Thermal power unit agc load regulation method based on modeling of boiler thermodynamic characteristics

Abstract

The application discloses a thermal power unit AGC load regulation method based on boiler thermodynamic characteristic modeling, relates to the technical field of thermal power generation, and comprises the following steps: collecting and preprocessing unit operation data, establishing a dynamic fouling heat resistance soft measurement model based on heat accumulation balance, and outputting a fouling heat resistance factor; a first-order inertia plus pure delay transfer function model is established in response to a main steam pressure to a boiler combustion rate instruction, the fouling heat resistance factor obtained in the first step is coupled into real-time calculation of an inertia time constant, and a time-varying dynamic inertia time constant driven by a fouling state is obtained; the time-varying inertia time constant is embedded into a dynamic matrix, a time-varying step response column vector is constructed, an AGC instruction is inversely solved to generate a heat accumulation compensation feedforward control sequence, a boiler combustion rate instruction is obtained by superimposing a current step of the feedforward sequence and a feedback regulation amount, coal and air volume are adjusted, a governing valve on a steam turbine side is adjusted in response to the AGC instruction, and the previous steps are executed on line and are rolled over to update the time-varying inertia time constant and the feedforward sequence in real time.

Description

Technical Field

[0001] This invention belongs to the field of thermal power generation technology, and in particular relates to an AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling. Background Technology

[0002] With the continuous increase in the proportion of new energy power generation, the power grid has placed more stringent requirements on the regulation rate and accuracy of automatic generation control (AGC) of thermal power units. Circulating fluidized bed boilers, as the main type of boiler for the clean utilization of low-quality and high-alkali coal, typically employ AGC load regulation methods based on coordinated control strategies. This method decomposes the grid load command into turbine-side valve opening commands and boiler-side combustion rate commands, with the generation of the boiler-side combustion rate command depending on the thermal dynamic model of the controlled object. Specifically, in engineering, a first-order inertial plus pure delay transfer function model is commonly established based on the response characteristics of the main steam pressure to the combustion rate command. A proportional-integral-derivative feedback controller or feedforward compensator is designed using a fixed inertial time constant and gain parameters to coordinate the combustion rate with the turbine valve action, maintaining the main steam pressure within a safe range while meeting load response requirements.

[0003] However, the aforementioned load regulation method based on fixed thermal inertia parameters faces the problem of deteriorated control quality caused by fuel characteristics when applied to circulating fluidized bed boilers that co-fire high-alkali coal. Specifically, the core source of interference in this scenario comes from the dense ash layer formed by the condensation of alkali metals on the heating surfaces after gasification during the combustion of high-alkali coal. After the alkali metal elements such as potassium and sodium in high-alkali coal are gasified in the high-temperature zone of the furnace, they condense and deposit as they flow through the convective and semi-radiative heating surfaces with the flue gas, forming a highly viscous, low thermal conductivity fouling layer with fly ash particles. This type of fouling is fundamentally different from the loose ash produced by the combustion of conventional coal: First, the fouling layer is tightly bonded to the metal pipe wall, making it difficult to completely remove by conventional soot blowing methods, and it exhibits the characteristic of accumulating and increasing over time; Second, the thermal conductivity of the fouling layer is much lower than that of the metal pipe wall, and its thermal resistance accounts for a continuously increasing proportion of the total heat transfer resistance as the fouling develops; Third, the formation rate of alkali metal fouling has a non-linear relationship with the alkali metal content in the coal, the furnace temperature level, and the blending ratio, and the degree of fouling will undergo unpredictable time-varying drift when the blending ratio is adjusted.

[0004] Under such fouling interference on the heating surface, existing fixed-parameter AGC load regulation methods will produce serious model mismatch and control imbalance problems. On the one hand, at the dynamic characteristic level, the accumulation of fouling layer is equivalent to introducing an increasing additional thermal resistance between flue gas and working fluid. Evaporating the heating surface metal requires a longer time to absorb or release heat to change the working fluid state, which, from a control perspective, manifests as a significant lengthening of the boiler's inertial time constant. Feedback controllers designed with a fixed reference inertial time constant cannot adapt to this time-varying characteristic, resulting in significant dynamic deviations in main steam pressure during load changes, and in severe cases, triggering pressure protection actions. On the other hand, at the feedforward compensation level, conventional dynamic matrix control or feedforward controllers based on fixed step response models have a step response coefficient sequence in their predictive models that remains unchanged after commissioning. When fouling causes the actual response characteristics of the object to drift, a mismatch occurs between the feedforward compensation amount and the actual required heat storage compensation amount. This not only fails to effectively suppress pressure fluctuations but may also introduce additional regulation disturbances due to improper compensation, directly leading to unqualified regulation rate and regulation accuracy in AGC performance indicators, and increasing the risk of unplanned unit shutdowns.

[0005] The main steam pressure deviation mentioned in the background section specifically refers to the dynamic pressure deviation caused by a short-term mismatch between steam flow and heat during AGC load changes, due to the boiler combustion rate response lagging behind the turbine control valve opening changes. If the amplitude and duration of this deviation exceed the safety threshold, it will jeopardize the lifespan of the heated surface tube wall metal and reduce the economic efficiency of the entire plant's thermal cycle. Therefore, the following solutions are proposed to address the above problems. Summary of the Invention

[0006] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0007] This invention provides a method for AGC load regulation of thermal power units based on boiler thermodynamic characteristics modeling, comprising the following steps:

[0008] Step S1: Obtain real-time operating data of the unit from the distributed control system. After performing consistency verification and low-pass filtering on the raw data, establish the dynamic balance relationship between metal heat storage and working fluid heat storage through mechanism analysis. Construct an online soft measurement model of dynamic fouling thermal resistance of the boiler steam-water side lumped parameter type and calculate the dynamic fouling thermal resistance factor characterizing the degree of fouling.

[0009] Step S2: Establish a lumped parameter continuous step response transfer function model for the heat storage process of the heating surface of the circulating fluidized bed boiler. Taking the response characteristics of the main steam pressure to the boiler combustion rate command as the object, the dynamic fouling thermal resistance factor obtained in step S1 is coupled into the real-time calculation of the inertial time constant to form a time-varying parameter model driven by the physical state, and the corrected dynamic inertial time constant is obtained.

[0010] Step S3: Embed the dynamic inertia time constant obtained in step S2 into the dynamic matrix generation process under the model predictive control framework, construct a dynamic column vector based on the time-varying step response coefficient, and use the dynamic column vector to perform inverse operation on the AGC load command to generate the feedforward control sequence for active heat storage compensation.

[0011] Step S4: The current step execution value in the feedforward control sequence generated in step S3 is superimposed with the feedback increment output by the boiler main control proportional-integral-derivative feedback regulator to obtain the overall boiler combustion rate command increment, which is applied to the coal feeder speed and primary air volume regulation system; at the same time, the turbine-side digital electro-hydraulic control system receives the change part of the AGC load command and adjusts the turbine control valve opening; steps S1 to S3 are executed online in a rolling manner for each sampling period to realize the real-time update of the dynamic inertia time constant and the feedforward control sequence.

[0012] Further, step S1 includes the following steps:

[0013] Step S11: With sampling period Continuously read the main steam flow rate from the distributed control system Steam drum pressure Main steam pressure Furnace outlet flue gas temperature Steam drum wall temperature and the total flow rate of desuperheating water in each stage of the superheater The original data undergoes consistency checks and low-pass filtering to remove outliers.

[0014] Step S12: By using the balance relationship between the current flue gas temperature and saturation temperature, the working fluid heat absorption flow rate, and the metal heat storage rate, the real-time total heat transfer coefficient is solved in reverse and the clean baseline thermal resistance is subtracted to calculate the dynamic fouling thermal resistance factor. The dynamic fouling thermal resistance factor is used to quantify the equivalent thermal resistance of the ash layer on the heated surface that hinders the transfer of heat from flue gas to the metal pipe wall. The formula is as follows:

[0015] ;

[0016] In the formula, The main steam specific enthalpy, based on the main steam pressure. Determined by water vapor property functions; The feedwater specific enthalpy is determined by the economizer inlet parameters; The specific heat capacity of the metal on the evaporation heating surface. The effective mass of the metal on the evaporating heated surface is a known constant. The rate of change of the average temperature of the metal is given by the pressure in the steam drum. corresponding saturation temperature It is obtained by difference approximation calculation; This is the reference thermal resistance under clean conditions.

[0017] Furthermore, step S2 includes the following steps:

[0018] Step S21: Establish main steam pressure Boiler combustion rate directive The response characteristics are based on the object's first-order inertia plus pure delay transfer function model. ,in For gain, To delay time, The reference inertial time constant represents the rate at which the boiler effectively releases its stored heat.

[0019] Step S22: Calculate the corrected dynamic inertial time constant by multiplying the ratio of the dynamic contamination thermal resistance factor to the clean reference thermal resistance by the contamination amplification factor, adding one, and then multiplying by the reference inertial time constant under clean conditions. The modified dynamic inertia time constant is used to reflect the hysteresis effect of the degree of fouling of the heating surface on the boiler heat storage release rate, and the formula is as follows:

[0020] ;

[0021] In the formula, This is the reference inertial time constant of the boiler in a clean state. The reference thermal resistance under clean conditions. The contamination amplification factor is determined by fitting contamination data under the designed coal blending ratio.

[0022] Furthermore, step S3 includes the following steps:

[0023] Step S31: Increment the plant load according to the power dispatch center's instructions. In addition, based on the preset coordinated load distribution strategy between the turbine side and the boiler side, the expected change in the boiler side combustion rate command is calculated. By substituting the real-time dynamic inertia time constant, control step size, and number of predicted steps into the exponential decay expression of the step response coefficient, the predicted main steam pressure response value for each future step is calculated. The predicted response values ​​are used to construct a time-varying dynamic column vector. The formula is as follows:

[0024] ;

[0025] In the formula, For the gain of the transfer function model, The dynamic inertia time constant obtained in step S2, To control the step size, To predict the length of the time domain;

[0026] Step S32: Generate the feedforward control sequence for active thermal storage compensation by minimizing the performance index constructed from the time-varying dynamic column vector and the desired pressure deviation trajectory. The feedforward control sequence is used to provide an incremental sequence of boiler combustion rate commands in advance to compensate for boiler response hysteresis caused by changes in fouling thermal resistance. The formula is as follows:

[0027] ;

[0028] In the formula, For the future The deviation trajectory between the expected pressure and the current actual pressure for each step is determined by the pressure target value corresponding to the AGC load command and the current main steam pressure. The difference is calculated; This is the error weight matrix; To define the control inhibition weight matrix, set... , The adjustable inhibition factor; the first increment of the feedforward control sequence Extract it and use it as the current step execution value of the AGC feedforward instruction.

[0029] Furthermore, step S4 includes the following steps:

[0030] The overall boiler combustion rate command increment is calculated by adding the first increment of the feedforward control sequence to the feedback increment output by the feedback regulator. The overall command is used to synchronously adjust the coal feeder speed and primary air volume to quickly track the AGC load command and maintain stable main steam pressure. The formula is as follows:

[0031] ;

[0032] In the formula, The current step execution value in the feedforward control sequence generated in step S3. This refers to the feedback increment output of the boiler main control proportional-integral-derivative feedback regulator.

[0033] Furthermore, in step S11, the consistency check and low-pass filtering employ weighted moving average filtering, which calculates the filtered measurement point value by summing the sampled values ​​within the filtering window according to preset weight coefficients. The filtered measurement point values ​​are used to smooth the measurement signal and suppress random noise, as shown in the following formula:

[0034] ;

[0035] In the formula, For the first The original measurement point value at the sampling time, This is the length of the filtering window. The weighting coefficients are and satisfy the following conditions: .

[0036] Furthermore, in step S12, the small variation rate of the metal's average temperature The saturation temperature difference corresponding to the steam drum pressure is used for approximate calculation. By dividing the difference between the saturation temperature at the current moment and the previous moment by the sampling period, an approximate value of the micro-variability rate is obtained. This approximate value of the micro-variability rate is used to calculate the change in metal heat storage online. The formula is as follows:

[0037] ;

[0038] in, and The pressure of the steam drum at the current time and the previous time are respectively... The saturation temperature is determined by the saturated vapor property function. The sampling period.

[0039] Furthermore, in step S22, the contamination amplification factor... The fouling amplification factor is determined by linear regression based on historical operating data under the designed coal blending ratio. It is calculated by multiplying the high-alkali coal blending ratio by the first regression coefficient and then adding the second regression coefficient. This fouling amplification factor is used to quantitatively correct the sensitivity of the inertial time constant to the degree of fouling. The formula is as follows:

[0040] ;

[0041] In the formula, The proportion of high-alkali coal to be blended is... and The regression coefficients are obtained by least-squares fitting of the data on the changes in fouling thermal resistance and inertial time constant recorded by the unit under different co-firing ratios.

[0042] Furthermore, in step S31, the prediction time domain is... and control step size The selection of the time domain satisfies the following constraints: by ensuring that the product of the prediction time domain and the control step size is greater than or equal to the sum of twice the pure delay time and three times the current dynamic inertia time constant, the lower limit of the prediction time domain length is determined to ensure that the dynamic matrix covers the main dynamic response stages of the process. The formula is as follows:

[0043] ;

[0044] In the formula, For the pure delay time in the transfer function model, This is the corrected dynamic inertial time constant for the current moment.

[0045] Furthermore, in step S32, the error weight matrix As a diagonal matrix, by using the exponentially increasing forgetting factor as diagonal elements, recent prediction errors are given greater weight, thereby improving the adaptability of the feedforward optimization objective to the current contamination state, i.e., the th... diagonal elements Take values, where It is a forgetting factor and satisfies The control action inhibition weight matrix , .

[0046] The present invention has the following beneficial effects:

[0047] 1. This invention identifies the fouling thermal resistance of the heating surface online and maps changes in the degree of fouling to a dynamic adjustment of the boiler's thermal inertia time constant in real time, ensuring that the control model always matches the actual thermal state. The feedforward control sequence generated based on this time-varying model can proactively compensate for the lag in heat storage release caused by fouling in the early stages of load command changes, avoiding significant deviations in main steam pressure caused by model mismatch in conventional fixed parameter control. The reduction in pressure deviation directly reduces the alternating stress amplitude of boiler pressure-bearing components, reducing the risk of triggering protection actions due to pressure exceeding limits, thereby improving the unit's ability to operate continuously and safely under high-alkali coal co-firing conditions.

[0048] 2. Under the dynamic matrix control framework, this invention reconstructs the step response prediction model using real-time updated thermal inertia parameters. The feedforward control sequence can accurately predict and apply combustion rate compensation actions in advance. This physical state-driven adaptive feedforward strategy effectively offsets the response sluggishness caused by fouling, making the coupling transition process of main steam pressure and power generation closer to the design performance when the heating surface is clean, thus improving the speed and accuracy of load regulation.

[0049] 3. This invention establishes a continuous mapping relationship between online soft measurement of fouling thermal resistance and correction of inertial time constant, and the fouling amplification factor can be corrected by regression through the blending ratio, enabling control parameters to automatically adjust in response to changes in fuel quality. This adaptive mechanism, directly driven by physical states, allows the boiler coordinated control system to adapt to frequent changes in the blending ratio, expanding the operating window for the unit to safely and economically consume high-alkali coal.

[0050] 4. Through precise feedforward heat storage compensation, the coordination between combustion rate command and turbine valve opening is more coordinated, the matching degree of steam flow and heat in the initial stage of load change is improved, and the amount of desuperheating water that has to be put in due to steam temperature fluctuations is reduced; thus, unnecessary heat loss of working fluid is avoided, and the economic operation level of the unit is taken into account while ensuring AGC performance.

[0051] 5. All required input variables are derived from existing conventional measurement points in the power plant's distributed control system, eliminating the need for additional hardware sensors or boiler modifications. The online calculation process primarily utilizes algebraic and recursive difference operations, avoiding complex iterative optimization processes, resulting in low computational load and easy encapsulation as a custom function block within the DCS system. Continuous calculation and display of intermediate physical state variables such as fouling thermal resistance and time-varying inertial constants provide operators with an intuitive reference for the development trend of fouling on heated surfaces, facilitating optimization of soot blowing strategies and demonstrating good engineering practicality.

[0052] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0053] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 This is a schematic flowchart of the AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling according to the present invention. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] Please see Figure 1 As shown, this invention is a method for AGC load regulation of thermal power units based on boiler thermodynamic characteristics modeling, including the following steps:

[0057] Step S1: Obtain real-time operating data of the unit from the distributed control system. After performing consistency verification and low-pass filtering on the raw data, establish the dynamic balance relationship between metal heat storage and working fluid heat storage through mechanism analysis. Construct an online soft measurement model of dynamic fouling thermal resistance of the boiler steam-water side lumped parameter type and calculate the dynamic fouling thermal resistance factor characterizing the degree of fouling.

[0058] Step S2: Establish a lumped parameter continuous step response transfer function model for the heat storage process of the heating surface of the circulating fluidized bed boiler. Taking the response characteristics of the main steam pressure to the boiler combustion rate command as the object, the dynamic fouling thermal resistance factor obtained in step S1 is coupled into the real-time calculation of the inertial time constant to form a time-varying parameter model driven by the physical state, and the corrected dynamic inertial time constant is obtained.

[0059] Step S3: Embed the dynamic inertia time constant obtained in step S2 into the dynamic matrix generation process under the model predictive control framework, construct a dynamic column vector based on the time-varying step response coefficient, and use the dynamic column vector to perform inverse operation on the AGC load command to generate the feedforward control sequence for active heat storage compensation.

[0060] Step S4: The current step execution value in the feedforward control sequence generated in step S3 is superimposed with the feedback increment output by the boiler main control proportional-integral-derivative feedback regulator to obtain the overall boiler combustion rate command increment, which is applied to the coal feeder speed and primary air volume regulation system; at the same time, the turbine-side digital electro-hydraulic control system receives the change part of the AGC load command and adjusts the turbine control valve opening; steps S1 to S3 are executed online in a rolling manner for each sampling period to realize the real-time update of the dynamic inertia time constant and the feedforward control sequence.

[0061] Step S1 includes the following steps:

[0062] Step S11: With sampling period Continuously read the main steam flow rate from the distributed control system Steam drum pressure Main steam pressure Furnace outlet flue gas temperature Steam drum wall temperature and the total flow rate of desuperheating water in each stage of the superheater The original data undergoes consistency checks and low-pass filtering to remove outliers.

[0063] Step S12: By using the balance relationship between the current flue gas temperature and saturation temperature, the working fluid heat absorption flow rate, and the metal heat storage rate, the real-time total heat transfer coefficient is solved in reverse and the clean baseline thermal resistance is subtracted to calculate the dynamic fouling thermal resistance factor. The dynamic fouling thermal resistance factor is used to quantify the equivalent thermal resistance of the ash layer on the heated surface that hinders the transfer of heat from flue gas to the metal pipe wall. The formula is as follows:

[0064] ;

[0065] In the formula, The main steam specific enthalpy, based on the main steam pressure. Determined by water vapor property functions; The feedwater specific enthalpy is determined by the economizer inlet parameters; The specific heat capacity of the metal on the evaporation heating surface. The effective mass of the metal on the evaporating heated surface is a known constant. The rate of change of the average temperature of the metal is given by the pressure in the steam drum. corresponding saturation temperature It is obtained by difference approximation calculation; This is the reference thermal resistance under clean conditions.

[0066] Step S2 includes the following steps:

[0067] Step S21: Establish main steam pressure Boiler combustion rate directive The response characteristics are based on the object's first-order inertia plus pure delay transfer function model. ,in For gain, To delay time, The reference inertial time constant represents the rate at which the boiler effectively releases its stored heat.

[0068] Step S22: Calculate the corrected dynamic inertial time constant by multiplying the ratio of the dynamic contamination thermal resistance factor to the clean reference thermal resistance by the contamination amplification factor, adding one, and then multiplying by the reference inertial time constant under clean conditions. The corrected dynamic inertia time constant is used to reflect the hysteresis effect of the degree of fouling of the heating surface on the boiler heat storage release rate, and the formula is as follows:

[0069] ;

[0070] In the formula, This is the reference inertial time constant of the boiler in a clean state. The reference thermal resistance under clean conditions. The contamination amplification factor is determined by fitting contamination data under the designed coal blending ratio.

[0071] Step S3 includes the following steps:

[0072] Step S31: Increment the plant load according to the power dispatch center's instructions. In addition, based on the preset coordinated load distribution strategy between the turbine side and the boiler side, the expected change in the boiler side combustion rate command is calculated. By substituting the real-time dynamic inertia time constant, control step size, and number of predicted steps into the exponential decay expression of the step response coefficient, the predicted main steam pressure response value for each future step is calculated. The predicted response values ​​are used to construct a time-varying dynamic column vector. The formula is as follows:

[0073] ;

[0074] In the formula, For the gain of the transfer function model, The dynamic inertia time constant obtained in step S2, To control the step size, To predict the length of the time domain;

[0075] Step S32: Generate the feedforward control sequence for active thermal storage compensation by minimizing the performance index constructed from the time-varying dynamic column vector and the desired pressure deviation trajectory. The feedforward control sequence is used to provide an incremental sequence of boiler combustion rate commands in advance to compensate for boiler response hysteresis caused by changes in fouling thermal resistance. The formula is as follows:

[0076] ;

[0077] In the formula, For the future The deviation trajectory between the expected pressure and the current actual pressure for each step is determined by the pressure target value corresponding to the AGC load command and the current main steam pressure. The difference is calculated; This is the error weight matrix; To define the control inhibition weight matrix, set... , The adjustable inhibition factor; the first increment of the feedforward control sequence Extract it and use it as the current step execution value of the AGC feedforward instruction.

[0078] Step S4 includes the following steps:

[0079] The overall boiler combustion rate command increment is calculated by adding the first increment of the feedforward control sequence to the feedback increment output by the feedback regulator. The overall command is used to synchronously adjust the feeder speed and primary air volume to quickly track the AGC load command and maintain stable main steam pressure. The formula is as follows:

[0080] ;

[0081] In the formula, The current step execution value in the feedforward control sequence generated in step S3. This refers to the feedback increment output of the boiler main control proportional-integral-derivative feedback regulator.

[0082] In step S11, the consistency check and low-pass filtering adopt a weighted moving average filter. The filtered measurement point value is calculated by summing the sampled values ​​within the filtering window according to preset weight coefficients. The filtered measurement point values ​​are used to smooth the measurement signal and suppress random noise, as shown in the following formula:

[0083] ;

[0084] In the formula, For the first The original measurement point value at the sampling time, This is the length of the filtering window. The weighting coefficients are and satisfy the following conditions: .

[0085] In step S12, the small change rate of the metal's average temperature The saturation temperature difference corresponding to the steam drum pressure is used for approximate calculation. By dividing the difference between the saturation temperature at the current moment and the previous moment by the sampling period, an approximate value of the micro-variability is obtained. This approximate value of the micro-variability is used to calculate the change in metal heat storage online. The formula is as follows:

[0086] ;

[0087] in, and The pressure of the steam drum at the current time and the previous time are respectively... The saturation temperature is determined by the saturated vapor property function. The sampling period.

[0088] In step S22, the contamination amplification factor The fouling amplification factor is determined by linear regression based on historical operating data under the designed coal blending ratio. It is calculated by multiplying the high-alkali coal blending ratio by the first regression coefficient and then adding the second regression coefficient. The fouling amplification factor is used to quantitatively correct the sensitivity of the inertial time constant to the degree of fouling. The formula is as follows:

[0089] ;

[0090] In the formula, The proportion of high-alkali coal to be blended is... and The regression coefficients are obtained by least-squares fitting of the data on the changes in fouling thermal resistance and inertial time constant recorded by the unit under different co-firing ratios.

[0091] In step S31, the time domain is predicted. and control step size The selection of the time domain satisfies the following constraints: by ensuring that the product of the prediction time domain and the control step size is greater than or equal to the sum of twice the pure delay time and three times the current dynamic inertia time constant, the lower limit of the prediction time domain length is determined to ensure that the dynamic matrix covers the main dynamic response stages of the process. The formula is as follows:

[0092] ;

[0093] In the formula, For the pure delay time in the transfer function model, This is the corrected dynamic inertial time constant for the current moment.

[0094] In step S32, the error weight matrix As a diagonal matrix, by using the exponentially increasing forgetting factor as diagonal elements, recent prediction errors are given greater weight, thereby improving the adaptability of the feedforward optimization objective to the current contamination state, i.e., the th... diagonal elements Take values, where It is a forgetting factor and satisfies Control effect inhibition weight matrix , .

[0095] One specific application of this embodiment is:

[0096] This embodiment focuses on a 300MW subcritical circulating fluidized bed thermal power unit. This unit is designed for high-proportion blending of Xinjiang high-alkali coal. The grid AGC regulation assessment requirements are: load change rate of 2% of rated load / min, and dynamic deviation of main steam pressure controlled within 0.3MPa. The basic parameters of the unit under rated operating conditions are as follows:

[0097] Rated main steam pressure: Rated main steam temperature: ;

[0098] Rated main steam flow rate: ;

[0099] Boiler reference inertia time constant under clean conditions: ;

[0100] Pure delay time of main steam pressure response: ;

[0101] Steady-state gain of transfer function: ;

[0102] Specific heat capacity of the metal in the evaporation heating surface: ;

[0103] Effective mass of the evaporation heating surface metal: ;

[0104] Clean-state reference thermal resistance: ;

[0105] Rated operating condition feedwater specific enthalpy: ;

[0106] The specific implementation steps of the AGC load adjustment method in this embodiment are as follows:

[0107] Step S1: Data Acquisition and Preprocessing of Unit Operation and Calculation of Dynamic Fouling Thermal Resistance Factor

[0108] Step S11: Perform data acquisition and preprocessing

[0109] With sampling period Continuously read real-time operating measurement data from the unit's distributed control system (DCS). The original values ​​of the core measurement points at each time point are as follows:

[0110] Main steam flow rate: ;

[0111] Steam drum pressure: ;

[0112] Main steam pressure: ;

[0113] Furnace outlet flue gas temperature: ;

[0114] The corresponding saturation temperature of the steam drum: ;

[0115] Saturation temperature at the previous moment: ;

[0116] Total flow rate of desuperheating water in each stage of the superheater: ;

[0117] Perform consistency checks and weighted moving average low-pass filtering on the original data to remove outliers. The filter window length is [not specified]. The weighting coefficient is set to , , , , ,satisfy The filter calculation formula is:

[0118] ;

[0119] In the formula, For the first The original measurement point value at the sampling time, These are the filtered measurement values. After filtering, the smoothness of the data at each measurement point is significantly improved, and random measurement noise is suppressed to within 0.2% of the full scale.

[0120] Step S12: Calculation of dynamic fouling thermal resistance factor

[0121] Based on the dynamic equilibrium relationship between metal heat storage and working fluid heat storage, the real-time total heat transfer coefficient is solved in reverse and the clean baseline thermal resistance is subtracted to calculate the dynamic fouling thermal resistance factor characterizing the degree of fouling on the heated surface. The calculation formula is:

[0122] ;

[0123] In the formula:

[0124] Main vapor specific enthalpy, based on Calculated using the water vapor property function ;

[0125] The rate of change of the average temperature of the metal is approximated using the saturation temperature difference approximation, and the formula is: Substituting the values, we get: .

[0126] Substituting all the above parameters into the fouling thermal resistance factor formula and unifying the dimensions, we get:

[0127] ;

[0128]

[0129] Step S2: Correction of dynamic inertia time constant based on contamination thermal resistance

[0130] Step S21: Establishment of the baseline transfer function model of the controlled object

[0131] Establish based on main steam pressure Boiler combustion rate directive The baseline model of the first-order inertial plus pure delay transfer function for response characteristics:

[0132] ;

[0133] Substituting the baseline parameters of this embodiment, we obtain the baseline model:

[0134] ;

[0135] Step S22: Calculation of time-varying dynamic inertia time constant

[0136] The dynamic fouling thermal resistance factor obtained in step S1 is coupled into the real-time calculation of the inertial time constant to first determine the fouling amplification factor. The proportion of high-alkali coal blending in this embodiment. The regression coefficients were obtained by least squares fitting based on the historical operating data of the unit. , The formula for calculating the contamination amplification factor is:

[0137] ;

[0138] Substituting the values, we get: .

[0139] Correction of dynamic inertia time constant based on contamination thermal resistance factor The calculation formula is:

[0140] ;

[0141] Substituting the values, we get:

[0142] ;

[0143] The corrected dynamic inertia time constant accurately reflects the hysteresis effect of heating surface fouling on the boiler's heat storage release rate.

[0144] Step S3: Construction of Time-Varying Dynamic Matrix and Generation of Feedforward Control Sequence for Heat Storage Compensation

[0145] Step S31: Construction of Time-Varying Step Response Coefficients and Dynamic Column Vectors

[0146] The power dispatch center issued an incremental AGC load command. Based on the boiler-turbine coordinated load allocation strategy, the expected change in the boiler-side combustion rate command was calculated. Rated combustion rate.

[0147] Set control step size Predicting the length of the time domain Constraints must be met:

[0148] ;

[0149] Substitute the lower bound of the numerical calculation constraint: Therefore, the prediction time domain is taken. The constraints are met.

[0150] Based on the real-time dynamic inertial time constant, the predicted response values ​​of the main steam pressure at each future step are calculated. Construct time-varying dynamic column vectors The formula for calculating the step response coefficient is:

[0151] ;

[0152] Substitute the values ​​into the numerical calculation of the response coefficients for a typical step size:

[0153] when hour:

[0154] when hour: ;

[0155] Depend on to Constructing a 9101-dimensional time-varying dynamic column vector This enables real-time updates of the dynamic matrix.

[0156] Step S32: Solving the active thermal storage compensation feedforward control sequence

[0157] The target main steam pressure corresponding to the AGC load command is: Current actual main steam pressure Total deviation Plan the future according to a 150s linear transition. The expected pressure deviation trajectory of each step .

[0158] Set the error weight matrix For a diagonal matrix, the forgetting factor is... , No. The diagonal elements are:

[0159]

[0160] Control effect inhibition weight matrix Among them, tunable inhibitory factors , It is an identity matrix.

[0161] The feedforward control sequence for active thermal storage compensation is generated by minimizing the performance index. The calculation formula is:

[0162] ;

[0163] After solving the matrix operation, the feedforward control sequence is obtained. The first increment of the sequence is taken as the execution value of the current step.

[0164]

[0165] Step S4: Coordinated Adjustment and Online Rolling Execution of Boiler-Turbine Side

[0166] The current step execution value of the feedforward control sequence is superimposed with the feedback increment output by the boiler main control PID feedback regulator to calculate the overall boiler combustion rate command increment. The formula is as follows:

[0167] ;

[0168] In this embodiment, the feedback increment of the boiler main control PID output... The rated combustion rate is used to calculate the total combustion rate instruction increment:

[0169]

[0170] The total command increment is applied to the coal feeder speed and primary air volume regulation system: the coal feeder speed is synchronously adjusted from 380 r / min to 420 r / min, and the primary air volume is adjusted from 180,000 m3 / h to 192,000 m3 / h, achieving rapid and precise adjustment of the combustion rate; at the same time, the turbine-side digital electro-hydraulic control system (DEH) receives the AGC load command and opens the turbine control valve from 68% to 74%, synchronously responding to load changes.

[0171] Steps S1 to S3 The sampling cycle is executed online on a rolling basis. The dynamic contamination thermal resistance factor, time-varying inertia time constant and feedforward control sequence are updated in real time in each cycle to realize full-condition adaptive control of AGC load regulation.

[0172] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0173] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for AGC load regulation of thermal power units based on boiler thermodynamic characteristics modeling, characterized in that, Includes the following steps: Step S1: Obtain real-time operating data of the unit from the distributed control system. After performing consistency verification and low-pass filtering on the raw data, establish the dynamic balance relationship between metal heat storage and working fluid heat storage through mechanism analysis. Construct an online soft measurement model of dynamic fouling thermal resistance of the boiler steam-water side lumped parameter type and calculate the dynamic fouling thermal resistance factor characterizing the degree of fouling. Step S2: Establish a lumped parameter continuous step response transfer function model for the heat storage process of the heating surface of the circulating fluidized bed boiler. Taking the response characteristics of the main steam pressure to the boiler combustion rate command as the object, the dynamic fouling thermal resistance factor obtained in step S1 is coupled into the real-time calculation of the inertial time constant to form a time-varying parameter model driven by the physical state, and the corrected dynamic inertial time constant is obtained. Step S3: Embed the dynamic inertia time constant obtained in step S2 into the dynamic matrix generation process under the model predictive control framework, construct a dynamic column vector based on the time-varying step response coefficient, and use the dynamic column vector to perform inverse operation on the AGC load command to generate the feedforward control sequence for active heat storage compensation. Step S4: The current step execution value in the feedforward control sequence generated in step S3 is superimposed with the feedback increment output by the boiler main control proportional-integral-derivative feedback regulator to obtain the overall boiler combustion rate command increment, which is applied to the coal feeder speed and primary air volume regulation system; at the same time, the turbine-side digital electro-hydraulic control system receives the change part of the AGC load command and adjusts the turbine control valve opening; steps S1 to S3 are executed online in a rolling manner for each sampling period to realize the real-time update of the dynamic inertia time constant and the feedforward control sequence.

2. The AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: With sampling period Continuously read the main steam flow rate from the distributed control system Steam drum pressure Main steam pressure Furnace outlet flue gas temperature Steam drum wall temperature and the total flow rate of desuperheating water in each stage of the superheater The original data undergoes consistency checks and low-pass filtering to remove outliers. Step S12: By using the balance relationship between the current flue gas temperature and saturation temperature, the working fluid heat absorption flow rate, and the metal heat storage rate, the real-time total heat transfer coefficient is solved in reverse and the clean baseline thermal resistance is subtracted to calculate the dynamic fouling thermal resistance factor. The dynamic fouling thermal resistance factor is used to quantify the equivalent thermal resistance of the ash layer on the heated surface that hinders the transfer of heat from flue gas to the metal pipe wall. The formula is as follows: ； In the formula, The main steam specific enthalpy, based on the main steam pressure. Determined by water vapor property functions; The feedwater specific enthalpy is determined by the economizer inlet parameters; The specific heat capacity of the metal on the evaporation heating surface. The effective mass of the metal on the evaporating heated surface is a known constant. The rate of change of the average temperature of the metal is given by the pressure in the steam drum. corresponding saturation temperature It is obtained by difference approximation calculation; This is the reference thermal resistance under clean conditions.

3. The AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Establish main steam pressure Boiler combustion rate directive The response characteristics are based on the object's first-order inertia plus pure delay transfer function model. ,in For gain, To delay time, The reference inertial time constant represents the rate at which the boiler effectively releases its stored heat. Step S22: Calculate the corrected dynamic inertial time constant by multiplying the ratio of the dynamic contamination thermal resistance factor to the clean reference thermal resistance by the contamination amplification factor, adding one, and then multiplying by the reference inertial time constant under clean conditions. The modified dynamic inertia time constant is used to reflect the hysteresis effect of the degree of fouling of the heating surface on the boiler heat storage release rate, and the formula is as follows: ； In the formula, This is the reference inertial time constant of the boiler in a clean state. The reference thermal resistance under clean conditions. The contamination amplification factor is determined by fitting contamination data under the designed coal blending ratio.

4. The AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Increment the plant load according to the power dispatch center's instructions. In addition, based on the preset coordinated load distribution strategy between the turbine side and the boiler side, the expected change in the boiler side combustion rate command is calculated. By substituting the real-time dynamic inertia time constant, control step size, and number of predicted steps into the exponential decay expression of the step response coefficient, the predicted main steam pressure response value for each future step is calculated. The predicted response values ​​are used to construct a time-varying dynamic column vector. The formula is as follows: ； In the formula, For the gain of the transfer function model, The dynamic inertia time constant obtained in step S2, To control the step size, To predict the length of the time domain; Step S32: Generate the feedforward control sequence for active thermal storage compensation by minimizing the performance index constructed from the time-varying dynamic column vector and the desired pressure deviation trajectory. The feedforward control sequence is used to provide an incremental sequence of boiler combustion rate commands in advance to compensate for boiler response hysteresis caused by changes in fouling thermal resistance. The formula is as follows: ； In the formula, For the future The deviation trajectory between the expected pressure and the current actual pressure for each step is determined by the pressure target value corresponding to the AGC load command and the current main steam pressure. The difference is calculated; This is the error weight matrix; To define the control inhibition weight matrix, set... , The adjustable inhibition factor; the first increment of the feedforward control sequence Extract it and use it as the current step execution value of the AGC feedforward instruction.

5. The AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling according to claim 1, characterized in that, Step S4 includes the following steps: The overall boiler combustion rate command increment is calculated by adding the first increment of the feedforward control sequence to the feedback increment output by the feedback regulator. The overall command is used to synchronously adjust the coal feeder speed and primary air volume to quickly track the AGC load command and maintain stable main steam pressure. The formula is as follows: ； In the formula, The current step execution value in the feedforward control sequence generated in step S3. This refers to the feedback increment output of the boiler main control proportional-integral-derivative feedback regulator.

6. The AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling according to claim 1, characterized in that, In step S11, the consistency check and low-pass filtering adopt a weighted moving average filter, which calculates the filtered measurement point value by summing the sampled values ​​within the filtering window according to a preset weight coefficient. The filtered measurement point values ​​are used to smooth the measurement signal and suppress random noise, as shown in the following formula: ； In the formula, For the first The original measurement point value at the sampling time, The length of the filter window. The weighting coefficients are and satisfy the following conditions: .

7. The AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling according to claim 1, characterized in that, In step S12, the small change rate of the average temperature of the metal The saturation temperature difference corresponding to the steam drum pressure is used for approximate calculation. By dividing the difference between the saturation temperature at the current moment and the previous moment by the sampling period, an approximate value of the micro-variability rate is obtained. This approximate value of the micro-variability rate is used to calculate the change in metal heat storage online. The formula is as follows: ； in, and The pressure of the steam drum at the current time and the previous time are respectively... The saturation temperature is determined by the saturated vapor property function. The sampling period.

8. The AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling according to claim 1, characterized in that, In step S22, the contamination amplification factor The fouling amplification factor is determined by linear regression based on historical operating data under the designed coal blending ratio. It is calculated by multiplying the high-alkali coal blending ratio by the first regression coefficient and then adding the second regression coefficient. This fouling amplification factor is used to quantitatively correct the sensitivity of the inertial time constant to the degree of fouling. The formula is as follows: ； In the formula, The proportion of high-alkali coal to be blended is... and The regression coefficients are obtained by least-squares fitting of the data on the changes in fouling thermal resistance and inertial time constant recorded by the unit under different co-firing ratios.

9. The AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling according to claim 1, characterized in that, In step S31, the prediction time domain is... and control step size The selection of the time domain satisfies the following constraints: by ensuring that the product of the prediction time domain and the control step size is greater than or equal to the sum of twice the pure delay time and three times the current dynamic inertia time constant, the lower limit of the prediction time domain length is determined to ensure that the dynamic matrix covers the main dynamic response stages of the process. The formula is as follows: ； In the formula, For the pure delay time in the transfer function model, This is the corrected dynamic inertial time constant for the current moment.

10. The AGC load regulation method for thermal power units based on boiler thermodynamic characteristics modeling according to claim 1, characterized in that, In step S32, the error weight matrix As a diagonal matrix, by using the exponentially increasing forgetting factor as diagonal elements, recent prediction errors are given greater weight, thereby improving the adaptability of the feedforward optimization objective to the current contamination state, i.e., the th... diagonal elements Take values, where It is a forgetting factor and satisfies The control action inhibition weight matrix , .

Images