Turbine cold start sliding parameter intelligent regulation method and system based on multi-parameter collaborative optimization

Abstract

This invention proposes an intelligent control method and system for sliding parameters of steam turbines during cold start based on multi-parameter collaborative optimization. The method includes: acquiring the upper and lower cylinder temperatures on the cylinder surface during the initial stage of cold start of the steam turbine to calculate the thermal inertia index and temperature rise symmetry factor, which serve as the current state; determining whether the current state meets the conditions for restart based on the current thermal inertia index and temperature rise symmetry factor; if the conditions for restart are met, entering the recommended state, and generating recommended sliding parameters under the current operating conditions by combining historical samples with weighted matching; constructing the target acceleration path of the steam turbine during cold start based on the recommended sliding parameters under the current operating conditions, and dynamically generating the valve control path; collecting actual feedback signals, and correcting the target acceleration path and valve control path in real time to form a closed-loop control. This invention significantly enhances the system's safety and environmental adaptability while ensuring control performance.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of intelligent control of sliding parameters for cold start of steam turbine, and particularly relates to an intelligent control method and system for sliding parameters for cold start of steam turbine based on multi-parameter collaborative optimization. BACKGROUND

[0002] With the increasing requirements of power system on rapid start-stop and peak regulation response capability of units, gas-steam combined cycle units are facing more and more frequent cold start tasks in actual operation, especially under the background of rising natural gas cost, unstable operation load and increasing start-stop frequency. Improving the control accuracy and system safety during the cold start stage of steam turbine has become a key problem in the industry. When the steam turbine starts under cold state, the high-pressure inner cylinder wall temperature is usually lower than 200℃, and the thermal inertia difference between the cylinder and the rotor is significant. If the main steam parameter control is unreasonable, it will easily induce severe thermal expansion difference changes, shaft vibration amplification, thermal stress concentration, and even cause cylinder cracking or trip accidents. Currently, the industry generally uses fixed templates to set the main steam pressure, temperature and superheat degree sliding parameters, such as setting the main steam door front pressure to 2.0 to 2.5 MPa, the main steam temperature to 300 to 350℃, and executing the rush rotation when the superheat degree is higher than 50℃, and at the same time, completing the speed-up process at a constant rate in sections. Although this method is operable, it relies heavily on empirical parameters and lacks awareness of the actual thermal response capability of the cylinder, and cannot dynamically match the thermal balance process under different initial cooling states or structural asymmetry. The speed-up path also does not fully consider the actual physical limitations such as resonance interval avoidance and valve execution lag.

[0003] In addition, most control strategies are one-way execution, only the sliding parameter setting and path planning are set in advance, and there is a lack of real-time feedback and closed-loop correction mechanism at the structural response level, which leads to the inability to correct deviations in time when the expansion difference changes, shaft vibration amplification or valve error occurs. Therefore, the existing control method cannot balance the dynamic balance between start-up speed and equipment protection, especially under non-standard operating conditions, which easily causes control instability, speed-up lag or system failure, and there is an urgent need to build a sliding parameter intelligent control method that is multi-parameter collaborative, structure-aware, control path adaptive and has execution feedback capability. SUMMARY

[0004] The purpose of the present application is to provide an intelligent control method and system for sliding parameters for cold start of steam turbine based on multi-parameter collaborative optimization, which solves the above problems.

[0005] In order to achieve the above purpose, in the first aspect of the present application, an intelligent control method for sliding parameters for cold start of steam turbine based on multi-parameter collaborative optimization is provided, which comprises the following steps:

[0006] S1, obtaining the upper cylinder temperature and the lower cylinder temperature of the cylinder surface at the initial stage of the cold start of the steam turbine to calculate a thermal inertia index and a temperature rise symmetry factor as a current state;

[0007] S2, judging whether the current state has met a rush rotating condition according to the current thermal inertia index and the temperature rise symmetry factor; if the rush rotating condition is met, entering a recommended state, and generating a recommended sliding parameter under the current working condition by combining a historical sample weighted matching; wherein the sliding parameter comprises a main steam pressure, a main steam temperature and a superheat value;

[0008] S3, constructing a target speed-up path of the steam turbine in the cold start process based on the recommended sliding parameter under the current working condition, and dynamically generating a governing valve control path;

[0009] S4, collecting actual feedback signals, and correcting the target speed-up path and the governing valve control path in real time to form a closed-loop control.

[0010] Further, the S1 comprises:

[0011] collecting the upper cylinder temperature and the lower cylinder temperature; wherein the error of the upper cylinder temperature and the lower cylinder temperature after field sampling point calibration is not more than ±1℃;

[0012] combining the upper cylinder temperature and the lower cylinder temperature with a sliding window time interval, and generating a thermal inertia index by taking a difference value to realize a first-order derivative approximation, for describing the heat absorption rate per unit time of the cylinder body;

[0013] and calculating a temperature rise symmetry factor based on the upper cylinder temperature and the lower cylinder temperature, for representing the degree of thermal distribution asymmetry.

[0014] Further, the judging whether the current state has met the rush rotating condition according to the current thermal inertia index comprises:

[0015] if the current thermal inertia index is greater than or equal to a minimum thermal response threshold of the cylinder body, and the temperature rise symmetry factor is less than or equal to a maximum temperature difference asymmetry allowance, the rush rotating condition is met, and a recommended process is performed.

[0016] Further, if the rush rotating condition is met, entering the recommended state, and generating the recommended sliding parameter under the current working condition by combining the historical sample weighted matching, comprises:

[0017] collecting the current thermal inertia index and the current temperature rise symmetry factor that meet the rush rotating condition, and combining a main steam pressure value, a main steam temperature value and a main steam superheat degree to generate a successful rush rotating sample five-tuple;

[0018] For the current state, a recommended sliding parameter is generated based on the five-tuple of the successful rollover samples using a weighted average method; wherein, the weights of the weighted average are calculated based on the current thermal inertia index and temperature rise symmetry factor, as well as the historical thermal inertia index and temperature rise symmetry factor.

[0019] Furthermore, the main steam superheat is the difference between the main steam temperature and the saturation temperature.

[0020] Further, S3 includes:

[0021] The system receives the recommended sliding parameters under the current operating conditions, combines the turbine target speed from past time steps with the critical resonance prediction factor, and iteratively updates the target speed sequence to ensure real-time matching with the cylinder block thermal response capability, thereby generating the target acceleration path.

[0022] Based on the turbine target speed, main steam pressure, and main steam temperature, a control path for the regulating valve is generated using a static steam flow-speed relationship model function.

[0023] The control path for the regulating valve also incorporates a regulating valve inertia compensation function, where is an empirical model fitting term, and the fitting result is:

[0024] If the valve inertia compensation function = 1, it indicates that the execution is synchronized with the setting;

[0025] If the valve inertia compensation function is less than 1, it means that the actual valve movement is lagging, and the acceleration slope needs to be slowed down synchronously.

[0026] Furthermore, the critical resonance prediction factor is calculated based on the current actual turbine speed and vibration-sensitive speed points identified during historical operation.

[0027] Furthermore, the static steam flow-speed relationship model function is calculated based on the target speed-speed path, main steam pressure, and main steam temperature to determine the corresponding valve opening. It is established from the turbine thermodynamic characteristic curve and implemented using a bivariate fitting table.

[0028] Further, S4 includes:

[0029] The system collects current rotational speed, current valve opening, current expansion differential, and current shaft vibration, performs state fitting, and calculates the time-dependent changes using the structural control deviation function. The structural control comprehensive deviation index at any given time is used to measure the degree of difference between the target control path and the actual execution path and structural response;

[0030] When the overall deviation index of the structure control exceeds the safety threshold, the path fine-tuning or valve response lag correction is performed in combination with the target acceleration path and the valve control path to release the valve execution pressure synchronously and avoid thermal shock to the cylinder.

[0031] A second aspect of the present invention provides an intelligent control system for sliding parameters of a steam turbine during cold start-up based on multi-parameter collaborative optimization, the system comprising:

[0032] The multi-parameter acquisition unit is used to acquire the upper cylinder temperature and lower cylinder temperature of the cylinder surface during the initial cold start of the steam turbine, in order to calculate the thermal inertia index and temperature rise symmetry factor as the current state.

[0033] The multi-parameter analysis unit is used to determine whether the current state meets the conditions for restart based on the current thermal inertia index and temperature rise symmetry factor. If the conditions for restart are met, the unit enters the recommended state and generates recommended sliding parameters for the current operating condition by combining historical samples with weighted matching. The sliding parameters include: main steam pressure, main steam temperature and superheat value.

[0034] The intelligent control unit is used to construct the target acceleration path of the steam turbine during cold start-up based on the recommended sliding parameters under the current operating conditions, and dynamically generate the control path of the control valve.

[0035] The control and optimization unit is used to collect actual feedback signals and correct the target acceleration path and the control path of the valve in real time to form a closed-loop control.

[0036] The beneficial technical effects of the present invention are at least as follows:

[0037] This invention proposes an intelligent control method for sliding parameters during cold start of a steam turbine based on a closed-loop linkage of state perception, parameter recommendation, path planning, and structural feedback. By constructing thermal inertia indices and temperature rise symmetry factors through the evolution of the temperature difference between the upper and lower walls of the cylinder, the method dynamically perceives the cylinder's heat absorption capacity and temperature rise uniformity, thereby determining whether safe start-up conditions are met and recommending a suitable combination of sliding parameters. Based on this, a speed-up path and valve control strategy are dynamically constructed in conjunction with the recommended main steam parameters. By introducing a nonlinear speed-up slope control function, a thermal overload enhancement term, a vibration prediction and suppression term, and a valve hysteresis compensation factor, a target path with physical consistency and thermal safety is generated. Subsequently, by real-time acquisition of feedback data such as actual speed, valve opening, expansion difference, and shaft vibration, a control deviation function integrating structural response is constructed, and the future path is locally adjusted to achieve joint optimization of speed-up control and structural protection. Compared to traditional solutions that only focus on path setting and parameter templates, this invention constructs a complete control closed loop of state recognition, parameter recommendation, path construction, and execution feedback. This enables sliding parameter control to have cross-layer data-driven and structural state correction capabilities for the first time, significantly enhancing the system's safety and environmental adaptability while ensuring control performance. Attached Figure Description

[0038] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0039] Figure 1 This is a flowchart of the intelligent control method for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization, according to the present invention. Detailed Implementation

[0040] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0041] like Figure 1 As shown in the embodiment of the present invention, the intelligent control method for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization includes:

[0042] S1. Obtain the upper and lower cylinder temperatures on the cylinder surface during the initial cold start of the steam turbine to calculate the thermal inertia index and temperature rise symmetry factor, which serve as the current state.

[0043] Specifically, this step aims to construct two structural indices based on the cylinder surface temperature changes during the initial cold start of the steam turbine: overall thermal inertia. and temperature rise symmetry factor These two indicators characterize the current thermal response and thermal distribution of the cylinder block. They will serve as the basis for determining the recommended sliding parameter combinations and control strategy paths in subsequent steps. This method avoids using static pressure and temperature as the basis for start-up judgment, instead using the actual dynamic thermal state of the cylinder block as the core judgment basis. Input:

[0044] This step inputs data from two sets of high-precision temperature sensors on the cylinder surface, as detailed below:

[0045] Upper cylinder temperature The temperature is collected in real time by a K-type thermocouple located on the upper wall of the cylinder (such as the upper flange of the front section of the first stage stator vane), and the signal is connected to the DEH temperature acquisition module with a sampling period of 1 second.

[0046] Lower cylinder temperature The data is collected in real time by thermocouples located symmetrically at the bottom of the cylinder (such as the flange below the center of the base) and synchronously connected to the DEH module.

[0047] The temperature value, after being calibrated at the on-site sampling points, has an error of no more than ±1℃. The sampling data is input into the temperature rise identification submodule in the controller in the form of a floating-point time series.

[0048] During the cold start phase of a steam turbine, the thermal response capability of the cylinder block metal structure is a core indicator for assessing the safety of the turbine during startup. Therefore, the "thermal inertia index" is proposed. The formula describing the heat absorption rate per unit time of the cylinder block is as follows:

[0049]

[0050] in, The sliding window time interval is fixed at 30 seconds. and These are the current temperatures of the upper and lower cylinder walls, respectively, both being time-series data. The formula uses the difference value to achieve a first-order derivative approximation, which is used to calculate the heating slope.

[0051] The physical meaning of this value is the average heating rate of the upper and lower surfaces of the cylinder block. The higher the value, the greater the heat absorption rate of the cylinder block. This value can be used to determine whether there is sufficient thermal buffer for the start-up.

[0052] However, structural differences in temperature may exist between the upper and lower walls of the cylinder block, especially during high-load startup, where temperature differences can easily concentrate in the upper or lower cylinder, causing localized structural expansion differences. Therefore, a "temperature rise symmetry factor" is introduced. "Indicates the degree of asymmetry in heat distribution, defined as:

[0053]

[0054] The larger the value of this factor, the more the cylinder temperature rise is biased to one side. If it exceeds the threshold (such as 0.1), it may cause a sudden change in expansion difference or thermal stress concentration in subsequent acceleration. The structural design facilitates subsequent controller adjustments to the damper path, achieving balanced heat distribution.

[0055] In actual engineering, if For 138, If it is 118, then , This indicates that the current cylinder block heating rate is relatively high and the heat distribution is basically symmetrical, which meets the prerequisite for entering the start-up judgment stage.

[0056] Both formulas rely on actual data from field sensors, requiring no complex modeling, resulting in low computational cost, good real-time performance, and easy embedded deployment and DCS linkage.

[0057] Final output: Thermal inertia index : Represents the overall temperature rise response rate of the cylinder structure; temperature rise symmetry factor : Indicates the symmetry of the cylinder block's heat load distribution.

[0058] S2. Determine whether the current state meets the conditions for restart based on the current thermal inertia index and temperature rise symmetry factor; if the conditions for restart are met, enter the recommended state, and generate the recommended sliding parameters under the current operating conditions by combining historical samples with weighted matching; wherein the sliding parameters include: main steam pressure, main steam temperature and superheat value.

[0059] Specifically, the core objective of this step is to dynamically generate a combination of sliding parameters that closely matches the current thermal boundary conditions under cold start conditions, based on the cylinder block's thermal response. This combination includes main steam pressure, main steam temperature, and superheat, serving as the basis for determining the control path before revving. Unlike traditional methods that set revving parameters based on empirical values, this step proposes a sliding parameter generation mechanism driven by physical states, directly using the cylinder block's thermal inertia... and temperature rise symmetry factor Using this as input and combining it with historical successful startup samples, a real-time mapping model is constructed to recommend the sliding parameter that offers the best thermal safety and temperature rise adaptability under the current conditions.

[0060] During cold starts of steam turbines, the thermal response of the cylinder block exhibits significant individual differences. The structural temperature rise capacity and temperature difference distribution vary depending on the season and the duration of the previous shutdown. Continuing to use a fixed template to set start-up parameters can easily lead to two types of problems: first, premature start-up when the cylinder block temperature rise is lagging, causing thermal shock; second, start-up delay when the temperature has reached a stable state but the parameters still do not meet the template. Therefore, this paper proposes a "state-aligned" sliding parameter recommendation mechanism, which consists of two core parts: first, safety window identification; and second, state mapping recommendation.

[0061] Furthermore, firstly, determine whether the current thermal state has met the reversibility condition, and construct the following physical safety constraint function:

[0062]

[0063] in This is the minimum thermal response threshold for the cylinder block, and a setting of 0.4 is recommended, meaning that the average temperature rise per minute should be at least 12°C. The maximum allowable temperature difference asymmetry is empirically set to 0.1. Judgment result. If it indicates that the recommended process has begun, otherwise the process will remain in the preheating state and return to step one to continue monitoring.

[0064] Once the recommended state is entered, a nearest neighbor mapping model based on state space is constructed using successful restart samples recorded in the plant's historical startup database as references. The sample structure is a quintuple:

[0065]

[0066] in, Indicates the first historical sample The thermal inertia index at the time of the first start-up reflects the average temperature rise rate of the entire cylinder at that time; This indicates the symmetry deviation of the cylinder block temperature rise under this sample, and characterizes the uniformity of the temperature distribution between the upper and lower cylinder walls; This is the main steam pressure value during that initial spool-up cycle; This is the main steam temperature value; It is the main steam superheat (i.e., the difference between the main steam temperature and the saturation temperature). To successfully reverse the sample quintuple.

[0067] Regarding the current state Using all historical samples as a reference, a weighted average is used to recommend sliding parameter combinations, as follows:

[0068]

[0069] In the above formula, These are the recommended main steam pressure, main steam temperature, and superheat values ​​under the current operating conditions. For the number of historical samples,

[0070] Among them, the weighting coefficient Defined as:

[0071]

[0072] This formula is one of the key innovations of this invention. It integrates physical state distance and start-up confidence factor :

[0073] - and It represents the variance of thermal inertia and temperature rise symmetry variables in historical samples, and is used to normalize the distance between different physical quantities to a comparable scale, preventing a certain dimension from dominating similarity judgment due to its large numerical range.

[0074] It is the thermal shock score of the historical sample initiation stage, which is derived from posterior statistical indicators (such as the maximum expansion difference and the normalized superposition of the maximum axial amplitude). The larger the value, the greater the risk of the sample initiation process.

[0075] It is a regulatory factor used to balance the weights between state matching and risk exclusion, and a recommended value is between [0.3, 0.6].

[0076] Compared with the traditional minimum distance nearest neighbor method, this formula significantly improves the thermal safety of the actual sliding parameter recommendation, and through... The system controls whether to favor historical samples with low impact risk, and avoids cases with matching physical states but thermal failures from being included in the recommendation results, forming a three-dimensional mapping model of "structural state-sliding parameter-risk weight".

[0077] After the calculation is complete, the recommended sliding parameter triplet is output. This serves as the initial setting point for subsequent acceleration path and tuning strategy planning.

[0078] S3. Based on the recommended sliding parameters under the current operating conditions, construct the target acceleration path of the steam turbine during the cold start process, and dynamically generate the control path of the regulating valve.

[0079] Specifically, the goal of this step is to combine the sliding parameters recommended in step 2. Constructing the target acceleration path of a steam turbine during cold start-up With the control path of the regulating valve opening Unlike traditional linear segmented acceleration schemes, this step proposes an adaptive path planning mechanism based on thermal-vibration state prediction and equipment inertia compensation. This mechanism can dynamically adjust the acceleration slope according to the cylinder's thermal response capability in the initial cold state, and actively slow down and feedforward adjust the valve strategy in the critical vibration speed range. This reduces the risk of thermal shock, decreases the rate of change of expansion difference, and significantly improves the stability and safety margin of the acceleration process.

[0080] Furthermore, to accommodate the thermal inertia hysteresis characteristics of the cylinder block, this step first defines the target acceleration path as a variable slope function. The function must satisfy:

[0081] In the early stages of acceleration, try to match the cylinder block's temperature rise capacity;

[0082] Slow down during the critical speed range (1050~2450 rpm) to avoid triggering resonance;

[0083] Smoothly transition to constant speed at the end of the acceleration phase to ensure balanced temperature rise.

[0084] Based on this, the following path generation function is proposed:

[0085]

[0086] in, Indicates time The target acceleration path calculated by the path generation function at each moment; Indicates time The target turbine speed is calculated by the path generation function at any given time. Based on the basic acceleration slope, according to and Initial settings; Superheat, reflecting the temperature difference between the main steam and the cylinder block, serves as a factor enhancing acceleration potential; The weight for adjusting overheating is empirically set to 0.02; The resonance suppression coefficient is empirically set to 0.04; To control the step size, it should be consistent with the DCS sampling period; The critical resonance predictor is defined below; The path modulation function represents the inertial compensation term of the tuning response, which is obtained by modeling the actuator. ,in, The hysteresis gain coefficient (empirical value 0.1) is based on historical gate response fitting. To adjust the target trajectory, This is the actual trajectory of the tuning gate.

[0087] This path formula transforms the traditional segmented acceleration strategy into a unified acceleration control model that is driven by heat, coupled with resonance control and mechanical hysteresis.

[0088] Among them, resonance predictor It is a structural vibration risk function constructed based on historical vibration response, and its definition is:

[0089]

[0090] in, This indicates vibration-sensitive speed points (e.g., 1670 rpm, 1840 rpm) identified during historical operation, which are automatically extracted from actual shaft vibration records. This indicates the influence width of each critical frequency band, typically set to 50~100 rpm. This function will... When approaching the resonance zone, the output peak value is automatically suppressed, thus forming a "vibration-avoiding deceleration zone" to avoid triggering resonance in the mechanical structure; The current actual speed of the steam turbine

[0091] In addition, to address the actuator lag issue, the control path Introducing the tuning inertia compensation function This function is the fitting term for the empirical model, and the fitting result is:

[0092] This indicates that execution and settings are synchronized;

[0093] This indicates that the actual action of the valve is lagging, and the acceleration slope needs to be slowed down accordingly.

[0094] This can be achieved by identifying the response curve of the control valve offline, or by receiving closed-loop updates from the DCS system in real time.

[0095] Target trajectory of the pitch This can be derived by reverse derivation of the acceleration path:

[0096]

[0097] in, Indicates time The target opening degree of the regulating gate at any given time is a control quantity calculated by the path planning module and sent to the regulating gate actuator. This is a static steam flow-speed relationship model function. The corresponding valve opening is calculated based on the target speed and main steam parameters. This model can be established using the turbine thermodynamic characteristic curves provided by the manufacturer. It is usually implemented using a bivariate fitting table. In time The target acceleration path at each time point is output by the path generation function; The currently recommended main steam pressure is calculated by the sliding parameter recommendation module in step 2; The currently recommended main steam temperature is also calculated from step 2.

[0098] S4. Collect actual feedback signals and correct the target acceleration path and control path in real time to form a closed-loop control.

[0099] Specifically, the control command is input into the on-site DCS system, and the on-site execution data is monitored in real time. It is continuously compared with the target path to calculate the structural deviation index function. The system proactively adjusts future path points before deviations exceed limits to prevent structural thermal shock, uncontrolled shaft vibration, or drastic changes in thermal expansion. Specific details are as follows:

[0100] Furthermore, the control execution module collects the following actual feedback signals from the DCS every second:

[0101] Current speed Feedback is provided by a high-speed tachometer motor, with a sampling accuracy of ±5 rpm;

[0102] Current valve opening Read from the position feedback potentiometer of the control valve, with an error of ±0.5%;

[0103] Current inflation Data is collected by a grating displacement sensor, with a reading interval of 2 seconds;

[0104] Current shaft vibration Data is collected by the bearing housing eddy current sensor and returned in real time at a frequency of 1Hz.

[0105] After performing state fitting on the above signals, the following structural control deviation function is calculated:

[0106]

[0107] in, Indicates time The structural control comprehensive deviation index at any given time is used to measure the degree of difference between the target control path and the actual execution path and structural response; For time The target rotational speed at time 1 is calculated using the path generation function in step 3; For time The actual rotational speed at any given moment is collected in real time by a high-speed shaft speed measuring device. For time The target valve opening at any given time is calculated by the valve trajectory generation model in step 3; For time The actual opening degree of the regulating gate at any given time is obtained from the position sensor of the regulating gate actuator. This is the cylinder expansion difference value, measured by the cylinder displacement sensor. This represents the rate of change of the expansion difference over time, calculated using a sliding time window (current value minus the value 10 seconds ago, then divided by the time interval). For time The shaft amplitude value at time t is acquired by an eddy current displacement sensor; These are empirical weighting coefficients used to balance the weights of different deviation components in the overall deviation; it is recommended to set them to [value missing]. The innovation of this function lies in its comprehensive consideration of speed tracking error, valve execution deviation, and two key structural responses: expansion difference and shaft vibration. It directly incorporates equipment safety into the deviation judgment process of closed-loop control, thereby realizing intelligent correction control based on structural state.

[0108] when Exceeding the safety threshold (Set based on historical operational data statistics, such as) The following adjustment mechanism will be implemented:

[0109] Path fine-tuning: without replanning the entire path or Only in the future Within seconds, the acceleration slope decreases by 10%, meaning the local target path becomes:

[0110]

[0111] This operation is smoothly inserted into the path controller, avoiding controller bounce caused by abrupt changes.

[0112] Regulator response hysteresis correction: If Then to Add feedback delay item The output is from the previously trained pitch lag time predictor, with a typical delay of 1.5 seconds.

[0113] For example:

[0114] Assuming the speed increases to 1800 rpm, , ,and , Rapid rise, calculated The system will automatically adjust subsequent steps. The rate of increase per second has been reduced to 90% of its original speed, and an update has been performed. The delayed trajectory allows for synchronized release of pressure during valve operation, preventing thermal shock to the cylinder block.

[0115] This invention also provides an intelligent control system for sliding parameters of a steam turbine during cold start-up based on multi-parameter collaborative optimization, the system comprising:

[0116] The multi-parameter acquisition unit is used to acquire the upper cylinder temperature and lower cylinder temperature of the cylinder surface during the initial cold start of the steam turbine, in order to calculate the thermal inertia index and temperature rise symmetry factor as the current state.

[0117] The multi-parameter analysis unit is used to determine whether the current state meets the conditions for restart based on the current thermal inertia index and temperature rise symmetry factor. If the conditions for restart are met, the unit enters the recommended state and generates recommended sliding parameters for the current operating condition by combining historical samples with weighted matching. The sliding parameters include: main steam pressure, main steam temperature and superheat value.

[0118] The intelligent control unit is used to construct the target acceleration path of the steam turbine during cold start-up based on the recommended sliding parameters under the current operating conditions, and dynamically generate the control path of the control valve.

[0119] The control and optimization unit is used to collect actual feedback signals and correct the target acceleration path and the control path of the valve in real time to form a closed-loop control.

[0120] Those skilled in the art will 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.

[0121] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or units may be electrical, mechanical, or other forms.

[0122] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion 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 this application. 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.

[0123] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for intelligent control of sliding parameters of steam turbine during cold start based on multi-parameter collaborative optimization, characterized in that, The method includes the following steps: S1. Obtain the upper cylinder temperature and lower cylinder temperature of the cylinder surface during the initial cold start of the steam turbine to calculate the thermal inertia index and temperature rise symmetry factor as the current state. S2. Determine whether the current state meets the conditions for restart based on the current thermal inertia index and temperature rise symmetry factor; if the conditions for restart are met, enter the recommended state, and generate the recommended sliding parameters under the current operating conditions by combining historical samples with weighted matching; wherein the sliding parameters include: main steam pressure value, main steam temperature value and superheat value; S3. Based on the recommended sliding parameters under the current operating conditions, construct the target acceleration path of the steam turbine during the cold start process, and dynamically generate the control path of the regulating valve; S4. Collect actual feedback signals and correct the target acceleration path and control path in real time to form a closed-loop control.

2. The intelligent control method for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization according to claim 1, characterized in that, S1 includes: The temperatures of the upper cylinder and the lower cylinder are collected; wherein, the error of the upper cylinder temperature and the lower cylinder temperature after calibration at the on-site sampling points does not exceed ±1℃; For the upper cylinder temperature and lower cylinder temperature, combined with the sliding window time interval, the first derivative approximation is achieved by taking the difference value, and a thermal inertia index is generated to describe the heat absorption rate of the cylinder per unit time. The temperature rise symmetry factor is calculated based on the upper cylinder temperature and the lower cylinder temperature to represent the degree of heat distribution asymmetry.

3. The intelligent control method for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization as described in claim 1, characterized in that, The step of determining whether the current state meets the conditions for reversibility based on the current thermal inertia index and temperature rise symmetry factor is as follows: If the current thermal inertia index is greater than or equal to the minimum thermal response threshold of the cylinder block, and the temperature rise symmetry factor is less than or equal to the maximum temperature difference asymmetry allowable degree, then the conditions for restarting are met, and the recommended state is entered.

4. The intelligent control method for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization as described in claim 3, characterized in that, If the conditions for reversal are met, the system enters the recommended state, and generates recommended sliding parameters for the current operating condition by combining historical samples with weighted matching. Specifically: Collect the current thermal inertia index and current temperature rise symmetry factor that meet the conditions for successful restart, and combine them with the main steam pressure value, main steam temperature value and superheat value at the current restart time to generate a five-tuple of successful restart samples; For the current state, a recommended sliding parameter is generated based on the five-tuple of the successful rollover samples using a weighted average method; wherein, the weights of the weighted average are calculated based on the current thermal inertia index and temperature rise symmetry factor, as well as the historical thermal inertia index and temperature rise symmetry factor.

5. The intelligent control method for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization according to claim 4, characterized in that, The superheat value is the difference between the main steam temperature and the saturation temperature.

6. The intelligent control method for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization according to claim 1, characterized in that, The S3 includes: The system receives the recommended sliding parameters under the current operating conditions, combines the turbine target speed and critical resonance prediction factor from the historical time step, and iteratively updates the target speed sequence to ensure real-time matching with the cylinder thermal response capability, thereby generating the target acceleration path. Combining the turbine target speed, main steam pressure, and main steam temperature, a control path for the regulating valve is generated using a static steam flow-speed relationship model function. The control path for the regulating valve also incorporates a regulating valve inertia compensation function, which is an empirical model fitting term. The fitting result is as follows: If the valve inertia compensation function = 1, it indicates that the execution is synchronized with the setting; If the valve inertia compensation function is less than 1, it means that the actual valve movement is lagging, and the acceleration slope needs to be slowed down synchronously.

7. The intelligent control method for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization according to claim 6, characterized in that, The critical resonance prediction factor is calculated based on the current actual turbine speed and vibration-sensitive speed points identified during historical operation.

8. The intelligent control method for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization according to claim 6, characterized in that, The static steam flow-acceleration relationship model function calculates the corresponding valve opening based on the target acceleration path, main steam pressure value, and main steam temperature value. It is established by the turbine thermodynamic characteristic curve and implemented using a bivariate fitting table.

9. The intelligent control method for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization according to claim 1, characterized in that, The S4 includes: The system collects current rotational speed, current valve opening, current expansion differential, and current shaft vibration, performs state fitting, and calculates the time-dependent changes using the structural control deviation function. The structural control comprehensive deviation index at any given time is used to measure the degree of difference between the target control path and the actual execution path and structural response; When the overall deviation index of the structure control exceeds the safety threshold, the path fine-tuning or valve response lag correction is performed in combination with the target acceleration path and the valve control path to release the valve execution pressure synchronously and avoid thermal shock to the cylinder. The path fine-tuning refers to adjusting the path without replanning the overall target speed path or target valve opening, only in the future. Within seconds, the acceleration slope decreased by 10%; The valve response lag correction: If the difference between the target valve opening and the current valve opening is greater than 0.3, a feedback delay term is added to the target valve opening. The typical value of the feedback delay term is a 1.5-second delay.

10. A smart control system for sliding parameters of a steam turbine during cold start based on multi-parameter collaborative optimization, characterized in that: The multi-parameter acquisition unit is used to acquire the upper cylinder temperature and lower cylinder temperature of the cylinder surface during the initial cold start of the steam turbine, in order to calculate the thermal inertia index and temperature rise symmetry factor as the current state; The multi-parameter analysis unit is used to determine whether the current state has met the conditions for restart based on the current thermal inertia index and temperature rise symmetry factor. If the conditions for reversal are met, the system enters the recommended state and generates recommended sliding parameters for the current operating conditions by combining historical samples with weighted matching. The sliding parameters mentioned above include: main steam pressure, main steam temperature, and superheat. The intelligent control unit is used to construct the target acceleration path of the steam turbine during cold start-up based on the recommended sliding parameters under the current operating conditions, and dynamically generate the control path of the control valve. The control and optimization unit is used to collect actual feedback signals and correct the target acceleration path and the control path of the valve in real time to form a closed-loop control.

Images