Two-step reconstruction anti-interference PID control method and control method of SCR flue gas denitration system

Abstract

The invention discloses a two-step reconstruction anti-interference PID control method and a control method of an SCR flue gas denitration system, and the method comprises the steps: carrying out cascade active-disturbance-rejection control reconstruction, decomposing a controlled object of a cascade PID control system into a determined part model and an uncertainty part model containing system disturbance through virtual decomposition, and carrying out the reconstruction of the cascade active-disturbance-rejection control reconstruction on the basis of the determined part model, constructing a cascade active-disturbance-rejection control system of which an inner ring adopts internal model control and an outer ring adopts active-disturbance-rejection control; carrying out anti-interference PID control reconstruction, equivalently reconstructing the cascade active-disturbance-rejection control system to form an anti-interference PID control system by adopting an improved disturbance observer, and decomposing system disturbance into negligible internal disturbance and external disturbance which generates influence on a controlled variable and meets a preset requirement by the improved disturbance observer; and performing two-step reconstructed anti-interference PID control based on the anti-interference PID control system.

Description

Technical Field

[0001] This invention belongs to the field of industrial control technology, and in particular relates to a two-step reconfiguration anti-disturbance PID control method and a control method for an SCR flue gas denitrification system. Background Technology

[0002] Selective catalytic reduction (SCR) technology is currently the most widely used technology in the field of flue gas denitrification in thermal power units. Its basic principle is to precisely control the flow rate of ammonia injected into the flue gas, causing it to react with nitrogen oxides (NOx) in the flue gas under the action of a catalyst. X Selective catalytic reduction reaction occurs, producing harmless nitrogen and water, thereby achieving the purpose of flue gas purification.

[0003] With the construction and development of new power systems, thermal power units are increasingly undertaking the task of deep and frequent peak shaving. Under this operating condition, the rapid and significant changes in unit load lead to changes in flue gas flow rate, temperature, and inlet NO. X The drastic and wide-ranging fluctuations in concentration place extremely high demands on the ammonia injection control of the SCR denitrification system. The control system must achieve a rapid and precise balance between ensuring high removal efficiency and preventing ammonia escape to ensure operational safety, making its disturbance resistance a core performance indicator.

[0004] Currently, the cascade PID control strategy based on distributed control systems (DCS) is still widely used in industrial settings. This strategy has a simple structure, mature configuration, and is familiar to engineers, and can meet basic control requirements under steady-state or small-range disturbance conditions. However, when faced with severe and complex disturbances such as deep peak shaving, traditional cascade PID control, due to its inherent control structure limitations (limited ability to suppress unmodeled dynamics and external disturbances), often exhibits problems such as slow setpoint tracking response, large overshoot, and long recovery time, making it difficult to meet high-standard control performance requirements.

[0005] Active disturbance rejection control (ADRC) technology, due to its ability to estimate and compensate for total system disturbances (including model uncertainties and external disturbances) in real time through extended state observers (ESOs), theoretically exhibits excellent disturbance rejection capability and robustness, providing a potential solution to the aforementioned problems. However, ADRC faces significant bottlenecks in practical industrial application: its algorithm structure is relatively complex, the physical meaning of parameters is not as intuitive as PID, and the tuning process is cumbersome; more importantly, its core modules (such as ESOs) lack native support in mainstream industrial control software platforms, resulting in poor engineering adaptability, and there is a contradiction between observer bandwidth and high-frequency measurement noise sensitivity. This leads to a prominent contradiction between the rigor of advanced control theory (ADRC) and the stringent requirements of industrial sites for controller simplicity, ease of use, and reliability, greatly limiting its widespread application in complex industrial processes such as SCR.

[0006] Therefore, there is an urgent need in this field for a control method that can inherit the advanced disturbance rejection and robust performance of ADRC, while maintaining the characteristics of simple structure, easy parameter tuning, and easy implementation and promotion on existing industrial control platforms, such as PID, so as to effectively solve the high-performance control problem faced by SCR flue gas denitrification system under wide load and deep peak shaving conditions. Summary of the Invention

[0007] To address the problems existing in the prior art, this invention proposes a two-step reconfiguration anti-disturbance PID control method and a control method for an SCR flue gas denitrification system.

[0008] The technical solution of the present invention is as follows:

[0009] A two-step reconfiguration disturbance-resistant PID control method, based on a cascade PID control system, includes:

[0010] Cascade active disturbance rejection control reconfiguration is performed by decomposing the controlled object of the cascade PID control system into a deterministic part model and an uncertain part model containing system disturbances through virtual decomposition. Based on the deterministic part model, a cascade active disturbance rejection control system with internal model control in the inner loop and active disturbance rejection control in the outer loop is constructed.

[0011] To reconstruct the disturbance rejection PID control, an improved disturbance observer is used to reconstruct the cascade active disturbance rejection control system into an equivalent disturbance rejection PID control system. The improved disturbance observer decomposes the system disturbance into a negligible internal disturbance and an external disturbance that affects the controlled variable and meets the preset requirements.

[0012] Disturbance-resistant PID control based on two-step reconfiguration of disturbance-resistant PID control system.

[0013] Furthermore, the specific steps of decomposing the controlled object of the cascade PID control system into a deterministic partial model and an uncertain partial model containing system disturbances through virtual decomposition include:

[0014] The controlled object is decomposed virtually. Decomposed into a series of deterministic parts of the model and uncertain part of the model containing system disturbances , represented as ;

[0015] Furthermore, the specific method for constructing a cascade active disturbance rejection control system based on the determined partial model includes:

[0016] A portion of the model was determined through mirror calculation. mirror model and based on the mirror model A cascaded active disturbance rejection control system is constructed, comprising an outer-loop ADRC controller and an inner-loop inner membrane controller. The ADRC controller consists of a tracking differentiator (TD), a state error feedback (SEF), and an extended state observer (ESO). The transfer function of the inner membrane controller is... Internal model controlled filter With a determined partial model The ratio, ;

[0017] Furthermore, in the cascaded active disturbance rejection control system, the control input... Simultaneously input to the mirror model and the actual controlled object ;

[0018] Mirror model output signal As one state observation input signal of the extended state observer ESO;

[0019] Controlled object output As another input signal to the extended state observer (ESO), the ESO output includes the output to the controlled object. The first estimate, the second estimate of the internal state of the system, and the third estimate of the total disturbance;

[0020] First estimate and reference signal The smoothed signal after the tracking differentiator TD is subtracted to obtain the output tracking error, which is then used as a state error feedback input signal to the state error feedback SEF.

[0021] The second estimate is used as another state error feedback input signal to the state error feedback SEF. The state error feedback SEF adjusts the control quantity based on the output tracking error and the second estimate, and outputs an intermediate reference signal.

[0022] The difference between the third estimate and the intermediate reference signal is used to obtain the internal model controller. Input Internal mold controller The output is the control quantity. ;

[0023] Disturbance signal Transfer function through perturbation input channel With control input Superimposed effect on the controlled object Get the output of the controlled object .

[0024] Furthermore, the specific steps of using an improved disturbance observer to equivalently reconstruct the cascade active disturbance rejection control system into a disturbance rejection PID control system, wherein the improved disturbance observer decomposes the system disturbance into negligible internal disturbances and external disturbances that affect the controlled variable and meet preset requirements, include:

[0025] Construct a low-pass filter A pre-compensator that ensures external disturbances meet preset requirements. And the expected closed-loop model that makes the internal disturbance effect negligible. An improved interference observer;

[0026] By mapping the structure of the improved disturbance observer to the structure of the disturbance-resistant PID control system, the disturbance-resistant PID control system is determined. With reference input filter The transfer function of the disturbance rejection PID control system is: The transfer function of the reference input filter is: Expected closed-loop model ;

[0027] The parameters of the disturbance rejection PID control system are related to the parameters of the cascade active disturbance rejection control system as follows:

[0028] ;

[0029] ;

[0030] In the formula, To expand the bandwidth of the state observer; For the feedback controller bandwidth; The reciprocal of the controller bandwidth; For controller gain; This is the differential gain coefficient.

[0031] Furthermore, in the disturbance rejection PID control system, the reference input The transient signal is obtained through a filter at the reference input. , and the output of the controlled object The difference is used to obtain the error signal. Error signal The control signal is obtained through an anti-disturbance PID control system. Disturbance signal Transfer function through perturbation input channel With control signals The superposition of the effects on the controlled object yields the output of the controlled object. .

[0032] Furthermore, the method also includes:

[0033] 1) Construct a hybrid sensitivity optimization problem and define the sensitivity function. Complementary sensitivity function Where G(s) is the transfer function of the controlled object and C(s) is the transfer function of the controller;

[0034] Based on the minimum Theoretically, the robustness index of a single-loop control system is measured as follows: ;

[0035] For single-loop control systems H is the transfer matrix from the disturbance signal to the system output y and the control input u. ∞ Norm; Angular frequency;

[0036] 2) Let the complement function be equal to the expected closed-loop transfer function, i.e. , combined Based on the characteristics and robustness constraints, determine the desired closed-loop controller bandwidth of the system. ;

[0037] 3) Select the time-multiplied error absolute value integral index ITAE for optimization solution. and To minimize the ITAE index, .

[0038] A control method for an SCR flue gas denitrification system includes:

[0039] Obtain SCR reactor outlet NO X The setpoint r and the measured value y of the concentration;

[0040] The setpoint r and the measured value y are input to the disturbance rejection PID controller obtained by the two-step reconstructed disturbance rejection PID control method, and the disturbance rejection PID controller calculates the control quantity based on the setpoint r and the measured value y.

[0041] Adjust the opening of the ammonia valve on the ammonia injection grid according to the control quantity to control the ammonia flow rate injected into the flue, thereby reducing the NO at the SCR reactor outlet. X The concentration is tracked to the set value.

[0042] Furthermore, the control objective of the disturbance-resistant PID controller is the outlet NO of the SCR reactor in the SCR flue gas denitrification system. X The concentration, the control output is used to adjust the ammonia flow rate entering the SCR reactor.

[0043] Furthermore, the control method of the SCR flue gas denitrification system is used to address situations where changes in the load of the thermal power unit lead to changes in inlet NO. XWhen the concentration fluctuates, maintain the NO concentration at the outlet of the SCR reactor. X Concentration stability.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] This invention proposes a two-step reconfiguration disturbance rejection PID control method. The first step, reconfiguration, introduces a cascaded active disturbance rejection control (CADRC) structure, utilizing an extended state observer (ESO) to estimate and compensate for the total system disturbance (including model uncertainties and external disturbances) in real time. This allows the system to inherit the strong disturbance rejection capability and robustness to model uncertainties of ADRC. The second step further transforms the complex CADRC system into a simpler disturbance rejection PID control system. This design allows the invention to retain advanced disturbance rejection performance while possessing the advantages of traditional PID controllers: simple structure, relatively intuitive parameter physical meaning, and ease of configuration and debugging in existing industrial distributed control systems (DCS). This significantly lowers the engineering application threshold for advanced control algorithms.

[0046] This invention provides a two-step reconfiguration disturbance-resistant PID control method, offering a systematic parameter tuning approach based on H∞ hybrid sensitivity optimization and ITAE performance indicators. This design transforms the controller's dynamic performance, disturbance rejection, and robustness requirements into specific optimization problems, guiding the tuning of controller parameters with superior overall performance and avoiding the drawbacks of over-reliance on experience in traditional PID parameter tuning.

[0047] This invention also proposes a control method for an SCR flue gas denitrification system. Addressing the complex industrial process problems of large inertia, large lag, and severe load disturbances inherent in SCR flue gas denitrification systems, the method utilizes a control system reconstructed using the two-step reconstructed disturbance-resistant PID control method of this invention to control the SCR flue gas denitrification system. This effectively improves the outlet NO₂ level. X The concentration control quality effectively tracks the setpoint while significantly suppressing NO ingress. X Output fluctuations caused by external disturbances such as concentration fluctuations meet the stringent control requirements for deep peak shaving of thermal power units. Attached Figure Description

[0048] Figure 1 A flowchart illustrating the two-step reconfiguration disturbance rejection PID control method.

[0049] Figure 2 This is a structural diagram of a cascaded active disturbance rejection control system.

[0050] Figure 3 Structure diagram of an anti-disturbance PID control system;

[0051] Figure 4 Diagram of the disturbance channel structure;

[0052] Figure 5 The experimental results are used to track the set values.

[0053] Figure 6 The figure shows the experimental results of the system step disturbance;

[0054] Figure 7 The figure shows the experimental results of high-frequency sinusoidal continuous perturbation.

[0055] Figure 8 The figure shows the experimental results of low-frequency sinusoidal continuous perturbation.

[0056] Figure 9 The results of the noise interference measurement experiment are shown in the figure.

[0057] Figure 10 The figure shows the experimental results of robustness testing under a continuous random disturbance signal (variance 0.5).

[0058] Figure 11 The figure shows the experimental results of robustness testing under a continuous random disturbance signal (variance 1). Detailed Implementation

[0059] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. After reading this invention, any modifications of the invention in various equivalent forms by those skilled in the art will fall within the scope defined by the appended claims.

[0060] Example 1:

[0061] This invention provides a two-step reconfiguration disturbance-resistant PID control method based on a cascade PID control system, such as... Figure 1 As shown, it includes:

[0062] Cascade active disturbance rejection control reconfiguration is carried out by decomposing the controlled object of the cascade PID control system into a deterministic part model and an uncertain part model containing system disturbances through virtual decomposition. Based on the deterministic part model, a cascade active disturbance rejection control system with internal model control in the inner loop and active disturbance rejection control in the outer loop is constructed.

[0063] To reconstruct the disturbance rejection PID control, an improved disturbance observer is used to reconstruct the cascade active disturbance rejection control system into an equivalent disturbance rejection PID control system. The improved disturbance observer decomposes the system disturbance into negligible internal disturbance and external disturbance that affects the controlled variable and meets the preset requirements.

[0064] Disturbance-resistant PID control based on two-step reconfiguration of disturbance-resistant PID control system.

[0065] Furthermore, the specific steps for decomposing the controlled object of a cascade PID control system into a deterministic partial model and an uncertain partial model containing system disturbances through virtual decomposition include:

[0066] The controlled object is decomposed virtually. Decomposed into a series of deterministic parts of the model and uncertain part of the model containing system disturbances , The system disturbances include internal disturbances and external disturbances, denoted as: ;

[0067] Furthermore, based on a deterministic partial model, specific methods for constructing a cascade active disturbance rejection control system include:

[0068] A portion of the model was determined through mirror calculation. mirror model And based on the mirror model A cascaded active disturbance rejection control system is constructed, consisting of an outer-loop ADRC controller and an inner-loop inner membrane controller. The ADRC controller comprises a tracking differentiator (TD), a state error feedback (SEF), and an extended state observer (ESO). The transfer function of the inner membrane controller is... Internal model controlled filter With a determined partial model The ratio, ;

[0069] Furthermore, in a cascade active disturbance rejection control system, the control input... Simultaneously input to the mirror model and the actual controlled object ;

[0070] Mirror model output signal As one state observation input signal of the extended state observer ESO; further, the mirror model output signal The inner loop error is obtained by subtracting it from the output of the internal model. And as an internal mold controller The feedback input signal, and due to the internal model and the mirror model Consistent, mirror model output signal The inner loop error is equal to the output of the internal model. It is 0.

[0071] Controlled object output As another input signal to the extended state observer (ESO), the ESO output includes the output to the controlled object. First estimate A second estimate of the internal state of the system And a third estimate of the total disturbance. ;

[0072] First estimate With reference signal The smoothed signal after the tracking differentiator TD is subtracted to obtain the output tracking error, which is then used as a state error feedback input signal to the state error feedback SEF.

[0073] Second estimate As another state error feedback input signal to the state error feedback SEF, the state error feedback SEF adjusts the control quantity based on the output tracking error and the second estimate, and outputs an intermediate reference signal.

[0074] Third estimate The difference between the internal model controller and the intermediate reference signal is used to obtain the internal model controller. Input Internal mold controller The output is the control quantity. ;

[0075] Disturbance signal Transfer function through perturbation input channel With control input Superimposed effect on the controlled object Get the output of the controlled object .

[0076] Furthermore, by employing an improved disturbance observer, the cascade active disturbance rejection control system is equivalently reconstructed into a disturbance rejection PID control system. The specific steps of the improved disturbance observer in decomposing system disturbances into negligible internal disturbances and external disturbances that affect the controlled variable and meet preset requirements include:

[0077] Construct a low-pass filter A pre-compensator that ensures external disturbances meet preset requirements. And the expected closed-loop model that makes the internal disturbance effect negligible. An improved interference observer;

[0078] By mapping the structure of the improved disturbance observer to the structure of the disturbance-resistant PID control system, the disturbance-resistant PID control system is determined. With reference input filter The transfer function of the disturbance rejection PID control system is: The transfer function of the reference input filter is: Expected closed-loop model ;

[0079] The parameters of the disturbance rejection PID control system are related to those of the cascade active disturbance rejection control system as follows:

[0080] ;

[0081] ;

[0082] In the formula, To expand the bandwidth of the state observer; For the feedback controller bandwidth; The reciprocal of the controller bandwidth; The controller gain is proportional to ; This is the differential gain coefficient.

[0083] Furthermore, in a disturbance rejection PID control system, the reference input The transient signal is obtained through a filter at the reference input. , and the output of the controlled object The difference is used to obtain the error signal. Error signal The control signal is obtained through an anti-disturbance PID control system. Disturbance signal Transfer function through perturbation input channel With control signals The superposition of the effects on the controlled object yields the output of the controlled object. .

[0084] Furthermore, the method also includes:

[0085] 1) Construct a hybrid sensitivity optimization problem and define the sensitivity function. Complementary sensitivity function Where G(s) is the transfer function of the controlled object and C(s) is the transfer function of the controller;

[0086] Based on the minimum Theoretically, the robustness index of a single-loop control system is measured as follows: ;

[0087] For single-loop control systems H is the transfer matrix from the disturbance signal to the system output y and the control input u. ∞ Norm; Angular frequency;

[0088] 2) Let the complement function be equal to the expected closed-loop transfer function, i.e. , combined Based on the characteristics and robustness constraints, determine the desired closed-loop controller bandwidth of the system. ;

[0089] 3) Select the integral index ITAE (Time-to-Error Absolute Value) and optimize the solution within the specified parameter range. and To minimize the ITAE index, .

[0090] A control method for an SCR flue gas denitrification system includes:

[0091] Obtain SCR reactor outlet NO X The setpoint r and the measured value y of the concentration;

[0092] The setpoint r and the measured value y are input to the disturbance rejection PID controller obtained by the two-step reconstructed disturbance rejection PID control method. The disturbance rejection PID controller calculates the control quantity based on the setpoint r and the measured value y.

[0093] Adjust the opening of the ammonia valve on the ammonia injection grid according to the control quantity to control the ammonia flow rate injected into the flue, thereby reducing the NO at the SCR reactor outlet. X Concentration tracking setpoint.

[0094] Furthermore, the control objective of the disturbance-resistant PID controller is to reduce the NO at the outlet of the SCR reactor in the SCR flue gas denitrification system. X Concentration, control output is used to adjust the ammonia flow rate entering the SCR reactor.

[0095] Furthermore, the control method of the SCR flue gas denitrification system is used to address situations where changes in the load of a thermal power unit lead to changes in inlet NO. x When the concentration fluctuates, maintain the NO concentration at the SCR reactor outlet. X Concentration stability.

[0096] Example 2:

[0097] This embodiment applies the method of the present invention to the selective catalytic reduction (SCR) denitrification system of a 300MW ultra-supercritical coal-fired power generation unit. The aim is to solve the technical problems of slow setpoint tracking response and weak anti-disturbance capability under complex disturbances in the existing cascade PID control system, and to verify the effectiveness of the anti-disturbance PID control scheme proposed in this application.

[0098] This example first provides a principled introduction to the reconstruction process of the present invention from cascade PID control to cascade active disturbance rejection control (CADRC):

[0099] The core working principle of the SCR flue gas denitrification system is as follows: NH3 generated by the ammonia generator is first mixed with dilution air delivered by the dilution fan; the mixed gas mass is then injected into the flue gas duct through the ammonia injection grid, and then fully mixed with the flue gas from the economizer outlet through the flow guiding device inside the flue gas duct; the above mixed flue gas then enters the SCR reactor with a built-in vanadium-titanium-based catalyst, where, under the catalytic action of the catalyst, NO in the flue gas is removed through a selective catalytic reduction reaction. X It is converted into N2 and H2O, ultimately achieving the technical objective of denitrification of flue gas in thermal power units.

[0100] This example of cascade active disturbance rejection control (CADRC) reconfiguration uses the commonly used cascade PID control structure of SCR flue gas denitrification systems as the direct modification carrier. The specific components of this ordinary cascade PID control structure include: a main controller, a secondary controller, a transfer function of the controlled object, and a disturbance channel transfer function. Its structure provides basic hardware and control logic support for subsequent control reconfiguration.

[0101] The main controller is C1(s), and the secondary controller is C2(s); the controlled objects include the transfer function G1(s) from the ammonia valve opening to the ammonia flow rate and the transfer function G2(s) of the SCR reactor, i.e., the ammonia flow rate from the SCR inlet to the NO outlet. X Concentration transfer function; the perturbation channel transfer function is denoted as D(s), corresponding to the inlet NO. X Concentration change to outlet NO X The perturbation propagation relationship of concentration changes.

[0102] In a cascaded PID control structure, r represents the output NO. X Concentration setpoint, y is the outlet NO X Measured concentration value, d is the system disturbance, u is the main controller output, u a It is the output of the secondary controller.

[0103] The controller uses an actual PID controller:

[0104] ;

[0105] Among them, K p K is the proportionality coefficient. i For integral gain, K d For differential gain, T d The time for integration.

[0106] The transfer function of the controlled object and the transfer function of the disturbance channel are as follows:

[0107] ; ; ;

[0108] Where K1, K2, and K3 are proportionality constants, and T1, T2, and T3 are inertial time constants. , It is the lag time.

[0109] The refactoring from cascade PID control to cascade active disturbance rejection control (CADRC) includes the following steps:

[0110] 1) The controlled object G(s) is decomposed into a series of deterministic models G1(s) and models G2(s) with uncertainties, satisfying the following conditions. For control systems, there are uncertainties in the internal model, i.e., internal disturbances, as well as external disturbances such as changes in unit load. The object model G(s) contains both deterministic and uncertain parts. In object analysis, G1(s) is generally considered deterministic, and the uncertain part is G2(s). This is consistent with engineering practice because valve characteristics can be determined.

[0111] 2) Reconstruct the controlled object so that the reconstructed model has the same information structure and variable data as the original controlled object. The reconstructed model includes a mirror model of G1(s). The actual process of the controlled object G(s).

[0112] 3) Based on the above reconstruction model, when When = G1(s), the implementation structure of the cascade active disturbance rejection control system (CADRC) can be obtained. In this system, the outer loop adopts an active disturbance rejection controller (ADRC, typically composed of a tracking differentiator TD, a state error feedback SEF, and an extended state observer ESO), while the inner loop adopts a controller based on the internal model control (IMC) principle, where the design of the internal model control filter depends on F(s). When , At this time, the cascade active disturbance rejection control system (CADRC) is transformed into a classical active disturbance rejection control (ADRC) structure, completing the reconstruction from cascade PID control to cascade active disturbance rejection control (CADRC), as follows. Figure 2 As shown.

[0113] Reconstructing a cascaded active disturbance rejection system into a disturbance rejection PID control system includes the following steps:

[0114] 1) Based on the principles of disturbance estimation and compensation, an improved disturbance observer structure is adopted, which includes a pre-compensator K(s), a low-pass filter Q(s), and a desired closed-loop model H. R (s) uses disturbance compensation to match the closed-loop system to the desired closed-loop model H. R (s), mapping the improved disturbance observer structure to an anti-disturbance PID control structure, and determining the controller C(s) and the reference input filter F.r (s).

[0115] For first-order or second-order objects, the disturbance rejection PID control structure is as follows: Figure 3 As shown, it includes a controller C(s) and a reference input filter F. r (s). Satisfies the following relation: ; ;

[0116] C(s) is the PID controller, and the parameters of the PID controller are determined by the parameters ω of the active disturbance rejection control. c ω o The PID control parameters are determined by the active disturbance rejection control parameters.

[0117] If selected ;

[0118] In the formula, ω c For the ADRC feedback controller bandwidth, ω o The bandwidth of the ADRC extended state observer is given, and a and b are the highest-order coefficients of the denominator and numerator polynomials of the controlled object G(s). , ), It is the reciprocal of the controller bandwidth, which can be used to speed up response and suppress overshoot. It is the gain coefficient.

[0119] The controller is further obtained as ; The differential action in the formula is an ideal differential element, which cannot be realized. Therefore, in practical applications, the following formula is used: It is the differential gain coefficient.

[0120] ; ;

[0121] 2) To further clarify the system disturbance, the system disturbance is reconstructed, such as... Figure 4 As shown, the total disturbance f of the system consists of two parts, namely f = f1 + f2, where f1 is the internal disturbance and f2 is the external disturbance. By selecting the parameters of the front-end dynamic compensator, the influence of the internal disturbance can be ignored, and the total disturbance of the system is mainly due to the external disturbance.

[0122] The reconstructed system architecture is as follows ; ; This indicates the uncertainty of the internal model. ; ;

[0123] If selected If the internal disturbance f1 of the system is negligible, then the total disturbance of the system is mainly the external disturbance f2, which is consistent with the dynamic performance of each control system of the thermal power unit under deep peak shaving conditions.

[0124] 3) Design a front-end dynamic compensator to analyze from the disturbance side, so that the influence of external disturbances on the controlled variable meets the preset requirements, while ensuring the dynamic performance and stability of the disturbance-resistant PID control system.

[0125] The pre-amplifier dynamic compensator K(s) consists of a proportional element and a differential element, and is designed as follows: ;

[0126] Tracking error can be expressed as ;

[0127] Error synthesis perturbation signal f * The expression is ;

[0128] From the equation, we can obtain the error synthesis perturbation signal f. * It consists of three parts: setpoint fluctuation, internal disturbance, and external disturbance.

[0129] In practical applications, control systems rarely experience sudden changes in setpoints, and can always be kept within acceptable limits. This makes the influence of internal disturbances approximately zero; therefore, the main factor affecting the system error e is the external disturbance d, i.e. ;

[0130] The pre-amplifier dynamic compensator K(s) directly determines the magnitude of the external disturbance d's impact on the system. Especially when 1 / K(s) is a first-order inertial element, it effectively delays the impact of the external disturbance d on the system. K(s) is the design reference input filter F. r (s), controller C(s), desired closed-loop transfer function H R (s) is a key component.

[0131] Furthermore, it can be seen from the action path of the external disturbance d to the controlled variable y(s) that... The role of K(s) is particularly important. The dynamic performance and stability of disturbance-resistant PID control are both related to K(s). Choosing an appropriate K(s) is crucial to satisfying the requirements. K(s) is the pre-dynamic compensator in the active disturbance rejection structure. K(s) can be selected in the following form. ;like K(s) is the leading element, if K(s) is an inertial element, and m and n can be... , A linear combination of .

[0132] System reconstruction based on perturbation decomposition is for second-order and lower-order systems. For higher-order systems, the model can be downgraded to second-order or first-order before analysis and application.

[0133] Besides analyzing the system's anti-interference ability, robustness is also an important indicator affecting system performance. Using robust methods to design controller parameters aims to ensure that the system maintains stability and expected performance under parameter perturbations, external disturbances, or model uncertainties. ∞ Control is a common method in robust design, and its main task is to maximize the gain (H) of the system's input disturbance to the output transfer function. ∞ The norm is kept to a minimum to ensure that the impact of disturbances on the system is limited to the desired range.

[0134] A hybrid sensitivity function is selected as the optimization method for controller design; that is, the sensitivity function is defined as follows: ; ;

[0135] S(s) is the sensitivity function, T(s) is the complementary sensitivity function, G(s) is the transfer function of the controlled object, and C(s) is the transfer function of the controller. The sensitivity function characterizes the system's response to low- and mid-frequency uncertainties; the complementary sensitivity function characterizes the system's response to mid- and high-frequency uncertainties.

[0136] Based on the minimum Theoretically, the robustness index of a single-loop control system is measured as follows: ; It is a comprehensive indicator for measuring the robustness and disturbance rejection of a control system. A larger value indicates better system immunity to disturbances. The smaller the value, the better the robustness of the system.

[0137] In H ∞ In control, if let This makes it easier to determine the desired closed-loop controller bandwidth of the system. If a and b are the higher-order coefficients of the denominator and numerator polynomials of the controlled object, then the disturbance rejection PID controller and the reference input filter F... r The tuning of (s) only requires determining and That's it. You can obtain it by selecting the integral index ITAE (Time Multiplied by the Absolute Value of Error). and , .

[0138] In practical implementation, the transfer functions for each channel in this example are set as follows:

[0139] The expression for the transfer function G1(s) from valve opening to ammonia injection flow rate is as follows: ;

[0140] Ammonia injection flow rate to SCR outlet NO X The expression for the concentration transfer function G2(s) is: ;

[0141] Disturbance channel entrance NO X NO concentration at outlet X The expression for the concentration transfer function D(s) is as follows: ;

[0142] The control system structure is shown below. Figure 3 Disturbance channel see Figure 4 Design a disturbance filter 1 / K(s), and control system parameters are shown in Table 1.

[0143] Table 1 Controller Parameters

[0144]

[0145] To verify the setpoint tracking performance of the controller, the setpoint tracking results of ordinary cascade PID and disturbance-resistant PID are as follows: Figure 5 When the allowable error is ±2%, the settling time of the disturbance-resistant PID is 73.9 seconds, while that of the cascaded PID is 265.4 seconds. The settling tracking capability of the disturbance-resistant PID is better than that of the cascaded PID.

[0146] During deep peak shaving of thermal power units, disturbance rejection is a key indicator for evaluating the performance of the SCR control system. The SCR control system was tested for three types of disturbances: step disturbance, high- and low-frequency continuous disturbance, and measurement noise.

[0147] Figure 6 The results are from the system step disturbance test. When a 100% step disturbance with an amplitude of 1 is added at a simulation time of 300 seconds, the maximum fluctuation of the disturbance-resistant PID is 0.0694, and the maximum fluctuation of the cascaded PID is 0.1634. The disturbance suppression capability of the disturbance-resistant PID is significantly better than that of the cascaded PID, and the system fluctuation is greatly reduced.

[0148] Rapid fluctuations in process parameters of the control system and high-frequency noise in the control and communication loops are considered high-frequency interferences, while cumulative errors in the controller's integral element, low-frequency vibrations and characteristic drift of the mechanical system are considered low-frequency interferences. Figure 7 , Figure 8 The test results are obtained by adding high-frequency and low-frequency sinusoidal continuous perturbations with amplitudes of 1 and frequencies of 2kHz and 2Hz respectively during a simulation time of 300 seconds. Their mathematical expressions are as follows: and Under continuous high-frequency disturbances, the maximum fluctuation of the anti-disturbance PID is 0.017, and the maximum fluctuation of the cascaded PID is 0.037. The impact of high-frequency interference signals on the system gradually weakens or even disappears over time. Low-frequency interference, however, gradually increases its impact on the system over time. This is mainly because the characteristic drift of components and mechanical systems caused by environmental changes is irreversible, which is consistent with actual production practices. Within 3000 seconds, the maximum fluctuation of the anti-disturbance PID under continuous low-frequency interference is 0.0506, while the maximum fluctuation of the cascaded PID is 0.1157.

[0149] NO for SCR exports X Both the measurement signal and the ammonia injection flow measurement signal were superimposed with a random white noise signal with a mean of 0, a variance of 0.05, and a sampling time of 0.1 seconds. The test results are as follows: Figure 9 The results showed that the outlet NO with white noise X The measurement signals and ammonia injection flow measurement signals did not have a significant impact on the system. The maximum fluctuation of the disturbance-resistant PID was 0.0158, while the maximum fluctuation of the cascaded PID was 0.016.

[0150] Based on the Monte Carlo experiment principle, the SCR disturbance rejection PID control system outputs the following result under the action of a continuous random disturbance signal with a mean of 0, a variance of 0.5, and a sampling time of 0.1: Figure 10 At this point, the system's maximum fluctuation is 0.0247. Under the influence of a continuous random disturbance signal with a mean of 0, a variance of 1, and a sampling time of 0.1, the system output is as follows. Figure 11 At this point, the system's maximum fluctuation was 0.0812. Even with increasing interference, the system remained stable.

[0151] The two-step reconfigurable disturbance-resistant PID control of the SCR flue gas denitrification system is an improvement on the ordinary cascade PID control, which can meet the needs of deep peak shaving of thermal power units in new power systems. Compared with the ordinary cascade PID control, the two-step reconfigurable disturbance-resistant PID control has stronger setpoint tracking ability, disturbance resistance and robustness. Multiple disturbance tests show that the two-step reconfigurable disturbance-resistant PID control has outstanding disturbance suppression ability.

[0152] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A two-step reconfiguration disturbance-resistant PID control method, based on a cascade PID control system, characterized in that, include: Cascade active disturbance rejection control reconfiguration is performed by decomposing the controlled object of the cascade PID control system into a deterministic part model and an uncertain part model containing system disturbances through virtual decomposition. Based on the deterministic part model, a cascade active disturbance rejection control system with internal model control in the inner loop and active disturbance rejection control in the outer loop is constructed. To reconstruct the disturbance rejection PID control, an improved disturbance observer is used to reconstruct the cascade active disturbance rejection control system into an equivalent disturbance rejection PID control system. The improved disturbance observer decomposes the system disturbance into a negligible internal disturbance and an external disturbance that affects the controlled variable and meets the preset requirements. Disturbance-resistant PID control based on two-step reconfiguration of disturbance-resistant PID control system.

2. The two-step reconfiguration disturbance-resistant PID control method according to claim 1, characterized in that, The specific steps for decomposing the controlled object of the cascade PID control system into a deterministic partial model and an uncertain partial model containing system disturbances through virtual decomposition include: The controlled object is decomposed virtually. Decomposed into a series of deterministic parts of the model and uncertain part of the model containing system disturbances , represented as .

3. The two-step reconfiguration disturbance-resistant PID control method according to claim 2, characterized in that, The specific method for constructing a cascade active disturbance rejection control system based on the determined partial model includes: A portion of the model was determined through mirror calculation. mirror model and based on the mirror model A cascaded active disturbance rejection control system is constructed, comprising an outer-loop ADRC controller and an inner-loop inner membrane controller. The ADRC controller consists of a tracking differentiator (TD), a state error feedback (SEF), and an extended state observer (ESO). The transfer function of the inner membrane controller is... Internal model controlled filter With a determined partial model The ratio, .

4. The two-step reconfiguration disturbance-resistant PID control method according to claim 3, characterized in that, In the cascaded active disturbance rejection control system, the control input Simultaneously input to the mirror model and the actual controlled object ; Mirror model output signal As one state observation input signal of the extended state observer ESO; Controlled object output As another input signal to the extended state observer (ESO), the ESO output includes the output to the controlled object. The first estimate, the second estimate of the internal state of the system, and the third estimate of the total disturbance; First estimate and reference signal The smoothed signal after the tracking differentiator TD is subtracted to obtain the output tracking error, which is then used as a state error feedback input signal to the state error feedback SEF. The second estimate is used as another state error feedback input signal to the state error feedback SEF. The state error feedback SEF adjusts the control quantity based on the output tracking error and the second estimate, and outputs an intermediate reference signal. The difference between the third estimate and the intermediate reference signal is used to obtain the internal model controller. Input Internal mold controller The output is the control quantity. ; Disturbance signal Transfer function through perturbation input channel With control input Superimposed effect on the controlled object Get the output of the controlled object .

5. The two-step reconfiguration disturbance-resistant PID control method according to claim 4, characterized in that, The specific steps of using an improved disturbance observer to equivalently reconstruct the cascade active disturbance rejection control system into a disturbance rejection PID control system, wherein the improved disturbance observer decomposes the system disturbance into negligible internal disturbances and external disturbances that affect the controlled variable and meet preset requirements, include: Construct a low-pass filter A pre-compensator that ensures external disturbances meet preset requirements. And the expected closed-loop model that makes the internal disturbance effect negligible. An improved interference observer; By mapping the structure of the improved disturbance observer to the structure of the disturbance-resistant PID control system, the disturbance-resistant PID control system is determined. With reference input filter The transfer function of the disturbance rejection PID control system is: The transfer function of the reference input filter is: Expected closed-loop model ; The parameters of the disturbance rejection PID control system are related to the parameters of the cascade active disturbance rejection control system as follows: ； ； In the formula, To expand the bandwidth of the state observer; For the feedback controller bandwidth; The reciprocal of the controller bandwidth; For controller gain; This is the differential gain coefficient.

6. The two-step reconfiguration disturbance-resistant PID control method according to claim 5, characterized in that, In the aforementioned disturbance-resistant PID control system, the reference input The transient signal is obtained through a filter at the reference input. , and the output of the controlled object The difference is used to obtain the error signal. Error signal The control signal is obtained through an anti-disturbance PID control system. Disturbance signal Transfer function through perturbation input channel With control signals The superposition of the effects on the controlled object yields the output of the controlled object. .

7. The two-step reconfiguration disturbance-resistant PID control method according to claim 6, characterized in that, The method further includes: 1) Construct a hybrid sensitivity optimization problem and define the sensitivity function. Complementary sensitivity function Where G(s) is the transfer function of the controlled object and C(s) is the transfer function of the controller; Based on the minimum Theoretically, the robustness index of a single-loop control system is measured as follows: ; For single-loop control systems H is the transfer matrix from the disturbance signal to the system output y and the control input u. ∞ Norm; Angular frequency; 2) Let the complement function be equal to the expected closed-loop transfer function, i.e. , combined Based on the characteristics and robustness constraints, determine the desired closed-loop controller bandwidth of the system. ; 3) Select the time-multiplied error absolute value integral index ITAE for optimization solution. and To minimize the ITAE index, .

8. A control method for an SCR flue gas denitrification system, characterized in that, include: Obtain SCR reactor outlet NO X The setpoint r and the measured value y of the concentration; The setpoint r and the measured value y are input to the disturbance rejection PID controller obtained by the two-step reconstructed disturbance rejection PID control method according to any one of claims 1 to 7, and the disturbance rejection PID controller calculates the control quantity based on the setpoint r and the measured value y. Adjust the opening of the ammonia valve on the ammonia injection grid according to the control quantity to control the ammonia flow rate injected into the flue, thereby reducing the NO at the SCR reactor outlet. X The concentration is tracked to the set value.

9. The control method for the SCR flue gas denitrification system according to claim 8, characterized in that, The control objective of the disturbance-resistant PID controller is to control the NO at the outlet of the SCR reactor in the SCR flue gas denitrification system. X The concentration, the control output is used to adjust the ammonia flow rate entering the SCR reactor.

10. The control method for the SCR flue gas denitrification system according to claim 9, characterized in that, The control method of the SCR flue gas denitrification system is used to address situations where changes in the load of a thermal power unit lead to inlet NO. X When the concentration fluctuates, maintain the NO concentration at the outlet of the SCR reactor. X Concentration stability.

Images