Coal-fired unit coordination level PID and model prediction control loop decoupling control method

Abstract

The invention discloses a coal-fired unit coordination level PID and model predictive control loop decoupling control method, which comprises the following steps of 1, analyzing multivariable coupling characteristics of a unit coordination control system, and determining unit multi-loop decoupling control requirements; 2, based on the multi-loop decoupling control requirement of the unit, a PID controller of a unit coordination system is designed for a power generation power control loop so as to control the opening degree of a main steam valve; and step 3, designing a model prediction controller stack of the unit coordination control system according to double loops of main steam pressure and water wall outlet temperature. The PID controller is responsible for ensuring rapid adjustment of the opening degree of a main steam valve, and the double-loop model prediction controller achieves accurate adjustment of the coal feeding amount and the water feeding amount through the prediction control function of the double-loop model prediction controller. Through the synergistic effect, the control precision and the response speed of the whole system are remarkably improved.

Description

Technical Field

[0001] The invention relates to the technical field of thermal control of thermal power plants, and in particular to a decoupling control method for a coordinated-level PID and model predictive control loop of a coal-fired unit. Background Art

[0002] Due to the intermittent and uncertain nature of renewable energy sources such as wind and solar power, coal-fired power plants are becoming increasingly important in ensuring stable power supply to the power grid and energy security. Against the backdrop of growing and changing electricity demand, especially the frequent occurrence of rapid load changes and complex operating conditions, higher requirements are placed on the control efficiency and flexibility of coal-fired units. Traditional coordinated control systems for supercritical coal-fired units mainly adopt a PID control strategy combined with a feedforward function. Although this strategy is simple and adaptable, its limitations are gradually emerging as the demand for grid flexibility increases and unit capacity expands. For example, PID control has difficulty in handling multivariable coupling and complex constraints, resulting in insufficient response and inaccurate prediction under rapidly changing conditions, and difficulty in adapting to rapid load adjustments. Model predictive control, due to its unique prediction mechanism and ability to handle multivariable coupling and complex constraints, has shown significant advantages in improving dynamic response and operating efficiency within a wide load range of the unit. Model predictive control can predict future system behavior and consider future constraints and goals in the control process, thereby achieving better control results. However, model predictive control also faces challenges of high computational complexity and high real-time requirements in practical applications. To this end, by utilizing the dual advantages of the rapidity of PID control and the multivariable predictive control of model predictive control, the control loop is decoupled from PID and model predictive control, which not only improves the response speed and prediction accuracy of the control system, but also reduces the computational complexity, thereby achieving more efficient and stable coordinated control of coal-fired units. Summary of the invention

[0003] The object of the present invention is to provide a decoupling control method for a coordinated PID and model predictive control loop of a coal-fired unit to solve the problems raised in the above-mentioned background technology.

[0004] To achieve the above object, the present invention provides the following technical solution: a method for decoupling control of a PID and model predictive control loop of a coal-fired unit coordination level, comprising the following steps:

[0005] Step 1: Analyze the multivariable coupling characteristics of the unit coordinated control system and determine the multi-loop decoupling control requirements of the unit;

[0006] Step 2: Based on the multi-loop decoupling control requirements of the unit, design the PID controller of the unit coordination system for the power generation control loop to control the main steam valve opening;

[0007] Step 3: Design the model predictive controller stack of the unit coordinated control system for the dual circuits of main steam pressure and water wall outlet temperature.

[0008] Preferably, the multivariable coupling characteristics in step 1 include establishing a relationship between the dynamic response speed of the valve opening and the response speeds of water supply and coal supply. The main steam valve opening has a rapid dynamic response, and its rapid changes will directly affect the power generation response. In contrast, the response speeds of water supply and coal supply are slower because they involve more complex physical processes, such as the transportation and combustion of fuel, and the transportation and evaporation of water. This difference in response speed leads to dynamic coupling between control loops, so that the control action of a single loop may have an adverse effect on other loops, thereby reducing the control efficiency and stability of the entire system.

[0009] Preferably, in step 2, the main steam valve opening is controlled, which is specifically expressed as:

[0010]

[0011] Where: K P is the proportionality coefficient, K I is the integration coefficient, K D is the differential coefficient, and e is the error between the target value and the feedback value.

[0012] Preferably, step 3 specifically includes:

[0013] Using field data, the general form of the state space model is obtained based on the subspace system identification method:

[0014]

[0015] Where: A c , B c , C c , D c , K c is a continuous state space matrix; u(t) is the input; y(t) is the output; e(t) is the disturbance; x(t) is the n-dimensional state vector. For dynamic systems, D defaults to zero;

[0016] In order to design the prediction model of the model predictive controller, the system model is discretized and differentiated to obtain the system augmented model:

[0017]

[0018] Where: 0 m ∈n x represents an n-dimensional zero matrix; u(k) represents the control input increment at time k;

[0019] For a dual-input dual-output system, the predicted output sequence Y and the control input sequence ΔU are defined as:

[0020]

[0021] Based on the prediction model, model predictive control obtains the predicted output sequence in the prediction time domain:

[0022]

[0023] Where: Y is the prediction time domain N P Intra-prediction output sequence; ΔU is the control time domain N C Internal control increment sequence; F and Φ are the constructed coefficient matrices respectively.

[0024] Preferably, for the dual-loop coordinated control of the unit, the objective function of the controller is designed with minimizing the control error as the optimization goal:

[0025]

[0026] Where: J is the optimization objective function; r i (k) is the i-th output reference value at time k; y i (k) is the i-th output value at time k; R is the reference value matrix; S, are the error weight matrix and the control weight matrix respectively.

[0027] Preferably, the optimization solution goal of the model predictive control is to find the optimal control sequence to minimize the objective function, and the output sequence in the objective function is expressed as a control sequence:

[0028]

[0029] In order to facilitate the optimization solution of quadratic programming, the objective function is further organized into the form of quadratic programming:

[0030]

[0031] Where: H is the quadratic term coefficient matrix, f is the linear term coefficient matrix;

[0032] Furthermore, control quantity constraints are set according to unit safety restrictions and operating conditions:

[0033] U min (k)≤U(k)≤U max (k)

[0034] ΔU min (k)≤ΔU(k)≤ΔU max (k)

[0035] Where: Umin , U max are the minimum and maximum constraint matrices of the input respectively; ΔU min , ΔU max are the minimum and maximum constraint matrices for the input increments, respectively.

[0036] Preferably, the control action of the PID controller is input into the model predictive controller as measurable disturbance information, and the model predictive control obtains the optimal control sequence of the system through real-time prediction optimization solution to achieve rolling optimization control. This strategy of combining the fast load response of the PID controller and the coupling control of the coal and water supply dual loops by MPC effectively realizes the decoupling control of the unit coordination level loop.

[0037] Compared with the prior art, the present invention has the following beneficial effects: (1) The PID controller is responsible for ensuring the rapid adjustment of the main steam valve opening, while the dual-loop model predictive controller achieves precise adjustment of the coal and water feed rates through its predictive control function. This synergy significantly improves the overall control accuracy and response speed of the system.

[0038] (2) By using the control action of the PID controller as the measurable disturbance input of the dual-loop model predictive controller, this strategy can more effectively deal with the multivariable coupling problem in the supercritical unit, thereby reducing the adverse impact of the coupling effect on system stability.

[0039] (3) The predictive control characteristics of the model predictive controller enable it to predict and respond to external disturbances and internal uncertainties in advance. At the same time, the fast response characteristics of the PID controller help to quickly adjust the system state when disturbances occur, significantly enhancing the system's anti-interference ability and robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a decoupling control structure diagram of the coordinated-level PID and model predictive control loop of a coal-fired unit provided by the present invention.

[0041] Figure 2 This is a comparison chart of the control effects of the coordinated-level PID and model predictive control loop decoupling control methods for coal-fired units provided by the present invention.

[0042] Figure 3 A comparison chart of the control effects of the coordinated PID and model predictive control loop decoupling control methods for coal-fired units provided by the present invention on disturbance conditions. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] like Figure 1-3 As shown, the present invention provides a technical solution: a decoupling control method for a coordinated PID and model predictive control loop of a coal-fired unit.

[0045] In the coordinated control system of coal-fired units, the multivariable coupling characteristics are a core challenge. There is a close interaction between the variables of main steam valve opening, coal feed, water feed, main steam pressure and water wall outlet temperature. Rapid changes in the main steam valve opening will directly affect the main steam pressure, and this fast response is required for grid load changes. In contrast, the response speed of water feed and coal feed is slower because they involve more complex physical processes such as fuel transportation and combustion, and water transportation and evaporation. This difference in response speed leads to dynamic coupling between control loops, so that the control action of a single loop may have an adverse effect on other loops, thereby reducing the control efficiency and stability of the entire system. Therefore, the need for multi-loop decoupling control becomes particularly important, which requires the control system to be able to independently adjust each loop and reduce the mutual interference between them to achieve more accurate and fast control.

[0046] like Figure 1 As shown, the present invention provides a decoupling control method for the coordinated PID and model predictive control loops of a coal-fired unit. A multi-loop decoupling controller is designed according to the response speed differences among the main steam valve opening, coal supply, and water supply. Specifically, the method includes:

[0047] For the power generation control loop, the present invention designs a PID controller to meet the multi-loop decoupling control requirements of the unit coordinated control system. The PID controller is specifically responsible for quickly adjusting the main steam valve opening to quickly respond to instant changes in the grid load. By accurately adjusting the PID parameters, the controller can achieve fast and precise control of the main steam valve opening, thereby reducing the impact on other control loops. This design enables the PID controller to effectively decouple the coupling between the power generation control loop and other control loops, improve the system's response speed and control accuracy, and reduce the system stability problem caused by the coupling effect. Specifically expressed as:

[0048]

[0049] Where: K P is the proportionality coefficient, KI is the integration coefficient, K D is the differential coefficient (the coefficient is adjusted according to the situation in the application), and e is the error between the target value and the feedback value.

[0050] Aiming at the precise control of main steam pressure and water wall outlet temperature, the present invention designs a model predictive controller as a two-input and two-output structure in a unit coordination system.

[0051] Using field data, the general form of the state space model is obtained based on the subspace system identification method:

[0052]

[0053] Where: A c , B c , C c , D c , K c is the continuous state space matrix; u(t) is the input; y(t) is the output; e(t) is the disturbance; x(t) is the n-dimensional state vector. For dynamic systems, D defaults to zero.

[0054] The model predictive controller uses the system's predictive model to predict the future system behavior. The predictive model takes the form of a state space:

[0055]

[0056] Where: 0 m ∈n x represents an n-dimensional zero matrix; u(k) represents the control input increment at time k.

[0057] For a dual-input dual-output system, the predicted output sequence Y and the control input sequence ΔU are defined as:

[0058] Y=[y(k+1|k) Τ y(k+2|k) Τ … y(k+N P |k) Τ ] Τ

[0059] ΔU=[Δu(k) Τ Δu(k+1) Τ … Δu(k+N C -1) Τ ]

[0060] The model predictive controller considers the system behavior in the future through the prediction model and obtains the predicted output sequence in the prediction time domain:

[0061]

[0062] Where: Y is the prediction time domain N P Intra-prediction output sequence; ΔU is the control time domain N C Internal control increment sequence; F and Φ are the constructed coefficient matrices respectively.

[0063] Furthermore, for the dual-loop coordinated control of the unit, the objective function of the controller is designed with minimizing the control error as the optimization goal:

[0064]

[0065] Where: J is the optimization objective function; r i (k) is the i-th output reference value at time k; y i (k) is the i-th output value at time k; R is the reference value matrix; S, are the error weight matrix and the control weight matrix respectively.

[0066] The optimization solution goal of model predictive control is to find the optimal control sequence to minimize the objective function, and express the output sequence in the objective function as a control sequence:

[0067]

[0068] In order to facilitate the optimization solution of quadratic programming, the objective function is further organized into the form of quadratic programming:

[0069]

[0070] Where: H is the quadratic term coefficient matrix, f is the linear term coefficient matrix.

[0071] Furthermore, control quantity constraints are set according to unit safety restrictions and operating conditions:

[0072] U min (k)≤U(k)≤U max (k)

[0073] ΔU min (k)≤ΔU(k)≤ΔU max (k)

[0074] Where U min , U max are the minimum and maximum constraint matrices of the input respectively; ΔU min , ΔU max are the minimum and maximum constraint matrices for the input increments, respectively.

[0075] Model predictive control optimizes control inputs to meet future constraints and goals. This design enables model predictive control to predict and respond to external disturbances and internal uncertainties in advance, and considering the slow response speed of coal and water feed, model predictive control can achieve precise regulation of these two variables.

[0076] The output of the PID controller is input to the model predictive controller as a measurable disturbance, which allows the model predictive control to take into account the influence of the PID controller when predicting and optimizing the control strategy, thereby achieving more coordinated control.

[0077] like Figure 2 As shown, the present invention also provides a control effect comparison diagram of the coordinated-level PID and model predictive control loop decoupling control method of the coal-fired unit, and compares and analyzes the step performance of the PID and model predictive control loop decoupling control method control strategy, the model predictive controller, and the traditional PID control combined with feedforward. The control effect comparison diagram shows that although the traditional feedforward-PID control strategy exhibits a small overshoot in the response process, its response time is significantly delayed, highlighting the lack of response speed. In contrast, both the PID and model predictive control loop decoupling control strategy and the model predictive controller have demonstrated excellent rapid response capabilities and can respond quickly and effectively to step signals.

[0078] like Figure 3 As shown, the present invention also provides a comparison diagram of the control effect of the disturbance condition of the coordinated level PID and model predictive control loop decoupling control method of the coal-fired unit. In order to evaluate the performance of different control strategies in dealing with uncertain interference, step interferences were introduced at 500 seconds, 4000 seconds and 7500 seconds for the three key control variables of coal feed rate, water feed rate and main steam valve. The performance of PID and model predictive control loop decoupling control strategy, classical model predictive controller and feedforward-PID control in dealing with the above interferences was compared and analyzed. Under the three disturbance scenarios, the feedforward-PID controller showed the largest overshoot, and its adjustment process was accompanied by significant fluctuations, and it converged to the set reference value at a relatively slow speed. In contrast, both the PID and model predictive control loop decoupling control strategy and the model predictive controller showed excellent anti-interference ability. They can quickly identify and respond to interference, significantly shorten the setting time, and effectively reduce the control deviation. The PID and model predictive control loop decoupling control strategy achieves control performance similar to that of the model predictive controller, while significantly reducing the computational complexity of the solution process, thereby improving real-time performance and reducing resource consumption in practical applications. This reflects its significant potential and practical application value in the field of control strategy optimization.

[0079] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A decoupling control method for the coordinated PID and model predictive control loop of a coal-fired unit, characterized by: The following steps are involved: Step 1: Analyze the multivariable coupling characteristics of the unit coordinated control system and determine the multi-loop decoupling control requirements of the unit; Step 2: Based on the multi-loop decoupling control requirements of the unit, design the PID controller of the unit coordination system for the power generation control loop to control the main steam valve opening; Step 3: Design a model predictive controller for the unit coordinated control system for the dual circuits of main steam pressure and water wall outlet temperature.

2. A method for decoupling control of a coordinated PID and model predictive control loop of a coal-fired unit according to claim 1, characterized in that: The multivariable coupling characteristics in step 1 include establishing the relationship between the dynamic response speed of valve opening and the response speed of water supply and coal supply.

3. A method for decoupling control of a coordinated PID and model predictive control loop of a coal-fired unit according to claim 2, characterized in that: In step 2, the main steam valve opening is controlled, which is specifically expressed as: Where: K P is the proportionality coefficient, K I is the integration coefficient, K D is the differential coefficient, and e is the error between the target value and the feedback value.

4. A method for decoupling control of a coordinated PID and model predictive control loop of a coal-fired unit according to claim 3, characterized in that: Step 3 The specific steps include: Using field data, the general form of the state space model is obtained based on the subspace system identification method: Where: A c , B c , C c , D c , K c is a continuous state space matrix; u(t) is the input; y(t) is the output; e(t) is the disturbance; x(t) is the n-dimensional state vector. For dynamic systems, D defaults to zero; In order to design the prediction model of the model predictive controller, the system model is discretized and differentiated to obtain the system augmented model: Where: Represents an n-dimensional zero matrix; represents the control input increment at time k; For a dual-input dual-output system, the predicted output sequence Y and the control input sequence ΔU are defined as: Y=[y(k+1k) Τ y(k+2k) Τ …y(k+N P k)Τ] Τ ΔU=[Δu(k) Τ Δu(k+1) Τ …Δu(k+N C -1) Τ ] Based on the prediction model, model predictive control obtains the predicted output sequence in the prediction time domain: Where: Y is the prediction time domain N P Intra-prediction output sequence; ΔU is the control time domain N C Internal control increment sequence; F and Φ are the constructed coefficient matrices respectively.

5. A method for decoupling control of a coordinated PID and model predictive control loop of a coal-fired unit according to claim 4, characterized in that: For the dual-loop coordinated control of the unit, the objective function of the controller is designed with minimizing the control error as the optimization goal: Where: J is the optimization objective function; ri ( k ) is the i-th output reference value at time k; yi ( k ) is the i-th output value at time k; R is the reference value matrix; S, are the error weight matrix and the control weight matrix respectively.

6. A method for decoupling control of a coordinated PID and model predictive control loop of a coal-fired unit according to claim 5, characterized in that: The optimization solution goal of model predictive control is to find the optimal control sequence to minimize the objective function, and express the output sequence in the objective function as a control sequence: In order to facilitate the optimization solution of quadratic programming, the objective function is further organized into the form of quadratic programming: Where: H is the quadratic term coefficient matrix, f is the linear term coefficient matrix; Furthermore, control quantity constraints are set according to unit safety restrictions and operating conditions: YOU min (k)≤U(k)≤U max (k) ΔU min (k)≤ΔU(k)≤ΔU max (k) Where: U min , U max are the minimum and maximum constraint matrices for the input respectively; ΔU min , ΔU max are the minimum and maximum constraint matrices for the input increments, respectively.

7. A method for decoupling control of a coordinated PID and model predictive control loop of a coal-fired unit according to claim 6, characterized in that: The control action of the PID controller is input into the model predictive controller as measurable disturbance information. The model predictive control obtains the optimal control sequence of the system through real-time predictive optimization and realizes rolling optimization control.

Images