A control method based on a feedforward adaptive periodic complex expert controller

Abstract

The application discloses a control method based on a feedforward adaptive periodic complex expert controller, and comprises the following steps: identifying a power load of a generator and determining the generator with added or reduced load; according to the change trend of a low-low pressure steam header pressure and a steam pressure at a generator outlet, corresponding state types and control output periods are given; according to the output conditions met by the steam pressure, controller control output is carried out according to the met output conditions. The application not only relies on the low-low pressure steam pressure as feedback for control, but also considers the interference of the generator load and the inlet steam pressure on the low-low pressure steam pressure, so that fluctuations caused by untimely adjustment of single feedback control can be avoided, and fluctuations caused by the generator load and the inlet steam pressure can be overcome; through the technology, the controller can intelligently judge the danger of the low-low pressure steam pressure at the current time, and dynamically adjust the output period.

Description

Technical Field

[0001] This invention belongs to the field of automatic control, and in particular relates to a control method based on a feedforward adaptive periodic complex expert controller. Background Technology

[0002] A steam turbine generator consists of a steam turbine and a generator. Its function is to use steam to rotate the turbine impeller, which drives the generator rotor to rotate in the magnetic field of the excitation winding, cutting magnetic lines of force to generate electricity. In the application scenario of main steam pressure control of a steam turbine generator, the main control target is the main steam pressure, which is achieved through a control system. Existing periodic expert controllers have trend control and fast response modes. They can adjust the value of the controlled variable in a timely manner through actuators based on the changing trend of the controlled variable, and overcome the limitation of the control cycle to achieve a fast response. However, its trend control mode only adjusts the control based on the rate of change of steam pressure P, resulting in a relatively simple output cycle that cannot be dynamically adjusted after the cycle is determined.

[0003] For example, a control method based on a periodic expert controller, disclosed in Chinese patent literature (publication number CN111930032A), includes: acquiring the controller's count value and the value of the controlled variable; comparing the value of the controlled variable with preset conventional control parameters; controlling the controlled variable based on the comparison result and a fast response parameter; simultaneously, controlling the controlled variable based on preset trend control parameters when the count value has not reached the preset control period; and resetting the count value to 1 when the count value reaches the control period. This patent adds a trend control mode and a fast response mode to the traditional periodic expert controller. Although it can adjust the value of the controlled variable in a timely manner through the actuator according to the changing trend of the controlled variable, improving the sensitivity and control effect of the periodic expert controller, it still suffers from the problem that its trend control mode only adjusts the control based on the rate of change of steam pressure P, resulting in a relatively simple output period that cannot be dynamically adjusted after the period is determined. Summary of the Invention

[0004] The present invention aims to overcome the problems of existing control methods that adjust control only based on the rate of change of steam pressure P, resulting in a relatively simple output cycle that cannot be dynamically adjusted after the cycle is determined. The invention provides a control method based on a feedforward adaptive periodic complex expert controller.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A control method based on a feedforward adaptive periodic complex expert controller includes the following steps: S1: Identify the generator power load and determine the generators whose loads are being increased or decreased; S2: Assign corresponding state types and control output cycles based on the changing trends of the low-pressure steam main pressure and steam pressure at the generator set outlet.

[0007] S3: Determine the output conditions satisfied by the steam pressure based on the state type, and control the output according to the satisfied output conditions; the satisfied output conditions include the relevant conditions calculated by the controller feedforward complex expert control and the relevant conditions of output predictive control. This invention provides a control method based on a feedforward adaptive periodic complex expert controller, which adjusts the control according to the changing trends of the low-pressure steam main pressure at the generator set outlet and the steam pressure. It does not rely solely on the low-pressure steam pressure as feedback for control, but also considers the interference of generator load and inlet steam pressure on the low-pressure steam pressure. This avoids fluctuations caused by untimely adjustments in single feedback control, and overcomes fluctuations caused by generator load and inlet steam pressure. This technology enables the controller to intelligently determine the current danger level of the low-pressure steam pressure and dynamically adjust the output cycle.

[0008] As a preferred embodiment of the present invention, S1 specifically includes the following steps: S11: Obtain the corresponding switch status, real-time load, and load limit of each generator; S12: Exclude generators with a switch status of 0 from the control process, retain generators with a switch status of 1 in the control process, and sort the generators according to their real-time load; S13: Compare the real-time load of the generators retained in the control process with their corresponding load limits according to the sorting order. If the real-time load of a generator is greater than its corresponding load limit, it is determined that the generator is not allowed to increase its load. Finally, the generator with the largest load needs to reduce its load according to the sorting order, and the generator with the smallest load needs to increase its load. The present invention automatically identifies the generator load and automatically allocates generators for load increase or decrease. This technology can reduce the situation of untimely and erroneous human judgment. The controller of the present invention uses a complex expert control algorithm that combines generator load and inlet steam pressure as feedforward with outlet steam pressure feedback to obtain the output value and compare it with the threshold for control output. This technology not only relies on low-pressure steam pressure as feedback for control, but also considers the interference of generator load and inlet steam pressure on low-pressure steam pressure. This can avoid fluctuations caused by untimely adjustments in single feedback control, and at the same time overcome fluctuations caused by generator load and inlet steam pressure.

[0009] In a preferred embodiment of the present invention, S2 specifically involves: dividing the controlled variable, steam pressure, into intervals and state types; determining the generator's state type based on the measured value P of the low-pressure steam main outlet pressure and the pressure trend; and assigning a corresponding control output cycle. The intervals include normal interval, previous danger interval, upper danger interval, next danger interval, and lower danger interval. The state types are nine types: E, H1, H2, H3, H4, L1, L2, L3, and L4. This invention further divides the real-time state of low-pressure steam pressure by analyzing the changes in the measured values ​​of low-pressure steam pressure within different intervals, and simultaneously assigns corresponding control output cycles. This technology enables the controller to intelligently determine the current danger level of the low-pressure steam pressure and dynamically adjust the strength and speed of the control effect, achieving a replacement for human intervention.

[0010] As a preferred solution of the present invention, said step S2 specifically comprises the following steps: S21: setting adaptive relevant parameters of the control output cycle, comprising a target value SP, a boundary value MH between the last dangerous interval and the normal interval, a boundary value ML between the next dangerous interval and the normal interval, a boundary value MHH between the upper dangerous interval and the last dangerous interval, and a boundary value MLL between the lower dangerous interval and the next dangerous interval; S22: acquiring an actually measured value P of the outlet extra-low pressure steam main pipe pressure of the generator set; S23: when ML < P < MH, the state is E, and the control output cycle is assigned as T4; when MH ≤ P < MHH, judging the trend of pressure, if the pressure trend is in the positive direction, the state is H3, and the control output cycle is assigned as T2; if the pressure trend is in the negative direction, the state is H4, and the control output cycle is assigned as T3; when MLL < P ≤ ML, judging the trend of pressure, if the pressure trend is in the positive direction, the state is L4, and the control output cycle is assigned as T3; if the pressure trend is in the negative direction, the state is L3, and the control output cycle is assigned as T2; when MHH ≤ P, judging the trend of pressure, if the pressure trend is in the positive direction, the state is H1, and the control output cycle is assigned as T1; if the pressure trend is in the negative direction, the state is H2, and the control output cycle is assigned as T2; when P ≤ MLL, judging the trend of pressure, if the pressure trend is in the positive direction, the state is L2, and the control output cycle is assigned as T2; if the pressure trend is in the negative direction, the state is L1, and the control output cycle is assigned as T1. According to the present invention, the range of the actually measured value of extra-low pressure steam pressure can be automatically divided into dangerous intervals, sub-dangerous intervals, normal intervals and the like according to production requirements. By means of this technology, dynamic multi-interval division can be performed for the fluctuation range of extra-low pressure steam pressure, the division is more detailed, the limitation of fixed interval division is avoided, and the physical significance of actual pressure fluctuation is taken into consideration; moreover, through the change condition of the actually measured value of extra-low pressure steam pressure in different intervals, the real-time state of extra-low pressure steam pressure is further divided, and corresponding control output cycles are assigned at the same time. By means of this technology, a controller can intelligently judge the current danger of extra-low pressure steam pressure, dynamically adjust the intensity and output speed of the control effect, and achieve the effect of replacing manual work.

[0011] As a preferred solution of the present invention, the specific process of performing controller control output when the relevant conditions for controller feedforward complex expert control calculation are satisfied in S3 is as follows: S31: For generators with state types E, H3, L3, H4, L4, calculate the outlet steam pressure feedback output value FK1_OUT1, the inlet steam pressure feedforward output value QK1_OUT1, and the generator load feedforward output value QK2_OUT1, and sum them to obtain the single-step controller output value ZJ_OUT1; S32: Calculate the cumulative expert control output value ZJ_Q1; S33: Compare the cumulative expert control output value ZJ_Q1 with the cumulative output threshold ZJ_Q_MAX1, if the cumulative expert control output value ZJ_Q1 is greater than the cumulative output threshold ZJ_Q_MAX1, perform controller control output, otherwise, return to S1.

[0012] As a preferred solution of the present invention, the specific process of performing controller control output when the relevant conditions for output predictive control are satisfied in S3 is as follows: for generators with state types H1, L1, H2, L2, adaptive periodic predictive control controller output is performed in combination with the model between load increase / decrease and generator load, and only the first step is output. The present invention identifies the model between pulse signals and steam pressure, predicts future multi-step pulse signal output and steam pressure change trend based on the model relationship, outputs only the first step and performs rolling prediction. Through this technology, the prediction of future control can be realized. When the first step is output, the subsequent change trend of the target variable can be intuitively understood, so that the control is reliable, and the unknown impact effect on the target variable after control output is avoided, and the influence of historical control output on the future is considered.

[0013] As a preferred solution of the present invention, the control output periods satisfy T1<T2<T3<T4. The T1 period has the most sensitive response, and the T4 period has the weakest response.

[0014] Therefore, the present invention has the following beneficial effects:

[0015] 1. The generator load condition can be automatically identified, and generators for load increase and decrease can be automatically allocated, which can reduce untimely manual judgment and errors;

[0016] 2. According to production requirements, the measured value range of low-low pressure steam can be automatically divided into danger intervals, sub-danger intervals, normal intervals, etc. Dynamic multi-interval division can be performed according to the fluctuation range of low-low pressure steam, and the division can be more detailed, which avoids the limitation of fixed interval division and takes into account the physical meaning of actual pressure fluctuation;

[0017] 3. By analyzing the changes in measured values ​​of low-pressure steam pressure within different ranges, the real-time state of low-pressure steam pressure is further divided, and corresponding control output cycles are assigned. This allows the controller to intelligently determine the current danger level of low-pressure steam pressure and dynamically adjust the strength and speed of control effects, thus replacing human intervention.

[0018] 4. The controller uses a complex expert control algorithm that combines generator load and inlet steam pressure as feedforward with outlet steam pressure feedback to obtain the output value and compare it with the threshold to control the output. It does not rely solely on low-pressure steam pressure as feedback for control, but also considers the interference of generator load and inlet steam pressure on low-pressure steam pressure. This can avoid fluctuations caused by untimely adjustment in single feedback control, and overcome fluctuations caused by generator load and inlet steam pressure.

[0019] 5. Identify the model between pulse signal and steam pressure, and predict the future multi-step pulse signal output and steam pressure change trend based on the model relationship. Only output the first step and perform rolling prediction. This can achieve prediction of future control. When outputting the first step, we can intuitively understand the subsequent change trend of the target variable, have confidence in the control, avoid the unknown impact of the control output on the target variable, and consider the impact of historical control output on the future. Attached Figure Description

[0020] Figure 1 This is a simplified process diagram of a generator set consisting of three units connected in parallel using existing technology;

[0021] Figure 2 This is a flowchart of the control method of the feedforward adaptive periodic complex expert controller of the present invention;

[0022] Figure 3 Logic diagram of generator load distribution structure;

[0023] Figure 4 A schematic diagram of adaptive interval division and output cycle control for controlling the output cycle;

[0024] Figure 5 Logic diagram of the complex expert control computational structure for the controller feedforward;

[0025] Figure 6 A computational logic diagram for a complex expert control feedforward controller for a certain type of state;

[0026] Figure 7 Output the structural logic diagram for the controller;

[0027] Figure 8 A graph showing the relationship between load increase / decrease pulse signals and outlet steam pressure;

[0028] Figure 9This is a schematic diagram of the adaptive periodic model prediction. Detailed Implementation

[0029] The present invention will now be further described with reference to the accompanying drawings and specific embodiments.

[0030] For a factory with three generator sets connected in parallel, the production control requirements are to allocate the load of the three generators, such as... Figure 1 The process flow is shown below.

[0031] Three identical 3MW generators are connected in parallel. Steam enters each generator from a low-pressure steam distributor, and the outlet steam is 0.18MPa low-pressure steam supplied to downstream units. The main control objective of this section is to ensure a stable 0.18MPa steam pressure while maintaining similar loads on the three generators. The most important and challenging part of the entire process is ensuring that the generators can automatically adjust their loads according to changes in steam pressure (automatically determining which generator to adjust the load) even when the generators are not at full load, thus maintaining stable steam pressure. The main control challenges are as follows: 1. When a load change is needed, the operator must manually assess the load status of the three generators, which can lead to untimely assessments; 2. The operator manually adjusts the load increase / decrease buttons based on pressure to achieve stable inlet steam pressure control, which can result in significant pressure fluctuations due to untimely adjustments; 3. The load increase / decrease buttons are pulse signals, not continuous variables, and cannot be automatically adjusted; 4. Traditional controllers have a fixed output cycle and cannot adjust the frequency of load increases / decreases based on steam pressure trends.

[0032] The present invention provides a control method based on a feedforward adaptive periodic complex expert controller, comprising the following steps: S1: identifying the power load of generators, and determining the generators to increase or decrease load; S1 specifically comprises the following steps: S11: acquiring the corresponding switching state, real-time load and load upper limit of each generator; S12: excluding generators with a switching state of 0 from the control process, retaining generators with a switching state of 1 in the control process, and sorting the generators according to the real-time load of the generators; S13: comparing the real-time load of the generators retained in the control process with their corresponding load upper limits according to the sorting order, if the real-time load of a generator is greater than its corresponding load upper limit, determining that the generator is not allowed to increase load; finally, determining according to the sorting order that the generator with the maximum load needs to reduce load, and the generator with the minimum load needs to increase load. S2: assigning corresponding state types and control output periods according to the pressure of the extra-low pressure steam main pipe at the outlet of the generator set and the change trend of the steam pressure; S2 is specifically as follows: dividing intervals and state types for the steam pressure which is the controlled variable, judging the state type of the generator and assigning a corresponding control output period according to the actually measured value P of the pressure of the extra-low pressure steam main pipe at the outlet of the generator set and the pressure trend, wherein the intervals comprise a normal interval, a previous warning interval, an upper warning interval, a next warning interval and a lower warning interval, and the state types comprise nine types: E, H1, H2, H3, H4, L1, L2, L3 and L4. S2 specifically comprises the following steps: S21: setting parameters related to control output period adaptation, comprising a target value SP, a boundary value MH between the previous warning interval and the normal interval, a boundary value ML between the next warning interval and the normal interval, a boundary value MHH between the upper warning interval and the previous warning interval, and a boundary value MLL between the lower warning interval and the next warning interval; S22: acquiring the actually measured value P of the pressure of the extra-low pressure steam main pipe at the outlet of the generator set; S23: when ML<P<MH, the state is E, and the assigned control output period is T4; when MH≤P<MHH, judging the pressure trend, if the pressure trend is in the positive direction, the state is H3, and the assigned control output period is T2; if the pressure trend is in the negative direction, the state is H4, and the assigned control output period is T3; when MLL<P≤ML, judging the pressure trend, if the pressure trend is in the positive direction, the state is L4, and the assigned control output period is T3; if the pressure trend is in the negative direction, the state is L3, and the assigned control output period is T2; when MHH≤P, judging the pressure trend, if the pressure trend is in the positive direction, the state is H1, and the assigned control output period is T1; if the pressure trend is in the negative direction, the state is H2, and the assigned control output period is T2; when P≤MLL, judging the pressure trend, if the pressure trend is in the positive direction, the state is L2, and the assigned control output period is T2; if the pressure trend is in the negative direction, the state is L1, and the assigned control output period is T1; wherein the control output periods satisfy T1<T2<T3<T4.S3: Determine the output conditions satisfied by the steam pressure based on the state type, and perform controller control output based on the satisfied output conditions; the satisfied output conditions include controller feedforward complex expert control calculation related conditions and output predictive control related conditions; the specific process of performing controller control output based on satisfying controller feedforward complex expert control calculation related conditions in S3 is as follows: S31: For generators with state types E, H3, L3, H4, and L4, calculate the outlet steam pressure feedback output value FK1_OUT1, the inlet steam pressure feedforward output value QK1_OUT1, and the generator load feedforward output value QK2_OUT1, and sum them to obtain single-step control. S32: Calculate the cumulative value of expert control output ZJ_Q1; S33: Compare the cumulative value of expert control output ZJ_Q1 with the cumulative output threshold ZJ_Q_MAX1. If the cumulative value of expert control output ZJ_Q1 is greater than the cumulative output threshold ZJ_Q_MAX1, then the controller outputs control; otherwise, return to S1. The specific process of the controller output in S3 that meets the relevant conditions for output predictive control is as follows: For generators with state types H1, L1, H2, and L2, combined with the model between load increase / decrease and generator load, adaptive periodic predictive control controller output is performed, and only the first step is output.

[0033] Example: In this example, a control method for a feedforward adaptive periodic complex expert controller is proposed, the process of which is as follows: Figure 2 As shown, it includes:

[0034] Step 1: Automatically identify the generator power load and determine the generator to be added or removed from the load.

[0035] For coordinated control of multiple generator sets, it is necessary to consider the load limit of each generator according to production requirements and reasonably allocate the load of each generator.

[0036] Generator load distribution logic is as follows Figure 3 As shown, first determine the switching status of each generator, then determine whether the load of each generator with switch 1 in the generator set has reached the upper limit. Using the generator set load distribution module, determine which generator to increase the load and which to decrease the load.

[0037] Taking three generator sets as an example, the specific process is as follows:

[0038] Step 11: Obtain the switches MSW1, MSW2 and MSW3 of the three generators, the real-time loads M1, M2 and M3, and the load limits MH1, MH2 and MH3 of the three generators respectively;

[0039] Step 12: Determine whether the switches of the three generators are equal to 1. If so, the corresponding generator will be considered in the control process; otherwise, it will be excluded from the control process.

[0040] Step 13: Compare whether the real-time load of the generator that meets the conditions of step 12 exceeds the corresponding upper load limit; if the real-time load is greater than or equal to the upper limit, it is determined that load increase is not allowed for the generator;

[0041] Step 14: Compare the magnitude relationship of the real-time loads of the generators that meet the conditions of step 12;

[0042] Step 15: Determine the generator with higher load as the one that needs to reduce load;

[0043] Step 16: For the generator with lower load, if it meets the conditions of step 13, recursively process the generator with the next lower load and repeat step 16; if it does not meet the conditions, determine the generator with lower load as the one that needs to increase load.

[0044] Step 2: Assign a corresponding control output period according to the change trend of steam pressure.

[0045] Taking into account the characteristic in the actual steam pressure regulation process that the greater the deviation from the target value, the more sensitive the control response, different control output periods are assigned for steam pressure in different interval states, the specific interval division and control output periods are as Figure 4 . It can be adjusted according to actual conditions.

[0046] The interval of the controlled variable steam pressure is divided into: a normal interval, upper and lower sub-danger intervals, and upper and lower danger intervals. The state types include: E, H1, H2, H3, H4, L1, L2, L3, L4. The control output periods are: T1, T2, T3, T4. Note: the intervals, state types and control output periods can be divided according to actual conditions.

[0047] Taking the above three generator units as an example, the specific process is as follows:

[0048] Step 21: Set relevant parameters of the adaptive control output period module: target value SP, upper sub-danger / normal interval boundary MH, lower sub-danger / normal interval boundary ML, upper danger / upper sub-danger boundary MHH and lower danger / lower sub-danger boundary MLL; although the target value SP does not participate in the assignment process of the control output period herein, the target value is the reference and basis for the setting of other participating values MH, ML, MHH and MLL;

[0049] Step 22: Obtain the measured value P of the low-low pressure steam main pipe pressure of the generator unit;

[0050] Step 23: When ML < P < MH, the state is E, and the assigned control output period is T4 (seconds);

[0051] Step 24: When MH≤P<MHH, the trend of the pressure is determined. If the pressure trend is in the positive direction, the state is H3, and the control output period is assigned as T2 (seconds); if the pressure trend is in the negative direction, the state is H4, and the control output period is assigned as T3 (seconds);

[0052] Step 25: When MLL<P≤ML, the trend of the pressure is determined. If the pressure trend is in the positive direction, the state is L4, and the control output period is assigned as T3 (seconds); if the pressure trend is in the negative direction, the state is L3, and the control output period is assigned as T2 (seconds);

[0053] Step 26: When MHH≤P, the trend of the pressure is determined. If the pressure trend is in the positive direction, the state is H1, and the control output period is assigned as T1 (seconds); if the pressure trend is in the negative direction, the state is H2, and the control output period is assigned as T2 (seconds);

[0054] Step 27: When P≤MLL, the trend of the pressure is determined. If the pressure trend is in the positive direction, the state is L2, and the control output period is assigned as T2 (seconds); if the pressure trend is in the negative direction, the state is L1, and the control output period is assigned as T1 (seconds);

[0055] Step 28: The control output periods satisfy T1<T2<T3<T4. The period T1 has the most sensitive response, and the period T4 has the weakest response.

[0056] Step 3: When the steam pressure satisfies the relevant conditions of the feedforward complex expert control calculation module of the controller, an output value is obtained by using the feedforward complex expert control calculation module, and it is determined whether the output value is greater than a threshold value; if not, the process returns to step 1; when the steam pressure satisfies the relevant prediction control conditions in the output module, the predicted output of multi-step future pulse signals is calculated based on the pulse signals and the steam pressure model, and only the current first step output is executed.

[0057] It can be known from the above that the generator set load distribution module determines the generator for loading and unloading, and the adaptive control output period module determines the control output period. This part is the core algorithm part of the control method of a feedforward variable-period complex expert controller proposed by the present invention, and determines the output effect of the controller. The target variable, low-pressure steam pressure, is converted into generator load through the feedforward complex expert control calculation module.

[0058] According to the division of the adaptive control output period module, the feedforward complex expert control calculation module is divided into three parts: state type E, state type H3 / L3 and state type H4 / L4, as Figure 5 shown.

[0059] The three state types mentioned above correspond to different controller parameters. Since the calculation and comparison process is the same, only state type E is used as an example here. The first set of control parameters includes: outlet steam pressure feedback output: FK1_OUT1, inlet steam pressure feedforward output: QK1_OUT1, generator load feedforward output: QK2_OUT1, complex expert control output: ZJ_OUT1, complex expert control cumulative output: ZJ_Q1, and complex expert control cumulative output threshold: ZJ_Q_MAX1.

[0060] Flowchart as follows Figure 6 As shown, the steps are as follows:

[0061] Step 31: Calculate the outlet steam pressure feedback output: FK1_OUT1, the inlet steam pressure feedforward output: QK1_OUT1, and the generator load feedforward output: QK2_OUT1;

[0062] Step 32: Sum the outputs of the three parts mentioned above to obtain the single-step controller output ZJ_OUT1;

[0063] Step 33: Calculate the cumulative expert control output ZJ_Q1;

[0064] Step 34: Compare ZJ_Q1 with the cumulative output threshold ZJ_Q_MAX1. If it is less than the threshold, return to step 31; if it is greater than or equal to the threshold, then control the output.

[0065] The main functions of the controller output module in this section are:

[0066] For the five states with state types E, H3 / L3, and H4 / L4, the control output of the feedforward complex expert control calculation module is combined with the control output period adaptive module to perform period adaptive output.

[0067] For the four states of H1 / L1 and H2 / L2, an adaptive periodic predictive control output is performed by combining the model of load increase / decrease and generator load.

[0068] The logic of the controller output module is as follows: Figure 7 As shown.

[0069] Adaptive periodic model predictive control first identifies the model relationship between the load increase / decrease pulse signals and the outlet steam pressure. The model relationship is as follows: Figure 8 As shown.

[0070] Based on the above model, the prediction of future multi-step pulse signals is illustrated in the following diagram. Figure 9 As shown.

[0071] Taking state types H1 and H2 as examples, when the outlet steam pressure enters state H1, the controller output cycle becomes T1. The predictive control output results based on the model for the next three load increases are as follows: the time interval between the first and second steps is T1, and the time interval between the second and third steps is T2.

[0072] Predictive control only outputs the first step of load reduction. Within the predicted future time interval T1, the outlet steam pressure will still be in state H1. Therefore, the time interval between the first and second steps is T1. After the second step predicts the load reduction, the outlet steam pressure will enter state H2 within the predicted future time interval T1. Therefore, the predicted output interval between the second and third steps becomes T2, and so on.

[0073] The entire adaptive cycle model predictive control is a rolling prediction and only outputs the first step.

[0074] Therefore, the present invention has the following beneficial effects:

[0075] 1. It can automatically identify generator load conditions and automatically allocate generators that need to be increased or decreased in load. This technology can reduce the possibility of untimely and erroneous human judgment.

[0076] 2. Based on production requirements, the measured range of low-pressure steam can be automatically divided into dangerous, secondary dangerous, and normal ranges, etc. This technology allows for dynamic multi-range division of the fluctuation range of low-pressure steam, enabling more detailed division and avoiding the limitations of fixed range division while taking into account the physical meaning of actual pressure fluctuations.

[0077] 3. By analyzing the changes in measured low-pressure steam pressure within different ranges, the real-time state of low-pressure steam pressure is further divided, and corresponding control output cycles are assigned. This technology enables the controller to intelligently assess the current danger level of the low-pressure steam pressure and dynamically adjust the strength and speed of the control effect, effectively replacing human intervention.

[0078] 4. The controller utilizes a complex expert control algorithm that combines generator load and inlet steam pressure as feedforward with outlet steam pressure feedback to obtain the output value and compare it with a threshold for control output. This technology not only relies on low-pressure steam pressure as feedback for control but also considers the interference of generator load and inlet steam pressure on low-pressure steam pressure. This avoids fluctuations caused by untimely adjustments in single feedback control and overcomes fluctuations caused by generator load and inlet steam pressure.

[0079] 5. Identify the model between the pulse signal and steam pressure, and predict future multi-step pulse signal output and steam pressure change trends based on the model relationship. Only the first step is output for rolling prediction. This technique enables prediction of future control. When outputting the first step, the subsequent trend of the target variable is intuitively understood, providing confidence in the control and avoiding the unknown impact of control output on the target variable. It also considers the influence of historical control output on the future.

[0080] Terminology Explanation:

[0081] Model predictive control: The current control action is obtained by solving a finite-time open-loop optimal control problem at each sampling instant. The current state of the process serves as the initial state of the optimal control problem, and only the first control action is implemented from the solved optimal control sequence.

[0082] 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 changes or substitutions conceived without creative effort should be included within the scope of protection of the present invention.

Claims

1. A control method based on a feedforward adaptive periodic complex expert controller, characterized in that, Comprising the following steps: S1: identifying the power load of generators, and determining generators that need to increase or decrease load; S2: assigning corresponding state types and control output periods according to the pressure of the low-low pressure steam main pipe at the outlet of the generator set and the change trend of the steam pressure; defining a target value SP, a boundary value MH between the previous danger interval and the normal interval, a boundary value ML between the next danger interval and the normal interval, a boundary value MHH between the upper danger interval and the previous danger interval, a boundary value MLL between the lower danger interval and the next danger interval, the pressure P of the low-low pressure steam main pipe at the outlet of the generator set, and state types E, H1, H2, H3, H4, L1, L2, L3, L4; when ML < P < MH, the state is E, and the control output period is T4; when MH ≤ P < MHH and the pressure trend is in the positive direction, the state is H3, and the output period is T2; when the pressure trend is in the negative direction, the state is H4, and the output period is T3; when MLL < P ≤ ML and the pressure trend is in the positive direction, the state is L4, and the output period is T3; when the pressure trend is in the negative direction, the state is L3, and the output period is T2; when MHH ≤ P and the pressure trend is in the positive direction, the state is H1, and the output period is T1; when the pressure trend is in the negative direction, the state is H2, and the output period is T2; when P ≤ MLL and the pressure trend is in the positive direction, the state is L2, and the output period is T2; when the pressure trend is in the negative direction, the state is L1, and the output period is T1; S3: judging the output condition satisfied by the steam pressure according to the state type, and performing control output of the controller according to the satisfied output condition; the satisfied output conditions include relevant conditions for controller feedforward complex expert control calculation and relevant conditions for output predictive control.

2. The control method based on a feedforward adaptive periodic complex expert controller according to claim 1, characterized in that, S1 specifically comprises the following steps: S11: acquiring the corresponding switching state, real-time load and load upper limit of each generator; S12: excluding generators with a switching state of 0 from the control process, retaining generators with a switching state of 1 in the control process, and sorting the generators according to their real-time loads; S13: comparing the real-time load of the generators retained in the control process with their corresponding load upper limits according to the sorting order, if the real-time load of a generator is greater than its corresponding load upper limit, determining that the generator is not allowed to increase load; finally determining, according to the sorting order, that the generator with the largest load needs to reduce load, and the generator with the smallest load needs to increase load.

3. The control method based on a feedforward adaptive periodic complex expert controller according to claim 1, characterized in that, S2 is specifically: dividing intervals and state types for the steam pressure, which is a controlled variable, and judging the state type of the generator and assigning the corresponding control output period according to the actually measured value P of the pressure of the low-low pressure steam main pipe at the outlet of the generator set and the pressure trend, wherein the intervals comprise a normal interval, a previous danger interval, an upper danger interval, a next danger interval and a lower danger interval.

4. The control method based on a feedforward adaptive periodic complex expert controller according to claim 3, characterized in that, The intervals comprise a normal interval, a previous danger interval, an upper danger interval, a next danger interval and a lower danger interval.

5. The control method based on a feedforward adaptive periodic complex expert controller according to claim 4, characterized in that, The specific process of performing controller control output by satisfying the relevant conditions of controller feedforward complex expert control calculation in S3 is as follows: S31: For generators with state types E, H3, L3, H4, and L4, calculate the outlet steam pressure feedback output value FK1_OUT1, the inlet steam pressure feedforward output value QK1_OUT1, and the generator load feedforward output value QK2_OUT1, and sum them to obtain the single-step controller output value ZJ_OUT1. S32: Calculate the cumulative value of expert control output ZJ_Q1; S33: Compare the cumulative value of the expert control output ZJ_Q1 with the cumulative output threshold ZJ_Q_MAX1. If the cumulative value of the expert control output ZJ_Q1 is greater than the cumulative output threshold ZJ_Q_MAX1, then the controller output is executed; otherwise, return to S1.

6. The control method based on a feedforward adaptive periodic complex expert controller according to claim 4, characterized in that, The specific process of controller control output in S3 that satisfies the relevant conditions for output predictive control is as follows: For generators with state types H1, L1, H2, and L2, an adaptive periodic predictive control controller output is performed by combining the model between load increase / decrease and generator load, and only the first step is output.

7. A control method based on a feedforward adaptive periodic complex expert controller according to claim 4, 5, or 6, characterized in that, The control output cycle T1 <T2<T3<T4。

Images