Method and system for predicting and controlling abnormal working conditions of rotational flow superheater

Abstract

The invention provides a rotational flow superheater abnormal working condition predictive control method and a rotational flow superheater abnormal working condition predictive control system. The method comprises the steps that the inlet and outlet pressure of the cyclone superheater is detected in real time; the inlet and outlet pressure and the rotating speed are input into a dynamic expression model, corrected through a reinforcement learning and differential evolution algorithm, of the rotational flow superheater, and the mass flow real-time change trend of the rotational flow superheater is obtained; according to the real-time change trend of the mass flow, predicting the time when the current mass flow reaches an abnormal flow threshold value, and obtaining trigger time, the abnormal flow threshold value comprising an upper limit and / or a lower limit of the normal mass flow; whether the triggering time exceeds a set maximum response time threshold value or not is judged, and if yes, the rotating speed is adjusted according to the triggering time and / or the real-time state and the real-time change trend of the mass flow. According to the abnormal working condition prediction control method and system for the rotational flow superheater, the stability and safety of the rotational flow superheater under the complex working condition are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of automatic control technology, and in particular to a prediction control method and a system for abnormal operating conditions of a swirl superheater. Background Art

[0002] Cyclone superheaters are widely used in various high-efficiency energy conversion equipment to increase the temperature of steam or gas, thereby improving thermal efficiency. However, during the operation of cyclone superheaters, there is often a risk of abnormal operating conditions, especially mass flow fluctuations and pressure abnormalities, which may cause unstable pressure inside the superheater and even cause equipment failure or performance degradation. Therefore, monitoring and warning of abnormal operating conditions of cyclone superheaters are crucial to improving the stability and safety of the system.

[0003] In the prior art, some methods identify abnormal conditions by monitoring the inlet and outlet pressure, temperature, flow rate and other parameters of the superheater. Common control methods include regulation strategies based on traditional PID control. However, these methods often show low prediction accuracy and response speed when faced with complex operating conditions and highly nonlinear systems. In addition, the existing abnormal operating condition detection methods lack adaptive adjustment capabilities and are difficult to effectively respond to rapid changes in operating conditions in real time.

[0004] In short, the traditional PID control algorithm responds slowly to rapidly changing abnormal conditions and cannot adjust the equipment status in the shortest time. In addition, the existing dynamic models are mostly based on simplified assumptions and lack accurate modeling for complex operating environments, resulting in low prediction accuracy. When the operating conditions of the swirl superheater change dramatically, the existing control methods often cannot accurately determine the time and cause of the abnormality, resulting in the abnormal flow threshold being ignored or delayed response.

[0005] These shortcomings make the traditional abnormal operating condition detection and control methods unable to effectively prevent and respond to the failure of the swirl superheater. Therefore, a new control method is urgently needed to achieve more accurate and real-time abnormal operating condition prediction and control. Summary of the invention

[0006] The present invention provides a method and system for predicting and controlling the abnormal operating condition of a swirl superheater. The method utilizes the inlet and outlet pressures and rotation speed of the swirl superheater monitored in real time, a dynamic expression model corrected by reinforcement learning and differential evolution algorithms, and the changing trend of the mass flow rate to accurately predict the triggering time of the abnormal flow threshold, and timely adjust the equipment rotation speed according to the time and the changing trend of the mass flow rate, thereby avoiding abnormal operating conditions of the superheater.

[0007] In a first aspect, the present invention provides a method for predicting and controlling abnormal operating conditions of a cyclone superheater, characterized in that the method comprises:

[0008] Real-time detection of the inlet and outlet pressures of the swirl superheater;

[0009] The inlet and outlet pressures and rotation speeds are input into a dynamic expression model of a swirl superheater corrected by reinforcement learning and differential evolution algorithms to obtain a real-time change trend of the mass flow rate of the swirl superheater;

[0010] Predicting the time when the current mass flow reaches an abnormal flow threshold according to the real-time change trend of the mass flow, and obtaining a trigger time, wherein the abnormal flow threshold includes an upper limit and / or a lower limit of the normal mass flow;

[0011] Determine whether the trigger time exceeds a set maximum response time threshold: If so, adjust the rotation speed according to the trigger time and / or the real-time state and real-time change trend of the mass flow.

[0012] In a second aspect, the present invention further provides a cyclone superheater abnormal operating condition prediction control system, characterized in that the system comprises:

[0013] A detection device, which is used to detect the inlet and outlet pressures of the swirl superheater in real time;

[0014] A flow generation device, which is used to input the inlet and outlet pressures and rotation speed into a dynamic expression model of a swirl superheater corrected by reinforcement learning and differential evolution algorithms to obtain a real-time change trend of the mass flow of the swirl superheater;

[0015] A prediction device, which is used to predict the time when the current mass flow reaches an abnormal flow threshold according to the real-time change trend of the mass flow, and obtain a trigger time, wherein the abnormal flow threshold includes an upper limit and / or a lower limit of the normal mass flow;

[0016] The judgment and adjustment device is used to judge whether the trigger time exceeds the set maximum response time threshold: if so, adjust the rotation speed according to the trigger time and / or the real-time status and real-time change trend of the mass flow.

[0017] The present invention provides a method and system for predicting and controlling abnormal operating conditions of a swirl superheater: first, through a dynamic expression model corrected by reinforcement learning and differential evolution algorithms, combined with actual operating data of the swirl superheater, the dynamic characteristics of the equipment can be accurately described, thereby improving the accuracy of predicting the real-time change trend of the mass flow rate; second, through real-time detection and intelligent prediction, the time of occurrence of abnormal conditions can be identified in advance, and when the maximum response time threshold is exceeded, an alarm is triggered and the speed is adjusted in time to ensure that the equipment can respond quickly and avoid failures; third, by dynamically adjusting the sampling frequency and speed, the control strategy can be automatically optimized according to the real-time state and real-time change trend of the swirl superheater, which has It has strong adaptability and flexibility and can work effectively under different working conditions; fourth, by predicting the trigger time of the abnormal flow threshold, early warning measures can be taken before the abnormal condition occurs, thereby providing the equipment with sufficient buffer time for adjustment and reducing the equipment failure rate; fifth, from detecting the inlet and outlet pressures, obtaining the real-time change trend of the mass flow and the trigger time, to judging whether the trigger time exceeds the set maximum response time threshold, and subsequent response operations, the present invention automates the entire process, reduces dependence on manual experience, thereby reducing human errors, and achieves accurate prediction and avoidance of various abnormal conditions, which not only improves the automation level of the system, but also reduces the complexity of operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0019] Figure 1 It is a flow chart of a method for predicting and controlling abnormal operating conditions of a swirl superheater provided in an embodiment of the present invention;

[0020] Figure 2 It is a block diagram of a swirl superheater abnormal operating condition prediction control system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. SUMMARY OF THE INVENTION

[0023] As mentioned above, the present invention provides a method and system for predicting and controlling abnormal operating conditions of a swirl superheater, which can effectively improve the stability and safety of the swirl superheater under complex operating conditions.

[0024] Exemplary Methods

[0025] Figure 1 : is a flow chart of a method for predicting and controlling abnormal operating conditions of a swirl superheater provided in an embodiment of the present invention, and the embodiment includes the following steps:

[0026] S101: Real-time detection of the inlet and outlet pressures of the swirl superheater.

[0027] Specifically, select a suitable pressure sensor according to the working environment and pressure range of the swirl superheater. The sensor should have the characteristics of high sensitivity, wide range, high temperature resistance, vibration resistance, corrosion resistance, etc. to adapt to the complex environment of the swirl superheater site.

[0028] Then install the pressure sensors on the inlet and outlet pipes of the swirl superheater respectively. It should be noted that the sensors should be installed in the area where the fluid flow is stable to avoid turbulence or vibration interference to ensure accurate measurement.

[0029] Then connect the pressure sensor to a data acquisition unit (such as a PLC, DAQ card, etc.) to ensure that the output signal of the pressure sensor can be read in real time.

[0030] The sampling frequency is set according to the operating characteristics of the swirl superheater. Usually, high-frequency sampling helps to capture rapidly changing pressure fluctuations. The operating characteristics are various dynamic behaviors and performance indicators exhibited by the swirl superheater during actual operation, including pressure fluctuations, temperature changes, flow changes, flow states, load changes, thermal efficiency changes and / or vibration characteristics.

[0031] Specifically, the sampling frequency can be dynamically adjusted according to the operating characteristics monitored in real time. For example, when the pressure fluctuation of the swirl superheater is fast, the sampling frequency is set to more than twice the fluctuation frequency to ensure that the details of the pressure change can be accurately captured; when the pressure fluctuation is slow (such as stable operation or slow change), the sampling frequency is set to a lower value to reduce data redundancy.

[0032] The sampling frequency may also be determined by the operating characteristic historical data.

[0033] Finally, the inlet and outlet pressures of the swirl superheater are detected using the pressure sensor according to the sampling frequency, and the inlet and outlet pressures are transmitted to the data acquisition unit via a data bus or wireless communication technology.

[0034] In addition, a visual interface can be designed in the control system or monitoring platform to display the inlet and outlet pressure data collected by the data acquisition unit in real time. For example, a chart (such as a line chart) can be used to dynamically display the pressure change trend, so that the operator can quickly observe the inlet and outlet pressure changes.

[0035] S102: Inputting the inlet and outlet pressures and rotation speed into a dynamic expression model of the swirl superheater corrected by reinforcement learning and differential evolution algorithms to obtain a real-time change trend of the mass flow rate of the swirl superheater.

[0036] The present invention predicts abnormal operating conditions by mass flow rate, mainly because mass flow rate is a key parameter characterizing the operating state of the cyclone superheater. Monitoring and analyzing mass flow rate is of great significance for timely discovering potential abnormal operating conditions (such as surge, overheating, overpressure, etc.). The specific reasons for selecting mass flow rate as the basis for predicting abnormal operating conditions in the present invention are as follows:

[0037] First, the mass flow directly affects the stability of the system. The mass flow is a measure of the flow of the fluid medium in the swirl superheater and is one of the basic parameters for the operation of the equipment. The mass flow of the fluid is directly related to the thermodynamic process, energy transfer and dynamic characteristics of the airflow of the system. When the mass flow changes abnormally (such as a sudden increase, a sudden decrease or excessive fluctuation), it may cause the swirl superheater to work unstably, thereby causing abnormal operating conditions such as surge, overheating, overpressure, etc.

[0038] Second, the sensitivity of mass flow to pressure fluctuations and surges. Surges are usually caused by factors such as too low air or steam flow and drastic pressure fluctuations. Mass flow is closely related to the kinetic energy of the fluid. Changes in flow rate usually cause changes in pressure, especially when the flow rate fluctuates greatly, which directly affects the operating stability of the system. In a swirl superheater, sudden changes or instability in flow rate often cause a series of phenomena such as pressure fluctuations and temperature changes. Therefore, by real-time monitoring of the real-time change trend of mass flow, potential abnormal risks can be discovered early, and measures can be taken in advance to avoid the occurrence of unstable conditions such as surges.

[0039] Third, the change in mass flow rate reflects the operating status of the equipment. The change in mass flow rate directly reflects the changes in key operating parameters such as the heat load, flow state, and pressure difference of the swirl superheater. For example, when the mass flow rate is low, it may mean that the system is in a low load state, which may cause energy instability; when the mass flow rate is too high, it may cause overheating, overpressure and other operating conditions. Therefore, by analyzing the trend of mass flow rate changes, the operating status of the equipment can be predicted and evaluated, and whether there is a potential risk of failure or abnormal operating conditions can be determined in time, and adjustments can be made through control algorithms.

[0040] Fourth, the time series analysis of mass flow helps to capture periodic and trend changes. Mass flow usually shows certain periodic or trend changes (such as load changes, equipment start and stop, external environment changes, etc.). By performing time series analysis on mass flow data, these laws can be effectively captured, so as to make a prediction of the abnormal working conditions of the equipment. The present invention models the mass flow through time series analysis methods (such as ARIMA, LSTM, etc.), which can extract potential laws and trends, and then predict the changes in mass flow in the future period of time, and timely warn of possible abnormal working conditions.

[0041] Specifically, the dynamic expression model of the swirl superheater in S102 is:

[0042]

[0043]

[0044]

[0045] Where p is the pressure in the static pressure chamber, m is the mass flow of the compressor, kτ is the valve opening of the air storage chamber, ω is the motor speed, a0 is the stagnation sound velocity of the intake air, V p is the volume of the static pressure chamber, A1 is the reference area of ​​the impeller, L c is the air duct length, J is the total inertia of the compressor, τ m is the torque generated by the drive motor, τ c is the torque produced by the compressor, and Ψ(ω,m) is the pressure ratio of the compressor outlet to the inlet air pressure.

[0046] Ψ(ω,m) can be determined by the following formula:

[0047] Among them, η(ω,m)Δh ideal is the actual enthalpy increase of the fluid, γ is the gas specific heat ratio, Ψ(ω,m) is the output characteristic function; p 01 and p 02 Respectively represent the inlet pressure and outlet pressure; T 01 is the steam inlet temperature; Δh ideal is the ideal specific enthalpy transferred to the fluid, σ is the slip coefficient, U2 is the tangential velocity at the impeller outlet, and the efficiency η(m,ω) is a function of the mass flow rate m and the rotation speed w.

[0048] The actual enthalpy increase of the fluid in the formula is η(ω,m)Δh ideal It is necessary to consider the incident loss and friction loss, which can be expressed as follows: η(ω,m)Δh ideal =Δht -Δh i -Δh f -Δh oth (2)

[0049] Where Δh t is the total enthalpy increment, Δh i and Δh f are the enthalpy changes caused by incident loss and friction loss, Δh oth It is the sum of other losses such as gap loss, backflow loss and leakage loss.

[0050] The fluid enthalpy change is calculated using the following expression: Δh t =μr2 2 ω 2 Δh f =k f m 2 Δh oth =εΔh t (3)

[0051] ε is the loss rate, generally 8%, r1 is the average radius of the guide vane, r2 is the impeller radius, k f is the fluid friction constant, μ and α are parameters, determined by the following formula:

[0052] Among them, β 1b and β 2b are the blade inlet angle and outlet angle respectively, ρ1 is the constant stagnation inlet density, and σ is the slip coefficient slightly less than 1.

[0053] By combining formulas 1 to 4, the relationship between the inlet and outlet pressure ratio, inlet flow rate and speed is obtained as follows:

[0054] Formula 5 is a mathematical expression of the dynamic characteristics of the swirl superheater, and its accuracy is crucial to the effective description of the dynamic model. The physical parameters in the compressor mechanism model are usually derived empirically. These physical parameters are theoretically derived based on idealized geometric models (such as standard cylinders), rather than directly derived from the measured data of the actual equipment. Therefore, simply substituting the actual measured parameters into the model cannot guarantee the complete accuracy of the results. In order to improve the accuracy of the model, reinforcement learning and differential evolution algorithms must be introduced to intelligently identify and optimize physical parameters and empirical parameters, combined with actual operating data for continuous correction. This method enables the model to more accurately reflect the true performance of the swirl superheater under different operating conditions.

[0055] Wherein, the reinforcement learning is a dual deep Q network (DDQN).

[0056] The corrected dynamic expression model is obtained through the following process:

[0057] By detecting the swirl superheater in a steady-state operation process, the operating parameters of the swirl superheater are obtained, wherein the operating parameters include fluid parameters and rotation speed;

[0058] Based on the operating parameters, the physical parameters and empirical parameters of the optimized swirl superheater are derived by a dual-depth Q network and a differential evolution algorithm, wherein the physical parameters include the installation angle of the blades, the reference cross-sectional area of ​​the inlet chamber, the average radius of the guide vane and / or the impeller radius, and the empirical parameters include the slip coefficient and / or the fluid friction constant;

[0059] The dynamic expression model is calibrated using the optimized physical parameters and empirical parameters.

[0060] The specific steps of deriving the optimized physical parameters and empirical parameters through the dual deep Q network and differential evolution algorithm are as follows:

[0061] Inputting operating parameters to set initial physical parameters and initial empirical parameters of the swirl superheater;

[0062] Initializing a population and key parameters of a differential evolution algorithm based on the initial physical parameters and the initial empirical parameters, wherein the key parameters include a population size, a scaling factor and / or a crossover rate;

[0063] Extracting state characteristics of the current population in the operation of the differential evolution algorithm, wherein the state characteristics include fitness, population diversity, stagnation and / or function evaluation times, and inputting them into a dual-depth Q network;

[0064] Based on the state characteristics of the current population, the main network in the dual-depth Q network is used to select the optimal mutation strategy;

[0065] Using the optimal mutation strategy to perform mutation operations on individuals in the current population to generate new solutions, and performing fitness calculations on the new solutions to obtain fitness changes, where the fitness changes are the fitness differences between the new solutions and the current optimal solution;

[0066] Calculate the instant reward according to the fitness change, calculate the target Q value using the target network in the dual-depth Q network, the target Q value instant reward and the sum of the future reward calculated by the discount factor, update the main network Q value according to the target Q value, and optimize the strategy based on the updated Q value, the main network Q value is the expected value of the long-term cumulative reward that can be obtained by the current strategy;

[0067] Repeat the optimization process until the convergence condition is met;

[0068] Output optimized physical parameters and empirical parameters of swirl superheater.

[0069] In summary, the combination of reinforcement learning and differential evolution algorithms can adaptively optimize the dynamic model of the swirl superheater. Under different operating conditions, the performance of the swirl superheater may be different. Traditional models based on fixed parameters often cannot accurately reflect the behavior of the equipment under complex conditions. Especially in complex dynamic environments, the model can identify potential abnormal conditions (such as surge, overheating, overpressure, etc.) in advance. Through reinforcement learning, the model can continuously learn, adjust and optimize parameters, so that the dynamic model is more in line with the actual operating conditions and the accuracy is greatly improved. The differential evolution algorithm can efficiently search the model space, optimize key physical parameters and empirical parameters, and further improve the accuracy of the model.

[0070] The reinforcement learning and differential evolution algorithms in the present invention are not only based on fixed physical parameters for modeling, but can also be dynamically corrected based on real-time collected operating parameters. When the equipment changes state or the operating conditions fluctuate, the model can quickly adapt and optimize. This adaptive optimization enables the dynamic model to continuously adapt to new operating conditions during long-term operation, avoiding the problems of model aging and inadaptability in traditional methods.

[0071] S102 specifically includes:

[0072] Through experiments, the corresponding relationship between the compressor flow rate and the inlet and outlet pressure ratio under different speed stable operating conditions is obtained, and based on the corresponding relationship, the optimal empirical parameters in the dynamic expression model are obtained through reinforcement learning and differential evolution algorithms, and the dynamic expression model is corrected using the optimal empirical parameters to achieve accurate prediction under all working conditions;

[0073] The corrected dynamic expression model uses the inlet and outlet pressure ratio, inlet flow speed and shaft power in actual working conditions to calculate the mass flow rate change rate, inlet and outlet pressure ratio change rate and speed change rate of the swirl superheater at any time;

[0074] The mass flow rate of change at any moment is used to perform a time series analysis on the mass flow detected in real time, so as to obtain the real-time change trend of the mass flow, the real-time change trend of the inlet and outlet pressure ratio, and the real-time change trend of the rotation speed.

[0075] As an optional embodiment, the mass flow detected in real time can be collected as time series data, and organized into continuous data points in chronological order to form time series data of mass flow; the time series data of mass flow is preprocessed to remove noise and abnormal data, and the preprocessing methods include but are not limited to data denoising, smoothing and / or missing value interpolation; based on the selected time series analysis method, the mass flow rate of change at any time is used to model the preprocessed time series data of mass flow to obtain the real-time change trend of the mass flow. Among them, according to the characteristics of real-time mass flow data, an appropriate time series analysis method can be selected, such as for flow data with seasonal, trend or periodic changes, the ARIMA model (autoregressive integrated moving average model) is used for analysis; for short-term analysis that does not require too much historical data, exponential smoothing is used; if the flow data has a strong time dependence or nonlinear trend, a deep learning method such as a long short-term memory network (LSTM) can be used.

[0076] The present invention uses time series analysis to obtain the real-time change trend of the mass flow rate, which can not only provide an effective basis for abnormal flow prediction, but also help identify the periodic or trend changes of the swirl superheater in normal operation, and provide important information for early warning and optimized control.

[0077] S103: Predicting the time when the current mass flow reaches an abnormal flow threshold according to the real-time change trend of the mass flow, and obtaining a trigger time, wherein the abnormal flow threshold includes an upper limit and / or a lower limit of the normal mass flow.

[0078] Specifically, the time when the current mass flow reaches the abnormal flow threshold can be predicted by using curve fitting and / or machine learning according to the real-time change trend of the mass flow to obtain the trigger time.

[0079] As an optional embodiment, a suitable fitting model is selected based on the historical mass flow data. For example, linear regression is suitable for the case where the data presents an approximate linear relationship, that is, the mass flow changes are relatively uniform and the change rate is relatively stable; polynomial fitting is suitable for the case where the data changes are relatively complex, there may be fluctuations or irregular changes, but the overall trend can be approximately described by the polynomial curve; exponential fitting is suitable for the case where the data presents an exponential growth or exponential decay trend. 2 The value (coefficient of determination) evaluates the accuracy of the fitted model. 2The closer it is to 1, the better the model fit is. Check the distribution of the residual (the gap between the actual value and the predicted value) to ensure that there is no obvious pattern or systematic error. Then fit the historical mass flow data in the real-time change trend of the mass flow through the selected fitting model to obtain a prediction curve. Finally, use the prediction curve to predict the time when the current mass flow reaches the abnormal flow threshold to obtain the trigger time, such as by analyzing the intersection of the prediction curve and the abnormal threshold to calculate the trigger time.

[0080] As another optional embodiment, key features are extracted from the real-time change trend of mass flow, such as historical flow data, mass flow rate of change, time-related features (such as time period, periodic fluctuations, etc.). The flow data is standardized or normalized so that the machine learning algorithm can better process it. Then, according to the requirements of the task, a suitable machine learning algorithm is selected to predict the time when the flow reaches the threshold, such as linear regression, support vector regression (SVR), decision tree regression, ARIMA, LSTM, etc. Then the historical flow data and its corresponding trigger time are input into the selected machine learning model as a training set. The model is trained using the training set, and the model parameters are adjusted to minimize the prediction error. The prediction ability of the trained machine learning model, such as mean square error or root mean square error, is evaluated using a validation set or a test set. The current mass flow data is predicted using the trained machine learning model to obtain the time when the current mass flow reaches the abnormal flow threshold (i.e., the trigger time). If the model accuracy is not high, the prediction performance can be improved by further optimizing feature selection, adjusting model parameters, or using more data for training.

[0081] Compared with curve fitting and machine learning, if the mass flow rate changes relatively smoothly and the trend is simple, the curve fitting method is preferred because it is easy to calculate and better adapts to regular data; if the mass flow rate fluctuates greatly, there are complex nonlinear patterns, or there are emergencies, the machine learning method is preferred because it has stronger processing capabilities for complex data.

[0082] In addition, the time when the current inlet and outlet pressure ratio reaches the abnormal pressure ratio threshold value can be predicted according to the real-time change trend of the inlet and outlet pressure ratio to obtain the trigger time.

[0083] S104: Determine whether the trigger time exceeds the set maximum response time threshold: If so, adjust the rotation speed according to the trigger time and / or the real-time status and real-time change trend of the mass flow to prevent the system from reaching an abnormal flow threshold and causing an unstable operating condition.

[0084] Predict potential abnormal risks based on trigger time, evaluate the system's operating status based on real-time status, and formulate appropriate adjustment strategies based on changing trends to accurately adjust the speed, avoid unnecessary risks, and optimize equipment operating efficiency and stability.

[0085] The maximum response time threshold is set according to the tolerance range or safety requirements of the system to ensure that the system can respond in time. Exceeding this time will easily cause abnormal operation or damage of the equipment.

[0086] Specifically, if so, the rotation speed may be gradually adjusted according to the trigger time and / or the changing trend of the mass flow rate to smoothly transition to a new operating range and avoid system instability caused by abrupt changes.

[0087] More specifically, a proportional control algorithm, a PID control algorithm or an adaptive control algorithm may be used to automatically adjust the speed change range according to the trigger time and / or the real-time state and real-time change trend of the mass flow rate.

[0088] When the mass flow deviation is small, the proportional control algorithm can determine the speed adjustment amplitude according to the mass flow deviation (i.e. the difference between the actual flow and the set flow). The amplitude is proportional to the deviation. The larger the deviation, the larger the speed adjustment amplitude.

[0089] When the mass flow rate changes are complex (for example, with significant fluctuations or hysteresis effects), the PID (proportional-integral-derivative) control algorithm can adjust the speed based on the mass flow rate deviation, rate of change (differential part) and historical accumulated deviation (integral part). PID control can provide higher accuracy and more stable control effects.

[0090] If the system exhibits complex nonlinear characteristics under different operating conditions, an adaptive control algorithm can be used. The adaptive control algorithm adjusts the control strategy and speed adjustment range in real time to adapt to different operating conditions and changing trends, ensuring that the speed of the swirl superheater is always within a stable range under various operating conditions. For example, if the trigger time is too early and the mass flow rate increases abnormally, the speed may need to be adjusted more quickly; conversely, if the trigger time is late, the adjustment range may be relatively small.

[0091] During the speed adjustment process, the adjusted mass flow rate can also be monitored in real time, and the speed adjustment range can be continuously corrected according to the adjusted mass flow rate to ensure that the mass flow rate is stable within the normal range. If the mass flow rate has stabilized and is close to the normal value, there is no need to make excessive speed adjustments. On the contrary, if the mass flow rate deviates from the normal range and the fluctuation increases, the speed adjustment will be accelerated to restore stability.

[0092] When the trigger time exceeds the set maximum response time threshold, an alarm notification can also be output simultaneously. The alarm notification can be issued through local warning devices (such as lights, sounds) and remote communication methods (such as text messages, emails, etc.).

[0093] In addition, the trigger time, speed adjustment process and actual adjustment results can be recorded in the database to facilitate subsequent analysis and optimization of control strategies.

[0094] Exemplary Systems

[0095] Accordingly, an embodiment of the present invention further provides a cyclone superheater abnormal operating condition prediction control system. Figure 2 is a block diagram of a prediction control system for abnormal operating conditions of a swirl superheater provided in an embodiment of the present invention, such as Figure 2 As shown, the system 100 provided in this embodiment includes:

[0096] A detection device 101, which is used to detect the inlet and outlet pressures of the cyclone superheater in real time;

[0097] A flow generation device 102, which is used to input the inlet and outlet pressures and rotation speed into a dynamic expression model of the swirl superheater corrected by reinforcement learning and differential evolution algorithms to obtain a real-time change trend of the mass flow of the swirl superheater;

[0098] A prediction device 103, which is used to predict the time when the current mass flow reaches an abnormal flow threshold according to the real-time change trend of the mass flow, and obtain a trigger time, wherein the abnormal flow threshold includes an upper limit and / or a lower limit of the normal mass flow;

[0099] The judging and adjusting device 104 is used to judge whether the trigger time exceeds the set maximum response time threshold: if so, adjust the rotation speed according to the trigger time and / or the real-time status and real-time change trend of the mass flow.

[0100] Specifically, the dynamic expression model is:

[0101]

[0102]

[0103]

[0104] Where p is the static pressure chamber pressure, m is the compressor mass flow rate, k τ is the valve opening of the air storage chamber, ω is the motor speed, a0 is the intake stagnation sound speed, V p is the volume of the static pressure chamber, A1 is the reference area of ​​the impeller, L c is the air duct length, J is the total inertia of the compressor, τm is the torque generated by the drive motor, τ c is the torque produced by the compressor, and Ψ(ω,m) is the pressure ratio of the compressor outlet to the inlet air pressure.

[0105] Ψ(ω,m) can be determined by the following formula:

[0106] Among them, η(ω,m)Δh ideal is the actual enthalpy increase of the fluid, γ is the gas specific heat ratio, Ψ(ω,m) is the output characteristic function; p 01 and p 02 Respectively represent the inlet pressure and outlet pressure; T 01 is the steam inlet temperature; Δh ideal is the ideal specific enthalpy transferred to the fluid, σ is the slip coefficient, U2 is the tangential velocity at the impeller outlet, and the efficiency η(m,ω) is a function of the mass flow rate m and the rotation speed w.

[0107] The actual enthalpy increase of the fluid in the formula is η(ω,m)Δh ideal It is necessary to consider the incident loss and friction loss, which can be expressed as follows: η(ω,m)Δh ideal =Δh t -Δh i -Δh f -Δh oth (2)

[0108] Where Δh t is the total enthalpy increment, Δh i and Δh f are the enthalpy changes caused by incident loss and friction loss, Δh oth It is the sum of other losses such as gap loss, backflow loss and leakage loss.

[0109] The fluid enthalpy change is calculated using the following expression: Δh t =μr2 2 ω 2 Δh f =k f m 2 Δh oth =εΔh t (3)

[0110] ε is the loss rate, generally 8%, r1 is the average radius of the guide vane, r2 is the impeller radius, k fis the fluid friction constant, μ and α are parameters, determined by the following formula:

[0111] Among them, β 1b and β 2b are the blade inlet angle and outlet angle respectively, ρ1 is the constant stagnation inlet density, and σ is the slip coefficient slightly less than 1.

[0112] By combining formulas 1 to 4, the relationship between the inlet and outlet pressure ratio, inlet flow rate and speed is obtained as follows:

[0113] The traffic generating device 102 further includes:

[0114] Used to obtain, through experiments, the corresponding relationship between the compressor flow rate and the inlet and outlet pressure ratio under different speed stable operating conditions, and to obtain the optimal empirical parameters in the dynamic expression model through reinforcement learning and differential evolution algorithm based on the corresponding relationship, and to calibrate the units of the dynamic expression model using the optimal empirical parameters;

[0115] The corrected dynamic expression model uses the inlet and outlet pressure ratio, inlet flow speed, and shaft power in actual working conditions to calculate the mass flow rate change rate, inlet and outlet pressure ratio change rate, and speed change rate of the swirl superheater at any time;

[0116] The analysis unit 105 is used to perform time series analysis on the mass flow detected in real time by using the mass flow rate of change at any time to obtain the real-time change trend of the mass flow, the real-time change trend of the inlet and outlet pressure ratio and the real-time change trend of the rotation speed.

[0117] The analysis unit 105 further includes:

[0118] A module for collecting the mass flow detected in real time as time series data and organizing it into continuous data points in chronological order to form time series data of the mass flow;

[0119] A module for preprocessing the time series data of the mass flow;

[0120] A module for modeling the pre-processed mass flow time series data based on a selected time series analysis method using the mass flow rate of change at any time to obtain the real-time change trend of the mass flow.

[0121] The prediction device 103 further includes:

[0122] A fitting unit 106, which is used to predict the time when the current mass flow reaches the abnormal flow threshold value by using curve fitting according to the real-time change trend of the mass flow, and obtain a trigger time; and / or

[0123] A unit for predicting the time when the current mass flow reaches an abnormal flow threshold value by using machine learning according to the real-time change trend of the mass flow, and obtaining a trigger time.

[0124] The fitting unit 106 further includes:

[0125] A module for obtaining a prediction curve by fitting the historical mass flow data in the real-time change trend of the mass flow;

[0126] A module for using the prediction curve to predict the time when the current mass flow reaches the abnormal flow threshold to obtain the trigger time.

[0127] The judging and adjusting device 104 further includes an alarm unit 107 and a gradual adjustment unit 108:

[0128] The judgment and adjustment device 104 is used to judge whether the trigger time exceeds the set maximum response time threshold: if so, the alarm unit 107 is used to output an alarm notification, and the gradual adjustment unit 108 is used to gradually adjust the rotation speed according to the change trend of the trigger time and / or the mass flow rate.

[0129] The step-by-step adjustment unit 108 further includes:

[0130] A module for automatically adjusting the speed change range according to the trigger time and / or the real-time status and real-time change trend of the mass flow rate using a PID control algorithm; or

[0131] A module for automatically adjusting the speed change range according to the trigger time and / or the real-time status and real-time change trend of the mass flow rate by using an adaptive control algorithm.

[0132] The detection device 101 further includes:

[0133] a unit for connecting the pressure sensor to a data acquisition unit;

[0134] A frequency setting unit 109, which is used to set the sampling frequency according to the operating characteristics of the swirl superheater, wherein the operating characteristics are various dynamic behaviors and performance indicators exhibited by the swirl superheater during actual operation, including pressure fluctuations, temperature changes, flow changes, flow states, load changes, thermal efficiency changes and / or vibration characteristics;

[0135] A unit for detecting the inlet and outlet pressures of the swirl superheater using the pressure sensor according to the sampling frequency, and transmitting the inlet and outlet pressures to the data acquisition unit.

[0136] The frequency setting unit 109 also includes:

[0137] A module for dynamically adjusting the sampling frequency according to the operating characteristics monitored in real time; and / or

[0138] A module is configured to determine the sampling frequency based on the operating characteristic historical data.

[0139] It should be noted that although the operations of the abnormal condition prediction control method of the swirl superheater of the present invention are described in a specific order in the drawings, this does not require or imply that these operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

[0140] In addition, although several devices, units, or modules of the abnormal operating condition prediction control system of the swirl superheater are mentioned in the above detailed description, this division is only exemplary and not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more modules described above can be embodied in one module. Conversely, the features and functions of one module described above can be further divided into multiple modules for embodiment.

[0141] Although the spirit and principle of the present invention have been described with reference to several specific embodiments, it should be understood that the present invention is not limited to the disclosed specific embodiments, and the division of various aspects does not mean that the features in these aspects cannot be combined to benefit, and such division is only for the convenience of expression. The present invention is intended to cover various modifications and equivalent arrangements included in the spirit and scope of the attached claims.

[0142] The present invention provides: 1. A method for predicting and controlling abnormal operating conditions of a cyclone superheater, characterized in that the method comprises: Real-time detection of the inlet and outlet pressures of the swirl superheater; The inlet and outlet pressures and rotation speeds are input into a dynamic expression model of a swirl superheater corrected by reinforcement learning and differential evolution algorithms to obtain a real-time change trend of the mass flow rate of the swirl superheater; Predicting the time when the current mass flow reaches an abnormal flow threshold according to the real-time change trend of the mass flow, and obtaining a trigger time, wherein the abnormal flow threshold includes an upper limit and / or a lower limit of the normal mass flow; Determine whether the trigger time exceeds a set maximum response time threshold: If so, adjust the rotation speed according to the trigger time and / or the real-time state and real-time change trend of the mass flow. 2. The abnormal condition prediction and control method for a swirl superheater according to item 1 is characterized in that the dynamic expression model is: Where p is the static pressure chamber pressure, m is the compressor mass flow rate, k t is the valve opening of the air storage chamber, ω is the motor speed, a0 is the intake stagnation sound speed, V p is the volume of the static pressure chamber, A1 is the reference area of ​​the impeller, L c is the air duct length, J is the total inertia of the compressor, τ m is the torque generated by the drive motor, τ c is the torque produced by the compressor, Ψ(ω,m) is the pressure ratio of the compressor outlet to the inlet air pressure; Ψ(ω,m) is determined by the following formula: Among them, η(ω,m)Δh ideal is the actual enthalpy increase of the fluid, γ is the gas specific heat ratio, Ψ(ω,m) is the output characteristic function; p 01 and p 02 Respectively represent the inlet pressure and outlet pressure; T 01 is the steam inlet temperature; Δh ideal is the ideal specific enthalpy transferred to the fluid, and the efficiency η(m,ω) is a function of the mass flow rate m and the rotation speed w; η(ω,m)Δh ideal It is expressed by the following formula: η(ω,m)Δh ideal =Δh t -Δh i -Δh f -Δh oth (2) Where Δh t is the total enthalpy increment, Δh i and Δh f are the enthalpy changes caused by incident loss and friction loss, Δh oth It is the sum of other losses such as gap loss, backflow loss and leakage loss; The fluid enthalpy change is calculated using the following expression: Δh t =μr2 2 ω 2 Δh f =kf m 2 Δh oth =εΔh t (3) Among them, ε is the loss rate, r1 is the average radius of the guide vane, r2 is the impeller radius, k f is the fluid friction constant, μ and α are parameters, determined by the following formula: Among them, β 1b and β 2b are the blade inlet angle and outlet angle respectively, ρ1 is the constant stagnation inlet density, and σ is the slip coefficient slightly less than 1. Combining formulas 1 to 4, we can obtain the relationship between the inlet and outlet pressure ratio, inlet flow rate and rotation speed: 3. The abnormal condition prediction and control method for a swirl superheater according to item 2 is characterized in that the inlet and outlet pressures and rotation speeds are input into a dynamic expression model of a swirl superheater corrected by reinforcement learning and differential evolution algorithms, and the steps of obtaining a real-time change trend of the mass flow rate of the swirl superheater are specifically as follows: Through experiments, the corresponding relationship between the compressor flow rate and the inlet and outlet pressure ratio under different speed stable operating conditions is obtained, and based on the corresponding relationship, the optimal empirical parameters in the dynamic expression model are obtained through reinforcement learning and differential evolution algorithm, and the dynamic expression model is corrected using the optimal empirical parameters; The corrected dynamic expression model uses the inlet and outlet pressure ratio, inlet flow speed and shaft power in actual working conditions to calculate the mass flow rate change rate, inlet and outlet pressure ratio change rate and speed change rate of the swirl superheater at any time; The mass flow rate of change at any moment is used to perform a time series analysis on the mass flow detected in real time, so as to obtain the real-time change trend of the mass flow, the real-time change trend of the inlet and outlet pressure ratio, and the real-time change trend of the rotation speed. 4. The abnormal condition prediction and control method for a swirl superheater according to item 3 is characterized in that the steps of performing time series analysis on the mass flow detected in real time by using the mass flow rate of change at any time to obtain the real-time change trend of the mass flow, the real-time change trend of the inlet and outlet pressure ratio and the real-time change trend of the speed are specifically as follows: The mass flow detected in real time is collected as time series data, and organized into continuous data points in time sequence to form time series data of the mass flow; Preprocessing the time series data of the mass flow; Based on the selected time series analysis method, the mass flow rate change rate at any moment is used to model the pre-processed time series data of the mass flow rate, so as to obtain the real-time change trend of the mass flow rate. 5. The abnormal condition prediction and control method for a swirl superheater according to any one of items 1 to 4 is characterized in that the time when the current mass flow reaches the abnormal flow threshold is predicted according to the real-time change trend of the mass flow, and the step of obtaining the trigger time is specifically: The time when the current mass flow reaches the abnormal flow threshold is predicted by using curve fitting and / or machine learning according to the real-time change trend of the mass flow, so as to obtain the trigger time. 6. The abnormal condition prediction and control method for a swirl superheater according to item 5 is characterized in that the time when the current mass flow reaches the abnormal flow threshold is predicted by curve fitting according to the real-time change trend of the mass flow, and the step of obtaining the trigger time is specifically as follows: Obtaining a prediction curve by fitting the historical mass flow data in the real-time change trend of the mass flow; The prediction curve is used to predict the time when the current mass flow rate reaches the abnormal flow rate threshold to obtain the trigger time. 7. The abnormal condition prediction and control method for a swirl superheater according to any one of items 1 to 4 is characterized in that it is judged whether the trigger time exceeds a set maximum response time threshold; if so, the steps of adjusting the rotation speed according to the trigger time and / or the real-time state and real-time change trend of the mass flow are specifically as follows: Determine whether the trigger time exceeds a set maximum response time threshold: if so, output an alarm notification, and gradually adjust the rotation speed according to the trigger time and / or the change trend of the mass flow rate. 8. The abnormal condition prediction and control method for a swirl superheater according to item 7 is characterized in that the step of gradually adjusting the rotation speed according to the change trend of the trigger time and / or the mass flow rate is specifically: The speed change range is automatically adjusted according to the trigger time and / or the real-time state and real-time change trend of the mass flow rate using a PID control algorithm or an adaptive control algorithm. 9. The abnormal operating condition prediction and control method of the cyclone superheater according to any one of items 1 to 4, characterized in that the real-time detection of the inlet and outlet pressures of the cyclone superheater specifically comprises: Connect the pressure sensor to the data acquisition unit; The sampling frequency is set according to the operating characteristics of the swirl superheater, wherein the operating characteristics are various dynamic behaviors and performance indicators exhibited by the swirl superheater during actual operation, including pressure fluctuations, temperature changes, flow changes, flow states, load changes, thermal efficiency changes and / or vibration characteristics; The inlet and outlet pressures of the swirl superheater are detected by using the pressure sensor according to the sampling frequency, and the inlet and outlet pressures are transmitted to the data acquisition unit. 10. The abnormal condition prediction and control method of the swirl superheater according to item 9 is characterized in that the step of setting the sampling frequency is specifically as follows according to the operating characteristics of the swirl superheater: Dynamically adjusting the sampling frequency according to the operating characteristics monitored in real time; and / or The sampling frequency is determined by the operating characteristic historical data. 11. A prediction and control system for abnormal operating conditions of a cyclone superheater, characterized in that the system comprises: A detection device, which is used to detect the inlet and outlet pressures of the swirl superheater in real time; A flow generation device, which is used to input the inlet and outlet pressures and rotation speed into a dynamic expression model of a swirl superheater corrected by reinforcement learning and differential evolution algorithms to obtain a real-time change trend of the mass flow of the swirl superheater; A prediction device, which is used to predict the time when the current mass flow reaches an abnormal flow threshold according to the real-time change trend of the mass flow, and obtain a trigger time, wherein the abnormal flow threshold includes an upper limit and / or a lower limit of the normal mass flow; The judgment and adjustment device is used to judge whether the trigger time exceeds the set maximum response time threshold: if so, adjust the rotation speed according to the trigger time and / or the real-time status and real-time change trend of the mass flow. 12. The abnormal operating condition prediction control system of the swirl superheater according to item 11 is characterized in that the dynamic expression model is: Where p is the static pressure chamber pressure, m is the compressor mass flow rate, k τ is the valve opening of the air storage chamber, ω is the motor speed, a0 is the intake stagnation sound speed, V p is the volume of the static pressure chamber, A1 is the reference area of ​​the impeller, L c is the air duct length, J is the total inertia of the compressor, τ m is the torque generated by the drive motor, τ c is the torque produced by the compressor, Ψ(ω,m) is the pressure ratio of the compressor outlet to the inlet air pressure; Ψ(ω,m) is determined by the following formula: Among them, η(ω,m)Δh idealis the actual enthalpy increase of the fluid, γ is the gas specific heat ratio, Ψ(ω,m) is the output characteristic function; p 01 and p 02 Respectively represent the inlet pressure and outlet pressure; T 01 is the steam inlet temperature; Δh ideal is the ideal specific enthalpy transferred to the fluid, and the efficiency η(m,ω) is a function of the mass flow rate m and the rotation speed w; η(ω,m)Δh ideal It is expressed by the following formula: η(ω,m)Δh ideal =Δh t -Δh i -Δh f -Δh oth (2) Where Δh t is the total enthalpy increment, Δh i and Δh f are the enthalpy changes caused by incident loss and friction loss, Δh oth It is the sum of other losses such as gap loss, backflow loss and leakage loss; The fluid enthalpy change is calculated using the following expression: Δh t =μr2 2 ω 2 Δh f =k f m 2 Δh oth =εΔh t (3) Among them, ε is the loss rate, r1 is the average radius of the guide vane, r2 is the impeller radius, k f is the fluid friction constant, μ and α are parameters, determined by the following formula: Among them, β 1b and β 2b are the blade inlet angle and outlet angle respectively, ρ1 is the constant stagnation inlet density, and σ is the slip coefficient slightly less than 1. Combining formulas 1 to 4, we can obtain the relationship between the inlet and outlet pressure ratio, inlet flow rate and rotation speed: 13. The abnormal condition prediction control system for a swirl superheater according to item 12, characterized in that the flow generating device further comprises: Used to obtain, through experiments, the corresponding relationship between the compressor flow rate and the inlet and outlet pressure ratio under different speed stable operating conditions, and to obtain the optimal empirical parameters in the dynamic expression model through reinforcement learning and differential evolution algorithm based on the corresponding relationship, and to calibrate the units of the dynamic expression model using the optimal empirical parameters; The corrected dynamic expression model uses the inlet and outlet pressure ratio, inlet flow speed, and shaft power in actual working conditions to calculate the mass flow rate change rate, inlet and outlet pressure ratio change rate, and speed change rate of the swirl superheater at any time; The analysis unit is used to perform time series analysis on the mass flow detected in real time by using the mass flow change rate at any time to obtain the real-time change trend of the mass flow, the real-time change trend of the inlet and outlet pressure ratio and the real-time change trend of the rotation speed. 14. The abnormal operating condition prediction control system of the swirl superheater according to item 13, characterized in that the analysis unit further comprises: A module for collecting the mass flow detected in real time as time series data and organizing it into continuous data points in chronological order to form time series data of the mass flow; A module for preprocessing the time series data of the mass flow; A module for modeling the pre-processed mass flow time series data based on a selected time series analysis method using the mass flow rate of change at any time to obtain the real-time change trend of the mass flow. 15. The abnormal operating condition prediction control system of a cyclone superheater according to any one of items 11 to 14, characterized in that the prediction device further comprises: A fitting unit, which is used to predict the time when the current mass flow reaches the abnormal flow threshold value by using curve fitting according to the real-time change trend of the mass flow, and obtain a trigger time; and / or A unit for predicting the time when the current mass flow reaches an abnormal flow threshold value by using machine learning according to the real-time change trend of the mass flow, and obtaining a trigger time. 16. The abnormal operating condition prediction control system of the swirl superheater according to item 15, characterized in that the fitting unit further comprises: A module for obtaining a prediction curve by fitting the historical mass flow data in the real-time change trend of the mass flow; A module for using the prediction curve to predict the time when the current mass flow reaches the abnormal flow threshold to obtain the trigger time. 17. The abnormal operating condition prediction control system of the cyclone superheater according to any one of items 11 to 14, characterized in that the judgment and adjustment device further comprises an alarm unit and a step-by-step adjustment unit: The judgment and adjustment device is used to judge whether the trigger time exceeds the set maximum response time threshold: if so, the alarm unit is used to output an alarm notification, and the gradual adjustment unit is used to gradually adjust the rotation speed according to the change trend of the trigger time and / or the mass flow rate. 18. The abnormal operating condition prediction control system of the swirl superheater according to item 17, characterized in that the step-by-step adjustment unit further comprises: A module for automatically adjusting the speed change range according to the trigger time and / or the real-time status and real-time change trend of the mass flow rate using a PID control algorithm; or A module for automatically adjusting the speed change range according to the trigger time and / or the real-time status and real-time change trend of the mass flow rate by using an adaptive control algorithm. 19. The abnormal operating condition prediction control system of a cyclone superheater according to any one of items 11 to 14, characterized in that the detection device further comprises: a unit for connecting the pressure sensor to a data acquisition unit; A frequency setting unit, which is used to set the sampling frequency according to the operating characteristics of the swirl superheater, wherein the operating characteristics are various dynamic behaviors and performance indicators exhibited by the swirl superheater during actual operation, including pressure fluctuations, temperature changes, flow changes, flow states, load changes, thermal efficiency changes and / or vibration characteristics; A unit for detecting the inlet and outlet pressures of the swirl superheater using the pressure sensor according to the sampling frequency, and transmitting the inlet and outlet pressures to the data acquisition unit. 20. The abnormal operating condition prediction control system of the swirl superheater according to item 19, characterized in that the frequency setting unit further comprises: A module for dynamically adjusting the sampling frequency according to the operating characteristics monitored in real time; and / or A module is configured to determine the sampling frequency based on the operating characteristic historical data.

Claims

1. A method for predicting and controlling abnormal operating conditions of a swirl superheater, characterized in that: The method comprises: Real-time detection of the inlet and outlet pressures of the swirl superheater; The inlet and outlet pressures and rotation speeds are input into a dynamic expression model of a swirl superheater corrected by reinforcement learning and differential evolution algorithms to obtain a real-time change trend of the mass flow rate of the swirl superheater; Predicting the time when the current mass flow reaches an abnormal flow threshold according to the real-time change trend of the mass flow, and obtaining a trigger time, wherein the abnormal flow threshold includes an upper limit and / or a lower limit of the normal mass flow; Determine whether the trigger time exceeds a set maximum response time threshold: If so, adjust the rotation speed according to the trigger time and / or the real-time state and real-time change trend of the mass flow.

2. The abnormal operating condition prediction and control method of a swirl superheater according to claim 1, characterized in that: The dynamic expression model is: Where p is the static pressure chamber pressure, m is the compressor mass flow rate, k t is the valve opening of the air storage chamber, ω is the motor speed, a0 is the intake stagnation sound speed, V p is the volume of the static pressure chamber, A1 is the reference area of ​​the impeller, L c is the duct length, J is the total inertia of the compressor, τ m is the torque generated by the drive motor, τ c is the torque produced by the compressor, Ψ(ω,m) is the pressure ratio of the compressor outlet to the inlet air pressure; Ψ(ω,m) is determined by the following formula: Among them, η(ω,m)Δh ideal is the actual enthalpy increase of the fluid, γ is the gas specific heat ratio, Ψ(ω,m) is the output characteristic function; p 01 and p 02 Respectively represent the inlet pressure and outlet pressure; T 01 is the steam inlet temperature; Δh ideal is the ideal specific enthalpy transferred to the fluid, and the efficiency η(m,ω) is a function of the mass flow rate m and the rotation speed w; η(ω,m)Δh ideal It is expressed by the following formula: η(ω,m)Δh ideal =Δh t -Δh i -Δh f -Δh oth (2) Where Δh t is the total enthalpy increment, Δh i and Δh f are the enthalpy changes caused by incident loss and friction loss, Δh oth It is the sum of other losses such as gap loss, backflow loss and leakage loss; The fluid enthalpy change is calculated using the following expression: Dh t =μr2 2 oh 2 Δh f =k f m 2 Dh oth =eΔh t (3) Among them, ε is the loss rate, r1 is the average radius of the guide vane, r2 is the impeller radius, k f is the fluid friction constant, μ and α are parameters, determined by the following formula: Among them, β 1b and β 2b are the blade inlet angle and outlet angle respectively, ρ1 is the constant stagnation inlet density, and σ is the slip coefficient slightly less than 1. Combining formulas 1 to 4, we can obtain the relationship between the inlet and outlet pressure ratio, inlet flow rate and rotation speed:

3. The abnormal condition prediction and control method of a swirl superheater according to claim 2, characterized in that: The inlet and outlet pressures and rotation speeds are input into the dynamic expression model of the swirl superheater corrected by the reinforcement learning and differential evolution algorithm to obtain the real-time change trend of the mass flow of the swirl superheater. Specifically, the steps are as follows: Through experiments, the corresponding relationship between the compressor flow rate and the inlet and outlet pressure ratio under different speed stable operating conditions is obtained, and based on the corresponding relationship, the optimal empirical parameters in the dynamic expression model are obtained through reinforcement learning and differential evolution algorithm, and the dynamic expression model is corrected using the optimal empirical parameters; The corrected dynamic expression model uses the inlet and outlet pressure ratio, inlet flow speed and shaft power in actual working conditions to calculate the mass flow rate change rate, inlet and outlet pressure ratio change rate and speed change rate of the swirl superheater at any time; The mass flow rate of change at any moment is used to perform a time series analysis on the mass flow detected in real time, so as to obtain the real-time change trend of the mass flow, the real-time change trend of the inlet and outlet pressure ratio, and the real-time change trend of the rotation speed.

4. The abnormal condition prediction and control method of a swirl superheater according to claim 3, characterized in that: The steps of performing time series analysis on the mass flow detected in real time by using the mass flow rate of change at any time to obtain the real-time change trend of the mass flow, the real-time change trend of the inlet and outlet pressure ratio, and the real-time change trend of the speed are specifically as follows: The mass flow detected in real time is collected as time series data, and organized into continuous data points in time sequence to form time series data of the mass flow; Preprocessing the time series data of the mass flow; Based on the selected time series analysis method, the mass flow rate change rate at any moment is used to model the pre-processed time series data of the mass flow rate, so as to obtain the real-time change trend of the mass flow rate.

5. The abnormal operating condition prediction and control method of a swirl superheater according to any one of claims 1 to 4, characterized in that: The step of predicting the time when the current mass flow reaches the abnormal flow threshold according to the real-time change trend of the mass flow and obtaining the trigger time is specifically as follows: The time when the current mass flow reaches the abnormal flow threshold is predicted by using curve fitting and / or machine learning according to the real-time change trend of the mass flow, so as to obtain the trigger time.

6. The abnormal operating condition prediction and control method of a swirl superheater according to claim 5, characterized in that: The step of obtaining the trigger time by using curve fitting to predict the time when the current mass flow reaches the abnormal flow threshold according to the real-time change trend of the mass flow is as follows: Obtaining a prediction curve by fitting the historical mass flow data in the real-time change trend of the mass flow; The prediction curve is used to predict the time when the current mass flow rate reaches the abnormal flow rate threshold to obtain the trigger time.

7. The abnormal operating condition prediction and control method of a swirl superheater according to any one of claims 1 to 4, characterized in that: Determine whether the trigger time exceeds the set maximum response time threshold: If so, the steps of adjusting the rotation speed according to the trigger time and / or the real-time state and real-time change trend of the mass flow are specifically as follows: Determine whether the trigger time exceeds a set maximum response time threshold: if so, output an alarm notification, and gradually adjust the rotation speed according to the trigger time and / or the change trend of the mass flow rate.

8. The abnormal condition prediction and control method of a swirl superheater according to claim 7, characterized in that: The step of gradually adjusting the rotation speed according to the trigger time and / or the change trend of the mass flow rate is specifically as follows: The speed change range is automatically adjusted according to the trigger time and / or the real-time state and real-time change trend of the mass flow rate using a PID control algorithm or an adaptive control algorithm.

9. The abnormal operating condition prediction and control method of a swirl superheater according to any one of claims 1 to 4, characterized in that: The real-time detection of the inlet and outlet pressures of the cyclone superheater specifically includes: Connect the pressure sensor to the data acquisition unit; The sampling frequency is set according to the operating characteristics of the swirl superheater, wherein the operating characteristics are various dynamic behaviors and performance indicators exhibited by the swirl superheater during actual operation, including pressure fluctuations, temperature changes, flow changes, flow states, load changes, thermal efficiency changes and / or vibration characteristics; The inlet and outlet pressures of the swirl superheater are detected by using the pressure sensor according to the sampling frequency, and the inlet and outlet pressures are transmitted to the data acquisition unit.

10. A prediction control system for abnormal operating conditions of a cyclone superheater, characterized in that: The system comprises: A detection device, which is used to detect the inlet and outlet pressures of the swirl superheater in real time; A flow generation device, which is used to input the inlet and outlet pressures and rotation speed into a dynamic expression model of a swirl superheater corrected by reinforcement learning and differential evolution algorithms to obtain a real-time change trend of the mass flow of the swirl superheater; A prediction device, which is used to predict the time when the current mass flow reaches an abnormal flow threshold according to the real-time change trend of the mass flow, and obtain a trigger time, wherein the abnormal flow threshold includes an upper limit and / or a lower limit of the normal mass flow; The judgment and adjustment device is used to judge whether the trigger time exceeds the set maximum response time threshold: if so, adjust the rotation speed according to the trigger time and / or the real-time status and real-time change trend of the mass flow.

Images