A decoupling control method and system for a direct-fired pulverizing system based on a reduced-order full-drive active disturbance rejection

Abstract

This invention discloses a decoupling control method and system for a direct-fired pulverizing system based on reduced-order full-drive self-disturbance rejection, belonging to the field of automatic control technology for pulverizing systems in thermal power units. The method collects signals such as coal feed rate, primary air volume, pulverizer outlet temperature, and cold and hot damper openings to construct a pipeline network model and a pulverizer material-thermal dynamic model and identify parameters. Using primary air volume and outlet temperature as controlled outputs, and target primary air volume and target inlet mixed air temperature as virtual boundary inputs, a reduced-order full-drive standard model is constructed. The coal powder output and lumped disturbance are estimated through an internal material state observer and a linearly expanded state observer to calculate the target virtual boundary input, which is then processed by a cold and hot damper analytical distributor to obtain the target damper opening. After applying constraints to the damper opening, the actual achievable virtual boundary input is calculated and substituted back into the observer. This invention can improve the decoupling control accuracy of primary air volume and outlet temperature, and enhance the system's stability under disturbance and constrained operating conditions.

Description

Technical Field

[0001] This invention belongs to the field of automatic control technology for pulverizing systems of thermal power units, and particularly relates to a decoupling control method and system for a direct-fired pulverizing system based on reduced-order full-drive self-disturbance rejection. Background Technology

[0002] The direct-fired pulverizing system is a crucial component of boiler combustion control in coal-fired power units. Its primary task is to prepare and transport pulverized coal according to the unit's load requirements, while maintaining key variables such as primary air volume and mill outlet temperature within safe ranges. Primary air volume affects pulverized coal transport and boiler combustion stability, while mill outlet temperature affects pulverized coal drying efficiency and the safety of the pulverizing system.

[0003] In actual operation, direct-fired pulverizing systems are characterized by multivariable coupling, large inertia, nonlinearity, time-varying parameters, and significant actuator constraints. Both cold and hot air dampers affect the primary air volume and inlet mixed air temperature. The inlet mixed air temperature, coal feed rate, material inventory inside the mill, and primary air volume collectively determine the mill outlet temperature. Traditional control schemes typically design the primary air volume loop and outlet temperature loop as independent PID control loops, then allocate the opening of the cold and hot air dampers through empirical logic. This method struggles to accurately handle the coupling relationship between the cold and hot air dampers and the actuator limiting and speed constraints. Under conditions of varying coal feed rate, coal quality disturbances, hot air temperature fluctuations, or damper jamming, problems such as decreased control accuracy, control saturation, or uncoordinated cold and hot air damper actions can easily occur.

[0004] Active disturbance rejection control (ADRC) can estimate and compensate for unknown disturbances using an extended state observer, making it adaptable to industrial process control. However, if ADRC is only used for single-loop compensation, problems such as inconsistency between the ideal virtual input and the actual damper input, disturbance estimation deviation, and actuator saturation accumulation may still occur. Summary of the Invention

[0005] The purpose of this invention is to provide a decoupling control method and system for a direct-fired pulverizing system based on reduced-order full-drive self-disturbance rejection, in order to solve the problems existing in the control of existing direct-fired pulverizing systems, such as strong multivariable coupling, difficulty in handling actuator constraints, inconsistency between ideal control input and actual damper input, and easy deviation of disturbance estimation.

[0006] To achieve the above objectives, this invention provides a decoupling control method for a direct-fired pulverizing system based on reduced-order all-drive self-disturbance rejection, comprising the following steps:

[0007] S1. Obtain the operating signals of the direct-fired pulverizing system, including coal feed rate, primary air volume, coal mill outlet temperature, cold air damper opening, hot air damper opening, cold air temperature, and hot air temperature;

[0008] S2. Based on the operating signals, construct a primary air volume-mixed air temperature pipeline network model and a coal mill material-thermal dynamic model, and use the primary air volume and coal mill outlet temperature as controlled outputs, and the target primary air volume and target inlet mixed air temperature as virtual boundary inputs.

[0009] S3. Based on historical operating data and composite fitness function, identify the parameters of the primary air volume-mixed air temperature pipeline network model and the coal mill material-thermal dynamic model;

[0010] S4. Construct an internal material state observer based on the coal mill material-thermal dynamic model to estimate the amount of raw coal, pulverized coal, and pulverized coal blowout inside the coal mill.

[0011] S5. The direct-blown pulverizing system is constructed as a reduced-order, all-drive standard type:

[0012]

[0013] in, , For primary air volume, The outlet temperature of the coal mill. , For the target primary air volume, For the target inlet mixed air temperature, Given a dynamic term, For the all-drive input matrix, For aggregated disturbances, This is the output of the internal material state observer. For the coal feed mass flow rate, col(·) indicates that a column vector is constructed based on the elements within the parentheses;

[0014] S6. Based on the primary air volume measurement, the coal mill outlet temperature measurement, and the real reachable virtual boundary input, a linear extended state observer is constructed to estimate the lumped disturbances in the air volume loop and the temperature loop.

[0015] S7. Calculate the target virtual boundary input based on the primary air volume reference value, the outlet temperature reference value, the known dynamic term, the reduced-order all-drive standard form, and the estimated value of the lumped disturbance;

[0016] S8. The target virtual boundary input is converted into the target opening degree of the cold air damper and the target opening degree of the hot air damper through the cold and hot air damper distributor;

[0017] S9. Apply amplitude and rate constraints to the target opening of the cold air door and the target opening of the hot air door to obtain the actual opening of the cold air door and the actual opening of the hot air door. Then, calculate the real reachable virtual boundary input based on the actual opening of the cold air door and the actual opening of the hot air door, and substitute the real reachable virtual boundary input back into the linear expansion state observer.

[0018] S10. The actual opening degree of the cold air damper and the actual opening degree of the hot air damper are used as control signals for the direct-fired pulverizing system to achieve decoupled control of primary air volume and coal mill outlet temperature.

[0019] The operating signals refer to the process variables collected in real time during the current sampling period or control process of the direct-fired pulverizing system. These variables are used for online state observation, disturbance estimation, virtual boundary input calculation, hot and cold damper analytical allocation, and actuator constraint back-substitution. The historical operating data refers to historical operating condition data pre-collected and stored during the parameter identification phase. This data can originate from a distributed control system, programmable logic controller, or simulation platform, and includes coal feed rate, primary air volume, pulverizer outlet temperature, cold damper opening, hot damper opening, cold air temperature, and hot air temperature over multiple sampling periods. The historical operating data is used to update the parameters of the primary air volume-mixed air temperature network model and the pulverizer material-thermal dynamic model offline or online. After parameter identification is completed, the online control phase performs control calculations based on the real-time operating signals and the identified parameters.

[0020] Based on the above technical solutions, the present invention may also employ the following further technical solutions, or combine these further technical solutions:

[0021] In step S3, parameter identification includes the following methods: The parameter to be identified is denoted as... A composite fitness function is constructed based on the actual output values ​​and model predicted output values ​​from historical operating data. :

[0022]

[0023] in, The parameter vector to be identified, It is a composite fitness function. The total number of sampling points. For sampling sequence number, For the first The actual output of each sampling point This is the model prediction output with parameter θ. , , Weighted by the rate of change deviation. Representing the 1-norm of a vector; minimizing based on particle swarm optimization algorithm. The model parameters are obtained.

[0024] In step S2, the primary air volume-mixed air temperature pipeline network model includes the primary air volume boundary relationship and the inlet mixed air temperature boundary relationship; the primary air volume boundary relationship is as follows:

[0025]

[0026] in, For the target primary air volume, For the opening of the cold air damper, For hot air damper opening, The equivalent flow coefficient from the cold air damper to the primary air volume. The equivalent flow coefficient from the hot air damper to the primary air volume. This represents the signed cross-coupling coefficient of the hot and cold air dampers. This is the primary air volume offset term; the inlet mixed air temperature boundary relationship is:

[0027]

[0028]

[0029] in, The temperature at the junction of hot and cold air. The equivalent heat weight of hot air relative to cold air. For hot air temperature, The temperature of the cold air. To prevent positive numbers with a denominator of zero, For the target inlet mixed air temperature, For ambient reference temperature, This is a correction factor for the temperature of the mixed air. This is a mixed air temperature offset term.

[0030] In step S2, the coal mill material-thermal dynamic model includes the following equivalent state equations:

[0031]

[0032]

[0033]

[0034]

[0035] in, This refers to the quantity and status of raw coal inside the coal mill. This refers to the quantity and status of pulverized coal inside the coal mill. For coal feed mass flow rate, The coefficient of the raw coal grinding rate. This refers to the amount of pulverized coal blown out. This is the primary air pulverized coal carrying capacity coefficient. For primary air mass flow rate, For reference material quantity, The outlet temperature of the coal mill. For the inlet mixed air temperature, For ambient reference temperature, The equivalent coefficient for convective heat transfer is... The sensible heat equivalent coefficient of pulverized coal leaving the machine. This is the equivalent coefficient for heat absorption during moisture evaporation in raw coal. This is a constant heat loss term. This is the heat loss coefficient related to temperature difference.

[0036] In step S4, the internal material state observer is updated as follows:

[0037]

[0038]

[0039]

[0040] in, This is an estimated value for the amount of raw coal. This is an estimated value for the amount of pulverized coal. This is an estimated value for the amount of pulverized coal blown out; the estimated value for the amount of pulverized coal blown out is used to calculate the known dynamic terms of the temperature loop.

[0041] In step S5, the known dynamic terms and the all-drive input matrix in the reduced-order all-drive standard form are:

[0042]

[0043]

[0044]

[0045]

[0046] in, For a primary air volume loop, the dynamic term is known. For the temperature loop, the dynamic terms are known. The virtual input gain for the primary air volume loop. For the virtual input gain of the temperature loop, The primary air volume dynamic time constant, Operators that represent diagonal matrices constructed from diagonal elements; when Greater than zero Greater than the preset minimum air volume and When greater than zero, and All are positive values, the total drive input matrix The full rank condition is met.

[0047] In step S6, the linearly extended state observer is updated discretely; for any loop The linearly extended state observer is:

[0048]

[0049]

[0050]

[0051] in, Number the loops. Indicates the primary air volume circuit. Indicates the temperature circuit. For sampling sequence number, The sampling period is For the first The controlled output of each loop For the first The known dynamic terms of each loop, For the first The input gain of each loop, The input is the real reachable virtual boundary obtained by back-calculation from the actual damper opening. For the first The output estimate of each loop, Let be the estimated lumped disturbance value for the i-th loop. and For observer gain, For the first The observer bandwidth of each loop.

[0052] In step S7, the target virtual boundary input is calculated according to the following all-drive closed-loop configuration law:

[0053]

[0054] in, Input the target virtual boundary. The target primary air volume is the virtual input. Virtual input for mixed air temperature at the target inlet. Let B be the inverse matrix of the total drive input matrix. As the reference vector, This is a reference value for primary air volume. This is a reference value for the coal mill outlet temperature. The first derivative of the reference vector, This is the closed-loop bandwidth matrix. For the primary air volume loop control bandwidth, For temperature loop control bandwidth, To output the estimated vector, This is the lumped disturbance estimation vector.

[0055] In step S8, the cold and hot air damper distributor obtains the target opening degree of the cold air damper and the target opening degree of the hot air damper in the following manner: first, the virtual boundary of the temperature loop target is input. Converted to target hybrid node temperature :

[0056]

[0057] in, For the target mixing node temperature, Virtual input for mixed air temperature at the target inlet. For ambient reference temperature, For mixed air temperature offset, This is the temperature correction factor for mixed airflow, and To ensure the validity of subsequent ratio calculations, Limiting the temperature range of cold and hot air to a safe level:

[0058]

[0059] in, To ensure a temperature safety margin, the opening ratio of the hot air damper to the cold air damper is then calculated. :

[0060]

[0061] in, The temperature of the cold air. For hot air temperature, The equivalent heat weight of hot air relative to cold air; then... Substituting the primary air volume boundary relationship, we obtain the information regarding the cold air damper opening. The quadratic equation of :

[0062]

[0063]

[0064] Where a, b, and c are the coefficients of the quadratic equation in one variable. This represents the signed cross-coupling coefficient of the hot and cold air dampers. The equivalent flow coefficient from the cold air damper to the primary air volume. The equivalent flow coefficient from the hot air damper to the primary air volume. This is the primary air volume offset term. The target primary air volume is used as a virtual input; the candidate cold air damper opening is obtained by solving the quadratic equation, and then... Candidate hot air damper openings are obtained; when the candidate openings do not meet the damper amplitude constraints or the quadratic equation does not have a real solution, the target openings of the cold air damper and the hot air damper are determined by the reachable domain projection method.

[0065] In step S9, the amplitude constraint and rate constraint are applied as follows:

[0066]

[0067]

[0068]

[0069] in, and The first The actual opening degree of the cold air damper and the actual opening degree of the hot air damper in each sampling period. and These represent the target opening of the cold air damper and the target opening of the hot air damper in the k-th sampling period, respectively. This represents the maximum allowable change in aperture during a single sampling period. The maximum operating speed of the actuator. The sampling period is and These are the lower and upper limits of the damper opening, respectively. The saturation function is used to define the actual damper opening after the restriction. and Substituting the primary air volume-mixed air temperature boundary model, the true reachable virtual boundary input is calculated. and Its expression is:

[0070]

[0071] in, For the first The actual achievable target primary air volume is calculated from the actual damper opening in each sampling period. For the first The actual reachable inlet mixed air temperature is calculated from the actual damper opening in each sampling period. and These are the equivalent flow coefficients of the cold air damper and the hot air damper for the primary air volume, respectively. This represents the signed cross-coupling coefficient of the hot and cold air dampers. This is the primary air volume offset term. This represents the equivalent heat weight of hot air relative to cold air. and The first The cold air temperature and hot air temperature for each sampling period. To prevent positive numbers with excessively small denominators; and As the actual input to the linearly extended state observer, it is used for observer updates in the next sampling period or the current sampling period to reduce the perturbation estimation offset caused by the inconsistency between the target virtual boundary input and the actual input after the actuator is constrained.

[0072] This invention also provides a decoupling control system for a direct-fired pulverizing system based on reduced-order all-drive self-disturbance rejection, employing the following technical solution:

[0073] It includes: a data acquisition module, a model building module, a parameter identification module, a state observation module, a disturbance estimation module, a virtual boundary control module, a damper allocation module, a constraint back-substitution module, and an execution output module;

[0074] The data acquisition module is used to acquire the coal feed rate, primary air volume, coal mill outlet temperature, cold air damper opening, hot air damper opening, cold air temperature, and hot air temperature of the direct-fired pulverizing system.

[0075] The model building module is used to build a primary air volume-mixed air temperature pipeline network model, a coal mill material-thermal dynamic model, and a reduced-order full-drive standard model.

[0076] The parameter identification module is used to identify model parameters based on historical operating data; the state observation module is used to estimate the amount of raw coal, pulverized coal, and pulverized coal blown out of the coal mill.

[0077] The disturbance estimation module is used to estimate the lumped disturbances in the primary air volume loop and the outlet temperature loop based on the linear expansion state observer.

[0078] The virtual boundary control module is used to calculate the target virtual boundary input based on the primary air volume reference value, the outlet temperature reference value, the known dynamic term, the total drive input matrix, and the lumped disturbance estimate.

[0079] The damper distribution module is used to convert the target virtual boundary input into the target opening degree of the cold damper and the target opening degree of the hot damper;

[0080] The constraint back-substitution module is used to apply amplitude constraints and rate constraints to the target opening of the cold air door and the target opening of the hot air door, and to substitute the real reachable virtual boundary input back to the disturbance estimation module;

[0081] The execution output module is used to output the actual opening degree of the cold air damper and the actual opening degree of the hot air damper, so as to achieve decoupled control of the primary air volume and the coal mill outlet temperature of the direct-fired pulverizing system.

[0082] The beneficial effects of this invention are as follows: First, by uniformly constructing the primary air volume and the coal mill outlet temperature as a reduced-order full-drive standard type with respect to the virtual boundary input, the structural decoupling capability of the multivariable system is improved. Second, by using the target primary air volume and the target inlet mixed air temperature as virtual boundary inputs and implementing physical execution through the hot and cold damper analyzer, the control law design layer and the actuator allocation layer are separated. Third, by using the internal material state observer to estimate the coal powder blowing rate and introducing it into the known dynamic term of the temperature loop, the accuracy of temperature loop disturbance compensation is improved. Fourth, by using the real reachable virtual boundary input for updating in the linear expansion state observer, the inconsistency between the target virtual boundary input and the actual input caused by actuator amplitude limiting, speed limiting, and reachability domain projection can be reduced, thus exhibiting an anti-saturation effect. Attached Figure Description

[0083] Figure 1 This is a flowchart of the method of the present invention;

[0084] Figure 2 This is a block diagram of the reduced-order all-drive active disturbance rejection control system of the direct-blown pulverizing system of the present invention;

[0085] Figure 3 This is a schematic diagram of the parameter identification process of the present invention;

[0086] Figure 4 This is a schematic diagram of the standard structure of the reduced-order all-wheel drive system of the present invention;

[0087] Figure 5 This is a schematic diagram illustrating the hot and cold air damper distribution of the present invention;

[0088] Figure 6 This is a block diagram of the system structure of the present invention;

[0089] Figure 7 This is a comparison chart of the control results of primary air volume, coal mill outlet temperature, and cold and hot air damper openings in an embodiment of the present invention. Detailed Implementation

[0090] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described below with reference to the accompanying drawings and embodiments. The following embodiments are for illustrative purposes only and are not intended to limit the scope of protection of this invention. Those skilled in the art can combine, substitute, or make equivalent modifications to the technical features of this invention without departing from the principles of this invention.

[0091] To avoid symbol ambiguity, this embodiment adopts... The mass flow rate of coal feed is expressed using... This indicates the outlet temperature of the coal mill; this document does not use superscript characters to indicate transpose; when column vectors are needed, use col(·) to avoid confusion with temperature symbols.

[0092] Example 1: Control Method

[0093] like Figure 1 and Figure 2 As shown in the figure, this embodiment provides a decoupling control method for a direct-fired pulverizing system based on reduced-order full-drive self-disturbance rejection.

[0094] First, the operating signals of the direct-fired pulverizing system are collected. These operating signals include time t, coal feed rate, and mass flow rate. Cold air door opening Hot air damper opening Primary air volume Cold air temperature Hot air temperature Inlet Mixed Air Temperature and the outlet temperature of the coal mill The cold air temperature and hot air temperature can be obtained by averaging multiple measuring points.

[0095] 1. Primary air volume-mixed air temperature duct network model

[0096] Construct a primary air volume-mixed air temperature piping network model. The primary air volume boundary relationships are as follows:

[0097]

[0098] in, For the target primary air volume, For the opening of the cold air damper, For hot air damper opening, The equivalent flow coefficient from the cold air damper to the primary air volume. The equivalent flow coefficient from the hot air damper to the primary air volume. This represents the signed cross-coupling coefficient of the hot and cold air dampers. This is the primary airflow offset term. When the value is negative, the cross term characterizes the coupled throttling effect when both hot and cold air dampers are opened simultaneously; when When the value is positive, the cross term represents the coupling enhancement effect.

[0099] The inlet mixed air temperature boundary relationship is as follows:

[0100]

[0101]

[0102] in, The temperature at the junction of hot and cold air. The equivalent heat weight of hot air relative to cold air. For hot air temperature, The temperature of the cold air. To prevent positive numbers with a denominator of zero, For the target inlet mixed air temperature, For ambient reference temperature, This is a correction factor for the temperature of the mixed air. This is a mixed air temperature offset term.

[0103] 2. Material-Thermodynamic Dynamic Model of Coal Mill

[0104] Construct a material-thermal dynamic model for the coal mill. This involves calculating the coal feed rate and mass flow rate. As the input of raw coal, the primary air volume Converted to primary air mass flow rate Define the state of raw coal quantity inside the coal mill. and coal powder quantity status Material dynamics are as follows:

[0105]

[0106]

[0107]

[0108] in, This refers to the quantity and status of raw coal inside the coal mill. This refers to the quantity and status of pulverized coal inside the coal mill. For coal feed mass flow rate, The coefficient of the raw coal grinding rate. This refers to the amount of pulverized coal blown out. This is the primary air pulverized coal carrying capacity coefficient. For primary air mass flow rate, For reference material quantity.

[0109] Temperature dynamics are as follows:

[0110]

[0111] in, The outlet temperature of the coal mill. For the inlet mixed air temperature, For ambient reference temperature, The equivalent coefficient for convective heat transfer is... The sensible heat equivalent coefficient of pulverized coal leaving the machine. This is the equivalent coefficient for heat absorption during moisture evaporation in raw coal. This is a constant heat loss term. This is the heat loss coefficient related to temperature difference.

[0112] 3. Parameter Identification

[0113] In one implementation, such as Figure 3As shown, historical operating data was used to identify the parameters of the primary air volume-mixing air temperature pipeline network model and the coal mill material-thermal dynamic model. The parameter identification employed a composite fitness function:

[0114]

[0115] in, The parameter vector to be identified, It is a composite fitness function. The total number of sampling points. For sampling sequence number, For the first The actual output of each sampling point This is the model prediction output with parameter θ. , , Weighted by the rate of change deviation. This represents the first norm of a vector.

[0116] The velocity and position updates of the particle swarm optimization algorithm can be expressed as:

[0117]

[0118]

[0119] Where j is the particle number and m is the iteration number. Let be the velocity vector of the j-th particle in the m-th iteration. Let be the position vector of the j-th particle at the m-th iteration and represent a set of parameters to be identified, where w is the inertia weight. and As a learning factor, and A random number between 0 and 1 Let be the historical best position vector of the j-th particle. This is the globally optimal position vector for the particle swarm. By minimizing... Obtain the model parameters.

[0120] 4. Internal material condition observer

[0121] Since the amount of raw coal and pulverized coal inside a coal mill is usually unmeasurable, this embodiment constructs an internal material state observer:

[0122]

[0123]

[0124]

[0125] in, This is an estimated value for the amount of raw coal. This is an estimated value for the amount of pulverized coal. This is an estimated value for the amount of pulverized coal blown out. , , , and The meaning is the same as described above. This observer uses the measurable coal feed rate and primary air volume to update the status and obtain an estimate of the pulverized coal blowout rate. .

[0126] 5. Downgraded all-wheel drive standard model

[0127] like Figure 4 As shown, let the controlled output vector and the virtual boundary input vector be respectively:

[0128]

[0129]

[0130] in, The controlled output vector, For primary air volume, The outlet temperature of the coal mill. The virtual boundary input vector, For the target primary air volume, For the target inlet mixed air temperature, This indicates that a column vector is constructed based on the elements within the parentheses.

[0131] The primary air volume circuit can be written as:

[0132]

[0133]

[0134]

[0135] in, For a primary air volume loop, the dynamic term is known. The virtual input gain for the primary air volume loop. This is the lumped disturbance of the primary air volume loop. The dynamic time constant of primary air volume.

[0136] The temperature loop can be written as:

[0137]

[0138]

[0139]

[0140] in, For the temperature loop, the dynamic terms are known. For the virtual input gain of the temperature loop, This represents the lumped disturbance in the temperature loop; the meanings of other symbols are the same as described above.

[0141] This leads to the downgraded all-wheel drive standard model:

[0142]

[0143]

[0144]

[0145]

[0146] in, Given a dynamic term vector, For the all-drive input matrix, Operators that represent diagonal matrices constructed from diagonal elements. This is the lumped perturbation vector. Because... hour ,and Greater than the preset minimum air volume hour ,therefore It is a full-rank matrix. This full-rank condition ensures that the target primary air volume and the target inlet mixed air temperature can be used as two independent virtual boundary inputs, and the two outputs, primary air volume and outlet temperature, can be configured for full drive.

[0147] 6. Linear Extended State Observer

[0148] Construct linear expansion state observers for the primary air volume loop and the outlet temperature loop, respectively. For any loop... The discrete update form is:

[0149]

[0150]

[0151]

[0152] in, Number the loops. Indicates the primary air volume circuit. Indicates the temperature circuit. For sampling sequence number, The sampling period is For the first The controlled output of each loop For the first The known dynamic terms of each loop, For the first The input gain of each loop, The input is the real reachable virtual boundary obtained by back-calculation from the actual damper opening. For the first The output estimate of each loop, Let be the estimated lumped disturbance value for the i-th loop. and For observer gain, For the first The observer bandwidth of each loop.

[0153] 7. Calculation of Virtual Boundary Input for Full Drive

[0154] According to the reference signal Calculate the target virtual boundary input:

[0155]

[0156] in, Input the target virtual boundary. The target primary air volume is the virtual input. Virtual input for mixed air temperature at the target inlet. Let B be the inverse matrix of the total drive input matrix. As the reference vector, This is a reference value for primary air volume. This is a reference value for the coal mill outlet temperature. The first derivative of the reference vector, This is the closed-loop bandwidth matrix. For the primary air volume loop control bandwidth, For temperature loop control bandwidth, To output the estimated vector, Let be the lumped disturbance estimation vector. For a constant reference, we can let This formula enables decoupling of the airflow loop and temperature loop at the virtual boundary input level, achieving order reduction for the entire drive.

[0157] 8. Analysis and allocation of hot and cold air dampers

[0158] like Figure 5 As shown, we obtain Next, hot and cold air damper allocation is performed. First, the target inlet mixing air temperature is converted into the target mixing node temperature:

[0159]

[0160] in, For the target mixing node temperature, Virtual input for mixed air temperature at the target inlet. For ambient reference temperature, For mixed air temperature offset, This is the temperature correction factor for mixed airflow, and To ensure the validity of subsequent ratio calculations, Limiting the temperature range of cold and hot air to a safe level:

[0161]

[0162] in, To ensure a temperature safety margin, the opening ratio of the hot air damper to the cold air damper is then calculated. :

[0163]

[0164] in, The temperature of the cold air. For hot air temperature, The equivalent heat weight of hot air relative to cold air; then... Substituting the primary air volume boundary relationship, we obtain the information regarding the cold air damper opening. The quadratic equation of :

[0165]

[0166]

[0167]

[0168]

[0169] Where a, b, and c are the coefficients of the quadratic equation in one variable. This represents the signed cross-coupling coefficient of the hot and cold air dampers. The equivalent flow coefficient from the cold air damper to the primary air volume. The equivalent flow coefficient from the hot air damper to the primary air volume. This is the primary air volume offset term. The target primary air volume is used as a virtual input; the candidate cold air damper opening is obtained by solving the quadratic equation, and then... Obtain candidate hot air damper opening .

[0170] If multiple candidate solutions satisfy the constraints, the one that minimizes the weighted sum of the primary air volume error, the inlet mixed air temperature error, and the damper opening change is selected. If no candidate solution satisfies the constraints, projection optimization can be performed within the damper's reachable region.

[0171]

[0172] in, To optimize the objective function for projection, and The reachable virtual boundary input obtained by substituting the candidate damper opening into the pipeline network model. and Assigning weights to airflow error and temperature error. The penalty weight for damper action, and These represent the actual opening degrees of the cold and hot air dampers in the previous sampling period, respectively.

[0173] 9. Actuator constraints and real input back substitution

[0174] Apply amplitude and rate constraints to the damper opening:

[0175]

[0176]

[0177]

[0178] in, and The first The actual opening degree of the cold air damper and the actual opening degree of the hot air damper in each sampling period. and The first The target opening of the cold air damper and the target opening of the hot air damper for each sampling period. This represents the maximum allowable change in aperture during a single sampling period. Maximum operating rate of the actuator The sampling period is and These are the lower and upper limits of the damper opening, respectively. It is a saturation function.

[0179] Next, the actual damper opening will be... and Substituting the primary air volume-mixed air temperature duct network model, the actual reachable virtual boundary input is calculated in reverse:

[0180]

[0181]

[0182] in, For the first The actual achievable target primary air volume is calculated from the actual damper opening in each sampling period. For the first The actual reachable inlet mixed air temperature is calculated from the actual damper opening in each sampling period. and These are the equivalent flow coefficients of the cold air damper and the hot air damper for the primary air volume, respectively. This represents the signed cross-coupling coefficient of the hot and cold air dampers. This is the primary air volume offset term. This represents the equivalent heat weight of hot air relative to cold air. and The first The cold air temperature and hot air temperature for each sampling period. To prevent positive numbers with excessively small denominators, we will... and Substituting back into the linear extended state observer can prevent the extended state observer from mistakenly believing that the ideal virtual input has been fully applied to the system, thereby reducing the estimation bias and integral saturation caused by actuator constraints.

[0183] Example 2: System Implementation

[0184] like Figure 6 As shown in the figure, this embodiment also provides a decoupling control system for a direct-fired pulverizing system based on reduced-order all-drive self-disturbance rejection. The system includes a data acquisition module, a model building module, a parameter identification module, a state observation module, a disturbance estimation module, a virtual boundary control module, a damper allocation module, a constraint back-substitution module, and an execution output module.

[0185] The data acquisition module acquires coal feed rate, primary air volume, pulverizer outlet temperature, cold air damper opening, hot air damper opening, cold air temperature, and hot air temperature from the DCS, PLC, or simulation platform. The model building module stores or updates online the primary air volume-mixed air temperature network model parameters and the pulverizer material-thermal dynamic model parameters, and constructs a reduced-order full-drive standard model based on these parameters. The parameter identification module identifies model parameters based on historical operating data and a composite fitness function. The state observation module estimates the amount of raw coal, pulverized coal, and pulverized coal blown out of the pulverizer. The disturbance estimation module executes a linearly extended state observer to estimate the lumped disturbances in the primary air volume loop and the outlet temperature loop. The virtual boundary control module calculates the target primary air volume and target inlet mixed air temperature based on the full-drive closed-loop configuration law. The damper allocation module converts the target primary air volume and target inlet mixed air temperature into target cold air damper openings and target hot air damper openings based on analytical allocation and reachability domain projection. The constraint back-substitution module is used to perform inverse calculations of the upper and lower limits of the damper opening, the rate constraint, and the input of the actual reachable virtual boundary, and then substitutes the inverse calculation results back to the disturbance estimation module. The execution output module is used to output the actual opening command to the cold damper actuator and the hot damper actuator.

[0186] The above methods and systems can be implemented by computer programs. These computer programs can be stored in computer-readable storage media, and when executed by a processor, they implement the decoupling control method for the direct-blown pulverizing system described in any embodiment of the present invention. The processor can be a central processing unit, digital signal processor, embedded controller, PLC, DCS controller, field-programmable gate array, or other programmable logic device.

[0187] Example 3: Comparison of Control Results

[0188] like Figure 7 As shown, under the same coal feed disturbance and actuator constraints, the method of the present invention and the comparative method are subjected to closed-loop simulation. Figure 7 The upper part shows the coal mill outlet temperature response curve, the middle part shows the primary air volume response curve, and the lower part shows the cold and hot air damper opening change curves.

[0189] Depend on Figure 7 As can be seen, compared with the comparative method, the method of the present invention enables the coal mill outlet temperature to operate more stably around the reference value, reducing the temperature deviation caused by fluctuations in coal feed rate; at the same time, the primary air volume can track the set value better, and the changes in the opening of the cold and hot dampers are more coordinated. This result demonstrates that the described reduced-order full-drive virtual input, cold and hot damper analytical allocation, and real reachable input back-substitution mechanism can improve the coupling effect between primary air volume and outlet temperature, and reduce the observation mismatch caused by actuator constraints.

[0190] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, all technical solutions obtained through equivalent substitution or transformation fall within the protection scope of the present invention.

Claims

1. A decoupling control method for a direct-fired pulverizing system based on reduced-order full-drive self-disturbance rejection, characterized in that, Includes the following steps: S1. Obtain the operating signals of the direct-fired pulverizing system, including coal feed rate, primary air volume, coal mill outlet temperature, cold air damper opening, hot air damper opening, cold air temperature, and hot air temperature; S2. Based on the operating signals, construct a primary air volume-mixed air temperature pipeline network model and a coal mill material-thermal dynamic model, and use the primary air volume and coal mill outlet temperature as controlled outputs, and the target primary air volume and target inlet mixed air temperature as virtual boundary inputs. S3. Based on historical operating data and composite fitness function, identify the parameters of the primary air volume-mixed air temperature pipeline network model and the coal mill material-thermal dynamic model; S4. Construct an internal material state observer based on the coal mill material-thermal dynamic model to estimate the amount of raw coal, pulverized coal, and pulverized coal blowout inside the coal mill. S5. The direct-blown pulverizing system is constructed as a reduced-order, all-drive standard type: in, , For primary air volume, The outlet temperature of the coal mill. , For the target primary air volume, For the target inlet mixed air temperature, Given a dynamic term, For the all-drive input matrix, For aggregated disturbances, This is the output of the internal material state observer. For the coal feed mass flow rate, col(·) indicates that a column vector is constructed based on the elements within the parentheses; S6. Based on the primary air volume measurement, the coal mill outlet temperature measurement, and the real reachable virtual boundary input, a linear extended state observer is constructed to estimate the lumped disturbances in the air volume loop and the temperature loop. S7. Calculate the target virtual boundary input based on the primary air volume reference value, the outlet temperature reference value, the known dynamic term, the reduced-order all-drive standard form, and the estimated value of the lumped disturbance; S8. The target virtual boundary input is converted into the target opening degree of the cold air damper and the target opening degree of the hot air damper through the cold and hot air damper distributor; S9. Apply amplitude and rate constraints to the target opening of the cold air door and the target opening of the hot air door to obtain the actual opening of the cold air door and the actual opening of the hot air door. Then, calculate the real reachable virtual boundary input based on the actual opening of the cold air door and the actual opening of the hot air door, and substitute the real reachable virtual boundary input back into the linear expansion state observer. S10. The actual opening degree of the cold air damper and the actual opening degree of the hot air damper are used as control signals for the direct-fired pulverizing system to achieve decoupled control of primary air volume and coal mill outlet temperature.

2. The decoupling control method for a direct-blown pulverizing system as described in claim 1, characterized in that, In step S3, parameter identification includes the following methods: The parameter to be identified is denoted as... A composite fitness function is constructed based on the actual output values ​​and model predicted output values ​​from historical operating data. : in, The parameter vector to be identified, It is a composite fitness function. The total number of sampling points. For sampling sequence number, For the first The actual output of each sampling point This is the model prediction output with parameter θ. , , Weighted by the rate of change deviation. Representing the 1-norm of a vector; minimizing based on particle swarm optimization algorithm. The model parameters are obtained.

3. The decoupling control method for a direct-blown pulverizing system as described in claim 1, characterized in that, In step S2, the primary air volume-mixed air temperature pipeline network model includes the primary air volume boundary relationship and the inlet mixed air temperature boundary relationship; the primary air volume boundary relationship is as follows: in, For the target primary air volume, For the opening of the cold air damper, For hot air damper opening, The equivalent flow coefficient from the cold air damper to the primary air volume. The equivalent flow coefficient from the hot air damper to the primary air volume. This represents the signed cross-coupling coefficient of the hot and cold air dampers. This is the primary air volume offset term; the inlet mixed air temperature boundary relationship is: in, The temperature at the junction of hot and cold air. The equivalent heat weight of hot air relative to cold air. For hot air temperature, The temperature of the cold air. To prevent positive numbers with a denominator of zero, For the target inlet mixed air temperature, For ambient reference temperature, This is a correction factor for the temperature of the mixed air. This is a mixed air temperature offset term.

4. The decoupling control method for a direct-blown pulverizing system as described in claim 1, characterized in that, In step S2, the coal mill material-thermal dynamic model includes the following equivalent state equations: in, This refers to the quantity and status of raw coal inside the coal mill. This refers to the quantity and status of pulverized coal inside the coal mill. The mass flow rate of coal feed The coefficient of the raw coal grinding rate. This refers to the amount of pulverized coal blown out. This is the primary air pulverized coal carrying capacity coefficient. For primary air mass flow rate, For reference material quantity, The outlet temperature of the coal mill. For the inlet mixed air temperature, For ambient reference temperature, The equivalent coefficient for convective heat transfer is... The sensible heat equivalent coefficient of pulverized coal leaving the machine. This is the equivalent coefficient for heat absorption due to moisture evaporation in raw coal. This is a constant heat loss term. This is the heat loss coefficient related to temperature difference.

5. The decoupling control method for a direct-blown pulverizing system as described in claim 4, characterized in that, In step S4, the internal material state observer is updated as follows: in, This is an estimated value for the amount of raw coal. This is an estimated value for the amount of pulverized coal. This is an estimated value for the amount of pulverized coal blown out; the estimated value for the amount of pulverized coal blown out is used to calculate the known dynamic terms of the temperature loop.

6. The decoupling control method for a direct-blown pulverizing system as described in claim 1, characterized in that, In step S5, the known dynamic terms and the all-drive input matrix in the reduced-order all-drive standard form are: in, For a primary air volume loop, the dynamic term is known. For the temperature loop, the dynamic terms are known. The virtual input gain for the primary air volume loop. For the virtual input gain of the temperature loop, The primary air volume dynamic time constant, Operators that represent diagonal matrices constructed from diagonal elements; when Greater than zero Greater than the preset minimum air volume and When greater than zero, and All are positive values, the total drive input matrix The full rank condition is met.

7. The decoupling control method for a direct-blown pulverizing system as described in claim 1, characterized in that, In step S6, the linearly extended state observer is updated discretely; for any loop The linearly extended state observer is: in, Number the loops. Indicates the primary air volume circuit. Indicates the temperature circuit. For sampling sequence number, The sampling period is For the first The controlled output of each loop For the first The known dynamic terms of each loop, For the first The input gain of each loop, The input is the real reachable virtual boundary obtained by back-calculation from the actual damper opening. For the first The output estimate of each loop, Let be the lumped disturbance estimate for the i-th loop. and For observer gain, For the first The observer bandwidth of each loop.

8. The decoupling control method for a direct-blown pulverizing system as described in claim 7, characterized in that, In step S7, the target virtual boundary input is calculated according to the following all-drive closed-loop configuration law: in, Input the target virtual boundary. The target primary air volume is the virtual input. Virtual input for mixed air temperature at the target inlet. Let B be the inverse matrix of the total drive input matrix. As the reference vector, This is a reference value for primary air volume. This is a reference value for the coal mill outlet temperature. The first derivative of the reference vector, This is the closed-loop bandwidth matrix. For the primary air volume loop control bandwidth, For temperature loop control bandwidth, To output the estimated vector, This is the lumped disturbance estimation vector.

9. The decoupling control method for a direct-blown pulverizing system as described in claim 1, characterized in that, In step S8, the cold and hot air damper distributor obtains the target opening degree of the cold air damper and the target opening degree of the hot air damper in the following manner: first, the virtual boundary of the temperature loop target is input. Converted to target hybrid node temperature : in, For the target mixing node temperature, Virtual input for mixed air temperature at the target inlet. For ambient reference temperature, For mixed air temperature offset, This is the temperature correction factor for mixed airflow, and To ensure the validity of subsequent ratio calculations, Limiting the temperature range of cold and hot air to a safe level: in, To ensure a temperature safety margin, the opening ratio of the hot air damper to the cold air damper is then calculated. : in, The temperature of the cold air. For hot air temperature, The equivalent heat weight of hot air relative to cold air; then... Substituting the primary air volume boundary relationship, we obtain the information regarding the cold air damper opening. The quadratic equation of : Where a, b, and c are the coefficients of the quadratic equation in one variable. This represents the signed cross-coupling coefficient of the hot and cold air dampers. The equivalent flow coefficient from the cold air damper to the primary air volume. The equivalent flow coefficient from the hot air damper to the primary air volume. This is the primary air volume offset term. The target primary air volume is used as a virtual input; the candidate cold air damper opening is obtained by solving the quadratic equation in one variable, and then... Candidate hot air damper openings are obtained; when the candidate openings do not meet the damper amplitude constraints or the quadratic equation does not have a real solution, the target openings of the cold air damper and the hot air damper are determined by the reachable domain projection method.

10. The decoupling control method for a direct-blown pulverizing system as described in claim 9, characterized in that, In step S9, the amplitude constraint and rate constraint are applied as follows: in, and The first The actual opening degree of the cold air damper and the actual opening degree of the hot air damper in each sampling period. and These represent the target opening of the cold air damper and the target opening of the hot air damper in the kth sampling period, respectively. This represents the maximum allowable change in aperture during a single sampling period. The maximum operating speed of the actuator. The sampling period is and These are the lower and upper limits of the damper opening, respectively. The saturation function is used to define the actual damper opening after the restriction. and Substituting the primary air volume-mixed air temperature boundary model, the true reachable virtual boundary input is calculated. and Its expression is: in, For the first The actual achievable target primary air volume is calculated from the actual damper opening in each sampling period. For the first The actual reachable inlet mixed air temperature is calculated from the actual damper opening in each sampling period. and These are the equivalent flow coefficients of the cold air damper and the hot air damper for the primary air volume, respectively. This represents the signed cross-coupling coefficient of the hot and cold air dampers. This is the primary air volume offset term. This represents the equivalent heat weight of hot air relative to cold air. and The first The cold air temperature and hot air temperature for each sampling period. To prevent positive numbers with excessively small denominators; and As the actual input to the linearly extended state observer, it is used for observer updates in the next sampling period or the current sampling period to reduce the perturbation estimation offset caused by the inconsistency between the target virtual boundary input and the actual input after the actuator is constrained.

11. A decoupling control system for a direct-fired pulverizing system based on reduced-order full-drive self-disturbance rejection, characterized in that, include: The module includes a data acquisition module, a model building module, a parameter identification module, a state observation module, a disturbance estimation module, a virtual boundary control module, a damper allocation module, a constraint back-substitution module, and an execution output module. The data acquisition module is used to acquire the coal feed rate, primary air volume, coal mill outlet temperature, cold air damper opening, hot air damper opening, cold air temperature, and hot air temperature of the direct-fired pulverizing system. The model building module is used to build a primary air volume-mixed air temperature pipeline network model, a coal mill material-thermal dynamic model, and a reduced-order full-drive standard model. The parameter identification module is used to identify model parameters based on historical operating data; the state observation module is used to estimate the amount of raw coal, pulverized coal, and pulverized coal blown out of the coal mill. The disturbance estimation module is used to estimate the lumped disturbances in the primary air volume loop and the outlet temperature loop based on the linear expansion state observer. The virtual boundary control module is used to calculate the target virtual boundary input based on the primary air volume reference value, the outlet temperature reference value, the known dynamic term, the total drive input matrix, and the lumped disturbance estimate. The damper distribution module is used to convert the target virtual boundary input into the target opening degree of the cold damper and the target opening degree of the hot damper; The constraint back-substitution module is used to apply amplitude constraints and rate constraints to the target opening of the cold air door and the target opening of the hot air door, and to substitute the real reachable virtual boundary input back to the disturbance estimation module; The execution output module is used to output the actual opening degree of the cold air damper and the actual opening degree of the hot air damper, so as to achieve decoupled control of the primary air volume and the coal mill outlet temperature of the direct-fired pulverizing system.

Images