In current calculation methods for the conventional island of concentrated solar power (CSP) plants, the main problems in the performance analysis and simulation of the conventional island of CSP systems include: (1) Limited to steady-state thermodynamic calculation / heat balance: It can be used for design conditions and some set conditions, but it is not adaptable to dynamic operation processes such as start-up, load change, bypass switching, and steam extraction fluctuations. (2) Insufficient simulation accuracy of dynamic mechanism model under changing working conditions: The physical interpretability of the simulation model is strong, but the model parameters (heat transfer coefficient, pressure loss, valve flow coefficient, efficiency correction, etc.) will change due to scaling, air leakage, wear, etc. during long-term operation in the field, and some parameters are difficult to measure directly, resulting in insufficient simulation accuracy of the model in the real-time simulation of changing working conditions. (3) Pure data-driven modeling has poor transferability: it is highly dependent on the amount and quality of data, making it difficult to guarantee the ability to extrapolate across operating conditions. Furthermore, it lacks physical constraints and interpretability, making it difficult to use for online analysis and closed-loop applications related to operational safety.

[0003] Therefore, existing technologies generally suffer from problems such as insufficient accuracy under dynamic operating conditions, untraceable parameters, high requirements for training data quality, and weak online correction capabilities for continuous on-site operation. Summary of the Invention

[0004] The purpose of this invention is to provide a dynamic simulation method for the fusion of mechanism data of conventional island in a solar thermal power plant. By providing a method for the fusion of mechanism data of conventional island in a solar thermal power plant, a fusion scheme of mechanism and data is formed, and a simulation scheme is formed based on the fusion scheme. This invention provides a solution to the problems of insufficient accuracy, untraceable parameters, high requirements for training data quality, and weak online correction capability for continuous on-site operation in existing technologies under dynamic operating conditions.

[0005] To achieve the above objectives, this invention provides a dynamic simulation method for the mechanistic data fusion of a conventional island in a solar thermal power plant, comprising the following steps: Step S1: Based on the layout architecture of the conventional island of the solar thermal power plant, establish a dynamic mechanism model of the conventional island; Step S2: Obtain real-time or historical data input from the regular island and preprocess the data to form measurement values; Step S3: Based on the dynamic mechanism model and measured values, perform online state estimation to obtain the unmeasurable state; Step S4: Perform online identification of slowly varying parameters to obtain key output variables; Step S5: Calculate the residual between the model output and the measured value, train and update the residual correction module, output the correction value and give the confidence level; Step S6: Output and record the corrected twin status, performance indicators and health indicators.

[0006] Preferably, in step S1, the process of establishing the dynamic mechanism model of a conventional island is as follows: The modeling objects of the conventional island dynamic mechanism model include the steam generation system, turbine unit, condenser or air-cooled system, deaerator, feedwater pump, high and low pressure heaters, bypass system, main steam and reheat steam pipelines, extraction steam pipelines, and related volumetric components. Based on the thermodynamic flow and equipment connection relationships of the conventional island system, the model formulas are as follows: ; ; In the above formula, This represents the rate of change of the internal state of a regular island with respect to time. Represents the state change function. This represents the calculation process of the mechanism model. These are the internal state variables of the conventional island, including the internal temperature, pressure, enthalpy, metal wall temperature, volumetric pressure, and liquid level of each piece of equipment; External input parameters include inlet pressure, inlet temperature, inlet flow rate, valve opening, and pump speed collected by the DCS. These are the model parameters, including heat exchanger UA, pressure loss coefficient, valve flow coefficient, turbine efficiency correction coefficient, and heat loss coefficient; The output of the mechanistic model includes the temperature, pressure, flow rate, and turbine power of the working fluid at the hot and cold sides of the equipment outlet.

[0007] Preferably, in step S2, the process of acquiring and preprocessing the input real-time or historical data is as follows: Read relevant operational data of the conventional island from the power plant's distributed control system or historical operation database. The data includes: pressure, temperature, and flow rate of the working fluid at the inlet and outlet of the steam generation system; pressure, temperature, and flow rate of the main steam; pressure, temperature, and flow rate of the reheat steam; pressure, temperature, and flow rate of each stage of extraction steam; turbine output power; regulating valve opening; bypass valve opening; feedwater pump speed; condenser pressure; and deaerator pressure and water level. Preprocess the above operational data, including time alignment, abnormal data identification and removal, supplementation of missing data, unit standardization, and reliability marking, to form measured values.

[0008] Preferably, the specific process for preprocessing the acquired runtime-related data is as follows: Time alignment processing: Due to the sampling period, data refresh time and communication delay of the measurement points of different distributed control systems on the conventional island, it is necessary to unify the data of each measurement point onto the same time axis. For data with high sampling frequency, resampling or moving average is used to unify the time step; for data with low sampling frequency, hold, linear interpolation or spline interpolation is used to fill in the gaps so that all data have the same time step. Abnormal data identification and removal: Data that clearly exceeds the physical boundaries of the equipment, data that changes drastically in a short period of time but has no corresponding operation instructions, data that remains unchanged for a long period of time, and data that strongly conflict with adjacent measurement points are marked as abnormal data; To supplement missing data, for short-term missing data, interpolation between adjacent time points and the mean of a sliding window are used to fill in the missing data; for long-term missing data that cannot be reliably recovered, the data is not directly used in online identification, but the weight of the measurement point is reduced, and for necessary data, soft measurement values ​​are constructed using energy conservation or mass conservation. Units are standardized; units for data of the same category are converted and standardized; temperature is standardized to ℃, pressure to MPa, flow rate to kg / h or t / h, and is standardized to kg / s before mass conservation and energy conservation calculations; power or heat is standardized to kW or MW. Reliability labeling is performed by combining threshold judgment, rate of change judgment, operating condition boundary judgment, and energy / mass conservation consistency judgment for each DCS measurement point. Set credibility level They are categorized into four types: high credibility, medium credibility, low credibility, and invalid. The specific classification criteria are as follows: (1) Threshold judgment criteria: The physical allowable range of the measuring point is determined based on the equipment design parameters, instrument range and historical normal operation data. For any measuring point The The formula for judging based on the basic threshold for each data point is as follows: ; like If the measurement point exceeds the physical allowable range, it is marked as an invalid measurement point. and These are the lower and upper physical limits for the measuring point, respectively. For measuring points where temperature, pressure, flow rate, and liquid level fluctuate, the formulas are as follows: When the measuring point is within the normal operating range, it is marked as high confidence; when the measuring point is outside the normal operating range but still within the physical allowable range of the equipment, it is marked as medium confidence; when the measuring point is outside the physical allowable range of the equipment or the instrument range, it is marked as invalid. For flow measurement points, if a negative value or a value of zero is observed under non-shutdown and non-valve-closed operating conditions, it is marked as low confidence or invalid. For pressure measurement points, if the value is lower than the physical lower limit under the corresponding working fluid condition or significantly higher than the upper limit of the equipment design pressure, it is marked as invalid. For temperature measurement points, if the relationship with the inlet and outlet temperatures of adjacent equipment is significantly contradictory, it enters the consistency verification stage. (2) Criteria for judging the rate of change, for any measuring point Calculate the rate of change between adjacent sampling times. : in, Indicates the sampling time interval, when When the rate of change at the measuring point is normal, it is considered that the rate of change at that point is normal. If there is no change in load command, valve action, pump speed change, bypass switching or heat source disturbance at the corresponding time, then the measuring point is marked as a suspected sudden change measuring point. For temperature, pressure, and flow measurement points, the maximum allowable rate of change should be set according to the dynamic response characteristics of the equipment; for measurement points with strong thermal inertia, the allowable rate of change should be less than that for flow and valve opening measurement points; actuator measurement points for valve opening and pump speed are allowed to change relatively quickly, but their changes should be consistent with the control commands. (3) Operating condition boundary judgment criteria: Based on the unit load, operating mode, valve status, pump operating status and bypass status, establish the parameter allowable boundaries under different operating conditions. For any measuring point The constraint formula under the current operating conditions is: in, and This represents the allowable range of the measuring point under the operating conditions. Under load increase conditions, the main steam flow, feedwater flow, and turbine power will increase as a whole with the load command; under load decrease conditions, the criterion is to decrease as a whole. When the change direction of a certain measuring point is continuously opposite to the load change direction, and there is no explanation for bypass action, steam extraction change, or regulating valve action, it is marked as medium or low confidence. For heat exchanger equipment, the hot and cold side temperatures must satisfy the basic heat exchange direction. During normal heat exchange, the hot side outlet temperature should not be higher than the hot side inlet temperature, and the cold side outlet temperature should not be lower than the cold side inlet temperature, as shown in the following formula: ; If the above formula is not satisfied and there are no start-stop, reverse flow, or measurement delay conditions, the relevant temperature measurement point is marked as low confidence. For valve and pipeline measurement points, the flow direction should be consistent with the pressure difference direction. Under normal unidirectional flow conditions, if the pressure before the valve is lower than the pressure after the valve but the flow rate still shows a large positive flow rate, the relevant pressure or flow measurement point is marked as suspected abnormal. (4) Quality conservation consistency judgment criteria: For any equipment or node, establish the quality conservation residual: in, The total mass flow rate entering the device or node. The total mass flow rate out of the device or node. Expressed as the rate of change of the working fluid mass within the equipment volume; Under steady-state or slowly changing operating conditions, if At that time, the mass conservation law is consistent; if at... If the value is 3%-8%, it is considered to have a slight deviation, and the relevant measuring points are marked as medium confidence; if If the value is greater than 8%, it is considered that there is a significant conflict with the mass conservation principle, and the relevant measurement points are marked as low confidence or suspected anomalies. Under highly dynamic operating conditions such as start-up, rapid load changes, and bypass switching, the mass conservation residual limit can be appropriately relaxed. For example, the medium confidence threshold can be relaxed to 5%, and the low confidence threshold can be relaxed to 10%, and the judgment can be made in combination with the equipment volume effect and the measurement point time delay. (5) Energy conservation consistency judgment criteria: For any heat exchanger, steam turbine, condenser, or heater equipment, establish the energy conservation residual: In the above formula, h is the specific enthalpy of the working substance. To input heat from the outside, For output power or power consumption, The rate of change of energy within the equipment; under steady-state or slowly varying operating conditions, when Then the energy conservation is consistent; when If the value range is 5%-10%, it is considered that there is a certain deviation, and the relevant measuring points are marked as medium confidence; if If the percentage is greater than 10%, it is considered that there is a significant conflict in energy conservation, and the relevant measurement points are marked as low confidence or suspected anomalies. (6) Criteria for judging the consistency of valve and turbine characteristics: For valve measuring points, calculate the theoretical flow rate based on the valve flow characteristics. in, The valve flow coefficient, For valve opening, For the working fluid density, To calculate the pressure difference across the valve, the theoretical flow rate is compared with the measured flow rate using the following formula: when When the valve measuring points are consistent, the consistency is good; when Values ​​between 10% and 20% are marked as moderately reliable; when... If so, it is marked as low trust. For the turbine power measurement point, the theoretical output power is calculated based on the enthalpy difference between the inlet and outlet steam and the flow rate: ; in, This represents the turbine efficiency correction factor. The main steam or corresponding cylinder section steam flow rate; when the deviation between the measured power and the calculated power is less than 5%, the power measurement point is considered reliable; when the deviation is between 5% and 10%, it is marked as medium reliable; when the deviation exceeds 10%, it is marked as low reliable. (7) Based on the comprehensive reliability grading standard and the above judgment results, a reliability level is given for each measuring point. If the measuring point simultaneously meets the threshold judgment, rate of change judgment, working condition boundary judgment and conservation consistency judgment, it is marked as high reliability. If the measuring point only has slight over-limit, slight fluctuation or conservation residual in the middle range, it is marked as medium reliability. If the measuring point obviously violates the working condition boundary or conservation relationship, but has not completely exceeded the instrument range, it is marked as low reliability. If the measuring point exceeds the physical boundary, instrument range, or has long-term stagnation, communication interruption, or obvious reverse logic error, it is marked as invalid. The confidence weight is used for subsequent online state estimation, slow-varying parameter identification, and residual correction. When the confidence of a measurement point decreases, its impact on state estimation and parameter identification decreases accordingly. When a measurement point is marked as invalid, it does not directly participate in model correction, but is replaced by adjacent measurement points, mechanistic model predictions, or mass / energy conservation soft measurement values.

[0009] Preferably, in step S3, online state estimation is performed based on the mechanism model and measured values ​​to obtain unmeasurable states. Unmeasurable states include: the average temperature of the working fluid inside the heat exchanger, the metal wall temperature, the pipe volume pressure, the internal state of the steam drum or deaerator volume components, the equivalent heat transfer state of the condenser, the interstage state of the steam turbine, and some extraction steam flow rates that cannot be directly measured. The specific processing procedure is as follows: First, the unpredictable state at the next moment is predicted based on the dynamic mechanism model, and the estimated state at the previous moment is: , with input quantity and current parameters Based on this, calculate the current time. Prior states: ; In the above formula, This represents the current state predicted by the mechanistic model, with the prior state corrected based on measured values, and the mechanistic model output is then presented. Compared with on-site measurements By comparison, the measurement residuals are obtained. : ; The online state estimator corrects the state variables based on the residual magnitude, measurement point reliability, and model uncertainty to obtain the posterior state. : ; in, The state correction gain is set according to the measurement point noise, model error and measurement point confidence. When a measurement point is judged to be abnormal or drifting, the weight of the corresponding measurement point is reduced to avoid abnormal measurement values ​​directly contaminating the twin state of the simulation system.

[0010] Preferably, in step S4, the online identification process of the slowly varying parameters is as follows: After obtaining the real-time state estimation results, key parameters in the conventional island that change slowly with operating time but are difficult to measure directly are identified. The slow-changing parameters mainly include: the equivalent heat transfer coefficient UA of the condenser or air-cooled system, the equivalent heat transfer coefficient or heat loss coefficient of the feedwater preheater, evaporator, superheater, and reheater, the valve flow coefficient of the main steam valve, bypass valve, and extraction valve, the pressure loss coefficient of the pipeline and heat exchanger, the efficiency correction coefficient of the high-pressure cylinder, intermediate-pressure cylinder, and low-pressure cylinder of the steam turbine, and the efficiency correction coefficient of the pump equipment. The online identification employs a dual-timescale mechanism. For rapidly changing states, online filtering or state estimation methods from S3 are used for rapid updates. For slowly varying parameters such as heat transfer coefficient, pressure loss coefficient, and efficiency correction, the parameters do not change drastically with each sampling point, but are slowly updated within a certain time window through recursive identification or rolling optimization. The calculation process is as follows: The objective function is minimized using either a constrained weighted recursive least squares method or a rolling window optimization method. At the k-th sampling time, an identification time window is formed using N sampling points. The formula for the slowly varying parameter identification optimization problem is as follows: The constraint formula is: in, These are the estimated values ​​of the slowly varying parameters identified at the current time. These are the estimated values ​​of the slowly varying parameters from the previous time step. Let be the DCS measurement value at the i-th sampling time. For given parameters The output of the mechanistic model under the given conditions As the confidence weight of the measurement points, For parameter smoothing coefficients, and These are the lower and upper physical limits of the slowly varying parameters, respectively. The maximum allowable change of the slowly varying parameter within a unit update time; if the identification result meets the requirements of parameter boundary, rate of change constraint and measurement point reliability, then the slowly varying parameter is updated; otherwise, the parameter at the previous time step is retained or only amplitude limiting correction is performed. The constraint formula includes physical upper and lower limits for the parameter and constraints on the rate of change of the parameter, ensuring that the slowly varying parameter to be identified remains within the physically feasible range of the equipment and avoids drastic changes due to instantaneous disturbances or abnormal measuring points; the specific identification process of the slowly varying parameter includes the following steps: Based on equipment design parameters, performance test data, or historical stable operation data, determine the initial value, physical upper and lower limits, and maximum permissible rate of change of the slowly varying parameters to be identified. Slowly varying parameters include the heat exchanger's equivalent heat transfer coefficient UA and the valve flow coefficient. Pressure loss coefficient K, turbine efficiency correction coefficient ; Secondly, for different slowly varying parameters, corresponding input and output quantities are selected, and residuals are constructed and calculated based on the mechanism relationship. For the equivalent heat transfer coefficient of the heat exchanger, heat transfer residuals are constructed based on the relationship between heat transfer and logarithmic mean temperature difference. For the valve flow coefficient, the flow residual is constructed based on the pressure before and after the valve, the valve opening, and the working fluid density. For the pressure loss coefficient, the pressure loss residual is constructed based on the relationship between flow rate and pressure difference; For the turbine efficiency correction factor, the power residual is constructed based on the steam enthalpy drop, flow rate, and output power: Then, minimize the above residual within the rolling time window to obtain the candidate value of the slow-varying parameter at the current time. If the candidate value meets the physical upper and lower limits of the parameter, the rate of change constraint and the reliability requirements of the measurement point, then update the slow-varying parameter; otherwise, keep the parameter value of the previous time or only use the correction amount after the amplitude limit. Finally, the updated slowly varying parameters are fed back into the dynamic mechanism model to recalculate the equipment outlet parameters, turbine power, and system state. This enables online correction of the mechanism model as equipment performance degrades and operating conditions change. The key output variables are selected based on safety relevance, performance relevance, parameter sensitivity, measurement point availability, and data reliability. Each key output variable must meet at least one of the following conditions: (1) These are the conventional island safety operation constraint variables, including main steam temperature, main steam pressure, reheat steam temperature, reheat steam pressure, condenser pressure, deaerator pressure, deaerator water level, steam drum water level or other volumetric liquid levels; these variables directly affect the equipment safety boundary and operation alarm judgment, and should be given priority in the residual correction objects. (2) Variables belonging to the unit performance evaluation include turbine output power, main steam flow rate, reheat steam flow rate, extraction steam flow rate at each stage, exhaust steam pressure, feedwater flow rate, feedwater temperature, and system heat rate related variables; such variables directly affect the unit performance calculation, operation economic evaluation and upper-level optimization results, and should be included in the residual correction object; (3) It belongs to the output variable that is sensitive to slowly varying parameters; for candidate output variables Calculate its slowly varying parameters to be identified Sensitivity: in, To normalize the sensitivity, To prevent extremely small positive numbers with a denominator of zero, when a certain output variable is highly sensitive to heat exchanger UA, valve flow coefficient, pressure loss coefficient, or turbine efficiency correction coefficient, it is considered a key output variable. For example, heat exchanger outlet temperature and heat transfer are sensitive to UA; valve downstream flow is sensitive to valve flow coefficient; pressure difference across the equipment is sensitive to pressure loss coefficient; turbine output power and exhaust parameters are sensitive to turbine efficiency correction coefficient. (4) Variables that can be directly measured in the DCS system or obtained by energy conservation and mass conservation soft measurement; data that cannot be measured, cannot be soft measured or has been missing for a long time shall not be used as the main target variables for residual correction. (5) The confidence level of the variable measurement points meets the requirements; if a measurement point is marked as high confidence or medium confidence within the current time window, it is allowed to participate in residual correction; if it is marked as low confidence, it is only used as an auxiliary reference; if it is marked as invalid, it will not participate in residual correction. According to the above standards, key output variables include, but are not limited to: the temperature, pressure, and flow rate of the working fluid at the outlet of the preheater, evaporator, superheater, and reheater in the steam generation system; the temperature, pressure, and flow rate of the main steam; the temperature, pressure, and flow rate of the reheat steam; the temperature, pressure, and flow rate of the extraction steam at each stage of the turbine; the turbine exhaust pressure and exhaust enthalpy; the turbine output power; the condenser pressure; the deaerator pressure and water level; and the feedwater flow rate, feedwater temperature, and feedwater pressure. When several candidate variables simultaneously meet the above conditions, priority should be given to selecting variables related to the safety boundary, variables with high sensitivity to slowly changing parameters, variables with high reliability of measurement points, and variables directly related to upper-level operation optimization and performance evaluation as residual correction objects. Preferably, in step S5, the specific process is as follows: The residuals between the model output and the measured values ​​are calculated, and the residual correction module is used to perform controlled corrections on systematic biases that are difficult for the mechanistic model to fully explain. Key output variables that need to be corrected are selected, and for each key variable, the residuals before correction are calculated: ; in, For model residuals, DCS measurement value For the output values ​​of the mechanistic model, a residual correction module is constructed accordingly. The inputs of the residual correction module include one or more of the following: current DCS measurement values, mechanistic model output, online estimation status, slowly varying parameter identification results, operating condition characteristics, and historical residual sequences; the output is the residual correction amount. The corrected twin output is: ; in, For the corrected twin output, amplitude and rate of change constraints are set for the residual correction amount. When the correction amount exceeds the physical reasonable range, the system does not directly use the correction result, but reduces the confidence of the correction result and prompts that there is an abnormal measurement point, the working condition exceeds the training range, or the model structure is missing. The process of outputting confidence level for the residual correction module is as follows: The confidence level is determined by a combination of factors, including whether the input operating condition is within the coverage of historical training data, the magnitude of residual fluctuation, the reliability of measurement points, the magnitude of correction, and the prediction error over a recent period. When the operating condition is within the familiar range of the model and the residual is stable, the confidence level is high; when start-up, shutdown, rapid load change, bypass switching, or measurement point abnormality occurs, the confidence level is low.

[0011] Preferably, the specific implementation process of step S6 is as follows: After completing state estimation, parameter identification, and residual correction, the corrected conventional island digital twin result is output, which mainly includes three categories: The first category is the corrected twin state, including the temperature, pressure, and flow rate of the working fluid at the outlet of each heat exchanger in the steam generation system, the temperature, pressure, and flow rate of the main steam and reheat steam, the temperature, pressure, and flow rate of the steam extracted at each stage, the turbine exhaust parameters, the condenser pressure, the deaerator status, the feedwater system status, and the status of the main pipelines and volumetric components. The second category consists of performance indicators, including turbine output power, unit heat rate, heat exchange efficiency of steam generation system, condenser vacuum performance, feedwater pump power consumption, load change rate, main steam parameter deviation, reheat steam parameter deviation, and system energy balance deviation, which are used to evaluate the current operating performance of the conventional island and the degree of deviation from the design conditions. The third category is health indicators, including the degree of UA degradation in heat exchangers, the degree of decline in heat exchange performance of condensers, the trend of pressure loss coefficient, the degree of deviation in valve flow coefficient, the trend of turbine efficiency correction coefficient, the confidence level of measuring points, suspected drift measuring point markings, abnormal operating condition markings, and residual correction confidence levels, which are used to distinguish between actual equipment degradation, measuring point anomalies, and model uncertainty. The recording process involves each record including at least a timestamp, operating condition number, input data, mechanistic model output, state estimation result, slow-varying parameter identification result, residual correction amount, corrected output, confidence level, and anomaly marker.

[0012] Therefore, the present invention employs the above-mentioned dynamic simulation method for the fusion of conventional island mechanism data in a solar thermal power plant, which has the following advantages: (1) For conventional islands under dynamic operating conditions such as start-up, shutdown, load change, and steam extraction disturbance, the twin output has higher consistency compared to steady-state calculation or mechanism model only; (2) This application makes equipment degradation quantifiable and traceable by identifying slowly changing parameters online, such as condenser scaling or leakage; (3) This application introduces measurement point reliability and residual controlled correction for measurement data points, which can distinguish between "equipment degradation" and "measurement point abnormality" and reduce false alarms; The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0013] Figure 1 This is a flowchart illustrating the implementation of the technical solution in this invention; Figure 2 This is a structural diagram of a conventional island of a solar thermal power plant in an embodiment of the present invention; Figure 3 This is a structural diagram illustrating different ways of fusing the mechanism and data in this invention; Figure 4 This is a table showing the relevant detection data of the main steam in this embodiment of the invention; Figure 5This is a table showing the relevant detection data of reheat steam in the embodiments of the present invention; Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Specific model specifications need to be selected and determined according to the actual specifications of the device, etc. The specific selection calculation method adopts existing technology in the art, and therefore will not be described in detail.

[0015] Example like Figure 1 As shown, this invention provides a dynamic simulation method for the fusion of mechanism data of a conventional island in a solar thermal power plant, comprising the following steps: Step S1: As Figure 2 The diagram illustrates a layout architecture. Based on the layout architecture of the conventional island in a concentrated solar power (CSP) plant, the process of establishing a dynamic mechanism model for the conventional island is as follows: The modeling objects of the conventional island dynamic mechanism model include the steam generation system, turbine unit, condenser or air-cooled system, deaerator, feedwater pump, high and low pressure heaters, bypass system, main steam and reheat steam pipelines, extraction steam pipelines, and related volumetric components. Based on the thermodynamic flow and equipment connection relationships of the conventional island system, the model formulas are as follows: ; ; In the above formula, This represents the rate of change of the internal state of a regular island with respect to time. Represents the state change function. This represents the calculation process of the mechanism model. These are the internal state variables of the conventional island, including the internal temperature, pressure, enthalpy, metal wall temperature, volumetric pressure, and liquid level of each piece of equipment; External input parameters include inlet pressure, inlet temperature, inlet flow rate, valve opening, and pump speed collected by the DCS. These are the model parameters, including heat exchanger UA, pressure loss coefficient, valve flow coefficient, turbine efficiency correction coefficient, and heat loss coefficient; The output of the mechanistic model includes the temperature, pressure, flow rate, and turbine power of the working fluid at the hot and cold sides of the equipment outlet.

[0016] Step S2: Acquire real-time or historical data input from the conventional island and preprocess the data to form measured values; specifically, the process of acquiring and preprocessing the input real-time or historical data is as follows: Read relevant operating data of the conventional island from the power plant's distributed control system or historical operation database. The data includes: pressure, temperature, and flow rate of the working fluid at the inlet and outlet of the steam generation system; pressure, temperature, and flow rate of the main steam; pressure, temperature, and flow rate of the reheat steam; pressure, temperature, and flow rate of each stage of extraction steam; turbine output power; regulating valve opening; bypass valve opening; feedwater pump speed; condenser pressure; deaerator pressure and water level; preprocess the above operating data, including time alignment, abnormal data identification and removal, supplementation of missing data, unit unification, and reliability marking, to form measured values.

[0017] The specific process for preprocessing the acquired operational data is as follows: Time alignment: Due to the sampling period, data refresh time, and communication delay of the measurement points of different distributed control systems on the conventional island, it is necessary to unify the data of each measurement point onto the same time axis. For data with high sampling frequency, resampling or moving average is used to unify the time step; for data with low sampling frequency, hold, linear interpolation, or spline interpolation is used to fill in the gaps so that all data have the same time step. Abnormal data identification and removal: Data that clearly exceeds the physical boundaries of the equipment, data that changes drastically in a short period of time but has no corresponding operation instructions, data that remains unchanged for a long period of time, and data that strongly conflict with adjacent measurement points are marked as abnormal data; To supplement missing data, for short-term missing data, interpolation between adjacent time points and the mean of a sliding window are used to fill in the missing data; for long-term missing data that cannot be reliably recovered, the data is not directly used in online identification, but the weight of the measurement point is reduced, and for necessary data, soft measurement values ​​are constructed using energy conservation or mass conservation. Units are standardized; units for data of the same category are converted and standardized; temperature is standardized to ℃, pressure to MPa, flow rate to kg / h or t / h, and is standardized to kg / s before mass conservation and energy conservation calculations; power or heat is standardized to kW or MW. Reliability labeling is performed by combining threshold judgment, rate of change judgment, operating condition boundary judgment, and energy / mass conservation consistency judgment for each DCS measurement point. Set credibility level They are categorized into four types: high credibility, medium credibility, low credibility, and invalid. The specific classification criteria are as follows: (1) Threshold judgment criteria: The physical allowable range of the measuring point is determined based on the equipment design parameters, instrument range and historical normal operation data. For any measuring point The The formula for judging based on the basic threshold for each data point is as follows: ; like If the measurement point exceeds the physical allowable range, it is marked as an invalid measurement point. and These are the lower and upper physical limits for the measuring point, respectively. For measuring points where temperature, pressure, flow rate, and liquid level fluctuate, the formulas are as follows: When the measuring point is within the normal operating range, it is marked as high confidence; when the measuring point is outside the normal operating range but still within the physical allowable range of the equipment, it is marked as medium confidence; when the measuring point is outside the physical allowable range of the equipment or the instrument range, it is marked as invalid. For flow measurement points, if a negative value or a value of zero is observed under non-shutdown and non-valve-closed operating conditions, it is marked as low confidence or invalid. For pressure measurement points, if the value is lower than the physical lower limit under the corresponding working fluid condition or significantly higher than the upper limit of the equipment design pressure, it is marked as invalid. For temperature measurement points, if the relationship with the inlet and outlet temperatures of adjacent equipment is significantly contradictory, it enters the consistency verification stage. (2) Criteria for judging the rate of change, for any measuring point Calculate the rate of change between adjacent sampling times. : in, Indicates the sampling time interval, when When the rate of change at the measuring point is normal, it is considered that the rate of change at that point is normal. If there is no change in load command, valve action, pump speed change, bypass switching or heat source disturbance at the corresponding time, then the measuring point is marked as a suspected sudden change measuring point. For temperature, pressure, and flow measurement points, the maximum allowable rate of change should be set according to the dynamic response characteristics of the equipment; for measurement points with strong thermal inertia, the allowable rate of change should be less than that for flow and valve opening measurement points; actuator measurement points for valve opening and pump speed are allowed to change relatively quickly, but their changes should be consistent with the control commands. (3) Operating condition boundary judgment criteria: Based on the unit load, operating mode, valve status, pump operating status and bypass status, establish the parameter allowable boundaries under different operating conditions. For any measuring point The constraint formula under the current operating conditions is: in, and This represents the allowable range of the measuring point under the operating conditions. Under load increase conditions, the main steam flow, feedwater flow, and turbine power will increase as a whole with the load command; under load decrease conditions, the criterion is to decrease as a whole. When the change direction of a certain measuring point is continuously opposite to the load change direction, and there is no explanation for bypass action, steam extraction change, or regulating valve action, it is marked as medium or low confidence. For heat exchanger equipment, the hot and cold side temperatures must satisfy the basic heat exchange direction. During normal heat exchange, the hot side outlet temperature should not be higher than the hot side inlet temperature, and the cold side outlet temperature should not be lower than the cold side inlet temperature, as shown in the following formula: ; If the above formula is not satisfied and there are no start-stop, reverse flow, or measurement delay conditions, the relevant temperature measurement point is marked as low confidence. For valve and pipeline measurement points, the flow direction should be consistent with the pressure difference direction. Under normal unidirectional flow conditions, if the pressure before the valve is lower than the pressure after the valve but the flow rate still shows a large positive flow rate, the relevant pressure or flow measurement point is marked as suspected abnormal. (4) Quality conservation consistency judgment criteria: For any device or node, establish the quality conservation residual: in, The total mass flow rate entering the device or node. The total mass flow rate out of the device or node. Expressed as the rate of change of the working fluid mass within the equipment volume; Under steady-state or slowly changing operating conditions, if At that time, the mass conservation law is consistent; if at... If the value is 3%-8%, it is considered to have a slight deviation, and the relevant measuring points are marked as medium confidence; if If the value is >8%, it is considered that there is a significant conflict with the mass conservation principle, and the relevant measurement points are marked as low confidence or suspected anomalies. Under highly dynamic operating conditions such as start-up, rapid load changes, and bypass switching, the mass conservation residual limit can be appropriately relaxed. For example, the medium confidence threshold can be relaxed to 5%, and the low confidence threshold can be relaxed to 10%, and the judgment can be made in combination with the equipment volume effect and the measurement point time delay. (5) Energy conservation consistency judgment criteria: For any heat exchanger, steam turbine, condenser, or heater equipment, establish the energy conservation residual: In the above formula, h is the specific enthalpy of the working substance. To input heat from the outside, For output power or power consumption, The rate of change of energy within the equipment; under steady-state or slowly varying operating conditions, when Then the energy conservation is consistent; when If the value range is 5%-10%, it is considered that there is a certain deviation, and the relevant measuring points are marked as medium confidence; if If the percentage is greater than 10%, it is considered that there is a significant conflict in energy conservation, and the relevant measurement points are marked as low confidence or suspected anomalies. (6) Criteria for judging the consistency of valve and turbine characteristics: For valve measuring points, calculate the theoretical flow rate based on the valve flow characteristics. in, For valve flow coefficient, For valve opening, For the working fluid density, To calculate the pressure difference across the valve, the theoretical flow rate is compared with the measured flow rate using the following formula: when When the valve measuring points are consistent, the consistency is good; when Values ​​between 10% and 20% are marked as moderately reliable; when... If so, it is marked as low trust. For the turbine power measurement point, the theoretical output power is calculated based on the enthalpy difference between the inlet and outlet steam and the flow rate: ; in, This represents the turbine efficiency correction factor. The main steam or corresponding cylinder section steam flow rate; when the deviation between the measured power and the calculated power is less than 5%, the power measurement point is considered reliable; when the deviation is between 5% and 10%, it is marked as medium reliable; when the deviation exceeds 10%, it is marked as low reliable. (7) Based on the comprehensive reliability grading standard and the above judgment results, a reliability level is given for each measuring point. If the measuring point simultaneously meets the threshold judgment, rate of change judgment, working condition boundary judgment and conservation consistency judgment, it is marked as high reliability. If the measuring point only has slight over-limit, slight fluctuation or conservation residual in the middle range, it is marked as medium reliability. If the measuring point obviously violates the working condition boundary or conservation relationship, but has not completely exceeded the instrument range, it is marked as low reliability. If the measuring point exceeds the physical boundary, instrument range, or has long-term stagnation, communication interruption, or obvious reverse logic error, it is marked as invalid. The confidence weight is used for subsequent online state estimation, slow-varying parameter identification, and residual correction. When the confidence of a measurement point decreases, its impact on state estimation and parameter identification decreases accordingly. When a measurement point is marked as invalid, it does not directly participate in model correction, but is replaced by adjacent measurement points, mechanistic model predictions, or mass / energy conservation soft measurement values.

[0018] Step S3: Based on the dynamic mechanism model and measured values, perform online state estimation to obtain unmeasurable states. Unmeasurable states include: the average temperature of the working fluid inside the heat exchanger, the metal wall temperature, the pipe volume pressure, the internal state of the steam drum or deaerator volume components, the equivalent heat transfer state of the condenser, the interstage state of the steam turbine, and some extraction steam flow rates that cannot be directly measured. The specific processing procedure is as follows: First, the unpredictable state at the next moment is predicted based on the dynamic mechanism model, and the estimated state at the previous moment is: , with input quantity and current parameters Based on this, calculate the current time. Prior states: ; In the above formula, This represents the current state predicted by the mechanistic model, with the prior state corrected based on measured values, and the mechanistic model output is then presented. Compared with on-site measurements By comparison, the measurement residuals are obtained. : ; The online state estimator corrects the state variables based on the residual magnitude, measurement point reliability, and model uncertainty to obtain the posterior state. : ; in, The state correction gain is set according to the measurement point noise, model error and measurement point confidence. When a measurement point is judged to be abnormal or drifting, the weight of the corresponding measurement point is reduced to avoid abnormal measurement values ​​directly contaminating the twin state of the simulation system.

[0019] Step S4: Perform online identification of slowly varying parameters to obtain key output variables. The process of performing online identification of slowly varying parameters is as follows: After obtaining the real-time state estimation results, key parameters in the conventional island that change slowly with operating time but are difficult to measure directly are identified. The slow-changing parameters mainly include: the equivalent heat transfer coefficient UA of the condenser or air-cooled system, the equivalent heat transfer coefficient or heat loss coefficient of the feedwater preheater, evaporator, superheater, and reheater, the valve flow coefficient of the main steam valve, bypass valve, and extraction valve, the pressure loss coefficient of the pipeline and heat exchanger, the efficiency correction coefficient of the high-pressure cylinder, intermediate-pressure cylinder, and low-pressure cylinder of the steam turbine, and the efficiency correction coefficient of the pump equipment. The online identification employs a dual-timescale mechanism. For rapidly changing states, online filtering or state estimation methods from S3 are used for rapid updates. For slowly varying parameters such as heat transfer coefficient, pressure loss coefficient, and efficiency correction, the parameters do not change drastically with each sampling point, but are slowly updated within a certain time window through recursive identification or rolling optimization. The calculation process is as follows: The objective function is minimized using either a constrained weighted recursive least squares method or a rolling window optimization method. At the k-th sampling time, an identification time window is formed using N sampling points. The formula for the slowly varying parameter identification optimization problem is as follows: The constraint formula is: in, These are the estimated values ​​of the slowly varying parameters identified at the current time. These are the estimated values ​​of the slowly varying parameters from the previous time step. Let be the DCS measurement value at the i-th sampling time. For given parameters The output of the mechanistic model under the given conditions As the confidence weight of the measurement points, For parameter smoothing coefficients, and These are the lower and upper physical limits of the slowly varying parameters, respectively. The maximum allowable change of the slowly varying parameter within a unit update time; if the identification result meets the requirements of parameter boundary, rate of change constraint and measurement point reliability, then the slowly varying parameter is updated; otherwise, the parameter at the previous time step is retained or only amplitude limiting correction is performed. The constraint formula includes physical upper and lower limits for the parameter and constraints on the rate of change of the parameter, ensuring that the slowly varying parameter to be identified remains within the physically feasible range of the equipment and avoids drastic changes due to instantaneous disturbances or abnormal measuring points; the specific identification process of the slowly varying parameter includes the following steps: Based on equipment design parameters, performance test data, or historical stable operation data, determine the initial value, physical upper and lower limits, and maximum permissible rate of change of the slowly varying parameters to be identified. Slowly varying parameters include the heat exchanger's equivalent heat transfer coefficient UA and the valve flow coefficient. Pressure loss coefficient K, turbine efficiency correction coefficient ; Secondly, for different slowly varying parameters, corresponding input and output quantities are selected, and residuals are constructed and calculated based on the mechanism relationship. For the equivalent heat transfer coefficient of the heat exchanger, heat transfer residuals are constructed based on the relationship between heat transfer and logarithmic mean temperature difference. For the valve flow coefficient, the flow residual is constructed based on the pressure before and after the valve, the valve opening, and the working fluid density. For the pressure loss coefficient, the pressure loss residual is constructed based on the relationship between flow rate and pressure difference; For the turbine efficiency correction factor, the power residual is constructed based on the steam enthalpy drop, flow rate, and output power: Then, minimize the above residual within the rolling time window to obtain the candidate value of the slow-varying parameter at the current time. If the candidate value meets the physical upper and lower limits of the parameter, the rate of change constraint and the reliability requirements of the measurement point, then update the slow-varying parameter; otherwise, keep the parameter value of the previous time or only use the correction amount after the amplitude limit. Finally, the updated slowly varying parameters are fed back to the mechanistic model to recalculate the equipment outlet parameters, turbine power, and system state. This enables online correction of the mechanistic model as equipment performance degrades and operating conditions change. Key output variables are selected based on safety relevance, performance relevance, parameter sensitivity, measurement point availability, and data reliability. Key output variables must meet at least one of the following conditions: (1) These are the conventional island safety operation constraint variables, including main steam temperature, main steam pressure, reheat steam temperature, reheat steam pressure, condenser pressure, deaerator pressure, deaerator water level, steam drum water level or other volumetric liquid levels; these variables directly affect the equipment safety boundary and operation alarm judgment, and should be given priority in the residual correction objects. (2) Variables belonging to the unit performance evaluation include turbine output power, main steam flow rate, reheat steam flow rate, extraction steam flow rate at each stage, exhaust steam pressure, feedwater flow rate, feedwater temperature, and system heat rate related variables; such variables directly affect the unit performance calculation, operation economic evaluation and upper-level optimization results, and should be included in the residual correction object; (3) It belongs to the output variable that is sensitive to slowly varying parameters; for candidate output variables Calculate its slowly varying parameters to be identified Sensitivity: in, To normalize the sensitivity, To prevent extremely small positive numbers with a denominator of zero, when a certain output variable is highly sensitive to heat exchanger UA, valve flow coefficient, pressure loss coefficient, or turbine efficiency correction coefficient, it is considered a key output variable. For example, heat exchanger outlet temperature and heat transfer are sensitive to UA; valve downstream flow is sensitive to valve flow coefficient; pressure difference across the equipment is sensitive to pressure loss coefficient; turbine output power and exhaust parameters are sensitive to turbine efficiency correction coefficient. (4) Variables that can be directly measured in the DCS system or obtained by energy conservation and mass conservation soft measurement; data that cannot be measured, cannot be soft measured or has been missing for a long time shall not be used as the main target variables for residual correction. (5) The confidence level of the variable measurement points meets the requirements; if a measurement point is marked as high confidence or medium confidence within the current time window, it is allowed to participate in residual correction; if it is marked as low confidence, it is only used as an auxiliary reference; if it is marked as invalid, it will not participate in residual correction. According to the above standards, key output variables include, but are not limited to: the temperature, pressure, and flow rate of the working fluid at the outlet of the preheater, evaporator, superheater, and reheater in the steam generation system; the temperature, pressure, and flow rate of the main steam; the temperature, pressure, and flow rate of the reheat steam; the temperature, pressure, and flow rate of the extraction steam at each stage of the turbine; the turbine exhaust pressure and exhaust enthalpy; the turbine output power; the condenser pressure; the deaerator pressure and water level; and the feedwater flow rate, feedwater temperature, and feedwater pressure. When several candidate variables simultaneously meet the above conditions, priority should be given to selecting variables related to the safety boundary, variables with high sensitivity to slowly changing parameters, variables with high reliability of measurement points, and variables directly related to upper-level operation optimization and performance evaluation as residual correction objects. Step S5: Calculate the residual between the model output and the measured value, train and update the residual correction module, output the correction value and provide the confidence level. The specific process is as follows: The residuals between the model output and the measured values ​​are calculated, and the residual correction module is used to perform controlled corrections on systematic biases that are difficult for the mechanistic model to fully explain. Key output variables that need to be corrected are selected, and for each key variable, the residuals before correction are calculated: ; in, For model residuals, DCS measurement value For the output values ​​of the mechanistic model, a residual correction module is constructed accordingly. The inputs of the residual correction module include one or more of the following: current DCS measurement values, mechanistic model output, online estimation status, slowly varying parameter identification results, operating condition characteristics, and historical residual sequences; the output is the residual correction amount. The corrected twin output is: ; in, For the corrected twin output, amplitude and rate of change constraints are set for the residual correction amount. When the correction amount exceeds the physical reasonable range, the system does not directly use the correction result, but reduces the confidence of the correction result and prompts that there is an abnormal measurement point, the working condition exceeds the training range, or the model structure is missing. For the mechanistic model and the slowly varying parameter identification process, there are series and parallel relationships, respectively, such as... Figure 3 As shown, in the series connection, the output of the mechanism model and the input data are used as training data, and the slowly varying parameter identification output is used to improve the error between the mechanism model and the actual running data; in the parallel connection, the mechanism model is used as a constraint for the slowly varying parameter identification, and the slowly varying parameter identification output is used to identify the key slowly varying parameters to correct the simulation error of the mechanism model. Finally, the simulation output of the model is obtained by mixing the two.

[0020] The process of outputting confidence level for the residual correction module is as follows: The confidence level is determined by a combination of factors, including whether the input operating condition is within the coverage of historical training data, the magnitude of residual fluctuation, the reliability of measurement points, the magnitude of correction, and the prediction error over a recent period. When the operating condition is within the familiar range of the model and the residual is stable, the confidence level is high; when start-up, shutdown, rapid load change, bypass switching, or abnormal measurement points occur, the confidence level is low.

[0021] Step S6: Output and record the corrected twin state, performance indicators, and health indicators. After completing state estimation, parameter identification, and residual correction, output the corrected conventional island digital twin results. The output mainly includes three categories: The first category is the corrected twin state, including the temperature, pressure, and flow rate of the working fluid at the outlet of each heat exchanger in the steam generation system, the temperature, pressure, and flow rate of the main steam and reheat steam, the temperature, pressure, and flow rate of the steam extracted at each stage, the turbine exhaust parameters, the condenser pressure, the deaerator status, the feedwater system status, and the status of the main pipelines and volumetric components. The second category consists of performance indicators, including turbine output power, unit heat rate, heat exchange efficiency of steam generation system, condenser vacuum performance, feedwater pump power consumption, load change rate, main steam parameter deviation, reheat steam parameter deviation, and system energy balance deviation, which are used to evaluate the current operating performance of the conventional island and the degree of deviation from the design conditions. The third category is health indicators, including the degree of UA degradation in heat exchangers, the degree of decline in heat exchange performance of condensers, the trend of pressure loss coefficient, the degree of deviation in valve flow coefficient, the trend of turbine efficiency correction coefficient, the confidence level of measuring points, suspected drift measuring point markings, abnormal operating condition markings, and residual correction confidence levels, which are used to distinguish between actual equipment degradation, measuring point anomalies, and model uncertainty. The recording process involves each record including at least a timestamp, operating condition number, input data, mechanistic model output, state estimation results, slow-varying parameter identification results, residual correction amount, corrected output, confidence level, and anomaly marker. The dynamic simulation process for the conventional island based on variable load operating data uses a trough or tower solar thermal power plant as the object, with a rated power of 50MW. The conventional island includes a steam generation system, turbine unit, condenser or air-cooled system, deaerator, feedwater pumps, high and low pressure heaters, main steam pipelines, reheat steam pipelines, extraction steam pipelines, and bypass system. The steam generation system includes a preheater, evaporator, superheater, and reheater. The working fluid on the heat source side is molten salt or thermal oil, and the working fluid on the cold source side is water and steam.

[0022] This embodiment provides an example as follows, and the results are as follows: Figure 4 and Figure 5 The input data was selected from the continuously operating variable load conditions in the power plant's DCS system. Variable load conditions included the process of the unit load increasing from low to high load, stabilizing, and then decreasing to medium-low load. The collected data included the inlet temperature, pressure, and flow rate of the heat source side; feedwater temperature, pressure, and flow rate; main steam temperature, pressure, and flow rate; reheat steam temperature, pressure, and flow rate; turbine regulating valve opening; bypass valve opening; feedwater pump speed; condenser pressure; deaerator pressure and water level; and turbine output power. The sampling period was set to 1 second.

[0023] First, the DCS data undergoes time alignment, outlier identification, missing value imputation, and unit standardization. For short-term missing data, linear interpolation is used to impute the data; for measurement points that clearly exceed the physical boundaries of the equipment or conflict with the upstream and downstream conservation relationships, their reliability weight is reduced or they are marked as invalid measurement points.

[0024] Then, a dynamic mechanism model was established based on the conventional island thermal process. For the steam generation system, mass conservation equations, energy conservation equations, and heat transfer equations were established for the preheater, evaporator, superheater, and reheater, respectively, considering the heat capacity of the metal wall and the volume effect of the working fluid. For the turbine unit, the relationships between inlet steam parameters, exhaust steam parameters, extraction steam parameters, regulating valve opening, and output power were established. For valves and pipelines, the flow-pressure difference relationship and the volumetric pressure change relationship were established. The model input is the boundary conditions collected by the DCS, and the model output is the outlet temperature, pressure, and flow rate of each device.

[0025] At each sampling time, the current state is first predicted using a mechanistic model, and then the predicted state is corrected online based on DCS measurements to obtain an estimate of the unmeasurable states inside the conventional island. Unmeasurable states include the average temperature of the working fluid inside the heat exchanger, the metal wall temperature, and the volumetric pressure of the pipes.

[0026] Based on the state estimation, the slowly varying parameters are further identified online. Taking the equivalent heat transfer coefficient UA_ of the superheater as an example, the inlet temperature, outlet temperature, and flow rate of the superheater on the heat source side are selected as identification data. A weighted error objective function between the model output and the DCS measurement value is established to ensure that the slowly varying parameters are within the physically feasible range of the equipment and to avoid drastic changes in parameters due to instantaneous disturbances or abnormal measurement points.

[0027] After identifying the slowly varying parameters, the residuals between the computer model output and the DCS measurements are calculated. In this embodiment, main steam temperature, main steam pressure, main steam flow rate, reheat steam temperature, reheat steam pressure, and reheat steam flow rate are selected as key output variables. For each key output variable, the uncorrected residuals are calculated: ; in, For model residuals, DCS measurement value The residual correction module takes the current DCS input, the mechanistic model output, the state estimation results, the slowly varying parameter identification results, the operating condition characteristics, and the historical residual sequence as inputs, and outputs the residual correction amount. The corrected output is obtained: ; in, The results are the dynamic simulation results of a conventional island after residual correction. The residual correction amount is subject to amplitude and rate of change constraints. When the correction amount exceeds the allowable range, or when the correction result significantly conflicts with the mass conservation and energy conservation relationships, the confidence level of the correction result is reduced, and the output of the mechanism model or the state estimation result is retained first.

[0028] Through the above process, this embodiment can continuously output dynamic simulation results, slow-varying parameter identification results, residual correction values, and confidence levels of key parameters of the conventional island during unit load changes. The output includes the outlet temperature, pressure, and flow rate of each heat exchanger in the steam generation system; main steam and reheat steam parameters; turbine output power; condenser pressure; deaerator status; feedwater system status; trends in the equivalent heat transfer coefficient of the heat exchangers; trends in the pressure loss coefficient; valve flow coefficient deviation; and trends in the turbine efficiency correction coefficient.

[0029] Therefore, this invention adopts a dynamic simulation method for the fusion of mechanism data of conventional island in a solar thermal power plant. By providing a method for the fusion of mechanism data of conventional island in a solar thermal power plant, a fusion scheme of mechanism and data is formed, and a simulation scheme is formed based on the fusion scheme. This provides a solution to the problems of insufficient accuracy, untraceable parameters, high requirements for training data quality, and weak online correction capability for continuous on-site operation in existing technologies under dynamic operating conditions.

[0030] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A dynamic simulation method for the fusion of data on the mechanism of a conventional island in a solar thermal power plant, characterized in that: Includes the following steps: Step S1: Based on the layout architecture of the conventional island of the solar thermal power plant, establish a dynamic mechanism model of the conventional island; Step S2: Obtain real-time or historical data input from the regular island and preprocess the data to form measurement values; Step S3: Based on the dynamic mechanism model and measured values, perform online state estimation to obtain the unmeasurable state; Step S4: Perform online identification of slowly varying parameters to obtain key output variables; Step S5: Calculate the residual between the model output and the measured value, train and update the residual correction module, output the correction value and give the confidence level; Step S6: Output and record the corrected twin status, performance indicators and health indicators.

2. The dynamic simulation method for fusion of conventional island mechanism data in a solar thermal power plant according to claim 1, characterized in that: In step S1, the process of establishing the dynamic mechanism model of a conventional island is as follows: The modeling objects of the conventional island dynamic mechanism model include the steam generation system, turbine unit, condenser or air-cooled system, deaerator, feedwater pump, high and low pressure heaters, bypass system, main steam and reheat steam pipelines, extraction steam pipelines, and related volumetric components. Based on the thermodynamic flow and equipment connection relationships of the conventional island system, the model formulas are as follows: ； ； In the above formula, This represents the rate of change of the internal state of a regular island with respect to time. Represents the state change function. This represents the calculation process of the mechanism model. These are the internal state variables of the conventional island, including the internal temperature, pressure, enthalpy, metal wall temperature, volumetric pressure, and liquid level of each piece of equipment; External input parameters include inlet pressure, inlet temperature, inlet flow rate, valve opening, and pump speed collected by the DCS. These are the model parameters, including heat exchanger UA, pressure loss coefficient, valve flow coefficient, turbine efficiency correction coefficient, and heat loss coefficient; The output of the mechanistic model includes the temperature, pressure, flow rate, and turbine power of the working fluid at the hot and cold sides of the equipment outlet.

3. The dynamic simulation method for fusion of conventional island mechanism data in a solar thermal power plant according to claim 1, characterized in that: In step S2, the process of acquiring and preprocessing the input real-time or historical data is as follows: Read relevant operational data of the conventional island from the power plant's distributed control system or historical operation database. This data includes: pressure, temperature, and flow rate of the working fluid at the inlet and outlet of the steam generation system; pressure, temperature, and flow rate of the main steam; pressure, temperature, and flow rate of the reheat steam; pressure, temperature, and flow rate of each stage of extraction steam; turbine output power; regulating valve opening; bypass valve opening; feedwater pump speed; condenser pressure; and deaerator pressure and water level. Preprocess the above operational data, including time alignment, abnormal data identification and removal, supplementation of missing data, unit standardization, and reliability marking, to form measured values.

4. The dynamic simulation method for fusion of conventional island mechanism data in a solar thermal power plant according to claim 3, characterized in that: The specific process for preprocessing the acquired runtime-related data is as follows: Time alignment processing: Due to the sampling period, data refresh time and communication delay of the measurement points of different distributed control systems on the conventional island, it is necessary to unify the data of each measurement point onto the same time axis. For data with high sampling frequency, resampling or moving average is used to unify the time step; for data with low sampling frequency, hold, linear interpolation or spline interpolation is used to fill in the gaps so that all data have the same time step. Abnormal data identification and removal: Data that clearly exceeds the physical boundaries of the equipment, data that changes drastically in a short period of time but has no corresponding operation instructions, data that remains unchanged for a long period of time, and data that strongly conflict with adjacent measurement points are marked as abnormal data; To supplement missing data, for short-term missing data, interpolation between adjacent time points and the mean of a sliding window are used to fill in the missing data; for long-term missing data that cannot be reliably recovered, the data is not directly used in online identification, but the weight of the measurement point is reduced, and for necessary data, soft measurement values ​​are constructed using energy conservation or mass conservation. Units are standardized; units for data of the same category are converted and standardized; temperature is standardized to ℃, pressure to MPa, flow rate to kg / h or t / h, and is standardized to kg / s before mass conservation and energy conservation calculations; power or heat is standardized to kW or MW. Reliability labeling is performed by combining threshold judgment, rate of change judgment, operating condition boundary judgment, and energy / mass conservation consistency judgment for each DCS measurement point. Set credibility level They are categorized into four types: high credibility, medium credibility, low credibility, and invalid. The specific classification criteria are as follows: (1) Threshold judgment criteria: The physical allowable range of the measuring point is determined based on the equipment design parameters, instrument range and historical normal operation data. For any measuring point The The formula for judging based on the basic threshold for each data point is as follows: ； like If the measurement point exceeds the physical allowable range, it is marked as an invalid measurement point. and These are the lower and upper physical limits for the measuring point, respectively. For measuring points where temperature, pressure, flow rate, and liquid level fluctuate, the formulas are as follows: When the measuring point is within the normal operating range, it is marked as high confidence; when the measuring point is outside the normal operating range but still within the physical allowable range of the equipment, it is marked as medium confidence; when the measuring point is outside the physical allowable range of the equipment or the instrument range, it is marked as invalid. For flow measurement points, if a negative value or a value of zero is observed under non-shutdown and non-valve-closed operating conditions, it is marked as low confidence or invalid. For pressure measurement points, if the value is lower than the physical lower limit under the corresponding working fluid condition or significantly higher than the upper limit of the equipment design pressure, it is marked as invalid. For temperature measurement points, if the relationship with the inlet and outlet temperatures of adjacent equipment is significantly contradictory, it enters the consistency verification stage. (2) Criteria for judging the rate of change, for any measuring point Calculate the rate of change between adjacent sampling times. : in, Indicates the sampling time interval, when When the rate of change at the measuring point is normal, it is considered that the rate of change at that point is normal. If there is no change in load command, valve action, pump speed change, bypass switching or heat source disturbance at the corresponding time, then the measuring point is marked as a suspected sudden change measuring point. For temperature, pressure, and flow measurement points, the maximum allowable rate of change should be set according to the dynamic response characteristics of the equipment; for measurement points with strong thermal inertia, the allowable rate of change should be less than that for flow and valve opening measurement points; actuator measurement points for valve opening and pump speed are allowed to change relatively quickly, but their changes should be consistent with the control commands. (3) Operating condition boundary judgment criteria: Based on the unit load, operating mode, valve status, pump operating status and bypass status, establish the parameter allowable boundaries under different operating conditions. For any measuring point The constraint formula under the current operating conditions is: in, and This represents the allowable range of the measuring point under the operating conditions. Under load increase conditions, the main steam flow, feedwater flow, and turbine power will increase as a whole with the load command; under load decrease conditions, the criterion is to decrease as a whole. When the change direction of a certain measuring point is continuously opposite to the load change direction, and there is no explanation for bypass action, steam extraction change, or regulating valve action, it is marked as medium or low confidence. For heat exchanger equipment, the hot and cold side temperatures must satisfy the basic heat exchange direction. During normal heat exchange, the hot side outlet temperature should not be higher than the hot side inlet temperature, and the cold side outlet temperature should not be lower than the cold side inlet temperature, as shown in the following formula: ； If the above formula is not satisfied and there are no start-stop, reverse flow, or measurement delay conditions, the relevant temperature measurement point is marked as low confidence. For valve and pipeline measurement points, the flow direction should be consistent with the pressure difference direction. Under normal unidirectional flow conditions, if the pressure before the valve is lower than the pressure after the valve but the flow rate still shows a large positive flow rate, the relevant pressure or flow measurement point is marked as suspected abnormal. (4) Quality conservation consistency judgment criteria: For any device or node, establish the quality conservation residual: in, The total mass flow rate entering the device or node. The total mass flow rate out of the device or node. Expressed as the rate of change of the working fluid mass within the equipment volume; Under steady-state or slowly changing operating conditions, if At that time, the mass conservation law is consistent; if at... If the value is 3%-8%, it is considered to have a slight deviation, and the relevant measuring points are marked as medium confidence; if If the value is greater than 8%, it is considered that there is a significant conflict with the mass conservation principle, and the relevant measurement points are marked as low confidence or suspected anomalies. Under highly dynamic operating conditions such as start-up, rapid load changes, and bypass switching, the mass conservation residual limit can be appropriately relaxed. For example, the medium confidence threshold can be relaxed to 5%, and the low confidence threshold can be relaxed to 10%, and the judgment can be made in combination with the equipment volume effect and the measurement point time delay. (5) Energy conservation consistency judgment criteria: For any heat exchanger, steam turbine, condenser, or heater equipment, establish the energy conservation residual: In the above formula, h is the specific enthalpy of the working substance. To input heat from the outside, For output power or power consumption, The rate of change of energy within the equipment; under steady-state or slowly varying operating conditions, when Then the energy conservation is consistent; when If the value range is 5%-10%, it is considered that there is a certain deviation, and the relevant measuring points are marked as medium confidence; if If the percentage is greater than 10%, it is considered that there is a significant conflict in energy conservation, and the relevant measurement points are marked as low confidence or suspected anomalies. (6) Criteria for judging the consistency of valve and turbine characteristics: For valve measuring points, calculate the theoretical flow rate based on the valve flow characteristics. in, The valve flow coefficient, For valve opening, For the working fluid density, To calculate the pressure difference across the valve, the theoretical flow rate is compared with the measured flow rate using the following formula: when When the valve measuring points are consistent, the consistency is good; when Values ​​between 10% and 20% are marked as moderately reliable; when... If so, it is marked as low trust. For the turbine power measurement point, the theoretical output power is calculated based on the enthalpy difference between the inlet and outlet steam and the flow rate: ； in, This represents the turbine efficiency correction factor. The main steam or corresponding cylinder section steam flow rate; when the deviation between the measured power and the calculated power is less than 5%, the power measurement point is considered reliable; when the deviation is between 5% and 10%, it is marked as medium reliable; when the deviation exceeds 10%, it is marked as low reliable. (7) Based on the comprehensive reliability grading standard and the above judgment results, a reliability level is given for each measuring point. If the measuring point simultaneously meets the threshold judgment, rate of change judgment, working condition boundary judgment and conservation consistency judgment, it is marked as high reliability. If the measuring point only has slight over-limit, slight fluctuation or conservation residual in the middle range, it is marked as medium reliability. If the measuring point obviously violates the working condition boundary or conservation relationship, but has not completely exceeded the instrument range, it is marked as low reliability. If the measuring point exceeds the physical boundary, instrument range, or has long-term stagnation, communication interruption, or obvious reverse logic error, it is marked as invalid. The confidence weight is used for subsequent online state estimation, slow-varying parameter identification, and residual correction. When the confidence of a measurement point decreases, its impact on state estimation and parameter identification decreases accordingly. When a measurement point is marked as invalid, it does not directly participate in model correction, but is replaced by adjacent measurement points, mechanistic model predictions, or mass / energy conservation soft measurement values.

5. The dynamic simulation method for fusion of conventional island mechanism data in a solar thermal power plant according to claim 3, characterized in that: In step S3, online state estimation is performed based on the mechanism model and measured values ​​to obtain unmeasurable states. Unmeasurable states include: the average temperature of the working fluid inside the heat exchanger, the metal wall temperature, the pipe volume pressure, the internal state of the steam drum or deaerator volume components, the equivalent heat transfer state of the condenser, the interstage state of the steam turbine, and some extraction steam flow rates that cannot be directly measured. The specific processing procedure is as follows: First, the unpredictable state at the next moment is predicted based on the dynamic mechanism model, and the estimated state at the previous moment is: , with input quantity and current parameters Based on this, calculate the current time. Prior states: ； In the above formula, This represents the current state predicted by the mechanistic model, with the prior state corrected based on measured values, and the mechanistic model output is then presented. Compared with on-site measurements By comparison, the measurement residuals are obtained. : ； The online state estimator corrects the state variables based on the residual magnitude, measurement point reliability, and model uncertainty to obtain the posterior state. : ； in, The state correction gain is set according to the measurement point noise, model error and measurement point confidence. When a measurement point is judged to be abnormal or drifting, the weight of the corresponding measurement point is reduced to avoid abnormal measurement values ​​directly contaminating the twin state of the simulation system.

6. The dynamic simulation method for fusion of conventional island mechanism data in a solar thermal power plant according to claim 5, characterized in that: In step S4, the process of performing online identification of slowly varying parameters is as follows: After obtaining the real-time state estimation results, key parameters in the conventional island that change slowly with operating time but are difficult to measure directly are identified. The slow-changing parameters mainly include: the equivalent heat transfer coefficient UA of the condenser or air-cooled system, the equivalent heat transfer coefficient or heat loss coefficient of the feedwater preheater, evaporator, superheater, and reheater, the valve flow coefficient of the main steam valve, bypass valve, and extraction valve, the pressure loss coefficient of the pipeline and heat exchanger, the efficiency correction coefficient of the high-pressure cylinder, intermediate-pressure cylinder, and low-pressure cylinder of the steam turbine, and the efficiency correction coefficient of the pump equipment. The online identification employs a dual-timescale mechanism. For rapidly changing states, online filtering or state estimation methods from S3 are used for rapid updates. For slowly varying parameters such as heat transfer coefficient, pressure loss coefficient, and efficiency correction, the parameters do not change drastically with each sampling point, but are slowly updated within a certain time window through recursive identification or rolling optimization. The calculation process is as follows: The objective function is minimized using either a constrained weighted recursive least squares method or a rolling window optimization method. At the k-th sampling time, an identification time window is formed using N sampling points. The formula for the slowly varying parameter identification optimization problem is as follows: The constraint formula is: in, These are the estimated values ​​of the slowly varying parameters identified at the current time. These are the estimated values ​​of the slowly varying parameters from the previous time step. Let be the DCS measurement value at the i-th sampling time. For given parameters The output of the mechanistic model under the given conditions As the confidence weight of the measurement points, For parameter smoothing coefficients, and These are the lower and upper physical limits of the slowly varying parameters, respectively. The maximum allowable change of the slowly varying parameter within a unit update time; if the identification result meets the requirements of parameter boundary, rate of change constraint and measurement point reliability, then the slowly varying parameter is updated; otherwise, the parameter at the previous time step is retained or only amplitude limiting correction is performed. The constraint formula includes physical upper and lower limits for the parameter and constraints on the rate of change of the parameter, ensuring that the slowly varying parameter to be identified remains within the physically feasible range of the equipment and avoids drastic changes due to instantaneous disturbances or abnormal measuring points; the specific identification process of the slowly varying parameter includes the following steps: Based on equipment design parameters, performance test data, or historical stable operation data, determine the initial value, physical upper and lower limits, and maximum permissible rate of change of the slowly varying parameters to be identified. Slowly varying parameters include the heat exchanger's equivalent heat transfer coefficient UA and the valve flow coefficient. Pressure loss coefficient K, turbine efficiency correction coefficient ; Secondly, for different slowly varying parameters, corresponding input and output quantities are selected, and residuals are constructed and calculated based on the mechanism relationship. For the equivalent heat transfer coefficient of the heat exchanger, heat transfer residuals are constructed based on the relationship between heat transfer and logarithmic mean temperature difference. For the valve flow coefficient, the flow residual is constructed based on the pressure before and after the valve, the valve opening, and the working fluid density. For the pressure loss coefficient, the pressure loss residual is constructed based on the relationship between flow rate and pressure difference; For the turbine efficiency correction factor, the power residual is constructed based on the steam enthalpy drop, flow rate, and output power: Then, minimize the above residual within the rolling time window to obtain the candidate value of the slow-varying parameter at the current time. If the candidate value meets the physical upper and lower limits of the parameter, the rate of change constraint and the reliability requirements of the measurement point, then update the slow-varying parameter; otherwise, keep the parameter value of the previous time or only use the correction amount after the amplitude limit. Finally, the updated slowly varying parameters are fed back into the dynamic mechanism model to recalculate the equipment outlet parameters, turbine power, and system state. This enables online correction of the mechanism model as equipment performance degrades and operating conditions change. Key output variables are selected based on safety relevance, performance relevance, parameter sensitivity, measurement point availability, and data reliability. Key output variables must meet at least one of the following conditions: (1) These are the conventional island safety operation constraint variables, including main steam temperature, main steam pressure, reheat steam temperature, reheat steam pressure, condenser pressure, deaerator pressure, deaerator water level, steam drum water level or other volumetric liquid levels; these variables directly affect the equipment safety boundary and operation alarm judgment, and should be given priority in the residual correction objects. (2) Variables belonging to the unit performance evaluation include turbine output power, main steam flow rate, reheat steam flow rate, extraction steam flow rate at each stage, exhaust steam pressure, feedwater flow rate, feedwater temperature, and system heat rate related variables; such variables directly affect the unit performance calculation, operation economic evaluation and upper-level optimization results, and should be included in the residual correction object; (3) It belongs to the output variable that is sensitive to slowly varying parameters; for candidate output variables Calculate its slowly varying parameters to be identified Sensitivity: in, To normalize the sensitivity, To prevent extremely small positive numbers with a denominator of zero, when a certain output variable is highly sensitive to heat exchanger UA, valve flow coefficient, pressure loss coefficient, or turbine efficiency correction coefficient, it is considered a key output variable. For example, heat exchanger outlet temperature and heat transfer are sensitive to UA; valve downstream flow is sensitive to valve flow coefficient; pressure difference across the equipment is sensitive to pressure loss coefficient; turbine output power and exhaust parameters are sensitive to turbine efficiency correction coefficient. (4) Variables that can be directly measured in the DCS system or obtained by energy conservation and mass conservation soft measurement; data that cannot be measured, cannot be soft measured or has been missing for a long time shall not be used as the main target variables for residual correction. (5) The confidence level of the variable measurement points meets the requirements; if a measurement point is marked as high confidence or medium confidence within the current time window, it is allowed to participate in residual correction; if it is marked as low confidence, it is only used as an auxiliary reference; if it is marked as invalid, it will not participate in residual correction. According to the above standards, key output variables include, but are not limited to: the temperature, pressure, and flow rate of the working fluid at the outlet of the preheater, evaporator, superheater, and reheater in the steam generation system; the temperature, pressure, and flow rate of the main steam; the temperature, pressure, and flow rate of the reheat steam; the temperature, pressure, and flow rate of the extraction steam at each stage of the turbine; the turbine exhaust pressure and exhaust enthalpy; the turbine output power; the condenser pressure; the deaerator pressure and water level; and the feedwater flow rate, feedwater temperature, and feedwater pressure. When several candidate variables simultaneously meet the above conditions, priority should be given to selecting variables related to the safety boundary, variables with high sensitivity to slowly varying parameters, variables with high reliability of measurement points, and variables directly related to upper-level operation optimization and performance evaluation as residual correction objects.

7. The dynamic simulation method for fusion of conventional island mechanism data in a solar thermal power plant according to claim 6, characterized in that: In step S5, the specific process is as follows: The residuals between the model output and the measured values ​​are calculated, and the residual correction module is used to perform controlled corrections on systematic biases that are difficult for the mechanistic model to fully explain. Key output variables that need to be corrected are selected, and for each key variable, the residuals before correction are calculated: ； in, For model residuals, DCS measurement value For the output values ​​of the mechanistic model, a residual correction module is constructed accordingly. The inputs of the residual correction module include one or more of the following: current DCS measurement values, mechanistic model output, online estimation status, slowly varying parameter identification results, operating condition characteristics, and historical residual sequences; the output is the residual correction amount. The corrected twin output is: ； in, For the corrected twin output, amplitude and rate of change constraints are set for the residual correction amount. When the correction amount exceeds the physical reasonable range, the system does not directly use the correction result, but reduces the confidence of the correction result and prompts that there is an abnormal measurement point, the working condition exceeds the training range, or the model structure is missing. The process of outputting confidence level for the residual correction module is as follows: The confidence level is determined by a combination of factors, including whether the input operating condition is within the coverage of historical training data, the magnitude of residual fluctuation, the reliability of measurement points, the magnitude of correction, and the prediction error over a recent period. When the operating condition is within the familiar range of the model and the residual is stable, the confidence level is high; when start-up, shutdown, rapid load change, bypass switching, or measurement point abnormality occurs, the confidence level is low.

8. The dynamic simulation method for fusion of conventional island mechanism data in a solar thermal power plant according to claim 7, characterized in that: The specific implementation process of step S6 is as follows: After completing state estimation, parameter identification, and residual correction, the corrected conventional island digital twin result is output, which mainly includes three categories: The first category is the corrected twin state, including the temperature, pressure, and flow rate of the working fluid at the outlet of each heat exchanger in the steam generation system, the temperature, pressure, and flow rate of the main steam and reheat steam, the temperature, pressure, and flow rate of the steam extracted at each stage, the turbine exhaust parameters, the condenser pressure, the deaerator status, the feedwater system status, and the status of the main pipelines and volumetric components. The second category consists of performance indicators, including turbine output power, unit heat rate, heat exchange efficiency of steam generation system, condenser vacuum performance, feedwater pump power consumption, load change rate, main steam parameter deviation, reheat steam parameter deviation, and system energy balance deviation, which are used to evaluate the current operating performance of the conventional island and the degree of deviation from the design conditions. The third category is health indicators, including the degree of UA degradation in heat exchangers, the degree of decline in heat exchange performance of condensers, the trend of pressure loss coefficient, the degree of deviation in valve flow coefficient, the trend of turbine efficiency correction coefficient, the confidence level of measuring points, suspected drift measuring point markings, abnormal operating condition markings, and residual correction confidence levels, which are used to distinguish between actual equipment degradation, measuring point anomalies, and model uncertainty. The recording process involves each record including at least a timestamp, operating condition number, input data, mechanistic model output, state estimation result, slow-varying parameter identification result, residual correction amount, corrected output, confidence level, and anomaly marker.

Images