One-key start-stop control system of single-shaft gas turbine

Abstract

The invention relates to the field of start-stop control, and discloses a one-key start-stop control system of a single-shaft gas turbine, which is used for solving the technical bottlenecks of the traditional control method in the aspects of state perception, service life management and the like. Comprising the steps that a gas turbine state parameter set containing rotor thermal stress, combustion chamber temperature and exhaust temperature distribution is generated; and generating a life loss evaluation result and an operation efficiency evaluation result by adopting the life evaluation model and the efficiency calculation model, and generating a start-stop control prediction sequence through a dynamic programming algorithm. The service life loss and the operation efficiency are quantitatively evaluated in real time, so that the operation cost problem that the service life and the operation efficiency of the equipment are difficult to consider in the existing start-stop technology is solved, and the full-life-cycle operation cost is remarkably reduced while the long-term reliability of the equipment is guaranteed; maintenance cost can be effectively reduced, fuel consumption is saved, and power grid response economy is improved.

Description

Technical Field

[0001] This invention relates to the field of start-stop control, and more particularly to a one-button start-stop control system for a single-shaft gas turbine. Background Technology

[0002] With the deepening of energy structure adjustment, gas turbines are playing an increasingly important role in peak shaving within the power system. Frequent start-ups and shutdowns have become the norm for gas turbine operation, making the economics of the start-up and shutdown process a key factor affecting power plant operating costs. Currently, the operating costs of gas turbine power plants mainly consist of equipment maintenance costs, fuel costs, labor costs, and grid assessment fees, and the quality of the start-up and shutdown process directly impacts these cost elements.

[0003] Existing one-button start-stop systems generally lack real-time assessment and optimization mechanisms for equipment lifespan degradation. During start-up and shutdown, various components of the gas turbine are subjected to significant thermal stress shocks, and the lifespan degradation caused by different start-stop strategies varies greatly. Existing systems mostly use start-stop curves with fixed parameters, failing to adaptively adjust according to the actual equipment condition. This leads to two disadvantages: overly conservative start-stop strategies, while protecting the equipment, prolong start-stop time and increase fuel costs; simultaneously, overly aggressive start-stop strategies, while shortening start-stop time, accelerate equipment wear and tear, resulting in increased maintenance costs. Existing technologies treat start-up and shutdown processes and operational efficiency optimization as separate processes. In known control systems, start-up and shutdown control and operational efficiency optimization are often separate subsystems, lacking a coordination mechanism. This leads to start-up and shutdown processes potentially sacrificing operational efficiency, or vice versa. In fact, the parameter settings of the start-up and shutdown process directly affect the unit's operational efficiency after reaching the target load, but existing technologies have failed to establish a correlation model between the two. Especially in scenarios where grid peak-shaving demand changes frequently, this separation prevents gas turbines from achieving a balance between lifespan degradation and operational efficiency, increasing the power plant's operating costs.

[0004] Therefore, we propose a one-button start-stop control system for a single-shaft gas turbine to solve the above problems. Summary of the Invention

[0005] This invention provides a one-button start-stop control system for a single-shaft gas turbine, which addresses the technical bottlenecks of traditional control methods in terms of state perception, lifespan management, and multi-objectives.

[0006] The first aspect of this invention provides a one-button start-stop control system for a single-shaft gas turbine, comprising: an acquisition module for real-time acquisition of gas turbine operating parameters and calculation of a gas turbine state parameter set via a state observer; an evaluation module for calculating the equipment life loss rate under the current operating condition based on the gas turbine state parameter set, generating a life loss evaluation result, and simultaneously generating a current operating efficiency evaluation result; a prediction module for generating a start-stop control prediction sequence based on the life loss evaluation result and the operating efficiency evaluation result, with the dual objectives of minimizing life loss and improving operating efficiency; and an allocation module for inputting the start-stop control prediction sequence into a multi-loop coordinated control system to generate control prediction commands for each subsystem.

[0007] Optionally, in a first implementation of the first aspect of the present invention, the method includes: inputting rotor thermal stress distribution and temperature gradient data from the gas turbine state parameter set into a life assessment model, calculating the real-time life loss rate of each key component using a material damage accumulation algorithm, and generating a life loss rate distribution map; inputting combustion chamber temperature field distribution and exhaust temperature gradient data from the gas turbine state parameter set into an efficiency calculation model based on the second law of thermodynamics, calculating the efficiency loss of each thermodynamic process, and generating an efficiency loss distribution map; inputting the life loss rate distribution map and the efficiency loss distribution map into a dynamic weight allocation system, generating life assessment weight coefficients and efficiency assessment weight coefficients based on the current operating conditions and operational objectives; performing weighted calculations on the life loss rate distribution map based on the life assessment weight coefficients to generate standardized life loss assessment results, and simultaneously performing weighted calculations on the efficiency loss distribution map based on the efficiency assessment weight coefficients to generate standardized efficiency assessment results; and fusing the standardized life loss assessment results and efficiency assessment results to generate a current operating efficiency assessment result using a multi-objective algorithm.

[0008] Optionally, in a second implementation of the first aspect of the present invention, metallographic structure characteristic parameters of key parts of the rotor are collected by an online material condition monitoring system to generate a real-time material microstructure dataset; a micro-damage evolution map is generated based on the real-time material microstructure dataset; the micro-damage evolution map is coupled and analyzed with the operating load spectrum to calculate the life loss rate under the current operating condition and generate a real-time life loss rate distribution map; the real-time life loss rate distribution map is input into a remaining life prediction system, and combined with historical operating data and material fatigue characteristics, a remaining life prediction report for key components is generated; the remaining life prediction report is input into a control parameter system to adjust the temperature change rate and load change rate during start-up and shutdown processes to generate life prediction control parameters.

[0009] Optionally, in a third implementation of the first aspect of the present invention, the method includes: generating a control strategy candidate set based on the life loss assessment results and the operating efficiency assessment results; inputting the control strategy candidate set into a state transition model based on dynamic programming, and calculating the optimal control strategy from the final state to the initial state stage by stage using a reverse recursive algorithm to generate a preliminary optimal control sequence; inputting the preliminary optimal control sequence into a constraint-satisfying system, and combining the gas turbine safe operation boundary conditions and equipment physical limitations to generate a feasible control sequence that satisfies all constraints; evaluating the sensitivity of the control sequence to each state parameter through parameter perturbation analysis based on the feasible control sequence to generate a control sequence; and inputting the control sequence into a sequence encoder to extract the rate of increase setpoint, load rate setpoint, and temperature change rate setpoint to form a start-stop control prediction sequence.

[0010] Optionally, in the fourth implementation of the first aspect of the present invention, atmospheric pressure, ambient temperature, and humidity parameters are collected in real time to generate an environmental condition feature dataset; environmental corrections are made to the rate of increase and rate of temperature change in the control sequence based on the environmental condition feature dataset to generate an environmental adaptive control sequence; based on the environmental adaptive control sequence, the optimal operating mode is identified according to the current grid demand and unit status, and a mode switching decision instruction is generated; the mode switching decision instruction is input into an economic evaluation model, and economic evaluation reports under different modes are generated through fuel cost calculation and efficiency analysis; based on the economic evaluation reports, an optimal balance point is sought between lifetime loss and economy through a multi-objective trade-off algorithm to generate a final control prediction sequence.

[0011] Optionally, in the fifth implementation of the first aspect of the present invention, the method includes: mapping the rate of increase setpoint, load rate setpoint, and temperature change rate setpoint to the corresponding control loops respectively using an instruction decomposition algorithm based on the start-stop control prediction sequence, generating fuel control loop setpoints, guide vane control loop setpoints, and excitation control loop setpoints; inputting the setpoints of each loop into an adaptive PID controller based on a gain scheduling strategy, dynamically adjusting the proportional, integral, and derivative parameters of the PID controller by real-time detection of the deviation and operating conditions of each loop, generating preliminary control prediction instructions; and performing feedforward compensation on measurable disturbances based on the preliminary control prediction instructions and the gas turbine dynamic characteristic model, generating compensated control prediction instructions. Based on the compensated control prediction command, the mutual influence between each control loop is eliminated through a multivariate decoupling algorithm to generate a decoupled coordinated control prediction command. Based on the coordinated control prediction command and combined with the physical limitations and safe operation requirements of each actuator, the amplitude and rate of the control prediction command are limited to generate control prediction commands for each subsystem.

[0012] Optionally, in the sixth implementation of the first aspect of the present invention, the proportional, integral, and derivative parameters of the PID controller are dynamically adjusted by real-time detection of the deviation and operating conditions of each loop, with the proportional coefficient Kp adjusted according to the magnitude of the deviation. in, Here, α is the baseline value, e(t) is the adaptive coefficient, and e(t) is the real-time deviation.

[0013] Optionally, in the seventh implementation of the first aspect of the present invention, the response characteristics of each actuator are tested online to generate a set of actuator dynamic characteristic parameters; the timing of the issuance of control prediction commands for each loop is calculated based on the set of actuator dynamic characteristic parameters to generate a control prediction command sequence; the control prediction command sequence is input into a dynamic compensator to establish a feedforward compensation model and generate dynamically compensated coordinated control prediction commands; the degree of coupling between each loop is quantified based on the dynamically compensated coordinated control prediction commands to generate a decoupling compensation matrix; the decoupling compensation matrix is ​​applied to the multi-loop control system to eliminate mutual interference between each control loop through decoupling operations, generating completely decoupled independent control prediction commands.

[0014] Optionally, in the eighth implementation of the first aspect of the present invention, an execution module is included, used to convert the control prediction command into specific operations through an execution mechanism, drive the gas turbine to complete the start-up or shutdown process according to a predetermined program, and dynamically adjust the control prediction command based on real-time parameter feedback during the execution process: according to the control prediction command of each subsystem, the control prediction command is sorted according to the urgency and timing requirements of execution by a priority scheduling algorithm to generate an ordered execution command sequence; the ordered execution command sequence is input into a multimodal actuator system, and the electrical signal is converted into the corresponding mechanical action by a signal conversion module to generate the actual execution action of each subsystem of the gas turbine; feedback signals of the execution action are collected in real time through a distributed monitoring network, including valve opening feedback, guide vane angle feedback and current feedback, to generate an execution status feedback dataset; the execution status feedback dataset is compared and analyzed with the expected execution status, and an execution deviation analysis report is generated by a deviation calculation algorithm, and a control prediction command correction amount is generated based on an adaptive correction strategy; the control prediction command correction amount is fed back to the multi-loop coordinated control system, and the operating parameters of the controller are adjusted in real time by an online parameter tuning module, and the generated control prediction command is re-executed.

[0015] Optionally, in the ninth implementation of the first aspect of the present invention, an early warning module is further included: a multi-source monitoring dataset of the gas turbine's operating status is generated by real-time acquisition of vibration spectrum, acoustic emission signal, and temperature field data during gas turbine operation via a distributed sensor network; the multi-source monitoring dataset is input into a fault prediction system based on a physical model, and potential fault characteristics are identified through a component failure mechanism analysis algorithm to generate a fault early warning level assessment report; the fault early warning level assessment report is input into a control strategy reconfiguration system, and a corresponding emergency control scheme is selected from a preset control strategy library through a pre-set matching algorithm to generate an adaptive control strategy adjustment instruction; the adaptive control strategy adjustment instruction is input into a human-machine interface, and key early warning information and recommended operation schemes are presented to operators through a visualization module to generate a confirmation execution instruction; the confirmation execution instruction is input into a control prediction instruction execution system, and a safety interlock protection mechanism ensures smooth switching of the control strategy, completing adaptive control under abnormal operating conditions.

[0016] Beneficial effects: By calculating and analyzing the collected gas turbine operating parameters through the state observer, a comprehensive set of gas turbine state parameters is generated, realizing in-depth perception and accurate calculation of the gas turbine operating status, and providing a reliable data foundation for subsequent evaluation; By calculating the equipment life loss rate and operating efficiency evaluation results, a dual-objective evaluation system for minimizing life loss and optimizing operating efficiency was established, which solved the limitations of single-objective optimization in traditional systems and achieved a balance between long-term equipment reliability and operating economy. By calculating the life loss rate in real time and using it as a control target, excessive stress impact on hot components during start-up and shutdown can be effectively avoided, and life loss can be controlled within the optimal range. This can significantly extend the equipment overhaul cycle and service life, and directly reduce overhaul costs and spare parts replacement costs. While ensuring the safe lifespan of the equipment, the system takes operating efficiency as a parallel optimization target, ensuring that the start-up and shutdown paths tend to be energy-efficient, avoiding unnecessary fuel waste and efficiency loss. Especially for peak-shaving units that start and stop daily, the cumulative effect of fuel savings from a single start-up and shutdown is huge. By generating optimal prediction sequences, the system can compress start-up and shutdown times in the most economical way, quickly respond to grid dispatching needs, make power plants more competitive in the peak-shaving ancillary services market, and more quickly capture high-price opportunities in the electricity market. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of one embodiment of the single-shaft gas turbine one-button start-stop control system in this invention. Figure 2 This is a schematic diagram illustrating the deployment of the one-button start / stop control system for a single-shaft gas turbine in an embodiment of the present invention. Figure 3 This is a schematic diagram of another embodiment of the one-button start / stop control system for a single-shaft gas turbine in this invention. Figure 4 This is a schematic diagram of one embodiment of the one-button start / stop control device for a single-shaft gas turbine in this invention. Detailed Implementation

[0018] This invention provides a one-button start-stop control system for a single-shaft gas turbine, addressing the technical bottlenecks of traditional control methods in areas such as state perception, lifespan management, and multi-objective control. The terms "first," "second," "third," "fourth," etc. (if present)," in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0019] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the single-shaft gas turbine one-button start / stop control system of the present invention includes: 101. Real-time acquisition of gas turbine operating parameters, including speed, temperature, pressure and vibration data. Based on the gas turbine thermodynamic model and rotor dynamics model, the gas turbine state parameter set is generated by the state observer. The gas turbine state parameter set includes rotor thermal stress, combustion chamber temperature and exhaust temperature distribution. It is understood that the executing entity of this invention can be a single-shaft gas turbine one-button start / stop control device, or it can be a terminal or a server; no specific limitation is made here. This embodiment of the invention will be described using a server as an example.

[0020] It should be noted that, based on a typical scenario of a 50MW single-shaft gas turbine during cold start-up, the system uses standard sensors and embedded control units.

[0021] After the gas turbine's one-button start-stop system is activated, sensors installed at various parts of the gas turbine collect operating parameters in real time. The speed sensor measures the current speed as 1500 rpm, the thermocouple at the compressor outlet measures the temperature as 300 degrees Celsius, the pressure transmitter at the combustion chamber inlet measures the pressure as 8 bar, and the vibration sensor at the bearing housing measures the vibration velocity as 0.05 mm / s. These data are transmitted to the server's real-time database via the data acquisition module at a sampling frequency of 10 times per second, ensuring data timeliness.

[0022] The process utilizes pre-defined thermodynamic and rotor dynamic models of the gas turbine. The thermodynamic model, based on the gas turbine's geometry and material properties, simulates the flow, heat transfer, and energy conversion processes of the working fluid (air and fuel gas), estimating flame propagation and heat release rate within the combustion chamber. The rotor dynamic model considers the rotor's mass, stiffness, and damping characteristics, analyzing the relationship between mechanical stress and vibration. These models are embedded numerically into the state observer, eliminating the need for direct formula usage and enabling rapid calculations through table lookups or interpolation.

[0023] The state observer (extended Kalman filter) serves as the core algorithm, integrating real-time acquired data and model output. It dynamically corrects estimation errors by comparing measured values ​​with model predictions. The observer inputs include real-time data such as engine speed of 1500 rpm and temperature of 300°C, combined with the initial combustion chamber temperature predicted by the model (set to 800°C). Through iterative calculations, it generates a more accurate set of gas turbine state parameters. Specific outputs include: a rotor thermal stress estimate of 120 MPa, reflecting the stress caused by the temperature difference between the rotor surface and the core; a combustion chamber temperature estimate of 950°C, taking into account the effects of fuel injection and airflow; and an exhaust temperature distribution represented as a gradient from the turbine outlet to the chimney, with a high-temperature zone of 600°C and a low-temperature zone of 400°C, reflecting heat loss along the flow path.

[0024] 102. Based on the gas turbine state parameter set, calculate the equipment life loss rate under the current operating condition through the life assessment model, generate life loss assessment results, and at the same time generate the current operating efficiency assessment results through the efficiency calculation model. It should be noted that in step 101, the server has calculated the current set of gas turbine state parameters: the rotor thermal stress is 120 MPa, the combustion chamber temperature is 950 degrees Celsius, and the exhaust temperature distribution shows the characteristics of a high temperature zone of 600 degrees Celsius and a low temperature zone of 400 degrees Celsius.

[0025] Step 102 initiates the process by inputting the rotor thermal stress of 120 MPa into the preset life assessment model. This model is essentially a knowledge base containing extensive experimental and historical operational data, establishing the correlation between different levels of thermal stress, temperature change rates, and fatigue losses of equipment materials (turbine blades, rotor metal). The inference is as follows: if the current thermal stress level of 120 MPa continues for 10 minutes, it is equivalent to subjecting the rotor material to a medium-to-high intensity thermal cycle. By consulting the built-in stress and life curve tables and combining this with the current cumulative number of gas turbine operations (1500 start-up cycles), the model calculates that under this specific operating condition, the rotor component's life loss rate within this minute is 0.002%. This means that, relative to its total design life, this minute of operation consumes 0.002% of its lifespan. At the same time, the model will also evaluate the creep damage of high-temperature components in the combustion chamber (transition section and first-stage turbine blades) at 950 degrees Celsius, and may find that its loss rate is 0.001% / minute. The model summarizes the rotor fatigue loss and the creep loss of high-temperature components to generate a life loss assessment result: "The current comprehensive life loss rate is 0.003% / minute".

[0026] Almost simultaneously, the server ran a parallel efficiency calculation model. This model received key parameters including a combustion chamber temperature of 950 degrees Celsius and exhaust temperature distribution. At its core was an efficiency graph built based on gas turbine thermodynamics, reflecting the efficiency with which the entire unit converts fuel chemical energy into electrical energy under different combustion temperatures, pressures, and exhaust parameters. The model compared the current combustion chamber temperature and pressure (from raw data, 8 bar) with the optimal values ​​under design conditions. Although the combustion temperature of 950 degrees Celsius was lower than the 1100 degrees Celsius corresponding to the design peak efficiency, this temperature was reasonable considering the current start-up and warm-up phase. The model further analyzed the exhaust temperature distribution; a distribution between 600°C and 400°C might indicate a certain gradient loss in heat exchange. Through comprehensive calculation (interpolation using a multi-dimensional efficiency table), the model generated a current operating efficiency assessment result: "Current operating efficiency is 28%," meaning that under the current conditions, 28% of the input fuel energy was effectively converted into output work.

[0027] The output of step 102 is two parallel evaluation results: lifetime loss evaluation result (overall lifetime loss rate 0.003% / minute) and operating efficiency evaluation result (current operating efficiency 28%).

[0028] 103. Based on the life loss assessment results and the operation efficiency assessment results, with the dual objectives of minimizing life loss and improving operation efficiency, a start-stop control prediction sequence is generated through a dynamic programming algorithm. The control sequence includes the setpoints for the rate of increase, load rate, and temperature change rate. It should be noted that this is based on the scenario of a 50MW single-shaft gas turbine during cold start-up.

[0029] In step 102, the server has generated two key evaluation results: a lifetime loss assessment result (the current overall lifetime loss rate is 0.003% / minute) and an operating efficiency assessment result (the current operating efficiency is 28%). Now, based on these two results, the server needs to plan the optimal path for the gas turbine to go from its current state to full grid connection and operation.

[0030] The entire startup process is divided into multiple consecutive time periods (the next 20 minutes, with each 10 seconds representing a decision point). At each decision point, the algorithm needs to weigh the sometimes conflicting objectives of "lifetime degradation" and "operational efficiency".

[0031] Enumerate all possible operational options in the current state. For the "rate of increase" in the next 10 seconds, there are three candidate options: aggressive (set to 200 rpm / min), moderate (150 rpm / min), and conservative (100 rpm / min). Similarly, there are different options for the "load rate" (once connected to the grid) and the implicit "rate of temperature change".

[0032] Using a built-in gas turbine model, predictions are made as follows: If an aggressive approach (increase rate of 200 rpm / min) is chosen, the gas turbine temperature will rise rapidly, potentially increasing operating efficiency to 35% within 5 minutes, which is beneficial to the efficiency target. However, this rapid temperature rise will cause a sharp increase in rotor thermal stress to 180 MPa, which the model predicts will cause the lifespan loss rate to spike to 0.01% / min during this phase, negatively impacting the lifespan target. If a conservative approach (increase rate of 100 rpm / min) is chosen, the rotor thermal stress is controlled below 130 MPa, and the lifespan loss rate can be maintained at a low level of 0.0035% / min. However, the slow temperature rise will result in a sluggish efficiency increase, with the efficiency potentially reaching only 30% after 10 minutes, which is uneconomical. The task of the dynamic programming algorithm is to evaluate countless possible paths and select an operational setting for each future time period to minimize the total lifespan loss from start-up to completion while maximizing the average operating efficiency. This is a complex multi-objective process.

[0033] After calculation and consideration, a start-stop control prediction sequence is finally generated. This sequence is not a single instruction, but a series of setpoints that change over time, outputting a sequence like this: Time point 0-2 minutes: Given that the rotor already has a certain thermal stress (120MPa), in order to avoid impact, a gentle acceleration rate of 140rpm / min is adopted. During this stage, the load rate is 0% (not connected to the grid).

[0034] Time point 2-5 minutes: After the rotor thermal stress stabilizes, in order to quickly pass through the low efficiency zone, increase the acceleration rate to 180 rpm / min and accelerate to the rated speed of 3000 rpm.

[0035] Time point 5-8 minutes: After grid connection, control the load rate to increase steadily from 0% to 50%, while limiting the temperature change rate to within 50℃ / minute, so as to ensure that the efficiency steadily increases to 40% while the life loss rate does not exceed 0.005% / minute.

[0036] 104. Input the start-stop control prediction sequence into the multi-loop coordinated control system, and generate control prediction commands for each subsystem through the adaptive PID controller. The subsystems include the fuel control system, the inlet guide vane system, and the excitation system. It should be noted that in step 103, the dynamic programming algorithm generates a start-stop control prediction sequence. Within the current time window, this sequence provides the following settings: acceleration rate 180 rpm / min, load rate 30% (already connected to the grid), and turbine exhaust temperature change rate not exceeding 60℃ / min.

[0037] Step 104 involves translating this high-level "target command" into precise, executable commands for the three subsystems (fuel control, inlet guide vanes, and excitation). This process is accomplished by a multi-loop coordinated control system, the core of which is an adaptive PID controller.

[0038] The multi-loop coordinated control system receives the above setpoint sequence. Internally, it contains three parallel control loops, each corresponding to a subsystem. The system understands that these three loops are interconnected: increasing fuel will raise the rotational speed and temperature; opening the inlet guide vanes will increase airflow and affect combustion efficiency; and adjusting the excitation will change the power generation load. Therefore, it needs to coordinate to avoid operating independently. To increase the rotational speed, both fuel and air (through the guide vanes) must be increased simultaneously, while the load rate is primarily controlled by the excitation system.

[0039] The adaptive PID controller begins operation. Taking the target of "controlling the engine speed to reach a rate of increase of 180 rpm" as an example: Speed ​​control loop (main loop): The adaptive PID controller continuously receives the actual value measured by the speed sensor (current speed is 2500 rpm) and compares it with the target value (the next target calculated based on the rate of increase, 2518 rpm), calculating the deviation. Parameter self-tuning: Unlike traditional fixed-parameter PID controllers, the adaptive PID controller can automatically adjust the strength of its control action (i.e., PID parameters) according to the current operating state of the gas turbine (current speed level, temperature). In the mid-speed range of 2500 rpm, it may use a set of parameters with medium gain to ensure response speed while avoiding overshoot.

[0040] Subsystem command generation: Based on the calculated control variables, the coordination system decomposes and issues commands to the sub-loops: Fuel control system command: The main PID outputs a command requesting that the fuel flow setpoint be increased from the current 3.5 tons per hour to 3.8 tons per hour. This command is sent to the slave PID controller in the fuel control loop, which precisely adjusts the opening of the fuel regulating valve. Inlet guide vane system command: To match the increased fuel and prevent the combustion chamber from being too rich or overheating, the coordination system synchronously issues a command to the inlet guide vane system to increase its blade angle from 45 degrees to 47 degrees to increase the intake air volume.

[0041] Excitation system command: In order to maintain the "load rate of 30%" setting, the coordination system will send a command to the excitation system to set the generator terminal voltage at 10.4 kV, thereby stabilizing the output power at about 15 MW (i.e. 30% of 50MW).

[0042] Throughout the process, the adaptive PID controller will fine-tune its output commands based on real-time feedback (whether the actual rotational speed accurately tracks the target and whether the temperature change rate exceeds the limit).

[0043] 105. The control prediction command is converted into a specific operation through the actuator, which drives the gas turbine to complete the start-up or shutdown process according to the predetermined program, and dynamically adjusts the control prediction command based on real-time parameter feedback during the execution process. It should be noted that in step 104, the multi-loop coordinated control system has generated specific subsystem control prediction commands: the fuel flow rate setpoint is 3.8 tons / hour, the inlet guide vane angle is 47 degrees, and the generator terminal voltage is set to 10.4 kV. The task of step 105 is to translate these electronic commands into actual physical actions and form a closed-loop control.

[0044] Control prediction commands are sent to the actuators of each subsystem via fieldbus. Fuel Control System: The command reaches the positioner of the fuel regulating valve. The positioner converts the flow setpoint of "3.8 tons / hour" into a corresponding valve opening signal (requiring a valve opening of 52%). Subsequently, the electro-hydraulic servo mechanism drives the valve core of the regulating valve to move, precisely adjusting the valve opening from the current 48% to 52%, thereby increasing the fuel supply. Inlet Guide Vane System: The command is sent to the servo controller of the guide vane actuator. The controller compares the "47 degrees" angle command with the current angle (45 degrees) fed back by the actuator, outputting a control signal to drive the hydraulic cylinder to rotate the adjustable guide vane ring at the compressor inlet, stabilizing the vane angle at 47 degrees and increasing the intake air volume.

[0045] Excitation system: The command is sent to the automatic voltage regulator. The regulator compares the voltage setpoint of "10.4 kV" with the actual voltage (10.35 kV) measured by the generator outlet transformer. By increasing the exciter field current, the magnetic field of the main generator is enhanced, and the output voltage is precisely increased to 10.4 kV, thereby stabilizing the output power at 15 MW (30% load).

[0046] The combined effect of these operations ensures that the gas turbine operates according to the optimal sequence planned in step 103: the speed steadily increases at a rate of 180 rpm / min, the load rate is maintained at 30%, and the entire heating process is stable.

[0047] Throughout the process, real-time parameters were continuously monitored. After adding fuel, the temperature sensor reported that the actual rate of increase in turbine exhaust temperature was 65°C / minute, slightly exceeding the preset limit of 60°C / minute.

[0048] This deviation is fed back in real time to the multi-loop coordinated control system in step 104. The adaptive PID controller responds immediately: it determines that the current heating rate is too fast, which may be detrimental to its lifespan. Therefore, it dynamically fine-tunes the output command, slightly reducing the fuel flow setpoint from 3.8 tons / hour to 3.77 tons / hour without significantly affecting the heating rate. At the same time, it may fine-tune the inlet guide vane angle to 46.8 degrees to maintain the air-fuel ratio.

[0049] The actuator then receives and executes the new instruction, and the fuel valve closes slightly.

[0050] In this embodiment of the invention, by fusing real-time acquired data and model output with a state observer, and dynamically correcting estimation errors through algorithms such as the extended Kalman filter, a set of gas turbine state parameters containing key parameters such as rotor thermal stress, combustion chamber temperature, and exhaust temperature distribution can be accurately generated. This provides an accurate basis for subsequent evaluation and control, effectively avoiding control deviations that may be caused by inaccurate data in traditional methods. Using a life assessment model and an efficiency calculation model, multiple factors such as thermal stress, temperature change rate, combustion chamber temperature, and exhaust temperature distribution are comprehensively considered to evaluate equipment life loss and operating efficiency. The life assessment model, combining extensive experimental and historical data, can accurately quantify life loss under different operating conditions; the efficiency calculation model, based on thermodynamic principles, constructs an efficiency graph, which can accurately assess current operating efficiency. Employing a dynamic programming algorithm, this system balances conflicting objectives: minimizing lifespan loss and improving operational efficiency. The startup process is divided into multiple time periods, enumerating all possible operational choices. A built-in gas turbine model predicts the impact of different choices on lifespan and efficiency, ultimately generating a start-stop control prediction sequence that includes setpoints for acceleration rate, load rate, and temperature change rate. This optimizes gas turbine startup and shutdown processes, improving operational efficiency and reducing operating costs while ensuring safe equipment operation. The multi-loop coordinated control system, through an adaptive PID controller, transforms the high-level start-stop control prediction sequence into precise executable commands for subsystems such as the fuel control system, inlet guide vane system, and excitation system. The adaptive PID controller automatically adjusts PID parameters based on the gas turbine's current operating state, achieving coordinated control between subsystems and avoiding independent operation. After the control prediction commands are translated into actual operations through the actuator, real-time parameters are continuously monitored during execution, and the control prediction commands are dynamically adjusted based on parameter feedback. When deviations occur, such as the rate of temperature rise exceeding the preset limit, the system can respond promptly by fine-tuning parameters such as fuel flow and inlet guide vane angle to ensure that the gas turbine always operates according to the predetermined program, while ensuring that the operating parameters are within a safe range, thus improving the stability and reliability of the system.

[0051] Please see Figure 2 , 3 Another embodiment of the single-shaft gas turbine one-button start / stop control system of the present invention includes: 201. Real-time acquisition of gas turbine operating parameters, including speed, temperature, pressure and vibration data. Based on the gas turbine thermodynamic model and rotor dynamics model, the gas turbine state parameter set is generated by the state observer. The gas turbine state parameter set includes rotor thermal stress, combustion chamber temperature and exhaust temperature distribution. Specifically, a distributed sensor network is used to collect real-time operating parameters of the gas turbine, generating a real-time operating parameter set that includes speed signals, temperature gradients, pressure distributions, and vibration spectra. This real-time operating parameter set is then input into a state observer based on an extended Kalman filter algorithm, combined with the gas turbine's thermodynamic conservation equations, to generate combustion chamber temperature field distribution and exhaust temperature gradient data. These data are then input into a rotor dynamics model, and based on thermoelasticity theory, the thermal stress distribution on the rotor surface and internally is calculated, generating a rotor thermal stress distribution map. The rotor thermal stress distribution map is then fused and analyzed with real-time vibration spectrum data, and modal analysis technology is used to identify rotor dynamic characteristics, generating a rotor dynamic response characteristic set. Finally, the combustion chamber temperature field distribution, exhaust temperature gradient data, rotor thermal stress distribution map, and rotor dynamic response characteristic set are integrated to form a complete gas turbine state parameter set.

[0052] It should be noted that a heavy-duty gas turbine, model GT-100, operates under standard environmental conditions during the startup phase.

[0053] A distributed sensor network is deployed at key parts of the gas turbine. This includes: a thermocouple array (8 measuring points, measuring temperature gradients from 200°C to 400°C) installed at the compressor outlet; pressure sensors (4 measuring points, pressure values ​​between 1.5 bar and 2.0 bar) arranged circumferentially around the combustion chamber; vibration sensors (collecting spectral data, with dominant frequencies of 50Hz and 100Hz, and amplitudes between 0.1mm / s and 0.5mm / s) installed at the rotor bearings; and speed sensors (accelerating the speed from 0 rpm to 3000 rpm in real time). These sensors collect data once per second, generating a real-time operating parameter set. At a given moment, with a speed of 1500 rpm, the temperature gradient data shows a combustion chamber inlet temperature of 300°C and an outlet temperature of 550°C, the pressure distribution shows a compressor outlet temperature of 1.8 bar, and the vibration spectrum shows a peak at 75Hz.

[0054] The real-time operating parameter set is input into a state observer based on the extended Kalman filter algorithm. This observer combines the gas turbine thermodynamic conservation equations (energy balance and mass conservation) to estimate the state that is not directly measured. Using temperature gradient data, the observer calculates the temperature field distribution inside the combustion chamber, showing that the high-temperature zone is in the center and the temperature gradient changes from 500°C to 800°C. At the same time, it generates exhaust temperature gradient data, from 600°C at the turbine outlet to 400°C in the flue.

[0055] These temperature data were input into the rotor dynamics model. Based on thermoelastic theory, the model calculated the thermal stress distribution on the rotor surface and inside. Considering that the rotor material is a nickel-based alloy, the model simulated the thermal expansion effect and generated a thermal stress distribution map, showing that the maximum stress occurs in the middle of the rotor, with a value of 150 MPa, and the minimum stress is at the end, with a value of 50 MPa.

[0056] The rotor thermal stress distribution map is fused with real-time vibration spectrum data for analysis. Modal analysis technology (identifying the rotor's natural frequencies and mode shapes) is used to identify the rotor's dynamic characteristics. Vibration data shows that the 75Hz component matches the rotor's first-order bending mode, generating a set of rotor dynamic response characteristics, including dynamic stress amplitude (fluctuating within ±10MPa) and critical speed information.

[0057] All data—combustion chamber temperature field distribution (two-dimensional temperature map), exhaust temperature gradient data (linear descent curve), rotor thermal stress distribution map (stress cloud map), and rotor dynamic response characteristic set (modal parameters)—are integrated to form a gas turbine state parameter set. This set is used in subsequent steps to provide a comprehensive view of the gas turbine's state. The entire process is completed within 1 second, ensuring real-time performance.

[0058] 202. Based on the gas turbine state parameter set, calculate the equipment life loss rate under the current operating condition through the life assessment model, generate life loss assessment results, and at the same time generate the current operating efficiency assessment results through the efficiency calculation model. Specifically, the rotor thermal stress distribution and temperature gradient data from the gas turbine state parameter set are input into a life assessment model based on the theory of creep and fatigue interaction. A material damage accumulation algorithm is used to calculate the real-time life loss rate of each key component, generating a life loss rate distribution map. The combustion chamber temperature field distribution and exhaust temperature gradient data from the gas turbine state parameter set are input into an efficiency calculation model based on the second law of thermodynamics. A multi-analysis algorithm is used to calculate the efficiency loss of each thermodynamic process, generating an efficiency loss distribution map. The life loss rate distribution map and the efficiency loss distribution map are input into a dynamic weight allocation system. Based on the current operating conditions and operational objectives, life assessment weight coefficients and efficiency assessment weight coefficients are generated. The life loss rate distribution map is weighted based on the life assessment weight coefficients to generate a standardized life loss assessment result. Simultaneously, the efficiency loss distribution map is weighted based on the efficiency assessment weight coefficients to generate a standardized efficiency assessment result. The standardized life loss assessment result and the efficiency assessment result are fused, and a multi-objective algorithm is used to generate a current operating efficiency assessment result that includes both life loss level and efficiency level. The life assessment model employs a real-time damage assessment method based on the theory of material microstructure evolution, including: collecting metallographic characteristic parameters of key rotor components through an online material condition monitoring system to generate a real-time material microstructure dataset; inputting the real-time material microstructure dataset into a material damage evolution model, calculating the material damage increment based on dislocation dynamics theory and grain boundary slip mechanism, and generating a micro-damage evolution map; coupling the micro-damage evolution map with the operating load spectrum, calculating the life loss rate under the current operating condition through a damage mechanics model, and generating a real-time life loss rate distribution map; inputting the real-time life loss rate distribution map into a remaining life prediction system, combining historical operating data and material fatigue characteristics to generate a remaining life prediction report for key components; and inputting the remaining life prediction report into a control parameter system, adjusting the temperature change rate and load change rate during start-up and shutdown through a life balancing allocation algorithm to generate life prediction control parameters. It should be noted that the description describes the evaluation process of the GT-100 gas engine when accelerating to 2500 rpm during startup.

[0059] The gas turbine state parameter set generated in step 201 is received, including: maximum rotor thermal stress of 180 MPa (located in the high-pressure turbine section), average combustion chamber temperature of 820°C, and exhaust temperature gradient from 650°C to 450°C. Life loss assessment: The rotor thermal stress distribution (peak value of 180 MPa) and temperature gradient data are input into a life assessment model based on creep and fatigue interaction theory. The model calculates the real-time life loss rate using a damage accumulation algorithm, based on the SN curve and creep curve of the material (IN718 alloy), combined with the current stress cycle amplitude (±20 MPa) and metal temperature (565°C). The calculated fatigue loss rate of the high-pressure turbine blade in this transient state is 0.002% (per cycle), and the creep loss rate is 0.001% (per hour). A life loss rate distribution map is generated, showing that the blade root is the critical damage area, with a total loss rate reaching a maximum of 0.003% / minute.

[0060] Efficiency Assessment: Combustion chamber temperature field distribution (showing localized high-temperature zones reaching 950℃) and exhaust temperature gradient data were simultaneously input into an efficiency calculation model based on thermal analysis. The model analyzed the irreversible losses of each thermodynamic process, finding that the thermal efficiency of the combustion process was 85%, the turbine expansion process was 78%, and exhaust heat loss accounted for the largest proportion. An efficiency loss distribution map was generated, quantifying that incomplete combustion loss accounted for 3%, exhaust heat loss accounted for 9%, and the overall current operating thermal efficiency was assessed at 34.5%.

[0061] Weighting and Fusion: The dynamic weighting system automatically generates weight coefficients based on the current operational objective of "prioritizing lifetime control in the initial startup phase": lifetime assessment weight is 0.7, and efficiency assessment weight is 0.3. The system uses these weights to perform weighted calculations on the loss rate distribution map and efficiency loss distribution map, generating standardized lifetime loss assessment results (mapping the loss rate to a 0-100 scale, with the current result being "scale 15," indicating low to medium loss) and standardized efficiency assessment results (efficiency scale "scale 65," indicating a medium level).

[0062] The two standardized results are fused using a multi-objective algorithm to generate the final current operational efficiency assessment result, outputting a lifetime loss level of "L-15" and an efficiency level of "E-65". The comprehensive assessment conclusion is given: "The lifetime loss during the current startup process is controllable, and there is room for efficiency improvement," providing a decision-making basis for the next step of generating a control sequence. The entire calculation and assessment cycle is less than 500 milliseconds.

[0063] 203. Based on the life loss assessment results and the operation efficiency assessment results, with the dual objectives of minimizing life loss and improving operation efficiency, a start-stop control prediction sequence is generated through a dynamic programming algorithm. The control sequence includes the setpoints for the rate of increase, load rate, and temperature change rate. Specifically, the life loss assessment results and operating efficiency assessment results are input into a multi-objective system, and a control strategy candidate set containing multiple non-dominated solutions is generated using the Pareto front analysis algorithm. The control strategy candidate set is then input into a dynamic programming-based state transition model, and the optimal control strategy is calculated stage by stage from the final state to the initial state using a reverse recursive algorithm to generate a preliminary optimal control sequence. The preliminary optimal control sequence is then input into a constraint-satisfying system, and combined with the gas turbine's safe operation boundary conditions and equipment physical limitations, a feasible control sequence satisfying all constraints is generated. The feasible control sequence is then input into a sensitivity analysis module, and parameter perturbation analysis is used to evaluate the sensitivity of the control sequence to each state parameter to generate a control sequence. Finally, the control sequence is input into a sequence encoder, and the rise rate setpoint, load rate setpoint, and temperature change rate setpoint are extracted to form a start-stop control prediction sequence. The dynamic programming algorithm employs an environmentally adaptive dynamic programming method, which includes: real-time acquisition of atmospheric pressure, ambient temperature, and humidity parameters through an environmental parameter monitoring system to generate an environmental condition characteristic dataset; inputting the environmental condition characteristic dataset into an environmental adaptation model, and using an aero-thermodynamic correction algorithm to perform environmental correction on the rate of ascent and rate of temperature change in the control sequence to generate an environmentally adaptive control sequence; inputting the environmentally adaptive control sequence into an operating mode recognition system, identifying the optimal operating mode based on current grid demand and unit status, and generating mode switching decision commands; inputting the mode switching decision commands into an economic evaluation model, and generating economic evaluation reports under different modes through fuel cost calculation and efficiency analysis; and using a multi-objective trade-off algorithm to seek the optimal balance between lifetime loss and economy in the economic evaluation reports, generating the final control prediction sequence.

[0064] It should be noted that, based on the output of step 202 (lifetime loss level L-15, efficiency level E-65), this step generates a start-stop control prediction sequence with the dual objectives of minimizing lifetime loss and improving operational efficiency. The implementation process is as follows: The lifetime loss assessment results (L-15) and efficiency assessment results (E-65) are input into the multi-objective system. The system generates three non-dominated solutions as a control strategy candidate set (see Table 1) in the trade-off space between lifetime loss rate (unit: % / min) and efficiency (unit: %) using the Pareto front analysis algorithm. Each solution corresponds to a different combination of control parameters.

[0065] Table 1: Candidate Set of Control Strategies (Example of Pareto Solution) The candidate set is input into a dynamic programming-based state transition model. The startup process is divided into 10 stages (30 seconds each), with state variables including rotational speed (currently 1500 rpm, target 3000 rpm), rotor thermal stress (currently 180 MPa), and exhaust temperature (currently 600℃). The optimal control action for each stage is calculated backward from the final state (3000 rpm) using a reverse recursive algorithm. The objective function is to minimize the total cost J, as shown in the formula: ; Among them, L k E is the lifetime attrition rate in stage k. k It's about efficiency, the weighting coefficient w L =0.6、w E =0.4 (based on operational objectives). After calculation, strategy B was selected as the preliminary optimal sequence because it strikes the best balance between lifetime and efficiency.

[0066] The initial sequence input constraints satisfy the system, and the gas turbine safety boundaries are verified: speed limit (max 3200 rpm) and turbine inlet temperature limit (max 900℃). The load rate in the sequence, at 10 MW / min, exceeds the limit of 8 MW / min, so it is adjusted to 8 MW / min. Subsequently, the sensitivity analysis module performs parameter perturbation (±5% speed fluctuation) on the adjusted sequence. The evaluation shows that the temperature change rate is most sensitive to thermal stress (sensitivity coefficient 0.8), therefore, the temperature change rate is reduced from 80℃ / min to 75℃ / min to enhance robustness.

[0067] The environmental parameter monitoring system collects atmospheric pressure (101.3 kPa) and ambient temperature (15℃) and inputs them into the environmental adaptation model. An aerodynamic-thermodynamic correction algorithm is used to adjust the rate of ascent and temperature change (the rate of ascent is adjusted from 600 rpm / min to 580 rpm / min due to high air density). Finally, the sequence encoder extracts the setpoints to form a start-stop control prediction sequence: rate of ascent 580 rpm / min, load rate 8 MW / min, and temperature change rate 75℃ / min. This sequence ensures that the start-up process is completed within 15 minutes, while maintaining a lifespan loss rate of less than 0.003% / min and an efficiency higher than 35%.

[0068] 204. The start-stop control prediction sequence is used to generate control prediction commands for each subsystem through an adaptive PID controller. The subsystems include the fuel control system, the inlet guide vane system, and the excitation system. Specifically, the start-stop control prediction sequence is input into the control sequence parser. An instruction decomposition algorithm maps the rate of increase setpoint, load rate setpoint, and temperature change rate setpoint to their respective control loops, generating fuel control loop setpoints, guide vane control loop setpoints, and excitation control loop setpoints. These loop setpoints are then input into an adaptive PID controller based on a gain scheduling strategy. By real-time detection of loop deviations and operating conditions, the proportional, integral, and derivative parameters of the PID controller are dynamically adjusted to generate preliminary control prediction instructions. These preliminary control prediction instructions are then input into a feedforward compensator. Combined with the gas turbine dynamic characteristic model, feedforward compensation is applied to measurable disturbances, generating compensated control prediction instructions. The compensated control prediction instructions are then input into a cross-decoupling controller. A multivariable decoupling algorithm eliminates the mutual influence between control loops, generating decoupled coordinated control prediction instructions. Finally, the coordinated control prediction instructions are input into an output limiting module. Based on the physical limitations and safe operation requirements of each actuator, the amplitude and rate of the control prediction instructions are limited, generating control prediction instructions for each subsystem. The multi-loop coordinated control system adopts an intelligent coordination strategy based on response characteristic matching, including: online testing of the response characteristics of each actuator through an actuator dynamic testing system to generate a set of actuator dynamic characteristic parameters containing response time, dead zone, and hysteresis characteristics; inputting the actuator dynamic characteristic parameter set into a timing prediction module to calculate the issuance timing of control prediction commands for each loop based on the response synchronization principle, generating a control prediction command sequence; inputting the control prediction command sequence into a dynamic compensator to establish a feedforward compensation model based on actuator characteristics, generating dynamically compensated coordinated control prediction commands; inputting the dynamically compensated coordinated control prediction commands into a coupling characteristic analysis system to quantify the coupling degree between loops using a relative gain analysis method, generating a decoupling compensation matrix; applying the decoupling compensation matrix to the multi-loop control system to eliminate mutual interference between control loops through decoupling operations, generating completely decoupled independent control prediction commands.

[0069] It should be noted that the start-stop control prediction sequence (acceleration rate 580 rpm / min, load rate 8 MW / min, temperature change rate 75℃ / min) generated in step 203 is converted into specific control prediction commands for fuel control, inlet guide vanes, and excitation systems. The implementation process is as follows: The control sequence parser maps global setpoints to each control loop using an instruction decomposition algorithm. Based on the dynamic characteristics of the gas turbine, the decomposition relationships are shown in Table 2 below: Table 2 After analysis, the setpoints for each loop are generated: the fuel control loop setpoint is a fuel flow rate of 0.6 kg / s. 2The acceleration increases, with a target value of 45 kg / s; the guide vane control circuit is set to open the guide vane angle to 55° at a rate of 5° / min; the excitation control circuit is set to increase the excitation current at a rate of 12 A / min, with a voltage target of 15 kV.

[0070] Each loop's setpoint is input to an adaptive PID controller based on a gain-based scheduling strategy. The controller detects the deviation in real time (actual fuel flow rate 43 kg / s vs. setpoint 45 kg / s, deviation 2 kg / s) and dynamically adjusts the PID parameters. The proportional gain Kp is adjusted according to the deviation magnitude, using the following formula: in, =0.8 is the baseline value, α=0.1 is the adaptive coefficient, and e(t) is the real-time deviation. When the deviation is 2kg / s, Simultaneously, the integral and derivative parameters are adjusted accordingly to generate preliminary control prediction commands (fuel valve opening command increased by 5%).

[0071] The feedforward compensator compensates for measurable disturbances (ambient temperature fluctuations ±2℃). Based on the gas turbine model, the impact of temperature disturbances on fuel flow rate is calculated (gain -0.05 kg / s / ℃), generating a compensation of +0.1 kg / s, which is then added to the initial command.

[0072] The compensated command is input to the cross-decoupled controller. Relative gain analysis identifies a high coupling degree between the fuel loop and the guide vane loop (coupling coefficient 0.7). The decoupling algorithm calculates the compensation matrix to eliminate mutual interference: the coupling effect of guide vane angle changes on fuel flow is canceled by feedforward, generating a decoupled coordinated control prediction command.

[0073] The output limiting module limits the amplitude and rate based on the physical limitations of the actuator: fuel valve opening is limited to 0-100% (rate limit 10% / s), guide vane angle is limited to 20-60° (rate limit 5° / s), and excitation current is limited to 0-2000A (rate limit 50A / s). It generates the final control prediction commands for each subsystem: fuel valve opening command 62%, guide vane angle command 52°, and excitation current command 1200A. The entire process is completed within 100 milliseconds, ensuring accurate and executable commands.

[0074] 205. The control prediction command is converted into a specific operation through the actuator, which drives the gas turbine to complete the start-up or shutdown process according to the predetermined program, and dynamically adjusts the control prediction command based on real-time parameter feedback during the execution process; Specifically, the control prediction commands of each subsystem are input into the command distributor. A priority scheduling algorithm sorts the control prediction commands according to their urgency and timing requirements, generating an ordered execution command sequence. This ordered execution command sequence is then input into the multimodal actuator system. A signal conversion module converts electrical signals into corresponding mechanical actions, generating the actual execution actions of each gas turbine subsystem. A distributed monitoring network collects feedback signals of the executed actions in real time, including valve opening feedback, guide vane angle feedback, and current feedback, generating an execution status feedback dataset. This dataset is compared and analyzed with the expected execution status. A deviation calculation algorithm generates an execution deviation analysis report, and an adaptive correction strategy is used to generate a control prediction command correction amount. The control prediction command correction amount is fed back to the multi-loop coordinated control system. An online parameter tuning module adjusts the controller's operating parameters in real time, generating the new control prediction commands, which are then re-executed.

[0075] It should be noted that the control prediction commands generated in step 204 (fuel valve opening command 62%, guide vane angle command 52°, excitation current command 1200A) are converted into actual mechanical actions and dynamically adjusted based on feedback during execution.

[0076] The commands are ordered according to safety priority and timing logic: fuel supply has the highest priority, followed by guide vane adjustment, and finally excitation control. The generated ordered execution command sequence is as follows: t=0s: fuel control prediction command executed (opening increased to 62%); t=2s: guide vane control prediction command executed (angle adjusted to 52°); t=4s: excitation control prediction command executed (current increased to 1200A). The sequence is sent to the multimodal actuator system. The electrical signals drive each actuator through the signal conversion module: the fuel command is converted into the linear displacement of the fuel regulating valve (target opening 62%) by the servo motor; the guide vane command is converted into the rotation angle of the inlet guide vane (target 52°) by the hydraulic actuator; and the excitation command adjusts the current of the generator excitation winding through the power amplifier (target 1200A).

[0077] After execution begins, the distributed monitoring network collects feedback signals in real time at 100-millisecond intervals. The feedback data collected at t=5s is compared with the expected value in Table 3 below: Table 3 Deviation calculation algorithm analysis revealed that the actual opening degree of the fuel valve was 2% lower than the commanded value, failing to reach the target flow rate within the expected time. There was also a slight lag in the guide vanes and excitation circuit.

[0078] Based on the deviation analysis report, the adaptive correction strategy was triggered. The system determined that the fuel circuit deviation had exceeded the allowable ±1% tolerance range and needed immediate correction. The correction strategy was as follows: The correction amount was calculated proportionally based on the magnitude and direction of the deviation. The correction amount for the fuel valve opening command was: .

[0079] Meanwhile, the online parameter tuning module fine-tuned the integral time of the PID controller based on the actuator's response speed to accelerate subsequent responses. At t=5.5s, the subsequent control prediction commands (fuel valve opening 63.6%, guide vane angle 52.5°, excitation current 1210A) were reissued for execution. By t=7s, feedback was collected again, showing that the actual values ​​were consistent with the commands, and the gas turbine smoothly accelerated to 2500rpm, completing this stage of the start-up task. The entire execution-monitoring-correction cycle was completed within 2.5 seconds, ensuring control accuracy.

[0080] 206. A distributed sensor network is used to collect vibration spectrum, acoustic emission signals, and temperature field data during gas turbine operation in real time, generating a multi-source monitoring dataset of gas turbine operating status. This dataset is then input into a physical model-based fault prediction system. A component failure mechanism analysis algorithm is used to identify potential fault characteristics and generate a fault warning level assessment report. This report is then input into a control strategy reconfiguration system. A pre-set matching algorithm selects the corresponding emergency control scheme from a pre-set control strategy library, generating an adaptive control strategy adjustment command. This command is input into the human-machine interface, where a visualization module presents key warning information and recommended operating schemes to operators, generating a confirmation execution command. Finally, the confirmation execution command is input into the control prediction command execution system, where a safety interlock protection mechanism ensures smooth switching of the control strategy, completing adaptive control under abnormal operating conditions.

[0081] It should be noted that when the GT-100 gas turbine speed reaches 2800 rpm, the equipment status is actively monitored, and safety adaptive control is implemented when an abnormality is detected.

[0082] A distributed sensor network acquires operational data at a frequency of 500 milliseconds. Vibration sensors detected that the high-pressure turbine side vibration acceleration increased from the normal value of 2.5 m / s². 2 The speed suddenly increased to 4.8 m / s 2 Spectral analysis showed a significant enhancement in the 200Hz component; the event count rate collected by the acoustic emission sensor at the bearing housing increased from 50 times / second to 120 times / second; and the infrared thermal imager showed a localized high-temperature zone on the third-stage turbine blade, with a temperature reaching 620℃ (45℃ higher than the adjacent blades). These data constitute a multi-source monitoring dataset.

[0083] After receiving data, the fault prediction system based on a physical model diagnoses the problem using a component failure mechanism analysis algorithm. The 200Hz component of the vibration spectrum matches the blade's passing frequency. Combined with localized high-temperature characteristics, the system identifies a potential fault as "localized overheating of the third-stage blade may lead to thermal fatigue cracks." An increase in acoustic emission events further corroborates the possibility of microcracks in the material. The system generates a fault warning level assessment report, comprehensively evaluating the warning level as "high risk."

[0084] After receiving a high-risk warning, the control strategy reconfiguration system selects a corresponding solution from the preset strategy library using a pre-defined matching algorithm. The matching principles include: fault type (blade overheating), severity (high risk), and current operating condition (startup phase, 2800 rpm). The system selects the "load reduction to mitigate thermal stress" solution and generates an adaptive control strategy adjustment command: temporarily maintain the target speed at 2800 rpm, reduce fuel flow by 5%, and lower the exhaust temperature control upper limit by 20°C.

[0085] The adjustment instructions and early warning information are displayed visually to operators through a human-machine interface. The interface clearly displays: alarm location (third-stage blade of the high-pressure turbine), early warning level (high risk, indicated in red), and current parameters (vibration 4.8 m / s²). 2 (Local temperature 620℃) and recommended operation (execute load reduction plan). After reviewing the data within 10 seconds, the operator clicks the confirmation button to generate a confirmation execution command.

[0086] The system confirms the execution of control commands and predictive command execution, with a safety interlock protection mechanism ensuring a smooth switching of control strategies. The system first verifies that the gas turbine is in a controllable state (stable speed, no other alarms), then completes the switching within 5 seconds according to a preset sequence: first, fine-tuning the guide vane angle to reduce the intake air volume; then, reducing the fuel flow rate by adjusting the slope; and finally, adjusting the cooling air flow rate. The entire process ensures no sudden jumps or surges.

[0087] Monitoring after the switch showed that the vibration value dropped to 3.0 m / s within 30 seconds. 2 The localized temperature of the blades dropped to 590℃, and the warning level was downgraded to "medium risk." The system continued to monitor until the risk was eliminated, after which it prompted operators to choose to restore the original startup procedure. This adaptive control intervention avoided potential equipment damage, and the entire process was completed automatically in less than 1 minute.

[0088] In this embodiment of the invention, a distributed sensor network is used to deploy various types of sensors at key parts of the gas turbine to comprehensively collect real-time operating parameters such as speed, temperature, pressure, and vibration. Data is collected once per second to ensure timeliness and accuracy, providing a reliable foundation for subsequent analysis. A multi-dimensional evaluation system covering lifespan assessment and operational efficiency assessment is constructed. The lifespan assessment model is based on creep and fatigue interaction theory, combined with the material microstructure evolution theory, to accurately calculate the lifespan loss rate of each key component in real time. The efficiency calculation model is based on the second law of thermodynamics, using a dynamic analysis algorithm to quantify the efficiency loss of each thermodynamic process, comprehensively evaluating the gas turbine's operating efficiency. With the dual objectives of minimizing lifespan loss and improving operational efficiency, a Pareto front analysis algorithm is used to generate a candidate set of control strategies. Then, a preliminary optimal control sequence is calculated through a dynamic programming-based state transition model, and adjustments are made using constraint satisfaction systems and sensitivity analysis modules to ensure that the generated control sequence meets safe operation requirements. While meeting the limitations of boundary conditions and equipment physical constraints, it possesses good robustness; the multi-loop coordinated control system adopts an intelligent coordination strategy based on response characteristic matching. Through multiple links such as control sequence parser, adaptive PID controller, feedforward compensator, cross-decoupling controller and output limiting module, it accurately converts the start-stop control prediction sequence into control prediction commands for each subsystem and eliminates the mutual influence between control loops, ensuring that the control prediction commands are accurate and executable; through a distributed sensor network, it collects multi-source monitoring data such as vibration spectrum, acoustic emission signal and temperature field data during the operation of the gas turbine in real time, providing rich information for fault prediction; The physical model-based fault prediction system accurately identifies potential fault characteristics using component failure mechanism analysis algorithms and generates a fault warning level assessment report. The control strategy reconfiguration system selects the corresponding emergency control scheme from a preset control strategy library based on the warning level, generates adaptive control strategy adjustment instructions, and allows operators to confirm execution through a human-machine interface. A safety interlock protection mechanism ensures smooth switching of the control strategy, achieving adaptive control under abnormal operating conditions, effectively avoiding potential equipment damage, and ensuring the safe operation of the gas turbine.

[0089] Figure 4 This is a schematic diagram of a one-button start / stop control device for a single-shaft gas turbine according to an embodiment of the present invention. The one-button start / stop control device 300 for a single-shaft gas turbine may include a processor 301 and a memory 302. The memory 302 is used to store program instructions and / or data, and the processor 301 is used to execute the program instructions stored in the memory 302, thereby implementing the method described in the above-described method embodiment.

[0090] Optionally, the memory 302 and the processor 301 are coupled. The coupling is an indirect coupling or communication connection between devices, units, or modules, and can be electrical, mechanical, or other forms, for information interaction between devices, units, or modules.

[0091] Optionally, the single-shaft gas turbine one-button start / stop control device 300 may further include a communication interface 303. The communication interface 303 is used to communicate with other devices via a transmission medium, for example, transmitting received signals from other communication devices to the processor 301, or transmitting signals from the processor 301 to other communication devices. The communication interface 303 may be a transceiver or an interface circuit, such as a transceiver circuit or a transceiver chip.

[0092] This application embodiment does not limit the specific connection medium between the processor 301, memory 302, and communication interface 303. This application embodiment... Figure 4 The processor 301, memory 302, and communication interface 303 are connected via a bus 304. Figure 4 The connections between other components are shown in bold and are for illustrative purposes only, not as limiting information. The bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0093] The present invention also provides a one-button start-stop control device for a single-shaft gas turbine. The one-button start-stop control device for a single-shaft gas turbine includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor performs the steps of the one-button start-stop control system for a single-shaft gas turbine in the above embodiments.

[0094] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the single-shaft gas turbine one-button start-stop control system.

[0095] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0096] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0097] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A one-button start / stop control system for a single-shaft gas turbine, characterized in that, include: The acquisition module is used to collect gas turbine operating parameters in real time and generate a gas turbine state parameter set through the state observer; The evaluation module is used to calculate the equipment life loss rate under the current operating conditions based on the gas turbine state parameter set, generate life loss evaluation results, and generate current operating efficiency evaluation results. The prediction module is used to generate a start-stop control prediction sequence based on the life loss assessment results and the operating efficiency assessment results, with the dual objectives of minimizing life loss and improving operating efficiency.

2. The single-shaft gas turbine one-button start / stop control system according to claim 1, characterized in that, include: The rotor thermal stress distribution and temperature gradient data from the gas turbine state parameter set are input into the life assessment model. The real-time life loss rate of each key component is calculated through the material damage accumulation algorithm, and a life loss rate distribution map is generated. The combustion chamber temperature field distribution and exhaust temperature gradient data from the gas turbine state parameter set are input into an efficiency calculation model based on the second law of thermodynamics to calculate the efficiency loss of each thermodynamic process and generate an efficiency loss distribution map. The life loss rate distribution map and efficiency loss distribution map are input into the dynamic weight allocation system, and life assessment weight coefficients and efficiency assessment weight coefficients are generated according to the current operating conditions and operating objectives. The life loss rate distribution map is weighted based on the life assessment weight coefficient to generate a standardized life loss assessment result. At the same time, the efficiency loss distribution map is weighted based on the efficiency assessment weight coefficient to generate a standardized efficiency assessment result. The standardized life loss assessment results and efficiency assessment results are fused together to generate the current operating efficiency assessment results.

3. The single-shaft gas turbine one-button start / stop control system according to claim 2, characterized in that, The metallographic structure characteristics of key parts of the rotor are collected by an online material condition monitoring system to generate a real-time material microstructure dataset. A micro-damage evolution map is generated based on the real-time material microstructure dataset; The micro-damage evolution spectrum is coupled with the operating load spectrum for analysis to calculate the life loss rate under the current operating condition and generate a real-time life loss rate distribution map. The real-time life loss rate distribution map is input into the remaining life prediction system, and combined with historical operating data and material fatigue characteristics, a remaining life prediction report for key components is generated. The life prediction control parameters are obtained based on the remaining life prediction report.

4. The single-shaft gas turbine one-button start / stop control system according to claim 2, characterized in that, It also includes an allocation module: A set of control strategy candidates is generated based on the life loss assessment results and the operating efficiency assessment results; The candidate set of control strategies is input into a state transition model based on dynamic programming. The optimal control strategy is calculated step by step from the final state to the initial state through a reverse recursive algorithm to generate a preliminary optimal control sequence. The preliminary optimal control sequence is input into the constraint-satisfying system, and combined with the gas turbine safe operation boundary conditions and equipment physical limitations, a feasible control sequence that satisfies all constraints is generated. Based on the feasible control sequence, the sensitivity of the control sequence to each state parameter is evaluated through parameter perturbation analysis, and a control sequence is generated; The start-stop control prediction sequence is formed by extracting the ramp rate setpoint, load rate setpoint, and temperature change rate setpoint from the control sequence.

5. The single-shaft gas turbine one-button start / stop control system according to claim 4, characterized in that, Real-time collection of atmospheric pressure, ambient temperature and humidity parameters to generate an environmental condition feature dataset; Based on the environmental condition feature dataset, the ramp rate and temperature change rate in the control sequence are modified for environmental conditions to generate an environmentally adaptive control sequence. Based on the environmental adaptive control sequence, the optimal operating mode is identified according to the current power grid demand and unit status, and a mode switching decision instruction is generated; The mode switching decision command is input into the economic assessment model, and economic assessment reports under different modes are generated through fuel cost calculation and efficiency analysis. Based on the aforementioned economic assessment report, a multi-objective trade-off algorithm is used to seek the optimal balance between lifetime loss and economic efficiency, generating the final control prediction sequence.

6. The single-shaft gas turbine one-button start / stop control system according to claim 4, characterized in that, include: Based on the start-stop control prediction sequence, the acceleration rate setpoint, load rate setpoint, and temperature change rate setpoint are mapped to the corresponding control loops through the instruction decomposition algorithm, generating fuel control loop setpoints, guide vane control loop setpoints, and excitation control loop setpoints. The setpoints of each loop are input into an adaptive PID controller based on a gain scheduling strategy. By detecting the deviation and operating conditions of each loop in real time, the proportional, integral and derivative parameters of the PID controller are dynamically adjusted to generate preliminary control prediction commands. Based on the preliminary control prediction command and the gas turbine dynamic characteristic model, feedforward compensation is performed on the measurable disturbance to generate the compensated control prediction command. Based on the compensated control prediction command, the mutual influence between each control loop is eliminated through a multivariable decoupling algorithm to generate a decoupled coordinated control prediction command. Based on the coordinated control prediction instructions and the physical limitations and safe operation requirements of each actuator, the amplitude and rate of the control prediction instructions are limited to generate control prediction instructions for each subsystem.

7. The single-shaft gas turbine one-button start / stop control system according to claim 6, characterized in that, By real-time detection of deviations and operating conditions in each loop, the proportional, integral, and derivative parameters of the PID controller are dynamically adjusted, with the proportional coefficient Kp adjusted according to the magnitude of the deviation. in, The baseline value is α, and the adaptive coefficient is α. This represents the real-time deviation.

8. The single-shaft gas turbine one-button start / stop control system according to claim 6, characterized in that, The response characteristics of each actuator are tested online to generate a set of dynamic characteristic parameters for the actuator; The timing sequence of control prediction commands for each loop is calculated based on the actuator dynamic characteristic parameter set, and a control prediction command sequence is generated. The control prediction command sequence is input into the dynamic compensator to establish a feedforward compensation model and generate the coordinated control prediction command after dynamic compensation. Based on the dynamically compensated coordinated control prediction command, the coupling degree between each loop is quantified, and a decoupling compensation matrix is ​​generated. The decoupling compensation matrix is ​​applied to a multi-loop control system. Through decoupling operations, mutual interference between control loops is eliminated, and completely decoupled independent control prediction commands are generated.

9. The one-button start / stop control system for a single-shaft gas turbine according to claim 1, characterized in that, It includes an execution module, used to convert the control prediction commands into specific operations through an actuator, drive the gas turbine to complete the start-up or shutdown process according to a predetermined program, and dynamically adjust the control prediction commands based on real-time parameter feedback during execution. Based on the control prediction instructions of each subsystem, the control prediction instructions are sorted according to their urgency and timing requirements using a priority scheduling algorithm to generate an ordered sequence of execution instructions; The ordered sequence of execution instructions is input into the multimodal actuator system, and the electrical signals are converted into corresponding mechanical actions by the signal conversion module to generate the actual execution actions of each subsystem of the gas turbine. The distributed monitoring network collects feedback signals of the executed actions in real time, including valve opening feedback, guide vane angle feedback and current feedback, and generates an execution status feedback dataset. The execution status feedback dataset is compared and analyzed with the expected execution status. An execution deviation analysis report is generated through a deviation calculation algorithm, and a control prediction instruction correction amount is generated based on an adaptive correction strategy. The control prediction command correction amount is fed back to the multi-loop coordinated control system, and the controller's operating parameters are adjusted in real time through the online parameter tuning module. The generated control prediction command is then re-executed.

10. The one-button start / stop control system for a single-shaft gas turbine according to claim 1, characterized in that, It also includes an early warning module: By collecting vibration spectrum, acoustic emission signal and temperature field data in real time during the operation of the gas turbine through a distributed sensor network, a multi-source monitoring dataset of the gas turbine's operating status is generated. The multi-source monitoring dataset is input into a physical model-based fault prediction system, and potential fault characteristics are identified through component failure mechanism analysis algorithms to generate a fault early warning level assessment report. The fault warning level assessment report is input into the control strategy reconstruction system, and the corresponding emergency control scheme is selected from the preset control strategy library through the pre-plan matching algorithm to generate an adaptive control strategy adjustment instruction; The adaptive control strategy adjustment command is input into the human-machine interface, and the key early warning information and recommended operation plan are presented to the operators through the visualization module, generating a confirmation execution command. The confirmation execution command is input into the control prediction command execution system, and the smooth switching of the control strategy is ensured through a safety interlock protection mechanism to complete adaptive control under abnormal operating conditions.

Images