Multi-parameter adaptive anti-surge control system for gas turbine unit

Abstract

This invention belongs to the field of gas turbine control technology. It discloses a multi-parameter adaptive anti-surge control system for gas turbine units, including a parameter acquisition device and an anti-surge master controller. By acquiring parameters in real time, it calculates the current surge margin and generates continuous fine-tuning commands. Based on a predefined nonlinear mapping model of the vent valve, the continuous commands are reconstructed into stepped control commands, effectively handling valve dead zones and step response characteristics. A strategy generation unit couples and evaluates the stepped commands with the unit's piping volume and gas column resonance frequency band, generating a comprehensive execution strategy that includes vent valve step actions and auxiliary fine-tuning compensation. The command distribution unit coordinates the vent valve's step action while simultaneously adjusting the compressor's variable guide vane opening and fuel flow, effectively mitigating pressure surges caused by valve actions. This invention improves the accuracy, stability, and safety of gas turbine anti-surge control, providing a guarantee for reliable unit operation under complex conditions.

Description

Technical Field

[0001] This invention relates to the field of gas turbine control technology, and more specifically, to a multi-parameter adaptive anti-surge control system for gas turbine units. Background Technology

[0002] Existing technologies in gas turbine anti-surge control face a technical contradiction: there is a mismatch between the fine-grained continuous adjustment pursued by multi-parameter adaptive control algorithms and the inherent nonlinear step characteristics of large-diameter anti-surge vent valves. When the unit approaches the surge boundary, advanced multi-parameter adaptive algorithms can accurately calculate the surge margin based on real-time data from multiple dimensions such as compressor outlet pressure, inlet temperature, and rotor speed, and output continuous fine-tuning commands to maintain an ideal safety margin. However, large-diameter fast vent valves in the field, due to the use of pneumatic-hydraulic linkage or high-thrust pneumatic actuators, generally exhibit significant mechanical dead zones and abrupt step response characteristics—the valve core may be completely unresponsive within a large opening range, but once the dead zone boundary is exceeded, a large-amplitude jump occurs instantaneously. This phenomenon of "the algorithm trying to fine-tune, but the valve is going through a bungee jump" manifests in actual operation as follows: When the surge margin gradually decreases, the algorithm continuously outputs progressive valve opening commands in an attempt to achieve "gentle" pressure relief. However, the valve remains motionless due to being in a dead zone, causing the surge margin to continue to deteriorate. As the algorithm continuously increases the control commands until it crosses the dead zone threshold, the valve suddenly and violently opens to a fixed position, causing an instantaneous drop in compressor outlet pressure. The surge margin then rapidly reverses and surges to an excessively safe state. The algorithm then issues a valve closing command, but then gets stuck in a dead zone loop in the valve closing direction. This periodic pressure pulsation not only seriously affects the unit's operational stability and combustion chamber conditions, but more seriously, when the frequency of frequent valve opening and closing is coupled with the natural frequency of the gas column in the gas turbine exhaust network, it can easily induce dangerous gas column resonance in the pipeline system, threatening the safe operation of the entire unit. Traditional control systems lack the ability to predict and avoid such complex coupling effects.

[0003] In view of this, the present invention proposes a multi-parameter adaptive anti-surge control system for gas turbine units to solve the above problems. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a multi-parameter adaptive anti-surge control system for gas turbine units, comprising: Parameter acquisition device and anti-surge main controller; The parameter acquisition device is used to acquire real-time operating multiple parameters of the gas turbine unit and transmit the real-time operating multiple parameters to the anti-surge main controller. The real-time operating multiple parameters include at least the compressor outlet pressure, inlet air temperature and rotor speed. The anti-surge controller includes a margin analysis unit, a nonlinear instruction reconstruction unit, a strategy generation unit, and an instruction distribution unit; The margin analysis unit is used to calculate the current surge margin based on real-time running multi-parameters, and generate continuous fine-tuning instructions based on the difference between the current surge margin and the preset target margin. The continuous fine-tuning instructions are the expected anti-surge vent valve opening adjustment amount. The nonlinear command reconstructing unit is used to reconstruct the continuous fine-tuning command based on the predefined nonlinear mapping model of the vent valve to obtain the stepped control command. The nonlinear mapping model of the vent valve is constructed based on the mechanical dead zone boundary and step response characteristics of the anti-surge vent valve. The stepped control command includes the target step position of the vent valve and the step margin compensation value. The strategy generation unit is used to couple and evaluate the stepped control commands with the gas turbine unit's pipeline volume and gas column resonance frequency band to determine the anti-surge execution strategy. The anti-surge execution strategy includes the vent valve step action command and the auxiliary fine-tuning compensation command. The auxiliary adjustment parameters controlled by the auxiliary fine-tuning compensation command include at least the compressor variable guide vane opening and fuel flow rate. The command distribution unit is used to control the anti-surge vent valve to perform the corresponding step action according to the anti-surge execution strategy, and simultaneously control the auxiliary adjustment parameters to perform the corresponding fine-tuning compensation action, so as to smooth the sudden change in compressor outlet pressure caused by the step action of the anti-surge vent valve.

[0005] The technical effects and advantages of the multi-parameter adaptive anti-surge control system for gas turbine units of this invention are as follows: This invention resolves the long-standing contradiction between the idealized output of algorithms and the realistic characteristics of actuators in the field of anti-surge control, achieving an organic unity between control precision and execution reliability. By mitigating the inherent dead zone and step characteristics of large-diameter vent valves, this invention transforms the theoretical requirement for continuous fine-tuning into practically executable precise control actions. This invention realizes the transformation of anti-surge control from single-valve pressure relief to multi-element coordinated regulation. By comprehensively considering the internal flow field regulation of the compressor and the response of the combustion system, it actively constructs a system buffer while performing pressure relief actions, reducing the impact of pressure surges on the stable operation of the unit. This invention also possesses the ability to intelligently predict and actively avoid the risk of pipeline gas column resonance, preventing resonance problems that may be caused by frequent valve operations from the source, providing a higher level of protection for the safe operation of gas turbines. Attached Figure Description

[0006] Figure 1 This is a schematic diagram of the multi-parameter adaptive anti-surge control system for gas turbine units according to the present invention. Detailed Implementation

[0007] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0008] Please see Figure 1 The embodiments of the present invention propose a multi-parameter adaptive anti-surge control system for gas turbine units, the system comprising: Parameter acquisition device and anti-surge main controller; The parameter acquisition device is used to acquire real-time operating multiple parameters of the gas turbine unit and transmit the real-time operating multiple parameters to the anti-surge main controller. The real-time operating multiple parameters include at least the compressor outlet pressure, inlet air temperature and rotor speed. Each parameter acquired in real time by the parameter acquisition device carries a precise acquisition timestamp, with millisecond-level accuracy to meet the requirements of rapid response control. Real-time operation involves multiple parameters, including at least compressor outlet pressure, inlet temperature, and rotor speed. Inlet pressure is also required as an auxiliary parameter when calculating the actual pressure ratio.

[0009] The anti-surge controller includes a margin analysis unit, a nonlinear instruction reconstruction unit, a strategy generation unit, and an instruction distribution unit; The margin analysis unit is used to calculate the current surge margin based on real-time running multi-parameters, and generate continuous fine-tuning instructions based on the difference between the current surge margin and the preset target margin. The continuous fine-tuning instructions are the expected anti-surge vent valve opening adjustment amount. The nonlinear command reconstructing unit is used to reconstruct the continuous fine-tuning command based on the predefined nonlinear mapping model of the vent valve to obtain the stepped control command. The nonlinear mapping model of the vent valve is constructed based on the mechanical dead zone boundary and step response characteristics of the anti-surge vent valve. The stepped control command includes the target step position of the vent valve and the step margin compensation value. The strategy generation unit is used to couple and evaluate the stepped control commands with the gas turbine unit's pipeline volume and gas column resonance frequency band to determine the anti-surge execution strategy. The anti-surge execution strategy includes the vent valve step action command and the auxiliary fine-tuning compensation command. The auxiliary adjustment parameters controlled by the auxiliary fine-tuning compensation command include at least the compressor variable guide vane opening and fuel flow rate. The command distribution unit is used to control the anti-surge vent valve to perform the corresponding step action according to the anti-surge execution strategy, and simultaneously control the auxiliary adjustment parameters to perform the corresponding fine-tuning compensation action, so as to smooth the sudden change in compressor outlet pressure caused by the step action of the anti-surge vent valve.

[0010] In this embodiment of the invention, the invention acquires real-time operating multiple parameters of the gas turbine unit and calculates the current surge margin to generate continuous fine-tuning commands. Based on the nonlinear mapping model of the vent valve, the continuous fine-tuning commands are reconstructed to obtain stepped control commands. The stepped control commands are coupled and evaluated with the pipeline volume and gas column resonance frequency band to avoid resonance. By using the fine-tuning compensation action of auxiliary adjustment parameters to smooth out the sudden pressure change at the compressor outlet, the invention effectively overcomes the technical problems in existing anti-surge control systems, such as the inability of continuous control commands to adapt to the mechanical dead zone characteristics of the vent valve, resulting in frequent valve jitter and ineffective execution, and the lack of a multi-parameter collaborative control mechanism due to the sole reliance on the pressure change and pipeline resonance caused by the vent valve action. This achieves the technical effects of accurate compensation for the nonlinear characteristics of the vent valve and reliable execution of step actions, smoothing out pressure change through collaborative control of auxiliary adjustment parameters, and avoiding harmful frequency excitation based on the pipeline resonance characteristics. Ultimately, the invention achieves the technical effects of stability and reliability in the anti-surge control of the gas turbine unit.

[0011] In a preferred embodiment of the present invention, the above-mentioned calculation of the current surge margin and generation of continuous fine-tuning instructions based on the difference between the current surge margin and the preset target margin includes: Step 1.1: Determine the current operating point of the compressor based on real-time multi-parameter operation and calculate the surge margin of the current operating point relative to the preset surge boundary line. Specifically, this includes deploying a multi-parameter acquisition device in the gas turbine unit control system. This device collects key operating parameters of the compressor in real time, including compressor outlet pressure, inlet temperature, and rotor speed. Each parameter carries a precise acquisition timestamp with millisecond-level accuracy to meet the requirements of rapid response control. After receiving the real-time multi-parameter operation, the compressor outlet pressure value is extracted, the inlet flow rate is extracted and temperature-corrected to obtain the converted flow rate, and the pressure ratio (compressor outlet pressure / inlet pressure) is calculated as the vertical axis parameter, and the converted flow rate as the horizontal axis parameter. The coordinate position of the current operating point is determined on the compressor characteristic graph. This operating point reflects the specific location of the compressor's current operating state on the flow-pressure plane. A pre-defined surge boundary line is marked on the compressor characteristic curve. This boundary line is drawn based on the surge characteristic curve provided by the compressor manufacturer, combined with the surge critical point at different rotor speeds. This boundary line represents the critical position where the compressor enters the unstable surge operating region. The flow margin between the current operating point and the pre-defined surge boundary line in the direction of the constant speed line is calculated. This flow margin serves as the basic metric for surge margin. The specific calculation formula is: Surge margin = (Surge boundary flow - Current operating point flow) / Surge boundary flow × 100%. Further considering the influence of rotor speed, normalization is used to unify the surge margin at different speeds to comparable dimensions. Finally, the surge margin value of the current operating point relative to the pre-defined surge boundary line is output. This value is usually expressed as a percentage; the larger the value, the farther the operating point is from the surge boundary and the safer the operation.

[0012] Step 1.2: Calculate the difference between the current surge margin and the preset target margin to obtain the margin deviation. Specifically, this includes: Based on the safety operation specifications and control strategy requirements of the gas turbine unit, a preset target margin value is pre-set, for example, 15%. This means the compressor should maintain a safety margin of at least 15% between the current operating point and the surge boundary line to ensure it does not enter the surge region during load fluctuations or environmental changes. After receiving the calculated current surge margin value, calculate the difference between it and the preset target margin: Margin Deviation = Current Surge Margin - Preset Target Margin. When the margin deviation is positive, it indicates that the current surge margin is greater than the target margin, the compressor operation is conservative, and there is an efficiency loss due to excessive opening of the anti-surge vent valve. When the margin deviation is negative, it indicates that the current surge margin is less than the target margin, the compressor operating point is too close to the surge boundary, there is a risk of surge, and the opening of the anti-surge vent valve needs to be increased to increase the margin. When the margin deviation is close to zero, it indicates that the current surge margin is basically consistent with the target margin, and the compressor is operating at an ideal balance point between safety and efficiency. The margin deviation, as an error signal of the control system, provides a precise control basis for the generation of subsequent continuous fine-tuning commands.

[0013] Step 1.3 involves calculating the margin deviation using an incremental adaptive control algorithm and outputting continuous fine-tuning commands. The range of these commands is a continuous valve opening percentage interval. Specifically, this includes: processing the margin deviation using an incremental adaptive control algorithm, which comprises proportional, integral, and derivative control loops. This algorithm generates precise control outputs based on the magnitude, trend, and historical accumulation of the deviation. The proportional loop determines the control gain based on the absolute value of the current margin deviation; a larger deviation results in a stronger control response. The integral loop accumulates historical deviations, eliminating steady-state errors and ensuring that the surge margin accurately converges to the target value during long-term operation. The derivative loop predicts the trend of deviation changes and adjusts in advance to improve control response speed. The adaptive mechanism dynamically adjusts control parameters according to the real-time operating conditions of the gas turbine unit: reducing control gain under low-load conditions to avoid overshoot, and increasing response speed under high-load, rapidly changing conditions to ensure safety. The algorithm outputs continuous fine-tuning instructions, expressed as a percentage of the anti-surge vent valve opening, ranging from 0% to 100%. For example, outputting "target opening 45.3%" indicates the desired adjustment of the anti-surge vent valve to 45.3% to achieve the target surge margin. These continuous fine-tuning instructions are smooth and continuous, allowing for precise expression of control requirements. However, due to the nonlinear characteristics of the vent valve, these continuous instructions cannot be executed directly and require subsequent nonlinear reconstruction processing.

[0014] In this embodiment of the invention, by employing a technique that uses real-time multi-parameter operation to determine the current operating point of the compressor and calculate its surge margin relative to a preset surge boundary line, then performing a difference calculation between the current surge margin and the preset target margin to obtain the margin deviation, and finally processing the margin deviation using an incremental adaptive control algorithm to output continuous valve opening percentage commands, the technical problems of existing surge control systems—such as using fixed threshold judgments, the inability to accurately quantify surge margin leading to insufficient control accuracy, and the lack of an adaptive control mechanism to cope with changes in control requirements under different operating conditions—are overcome. This achieves the technical effect of accurately quantifying and tracking the surge margin in real time, dynamically adjusting control parameters according to changes in operating conditions through an adaptive control algorithm, providing precise and continuous control commands for the anti-surge vent valve, and effectively improving the accuracy and adaptability of surge control.

[0015] In a preferred embodiment of the present invention, the above-mentioned continuous fine-tuning command is reconstructed based on a predefined nonlinear mapping model of the vent valve to obtain a stepped control command, including: Step 2.1: Obtain the current valve position feedback value of the anti-surge vent valve and compare the target opening value corresponding to the continuous fine-tuning command with the dead zone boundary and step transition point of the current valve position interval in the vent valve nonlinear mapping model. Specifically, this includes: acquiring the current valve position feedback value of the anti-surge vent valve in real time through a valve position sensor. This feedback value accurately reflects the current actual position of the vent valve actuator. For example, a current valve position feedback value of 32.5% indicates that the vent valve is at a 32.5% opening position. Read the continuous fine-tuning command generated in the previous step and extract its corresponding target opening value. For example, a target opening value of 38.7% indicates that the vent valve is expected to be adjusted to a 38.7% opening. Query the valve position interval where the current valve position feedback value of 32.5% is located from the predefined vent valve nonlinear mapping model. The vent valve nonlinear mapping model divides the 0-100% valve position range into multiple basic valve position intervals, each interval having independently calibrated dead zone boundaries and step response characteristics. Assuming the valve position falls within interval 2 (20%-40%) at 32.5%, the dead zone boundary parameters corresponding to this interval are read. The dead zone width for this interval is 2.5%, so the lower limit of the dead zone is the current valve position - 1.25%, i.e., 31.25%, and the upper limit is the current valve position + 1.25%, i.e., 33.75%. Simultaneously, the step transition point for this interval is read; for example, if the step transition point is the current valve position + 3.0%, i.e., 35.5%, it indicates that the valve will perform a step action when the target opening value exceeds 35.5%. The target opening value of 38.7% is compared numerically with both the dead zone boundary (30.5%-35.0%) and the step transition point of 35.5% to determine the relative relationship between the target opening value and the current valve position, providing a basis for subsequent instruction reconfiguration strategies.

[0016] Step 2.2: If the target opening value falls within the dead zone corresponding to the current valve position, the target step position of the vent valve is locked to the current valve position feedback value. The difference between the target opening value and the upper limit of the dead zone boundary is used as the step margin compensation value. Specifically, when the comparison result shows that the target opening value falls within the dead zone, for example, if the target opening value is 33.8% within the dead zone range of 30.5%-35.0%, it is determined that the adjustment of the target opening value has failed to break through the mechanical dead zone limit of the vent valve. Directly sending the opening command will cause the valve to fail to produce effective action or produce mechanical vibration. To avoid ineffective control and mechanical wear, the command reconstruction unit locks the target step position of the vent valve to the current valve position feedback value of 32.5%, that is, it does not send a step action command to the vent valve and maintains the current valve position. At the same time, the opening difference between the target opening value and the upper limit of the dead zone boundary is calculated: Step margin compensation value = Target opening value - Upper limit of dead zone boundary = 33.8% - 35.0% = -1.2%. This negative value indicates that the target opening value has not yet reached the upper limit of the dead zone, and there is a 1.2% opening adjustment requirement that cannot be achieved through the vent valve. This step margin compensation value is extracted as the remaining control requirement and will be compensated in the future through auxiliary adjustment parameters (compressor variable guide vane opening and fuel flow rate) to achieve the overall surge margin control target and avoid the control blind zone caused by the dead zone characteristics of the vent valve.

[0017] Step 2.3: If the target opening value exceeds the upper limit of the dead zone boundary, the actual landing point opening of the anti-surge vent valve after exceeding the dead zone is determined based on the step response characteristics as the target step position of the vent valve. The opening difference between the actual landing point opening and the target opening value is calculated as the step margin compensation value. Specifically, when the comparison result shows that the target opening value exceeds the upper limit of the dead zone boundary, for example, the target opening value is 38.7%, exceeding the upper limit of the dead zone of 35.0%, it is determined that the adjustment amount is sufficient to break through the mechanical dead zone limit, and the vent valve will generate a step action. According to the step response characteristics calibrated in the current valve position range in the nonlinear mapping model of the vent valve, this characteristic records the actual action behavior of the vent valve after receiving the command exceeding the dead zone, including the step overshoot and the final stable position. For example, the step response characteristics calibrated in the model show that when the command continues to increase from the upper limit of the dead zone (35.0%), the valve will first jump to 35.0% + step overshoot 5.5% = 40.5%, then fall back and stabilize at the actual landing point opening of 39.2%. This actual landing point opening of 39.2% serves as the target step position for the vent valve, and the command distribution unit will send a control command to the anti-surge vent valve to step to an opening of 39.2%. Since there is a difference between the actual landing point opening of 39.2% and the original target opening value of 38.7%, the opening difference is calculated as follows: Step margin compensation value = Target opening value - Actual landing point opening = 38.7% - 39.2% = -0.5%. This negative value indicates that the actual landing point opening of the valve exceeds the original control requirement by 0.5%, indicating overshoot. The step margin compensation value also needs to be compensated in reverse through auxiliary adjustment parameters to ensure that the final surge margin accurately reaches the target value and eliminates the control error introduced by the nonlinear characteristics of the vent valve step response.

[0018] In this embodiment of the invention, by acquiring the current valve position feedback value of the anti-surge venting valve and comparing the target opening value of the continuous fine-tuning command with the dead zone boundary and step transition point in the nonlinear mapping model of the venting valve, locking the current valve position when the target opening value falls into the dead zone interval and extracting the difference with the upper limit of the dead zone as a compensation value, and determining the actual landing point opening based on the step response characteristics and calculating the difference with the target value as a compensation value, the technical problems of the existing valve control system, such as the valve frequently jittering or ineffective action caused by the direct sending of continuous control commands, and the lack of compensation for the valve's nonlinear characteristics leading to a decrease in control accuracy, are overcome. This achieves the technical effect of accurately identifying the dead zone interval and step transition characteristics based on the nonlinear mapping model of the venting valve, avoiding ineffective control and mechanical jitter through step-based command reconstruction, and extracting the step margin compensation value to provide accurate compensation basis for auxiliary adjustment parameters, effectively improving the reliability and accuracy of anti-surge control.

[0019] In a preferred embodiment of the present invention, the above-mentioned coupling evaluation of the stepped control command with the gas turbine unit's pipeline volume and gas column resonant frequency band to determine the anti-surge execution strategy includes: Step 3.1: Based on the frequency of step actions of the anti-surge vent valve within a historical time period, determine the estimated frequency of step actions. Specifically, this includes: setting a step action history record module in the anti-surge main controller. This module continuously records the occurrence time, step amplitude, and step direction of each step action of the anti-surge vent valve, forming a complete step action time series data. A preset historical time period is selected as the statistical window, such as the most recent 60-second time window. The occurrence time records of all anti-surge vent valve step actions within this time period are extracted. The total frequency of step actions within the historical time period is calculated; for example, 6 step actions occurred within 60 seconds. The average time interval of the step actions is calculated: Average time interval = Historical time period length / Step action frequency = 60 seconds / 6 times = 10 seconds / action. Based on the average time interval, the estimated frequency of step actions is determined: Estimated frequency of step actions = 1 / Average time interval = 1 / 10 seconds = 0.1Hz. To further consider the temporal distribution characteristics of step movements, a sliding window statistical method is used to analyze whether there is a periodic clustering phenomenon in step movements. If periodic characteristics are detected, the dominant period is extracted as a correction parameter for the predicted frequency of step movements, ensuring that the frequency prediction can accurately reflect the actual occurrence pattern of step movements.

[0020] Step 3.2 compares the estimated frequency of the step action with the resonant frequency band of the gas column in the gas turbine exhaust network. Specifically, this involves pre-determining the resonant frequency band of the gas column in the exhaust network based on the geometry, pipe material, and gas medium characteristics through fluid dynamics calculations or experimental testing. The gas column resonant frequency band refers to the frequency range within which the gas column in the network exhibits a resonant amplification effect under specific frequency excitation; for example, the gas column resonant frequency band of a certain gas turbine exhaust network is 8-12Hz. The calculated estimated frequency of the step action (0.083Hz) is compared with the gas column resonant frequency band of 8-12Hz to determine whether the estimated frequency falls within the gas column resonant frequency band. If the estimated frequency of the step action is significantly lower or higher than the gas column resonant frequency band, it is determined that the step action will not induce gas column resonance in the network, and the step control strategy can be executed normally. If the estimated frequency of the step action falls into or is close to the air column resonance frequency band, it is determined that there is a risk of inducing pipeline resonance. Frequent step actions may induce air column resonance in the exhaust pipeline network, leading to drastic fluctuations in pipeline pressure, noise and vibration, and even fatigue damage to the pipeline structure. In the current example, the estimated frequency of the step action at 0.083Hz is far below the lower limit of the air column resonance frequency band of 8-12Hz, and it is determined that there is no risk of resonance. However, in high-frequency operating scenarios, the estimated frequency of the step action may increase significantly, requiring real-time monitoring and evaluation.

[0021] Step 3.3: If the estimated frequency of the step action falls within the gas column resonance frequency band, the target step position of the vent valve is attenuated and suppressed. The reduced pressure relief is then transferred to the auxiliary fine-tuning compensation command to form an anti-surge execution strategy to avoid resonance. Specifically, this includes: when the estimated frequency of the step action is detected to fall within the gas column resonance frequency band, the resonance avoidance mechanism is activated. First, the target step position of the vent valve is attenuated and suppressed. An attenuation coefficient α is designed, with a value ranging from 0 to 1. Its value is determined based on the proximity of the estimated frequency of the step action to the center frequency of the gas column resonance frequency band: when the estimated frequency of the step action is exactly at the center of the resonance frequency band, the attenuation coefficient α takes the minimum value (e.g., 0.3), indicating a significant attenuation of the step amplitude; when the estimated frequency of the step action is at the edge of the resonance frequency band, the attenuation coefficient α takes a larger value (e.g., 0.7), indicating moderate attenuation. The attenuated target step position of the vent valve is calculated as follows: Attenuated step position = current valve position + (target step position of vent valve - current valve position) × α. For example, if the original target step position of the vent valve is 45%, the current valve position is 35%, and the attenuation coefficient α = 0.5, then the attenuated step position = 35% + (45% - 35%) × 0.5 = 40%. Attenuation suppression reduces the actual step amplitude of the vent valve, and correspondingly reduces the pressure relief. The reduction in pressure relief is calculated as follows: Reduction in pressure relief = (Original target step position - Attenuated step position) × Pressure relief per unit opening = (45% - 40%) × Pressure relief per unit opening. To ensure that the overall surge margin control target is not affected, the reduction in pressure relief is converted into the adjustment requirement of auxiliary adjustment parameters and transferred to the auxiliary fine-tuning compensation command: increasing the compressor variable guide vane closing angle to reduce intake airflow, or reducing fuel flow to lower combustion chamber pressure, thereby achieving part of the pressure relief function originally undertaken by the vent valve through auxiliary adjustment parameters. The resulting anti-surge execution strategy includes: a damped vent valve step action command (stepping to 40% instead of 45%) and an enhanced auxiliary fine-tuning compensation command. This strategy achieves the surge margin control target while effectively avoiding the risk of air column resonance.

[0022] In this embodiment of the invention, the method of determining the estimated frequency of step action based on the frequency of step action of the anti-surge vent valve within a historical period, comparing the estimated frequency with the resonant frequency band of the gas column in the gas turbine exhaust network, and attenuating and suppressing the target step position of the vent valve when the estimated frequency falls into the resonant frequency band and transferring the reduced pressure relief to the auxiliary fine-tuning compensation command, overcomes the technical problems in existing anti-surge control systems where frequent step actions of the vent valve induce resonance in the gas column of the exhaust network, leading to pipeline pressure fluctuations and structural vibrations, and the lack of a resonance avoidance mechanism makes it impossible to achieve a balance between control performance and system safety. This achieves the technical effect of intelligent resonance avoidance based on the estimated step action frequency and the resonant characteristics of the pipeline network, and a synergistic strategy of step amplitude attenuation and auxiliary parameter compensation, effectively protecting the structural safety of the exhaust network while ensuring surge control, and improving the overall reliability of the gas turbine anti-surge system.

[0023] In a preferred embodiment of the present invention, the above-mentioned control of the anti-surge venting valve to perform a corresponding step action according to the anti-surge execution strategy, and simultaneous control of the auxiliary adjustment parameters to perform a corresponding fine-tuning compensation action, includes: Step 4.1, based on the auxiliary fine-tuning compensation command, converts the step margin compensation value into the first angle adjustment of the compressor variable guide vane and the second flow rate adjustment of the fuel flow. Specifically, this includes: receiving the step margin compensation value generated in the preceding steps. This compensation value represents the control quantity that needs to be compensated by the auxiliary adjustment parameters in the form of an equivalent opening percentage. For example, a step margin compensation value of -1.5% indicates that a 1.5% opening pressure relief effect needs to be compensated. A mapping relationship model between the step margin compensation value and the auxiliary adjustment parameters is established. This model is pre-calibrated based on the compressor characteristic curve and the combustion chamber thermodynamic characteristics. For the compressor variable guide vane, the influence coefficient of the unit angle adjustment on the surge margin is calibrated. For example, closing the variable guide vane by 1° can produce a margin gain effect equivalent to an increase of 0.3% in the vent valve opening. For the fuel flow, the influence coefficient of the unit flow rate adjustment on the surge margin is calibrated. For example, reducing the fuel flow by 10 kg / h can produce a margin gain effect equivalent to an increase of 0.2% in the vent valve opening, i.e., the influence coefficient is 0.02% / (kg / h). Based on the mapping model, the step margin compensation value is converted into the adjustment amounts of two auxiliary adjustment parameters: the first angle adjustment amount = step margin compensation value / variable guide vane influence coefficient = -1.5% / 0.3% = -5°, a negative value indicates that the variable guide vane needs to be closed by 5°; the second flow rate adjustment amount = step margin compensation value / fuel flow rate influence coefficient = -1.5% / 0.2% = -7.5 kg / h, a negative value indicates that the fuel flow rate needs to be reduced by 7.5 kg / h. During the conversion process, the physical constraints of the auxiliary adjustment parameters are considered, such as the variable guide vane angle range and fuel flow rate adjustment rate limits, to ensure that the conversion result is within the operable range of the actuator.

[0024] Step 4.2: Within the pre-set time frame before the anti-surge vent valve's step action, simultaneously adjust the compressor's variable guide vane opening according to the first angle adjustment amount and adjust the fuel flow rate according to the second flow rate adjustment amount to pre-build a flow and pressure buffer within the compressor. Specifically, this includes setting a pre-set time parameter, which is determined based on the response delay time of the auxiliary adjustment parameters. For example, the mechanical response delay of the compressor's variable guide vane is approximately 0.8 seconds, and the combustion response delay of the fuel flow rate adjustment is approximately 1.5 seconds. Taking all factors into consideration, the pre-set time is set to 1.8 seconds to ensure that the auxiliary adjustment parameters can effectively function before the vent valve's step action occurs. 1.8 seconds before the planned step action of the anti-surge vent valve, the command distribution unit simultaneously sends an adjustment command to the compressor's variable guide vane actuator: adjust the variable guide vane opening by the first angle adjustment amount -5°, i.e., close the variable guide vane by 5°, reducing the compressor's intake flow rate. Simultaneously, a regulation command is sent to the fuel control system: adjust the fuel flow rate according to the second flow rate regulation amount of -7.5 kg / h, i.e., reduce the fuel supply by 7.5 kg / h, thereby reducing the thermal power output of the combustion chamber. The closure of the variable guide vanes reduces the compressor intake flow rate, shifting the operating point to the left on the characteristic graph, away from the surge boundary; the reduced fuel flow rate leads to a decrease in combustion chamber pressure, reducing the back pressure effect on the compressor and further increasing the surge margin. These auxiliary regulation actions pre-establish flow and pressure buffers within the compressor, improving the surge margin even before the vent valve performs a step action, creating safer initial conditions for the subsequent vent valve step action. When the vent valve performs a step action 1.2 seconds later, because the auxiliary regulation parameters have already taken effect, the magnitude of the compressor outlet pressure change caused by the vent valve step action is significantly reduced. The auxiliary regulation parameters and the vent valve action form a timing synergy effect, effectively smoothing out pressure changes, avoiding drastic jumps in the compressor operating point on the characteristic graph, and improving the smoothness of the control process.

[0025] In this embodiment of the invention, because the step margin compensation value is converted into the first angle adjustment amount of the compressor variable guide vane and the second flow rate adjustment amount of the fuel flow based on the auxiliary fine-tuning compensation command, and the auxiliary parameters are adjusted in advance within the preset time before the anti-surge vent valve executes the step action to build a buffer for the flow and pressure inside the compressor, the technical problem of the existing anti-surge control system, which relies solely on the vent valve action and causes drastic changes in compressor outlet pressure and lacks a multi-parameter collaborative control mechanism to suppress pressure fluctuations, is overcome. Thus, the technical effect of achieving a smooth pressure transition through the timing coordination of auxiliary adjustment parameters and vent valve action, and building a buffer inside the compressor in advance based on the pre-adjustment strategy, effectively reducing the pressure change amplitude caused by the step action of the vent valve, and improving the stability of the anti-surge control process and the operating stability of the compressor, is achieved.

[0026] In a preferred embodiment of the present invention, the above-mentioned conversion of the step margin compensation value into a first angle adjustment amount of the compressor variable guide vane and a second flow rate adjustment amount of the fuel flow based on the auxiliary fine-tuning compensation command includes: Step 4.3 involves obtaining the first response delay time of the compressor variable guide vane and the second response delay time of the fuel flow. Specifically, this includes obtaining the response characteristic parameters of the compressor variable guide vane actuator through experimental testing or historical operating data analysis. The first response delay time is defined as the time required from the variable guide vane receiving the adjustment command to the variable guide vane angle actually starting to change and producing a measurable flow effect. This time includes the mechanical response delay of the actuator, the pressure build-up time of the pneumatic system, and the flow propagation delay, etc. For example, the tested first response delay time is 0.8 seconds. Similarly, the response characteristic parameters of the fuel flow control system are obtained through experimental testing or data analysis. The second response delay time is defined as the time required from the fuel control system receiving the flow adjustment command to the combustion chamber pressure change caused by the fuel flow change propagating to the compressor outlet and producing a measurable effect. This time includes the fuel valve response delay, the thermodynamic delay of the combustion process, and the pressure wave propagation delay, etc. For example, the tested second response delay time is 1.5 seconds. The response delay time parameters reflect the difference in response speed between different auxiliary adjustment parameters: a shorter first response delay time indicates that the variable guide vane adjustment is effective quickly, while a longer second response delay time indicates that the fuel flow adjustment is effective slowly. These parameters provide an important basis for subsequent dynamic weight allocation calculations.

[0027] Step 4.4: Based on the ratio of the reciprocal of the first response delay time and the second response delay time, determine the dynamic allocation weights for the compressor variable guide vanes and fuel flow. Specifically, this includes: using the reciprocal of the response delay time as a measure of response speed, calculating the response speed of the compressor variable guide vanes: Response speed 1 = 1 / first response delay time = 1 / 0.8 seconds = 1.25 times / second. Calculating the response speed of the fuel flow: Response speed 2 = 1 / second response delay time = 1 / 1.5 seconds ≈ 0.67 times / second. Determining the dynamic allocation weights based on the ratio of the two response speeds, and using normalization to ensure the sum of the weights is 1: Dynamic allocation weight W1 for the compressor variable guide vanes = Response speed 1 / (Response speed 1 + Response speed 2) = 1.25 / (1.25 + 0.67) ≈ 0.65; Dynamic allocation weight W2 for the fuel flow = Response speed 2 / (Response speed 1 + Response speed 2) = 0.67 / (1.25 + 0.67) ≈ 0.35. The physical significance of dynamically allocating weights lies in the following: auxiliary control parameters with faster response speeds are assigned higher weights, undertaking more compensation tasks and fully leveraging their rapid response advantage; auxiliary control parameters with slower response speeds are assigned lower weights, undertaking less compensation tasks and avoiding control lag caused by response delays. This weight allocation strategy ensures that auxiliary fine-tuning compensation can achieve maximum effect in the shortest time, improving the overall response speed of the control system.

[0028] Step 4.5: Decompose the step margin compensation value according to the dynamic allocation weight, and map the decomposed compensation amount to the first angle adjustment amount and the second flow rate adjustment amount respectively. Specifically, this includes: receiving the step margin compensation value and the calculated dynamic allocation weight, and performing compensation amount decomposition calculation. The compensation amount allocated to the compressor variable guide vane: Compensation amount 1 = Step margin compensation value × W1 = -1.5% × 0.65 = -0.975%. The compensation amount allocated to the fuel flow rate: Compensation amount 2 = Step margin compensation value × W2 = -1.5% × 0.35 = -0.525%. Based on the pre-calibrated mapping relationship model, the decomposed compensation amount is converted into actual adjustment units. For the compressor variable guide vane, compensation amount 1 = -0.975% is converted to the first angle adjustment amount: First angle adjustment amount = Compensation amount 1 / Variable guide vane influence coefficient = -0.975% / 0.3% = -3.25°, indicating that the variable guide vane needs to be closed by 3.25°. For fuel flow, the compensation amount 2 = -0.525% is converted into the second flow regulation amount: Second flow regulation amount = Compensation amount 2 / Fuel flow influence coefficient = -0.525% / 0.02% = -26.25 kg / h, indicating that the fuel flow needs to be reduced by 26.25 kg / h. The splitting and mapping process ensures that the total compensation effect equals the control requirement of the original step margin compensation value, and at the same time optimizes the allocation ratio according to the response characteristics of each auxiliary regulation parameter, realizing an organic combination of fast response and precise control.

[0029] In this embodiment of the invention, by acquiring the first response delay time of the compressor variable guide vane and the second response delay time of the fuel flow, determining the dynamic allocation weight based on the reciprocal ratio of the two, and splitting the step margin compensation value according to the dynamic allocation weight and mapping it to the first angle adjustment amount and the second flow adjustment amount respectively, the technical problems of existing multi-parameter control systems, such as using fixed weight allocation, being unable to optimize the allocation ratio according to the response characteristics of each adjustment parameter leading to insufficient control response speed, and lacking a dynamic weight adjustment mechanism to fully utilize the advantages of fast response parameters, are overcome. Thus, the invention achieves the technical effect of intelligent dynamic allocation of auxiliary adjustment parameters based on the difference in response delay time, prioritizing the use of parameters with fast response speed to improve the overall response speed of the control system, while ensuring that the total compensation effect meets the control requirements, effectively improving the response speed and control accuracy of anti-surge auxiliary adjustment.

[0030] In a preferred embodiment of the present invention, after controlling the anti-surge venting valve to perform the corresponding step action according to the anti-surge execution strategy, the method further includes: Step 5.1: Collect the actual change in compressor outlet pressure after the anti-surge vent valve's step action. Specifically, this includes: collecting pressure data in real-time using the compressor outlet pressure sensor after the anti-surge vent valve performs its step action. Record the baseline compressor outlet pressure value before the step action, for example, 1.85 MPa before the step. Continuously monitor the pressure change process after the step action, waiting for the pressure to reach a new stable state, and record the stable compressor outlet pressure value, for example, 1.80 MPa after the step. Calculate the actual change in compressor outlet pressure: Actual change = Stable pressure after step - Baseline pressure before step = 1.80 MPa - 1.85 MPa = -0.05 MPa. A negative value indicates a decrease in compressor outlet pressure, consistent with the expected effect of the anti-surge vent valve opening and pressure relief. The actual pressure change reflects the actual pressure relief effect of the vent valve's step action, providing real feedback data for subsequent model correction.

[0031] Step 5.2 compares the actual change in compressor outlet pressure with the theoretical change in pressure estimated based on the anti-surge execution strategy to obtain the pressure deviation. Specifically, this includes: when generating the anti-surge execution strategy, the theoretical pressure change is pre-calculated based on the nonlinear mapping model of the vent valve, the target step position of the vent valve, and the expected effects of the auxiliary adjustment parameters. For example, based on the target step position of the vent valve jumping from 35% to 40%, the theoretical calculation based on the pre-calibrated unit opening pressure relief coefficient (e.g., 0.008MPa / %) shows that this 5% opening increment will produce a pressure drop of 0.04MPa; simultaneously considering the auxiliary pressure relief effect of the auxiliary adjustment parameters (variable guide vane closed at 3.25°, fuel flow reduced by 2.625kg / h) of approximately 0.01MPa, the combined theoretical pressure change is -0.04MPa - 0.01MPa = -0.05MPa. The actual pressure change of -0.07 MPa is compared with the theoretical pressure change of -0.065 MPa to calculate the pressure deviation: Pressure deviation = Actual pressure change - Theoretical pressure change = -0.07 MPa - (-0.065 MPa) = -0.005 MPa. The absolute value of the pressure deviation, |0.005 MPa|, reflects the degree of difference between the actual pressure relief effect and the theoretical estimate. This difference may stem from factors such as parameter deviations in the nonlinear mapping model of the vent valve, changes in environmental conditions, or aging of the actuator.

[0032] Step 5.3: If the pressure deviation exceeds the preset tolerance range, the step transition point and step response characteristics in the nonlinear mapping model of the vent valve are corrected online based on the pressure deviation to update the actual dead zone and overshoot parameters of the anti-surge vent valve. Specifically, this includes: pre-setting a preset tolerance range for the pressure deviation, for example, ±0.01 MPa. This means that when the absolute value of the pressure deviation is less than 0.01 MPa, the model prediction is considered accurate and no correction is needed; when the absolute value of the pressure deviation is greater than or equal to 0.01 MPa, the model is determined to have a deviation and online correction is required. In the current example, the absolute value of the pressure deviation |0.005 MPa| is less than the upper limit of the tolerance range of 0.01 MPa, so the model prediction is considered to be within an acceptable range, and correction is not triggered temporarily. If, under other circumstances, the pressure deviation is -0.015 MPa, and the absolute value exceeds the tolerance range, the online correction mechanism is activated. The sign and magnitude of the pressure deviation are analyzed: a negative deviation indicates that the actual pressure relief effect is stronger than the theoretical prediction, possibly due to the actual opening degree of the vent valve exceeding expectations or improved pressure relief efficiency; a positive deviation indicates that the actual pressure relief effect is weaker than the theoretical prediction, possibly due to jamming of the vent valve or blockage of the pressure relief channel. Based on the pressure deviation, an adaptive learning algorithm is used to correct the parameters of the nonlinear mapping model of the vent valve. For the correction of the step transition point: if the actual pressure relief effect is stronger than expected, it indicates that the opening threshold at which the step occurs may be lower than the current model setting value, and the step transition point parameter is appropriately reduced; if the actual pressure relief effect is weaker than expected, the step transition point parameter is appropriately increased. For the correction of the step response characteristics: if there is a difference between the actual landing point opening and the expectation, the step overshoot parameter is updated; if the pressure relief effect after stabilization does not match the expectation, the unit opening pressure relief parameter is updated. The correction algorithm adopts an incremental update strategy to avoid overcorrection caused by a single deviation. First, the pressure deviation is converted into an equivalent opening deviation: Equivalent opening deviation = Pressure deviation / Unit opening pressure relief coefficient. For example, a pressure deviation of -0.015 MPa divided by a unit opening pressure relief coefficient of 0.008 MPa / % yields an equivalent opening deviation of -1.875%. Then, the parameters are updated: New parameter value = Old parameter value + Learning rate × Equivalent opening deviation × Correction coefficient. The learning rate (e.g., 0.1) controls the correction speed, and the correction coefficient is determined based on the current compressor operating conditions. The updated parameters of the vent valve nonlinear mapping model are stored in the model database. In subsequent control cycles, the corrected parameters are used to achieve adaptive optimization of the model, ensuring that the vent valve nonlinear mapping model can continuously and accurately reflect the actual operating characteristics of the anti-surge vent valve, thus improving control accuracy during long-term operation.

[0033] In this embodiment of the invention, the actual change in compressor outlet pressure after the step action of the anti-surge vent valve is collected, and the actual change is compared with the theoretical pressure change estimated based on the anti-surge execution strategy to obtain the pressure deviation. When the pressure deviation exceeds the preset tolerance range, the step transition point and step response characteristics in the nonlinear mapping model of the vent valve are corrected online based on the deviation. Therefore, this invention overcomes the technical problems of existing anti-surge control systems, such as the use of fixed parameters in the nonlinear mapping model of the vent valve, the inability to adapt to valve aging or changes in operating conditions leading to a gradual decrease in control accuracy, and the lack of a model adaptive correction mechanism that prevents long-term stable operation. Thus, it achieves the technical effect of online adaptive correction of the nonlinear mapping model of the vent valve based on actual operation feedback, continuous optimization of model parameters through a closed-loop learning mechanism, ensuring that the model can accurately reflect the real action characteristics of the anti-surge vent valve, and effectively improving the long-term operating accuracy and adaptability of the anti-surge control system.

[0034] In a preferred embodiment of the present invention, after reconstructing the continuous fine-tuning command based on a predefined nonlinear mapping model of the vent valve to obtain a stepped control command, the method further includes: Step 5.4 involves monitoring the actual valve position feedback value after the anti-surge vent valve performs a step action. Specifically, this includes: after the command distribution unit sends a step action command to the anti-surge vent valve, continuous monitoring of the actual valve position feedback value is initiated. A high-precision valve position sensor is used to collect the actual valve position feedback value of the anti-surge vent valve in real time at a fixed sampling frequency (e.g., 100Hz), forming a valve position change time series data. The moment the step action command is sent is recorded as the monitoring starting point; for example, a command to step to 40% opening is sent at t=0. The actual valve position feedback value is continuously collected at subsequent time points, for example, valve position is 35.2% at t=0.1 seconds, 36.8% at t=0.2 seconds, 38.1% at t=0.3 seconds, and 39.5% at t=0.5 seconds. Under normal circumstances, the actual valve position feedback value should gradually approach the target step position of 40% of the vent valve and stabilize within a reasonable time. The valve position change rate is recorded synchronously during the monitoring process to determine whether the valve's movement state is normal.

[0035] Step 5.5: If the actual valve position feedback value does not reach the target step position of the vent valve, and the stagnation time exceeds the preset jamming judgment threshold, the anti-surge vent valve is determined to be in a step jamming state. Specifically, this includes: comparing the actual valve position feedback value with the target step position of the vent valve in real time, and calculating the valve position tracking error: Valve position tracking error = Target step position of vent valve - Actual valve position feedback value. A valve position arrival judgment threshold is set; for example, when the absolute value of the valve position tracking error is less than 0.5%, the valve position is determined to be in place. If, during continuous monitoring, the valve position tracking error is consistently greater than the arrival judgment threshold, for example, the actual valve position feedback value stagnates at 38.1% and cannot continue moving towards the 40% target, the valve may have abnormal movement. Further analysis of the valve position change rate is performed; when the valve position change rate approaches zero within multiple consecutive sampling periods (e.g., the valve position change is less than 0.1% within one second), the valve is determined to be in a stagnant state. A preset jamming threshold is set, defined as the maximum allowable stall time. For example, setting it to 2 seconds means that if the valve cannot reach the target position within 2 seconds and remains stalled, it is judged as a jamming fault. When the actual stall time exceeds the preset jamming threshold, for example, if the valve stalls at the 38.1% position for 2.5 seconds without continuing to move, the anti-surge vent valve is judged to be in a step-jamming state. Jamming may be caused by mechanical wear, foreign object obstruction, actuator malfunction, or poor lubrication, causing the valve to fail to complete the expected step action, severely affecting the anti-surge control effect.

[0036] Step 5.6: In response to the step jamming state, expand the adjustment step size and amplitude of the auxiliary adjustment parameters in the auxiliary fine-tuning compensation command to replace the pressure relief function of the anti-surge vent valve with active adjustment of the compressor variable guide vanes and fuel flow. Specifically, this includes: when the anti-surge vent valve is detected to be in a step jamming state, immediately activate the fault response mechanism. Calculate the uncompleted pressure relief task of the vent valve: Uncompleted pressure relief amount = (vent valve target step position - actual valve position feedback value) × pressure relief amount per unit opening = (40% - 38.1%) × pressure relief amount per unit opening. This uncompleted pressure relief amount should have been completed by the vent valve, but due to valve jamming, it cannot be achieved and needs to be completely transferred to the auxiliary adjustment parameters. Expand the adjustment step size and amplitude of the auxiliary fine-tuning compensation command to fully allocate the uncompleted pressure relief amount to the compressor variable guide vanes and fuel flow. The expanded adjustment amount is calculated based on dynamically allocated weights: Expanded first angle adjustment amount = uncompleted pressure relief amount × W1 / variable guide vane influence coefficient; Expanded second flow adjustment amount = uncompleted pressure relief amount × W2 / fuel flow influence coefficient. For example, if the uncompleted pressure relief amount is 1.9%, W1 = 0.65, and the variable guide vane influence coefficient = 0.3%, then the expanded first angle adjustment amount = (1.9% × 0.65) / (0.3% / °) ≈ 4.1°, indicating that the variable guide vane needs to be closed an additional 4.1°. The command distribution unit immediately sends the expanded adjustment command to the compressor variable guide vane and fuel control system, completely replacing the pressure relief function of the anti-surge vent valve through the active adjustment of auxiliary adjustment parameters, ensuring that the surge margin control target can still be achieved. At the same time, the system records the vent valve jamming fault event, triggers equipment maintenance early warning, and notifies maintenance personnel to inspect the anti-surge vent valve to prevent the fault from escalating. This fault response mechanism demonstrates the fault tolerance capability of multi-parameter collaborative control. Even if the main actuator (vent valve) fails, the anti-surge protection function can still be maintained by adjusting auxiliary parameters, significantly improving system reliability.

[0037] In this embodiment of the invention, by monitoring the actual valve position feedback value after the anti-surge vent valve performs a step action, when the actual valve position fails to reach the target position and the stagnation time exceeds the preset jamming judgment threshold, the valve is determined to be in a step jamming state. In response to the jamming state, the adjustment step size and amplitude of the auxiliary adjustment parameter in the auxiliary fine-tuning compensation command are expanded to replace the vent valve's pressure relief function. Therefore, this overcomes the technical problems in the existing anti-surge control system where the vent valve cannot complete the pressure relief task when a jamming fault occurs, and the lack of a fault response mechanism leads to the failure of the surge protection function. Thus, it achieves the technical effect of rapid diagnosis of vent valve jamming fault based on valve position monitoring, and fault-tolerant replacement of the pressure relief function by expanding the adjustment amplitude through auxiliary adjustment parameters. This ensures that anti-surge protection can still be maintained even if the main actuator fails, effectively improving the fault tolerance capability and operational reliability of the anti-surge control system.

[0038] In a preferred embodiment of the present invention, nonlinear processing logic for the valve closing direction is further included: Step 6.1: When the current surge margin indicates the unit is in an over-margin state, generate continuous fine-tuning commands for the valve closing direction. Specifically, during the anti-surge control process, continuously monitor the relationship between the current surge margin and the preset target margin. When the current surge margin is significantly greater than the preset target margin, for example, the current surge margin is 25% while the preset target margin is 15%, the margin deviation is +10%, and the unit is determined to be in an over-margin state. An over-margin state indicates that the compressor operating point is too far from the surge boundary, and the anti-surge vent valve opening is too large, leading to unnecessary pressure relief losses and reducing the operating efficiency of the gas turbine unit. To improve operating efficiency, it is necessary to reduce the opening of the anti-surge vent valve to bring the surge margin back to near the preset target margin. The incremental adaptive control algorithm generates continuous fine-tuning commands in the valve-closing direction based on positive margin deviations. These commands are negative, indicating a desired reduction in the vent valve opening. For example, generating a continuous fine-tuning command of -5.2% in the valve-closing direction means a desired 5.2% closure of the anti-surge vent valve from its current opening. These continuous fine-tuning commands in the valve-closing direction also need to be processed by a nonlinear mapping model to accommodate the dead-zone characteristics of the vent valve in the closing direction.

[0039] Step 6.2: Based on the nonlinear mapping model of the vent valve, if the target opening value corresponding to the continuous fine-tuning command in the valve closing direction falls into the valve closing dead zone interval corresponding to the current valve position, the generation of the valve closing step command is temporarily suspended, and the current valve position is maintained. Specifically, this includes reading the valve closing dead zone interval parameter corresponding to the current valve position from the nonlinear mapping model of the vent valve. The valve closing dead zone interval and the valve opening dead zone interval may be asymmetrical because the mechanical friction characteristics of the valve differ in the closing and opening directions. For example, if the current valve position is 42%, the corresponding valve closing dead zone interval is 39.5%-42%, indicating that when the target opening value is higher than 39.5%, the valve will not close. Calculate the target opening value corresponding to the continuous fine-tuning command in the valve closing direction: Target opening value = Current valve position + Continuous fine-tuning command in the valve closing direction = 42% + (-5.2%) = 36.8%. Comparing the target opening value of 36.8% with the valve closing dead zone range of 39.5%-42%, it was found that the target opening value is lower than the lower limit of the valve closing dead zone of 39.5%, indicating that the command has exceeded the dead zone. However, if the continuous fine-tuning command in the valve closing direction is -1.8%, then the target opening value = 42% + (-1.8%) = 40.2%, which falls within the valve closing dead zone range of 39.5%-42%. When the target opening value falls within the valve closing dead zone range, it is determined that the valve closing command cannot exceed the mechanical dead zone, and sending it directly will result in invalid valve action or valve jitter. A conservative strategy is adopted: the generation of the valve closing step command is temporarily suspended, and the current valve position of 42% is maintained to avoid invalid control. However, the over-margin state still exists at this time, and other means are needed to reduce the surge margin.

[0040] Step 6.3: Simultaneously, adjust the auxiliary control parameters to actively increase the compressor load, thereby reducing the surge margin, until the continuous fine-tuning command in the valve closing direction crosses the valve closing dead zone boundary. Specifically, this includes: while maintaining the current valve position of the anti-surge vent valve, activating the active adjustment mechanism of the auxiliary control parameters to reduce the surge margin by increasing the compressor load. Adjustment strategies for reducing the surge margin include: increasing the compressor variable guide vane angle to increase the intake air flow, causing the compressor operating point to move to the right closer to the surge boundary; or increasing the fuel flow, increasing the combustion chamber outlet temperature and pressure, increasing turbine work, and thus increasing the compressor speed, causing the compressor operating point to move along the constant speed line towards the surge boundary. Calculate the required adjustment amount of the auxiliary control parameters to gradually reduce the surge margin from the current value of 25% to close to the target value of 15%. For example, increasing the variable guide vane angle by 3° and increasing the fuel flow by 5 kg / h reduces the surge margin to 18%. As the surge margin decreases, the margin deviation decreases to +3%. The incremental adaptive control algorithm generates a new continuous fine-tuning command for the valve closing direction, for example, -0.8%. The new target opening value is calculated as 42% + (-0.8%) = 41.2%, which still falls within the valve closing dead zone of 39.5%-42%. The current valve position is maintained, and the auxiliary control parameters are adjusted. This process continues until the surge margin further decreases to 16%, and the margin deviation decreases to +1%. At this point, the generated continuous fine-tuning command for the valve closing direction is -0.3%, and the target opening value is 42% + (-0.3%) = 41.7%, still within the dead zone. When the surge margin decreases to close to the target value of 15%, the margin deviation approaches zero. The system stops generating continuous fine-tuning commands for the valve closing direction and maintains the current state. This achieves the elimination of the excess margin state through active adjustment of auxiliary control parameters without triggering the vent valve's valve closing dead zone, thus improving operating efficiency and avoiding ineffective vent valve operation. In some cases, if the surge margin cannot be sufficiently reduced after adjusting the auxiliary adjustment parameters, the large margin deviation will result in a large amplitude of the continuous fine-tuning command for the valve closing direction. Ultimately, the target opening value will exceed the lower limit of the valve closing dead zone. At this time, the vent valve will generate a valve closing step action to complete the valve position adjustment.

[0041] In this embodiment of the invention, by employing a technique of generating continuous fine-tuning commands for the valve closing direction when the current surge margin indicates that the unit is in an over-margin state, and temporarily suspending the generation of valve closing step commands and maintaining the current valve position when the target opening value corresponding to the command falls into the valve closing dead zone, while simultaneously adjusting auxiliary adjustment parameters to actively increase the compressor load and reduce the surge margin, the technical problems of existing anti-surge control systems, such as valve closing direction control also being limited by dead zone, inability to effectively reduce the over-margin state leading to operating efficiency loss, and lack of a nonlinear processing mechanism for valve closing dead zone leading to frequent valve jitter or invalid action, are overcome. This achieves the technical effect of avoiding nonlinear dead zone in the valve closing direction, actively adjusting auxiliary adjustment parameters to replace the vent valve closing action, effectively eliminating the over-margin state while avoiding invalid valve action, and improving the operating efficiency of the gas turbine unit and the all-condition adaptability of the anti-surge control system.

[0042] In a preferred embodiment of the present invention, the above-mentioned nonlinear mapping model of the vent valve is pre-constructed in the following manner: Step 7.1: During the cold purging phase or low-load warm-up phase before gas turbine unit startup, send multiple tentative step opening commands to the anti-surge vent valve. Specifically, select the cold purging phase or low-load warm-up phase before gas turbine unit startup as the calibration time for the vent valve's nonlinear mapping model. During this phase, the unit is in an abnormal operating state, allowing for exploratory control experiments without affecting normal production. The cold purging phase refers to the phase before unit startup where an external air source (such as a starter motor driving the compressor or external compressed air) is used to purge the combustion chamber and exhaust pipes. At this time, the combustion chamber is not ignited, the compressor speed is low (usually not exceeding 30% of the rated speed), the system pressure is low, operational safety is high, and the tentative action of the vent valve will not affect unit operation. The low-load warm-up phase refers to the initial phase after unit ignition, where the temperature and load gradually increase. At this time, the load is low, and changes in operating conditions have little impact on the unit. During calibration, a series of tentative step-change commands were designed to cover the full opening range of the anti-surge venting valve, from 0% to 100%. The opening range was divided into several basic valve position intervals, such as 0-20%, 20-40%, 40-60%, 60-80%, and 80-100%. Within each basic valve position interval, a sequence of tentative step-change commands was designed. For example, in the 20-40% interval, the following commands were sent sequentially: after stabilizing the valve position at 25%, a command was sent to step to 28%, and the valve response was observed; after stabilization, a command was sent to step to 32%, and observation continued; this process was repeated to cover multiple step amplitudes within the interval. The design of the tentative step-change commands considered different step amplitudes (e.g., 1%, 3%, 5%, 10%) to comprehensively obtain the valve's response characteristics under different adjustment values.

[0043] Step 7.2 records the actual dead zone width and step overshoot of the anti-surge vent valve in different basic valve position ranges. Specifically, after sending each tentative step command, the response process of the anti-surge vent valve is recorded in real time using a high-precision valve position sensor. For dead zone width measurement: at a stable valve position (e.g., 25%), small incremental commands (e.g., +0.5%, +1.0%, +1.5%) are sent, and the actual valve position feedback value is observed to see if it changes. When the incremental command is less than a certain threshold, the actual valve position feedback value remains unchanged or the change is less than the measurement accuracy (e.g., 0.1%), indicating that the command falls into the dead zone; when the incremental command exceeds this threshold, the actual valve position feedback value changes significantly (the change is greater than 0.5%), indicating that the command breaks through the dead zone boundary. Record the minimum incremental command required to break through the dead zone as the single-sided dead zone width of the valve position. For example, if the single-sided dead zone width in the opening direction of the 25% valve position is measured to be 1.4%, then the total dead zone width of this valve position range is approximately 2.8% (assuming the dead zones in the opening and closing directions are basically symmetrical). For measuring step overshoot: send a step command to break through the dead zone (e.g., from 25% to 30%) and record the valve position response curve. Observe whether there is overshoot in the valve position, i.e., the valve position first exceeds the target value and then falls back to stabilize. Measure the difference between the peak overshoot and the target value as the step overshoot. For example, if the valve position jumps from 25% to 30%, first reaching 32.5%, then falling back and stabilizing at 30.8%, then the step overshoot is 32.5% - 30% = 2.5%, the actual landing point opening is 30.8%, and the actual landing point deviation is 30.8% - 30% = +0.8%, indicating that the final stable position exceeds the target opening by 0.8%. The above measurements were repeated in different basic valve position ranges to obtain the actual dead zone width and step overshoot data for the full opening range. The test data showed that the dead zone width and overshoot were not constant in different valve position ranges: the dead zone width in the low opening range (0-20%) may be smaller (e.g., 1.5%) because the initial friction of the valve is small; the dead zone width in the medium-high opening range (40-80%) may be larger (e.g., 3.5%) because the valve is subjected to greater aerodynamic forces; the overshoot may increase when the valve is close to the fully open position because the buffering effect of the limiting mechanism is different.

[0044] Step 7.3: Based on the actual dead zone width and step overshoot, construct a nonlinear mapping model of the vent valve indexed by the basic valve position intervals. This includes: organizing all data obtained from calibration tests and establishing a parameter database indexed by the basic valve position intervals. For each basic valve position interval, store the following key parameters: lower dead zone limit = current valve position - dead zone width / 2, upper dead zone limit = current valve position + dead zone width / 2, step transition point = upper dead zone limit + trigger threshold, step overshoot, actual landing point deviation, pressure relief per unit opening, etc. Using piecewise linear or nonlinear interpolation methods, construct a continuous nonlinear mapping model of the vent valve covering the full opening range. This model accepts the current valve position and target opening value as input and outputs stepped control commands (including the target step position and step margin compensation value of the vent valve). The model incorporates conditional judgment logic: based on the current valve position, it queries the corresponding basic valve position interval index, extracts the dead zone boundary and step response parameters of that interval; compares the target opening value with the dead zone boundary to determine whether a step action is required; if a step action is required, it predicts the stable valve position after the step based on the step overshoot and the actual landing point deviation; and calculates the step margin compensation value to quantify the prediction error. The constructed nonlinear mapping model of the vent valve is stored in the model database of the anti-surge master controller and is called in real time during actual operation to guide the generation of stepped control commands. The model also supports online correction, continuously optimizing parameters based on actual operation feedback to ensure long-term accuracy.

[0045] In this embodiment of the invention, by sending multiple tentative step opening commands to the anti-surge vent valve during the cold purging phase or low-load warm-up phase before the gas turbine unit starts up, recording the actual dead zone width and step overshoot under different basic valve position intervals, and constructing a nonlinear mapping model of the vent valve based on the measurement data with the basic valve position interval as the index, the technical problem of existing anti-surge control systems lacking accurate modeling of the nonlinear characteristics of the vent valve and using theoretical models or empirical parameters that cannot accurately reflect the actual valve characteristics is overcome. Thus, it achieves the technical effect of obtaining the real valve dead zone and overshoot characteristics through actual testing, constructing a refined nonlinear mapping model in intervals, providing an accurate model basis for the generation of step-based control commands, and effectively improving the anti-surge control system's ability to compensate for the nonlinear characteristics of the vent valve and the control accuracy.

[0046] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0047] It should be noted that all formulas in this manual are calculated by removing dimensions and taking their numerical values. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0048] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A multi-parameter adaptive anti-surge control system for gas turbine units, characterized in that, include: Parameter acquisition device and anti-surge main controller; The parameter acquisition device is used to acquire real-time operating multiple parameters of the gas turbine unit and transmit the real-time operating multiple parameters to the anti-surge main controller. The real-time operating multiple parameters include at least the compressor outlet pressure, inlet air temperature and rotor speed. The anti-surge controller includes a margin analysis unit, a nonlinear instruction reconstruction unit, a strategy generation unit, and an instruction distribution unit. The margin analysis unit is used to calculate the current surge margin based on the real-time running multi-parameters, and generate a continuous fine-tuning instruction based on the difference between the current surge margin and the preset target margin. The continuous fine-tuning instruction is the expected adjustment amount of the anti-surge vent valve opening. The nonlinear command reconstruction unit is used to reconstruct the continuous fine-tuning command based on a predefined nonlinear mapping model of the vent valve to obtain a stepped control command. The nonlinear mapping model of the vent valve is constructed based on the mechanical dead zone boundary and step response characteristics of the anti-surge vent valve. The stepped control command includes the target step position of the vent valve and the step margin compensation value. The strategy generation unit is used to couple and evaluate the stepped control command with the pipeline volume and gas column resonance frequency band of the gas turbine unit to determine the anti-surge execution strategy. The anti-surge execution strategy includes a vent valve step action command and an auxiliary fine-tuning compensation command. The auxiliary adjustment parameters controlled by the auxiliary fine-tuning compensation command include at least the compressor variable guide vane opening and fuel flow rate. The instruction distribution unit is used to control the anti-surge vent valve to perform a corresponding step action according to the anti-surge execution strategy, and simultaneously control the auxiliary adjustment parameters to perform a corresponding fine-tuning compensation action, so as to smooth out the compressor outlet pressure change caused by the step action of the anti-surge vent valve.

2. The gas turbine unit multi-parameter adaptive anti-surge control system as described in claim 1, characterized in that, The calculation of the current surge margin and the generation of continuous fine-tuning instructions based on the difference between the current surge margin and the preset target margin include: The current operating point of the compressor is determined based on the real-time operating multi-parameters, and the surge margin of the current operating point relative to the preset surge boundary line is calculated. The difference between the current surge margin and the preset target margin is calculated to obtain the margin deviation. The margin deviation is calculated using an incremental adaptive control algorithm, and the continuous fine-tuning command is output. The value range of the continuous fine-tuning command is a continuous valve opening percentage range.

3. The gas turbine unit multi-parameter adaptive anti-surge control system as described in claim 1, characterized in that, The predefined nonlinear mapping model of the vent valve is used to reconstruct the continuous fine-tuning command to obtain a stepped control command, including: Obtain the current valve position feedback value of the anti-surge vent valve, and compare the target opening value corresponding to the continuous fine-tuning command with the dead zone boundary and step transition point of the current valve position interval in the nonlinear mapping model of the vent valve. If the target opening value falls within the dead zone interval corresponding to the current valve position, the target step position of the vent valve is locked as the current valve position feedback value, and the difference between the target opening value and the upper limit of the dead zone boundary is used as the step margin compensation value. If the target opening value exceeds the upper limit of the dead zone boundary, the actual landing point opening of the anti-surge vent valve after crossing the dead zone is determined based on the step response characteristics as the target step position of the vent valve, and the opening difference between the actual landing point opening and the target opening value is calculated as the step margin compensation value.

4. The gas turbine unit multi-parameter adaptive anti-surge control system as described in claim 1, characterized in that, The step-wise evaluation of the stepped control commands with the gas turbine unit's pipeline volume and gas column resonant frequency band, to determine the anti-surge execution strategy, includes: Based on the frequency of step action of the anti-surge vent valve during historical periods, the estimated frequency of step action is determined. The estimated frequency of the step action is compared with the resonant frequency band of the gas column in the exhaust pipe network of the gas turbine unit; If the estimated frequency of the step action falls within the resonant frequency band of the air column, the target step position of the vent valve is attenuated and suppressed, and the reduced pressure relief amount is transferred and allocated to the auxiliary fine-tuning compensation command to form the anti-surge execution strategy to avoid resonance.

5. The gas turbine unit multi-parameter adaptive anti-surge control system as described in claim 1, characterized in that, The step of controlling the anti-surge venting valve to perform a corresponding step action according to the anti-surge execution strategy, and simultaneously controlling the auxiliary adjustment parameters to perform corresponding fine-tuning compensation actions, includes: Based on the auxiliary fine-tuning compensation command, the step margin compensation value is converted into the first angle adjustment amount of the compressor variable guide vane and the second flow rate adjustment amount of the fuel flow; Within a preset time before the anti-surge vent valve step action is executed, the compressor variable guide vane opening is simultaneously adjusted according to the first angle adjustment amount, and the fuel flow rate is adjusted according to the second flow rate adjustment amount, so as to build a flow and pressure buffer inside the compressor in advance.

6. The gas turbine unit multi-parameter adaptive anti-surge control system as described in claim 5, characterized in that, The step margin compensation value is converted into a first angle adjustment of the compressor variable guide vane and a second flow rate adjustment of the fuel flow based on the auxiliary fine-tuning compensation command, including: The first response delay time of the compressor variable guide vane and the second response delay time of the fuel flow rate are obtained; The dynamic allocation weights of the compressor variable guide vanes and fuel flow are determined based on the ratio of the first response delay time to the reciprocal of the second response delay time. The step margin compensation value is split according to the dynamically allocated weight, and the split compensation amounts are mapped to the first angle adjustment amount and the second flow rate adjustment amount, respectively.

7. The gas turbine unit multi-parameter adaptive anti-surge control system as described in claim 1, characterized in that, After controlling the anti-surge vent valve to perform the corresponding step action according to the anti-surge execution strategy, the method further includes: Collect the actual change in compressor outlet pressure after the step action of the anti-surge vent valve; The actual change in compressor outlet pressure is compared with the theoretical change in pressure estimated based on the anti-surge execution strategy to obtain the pressure deviation. If the pressure deviation exceeds the preset tolerance range, the step transition point and step response characteristics in the nonlinear mapping model of the vent valve are corrected online based on the pressure deviation to update the actual dead zone and overshoot parameters of the anti-surge vent valve.

8. The gas turbine unit multi-parameter adaptive anti-surge control system as described in claim 3, characterized in that, After reconstructing the continuous fine-tuning command based on the predefined nonlinear mapping model of the vent valve to obtain the stepped control command, the following steps are also included: Monitor the actual valve position feedback value after the anti-surge vent valve performs a step action; If the actual valve position feedback value does not reach the target step position of the vent valve and the stagnation time exceeds the preset jamming judgment threshold, then the anti-surge vent valve is determined to be in a step jamming state. In response to the step jamming state, the adjustment step size and amplitude of the auxiliary adjustment parameters in the auxiliary fine-tuning compensation command are expanded to replace the pressure relief function of the anti-surge vent valve through active adjustment of the compressor variable guide vanes and fuel flow.

9. The multi-parameter adaptive anti-surge control system for gas turbine units as described in claim 1, characterized in that, It also includes non-linear processing logic for the valve closing direction: When the current surge margin indicator unit is in an over-margin state, a continuous fine-tuning command for the valve closing direction is generated; Based on the nonlinear mapping model of the vent valve, if the target opening value corresponding to the continuous fine-tuning command in the valve closing direction falls into the valve closing dead zone interval corresponding to the current valve position, the generation of the valve closing step command is temporarily suspended, and the current valve position is maintained. Simultaneously, the auxiliary adjustment parameters are adjusted to actively increase the compressor load, thereby reducing the surge margin, until the continuous fine-tuning command in the valve closing direction crosses the valve closing dead zone boundary.

10. The multi-parameter adaptive anti-surge control system for gas turbine units as described in claim 1, characterized in that, The nonlinear mapping model of the vent valve is pre-constructed in the following manner: During the cold purging phase or low-load warm-up phase before the gas turbine unit starts up, send multiple trial step commands to the anti-surge vent valve. Record the actual dead zone width and step overshoot of the anti-surge vent valve in different basic valve position ranges; Based on the actual dead zone width and step overshoot, a nonlinear mapping model of the vent valve is constructed with the basic valve position range as the index.

Images