Load feedback-based automatic energy-saving control method and device for ring cooling fan

Abstract

The invention provides an automatic energy-saving control method and device for a ring cooling fan based on load feedback, relates to the field of ring cooling fans, and solves the technical problem of delay of regulation and control in an energy-saving working state. The method comprises the steps that working condition data are input into a preset powder box model, and a feed-forward air volume instruction is obtained through calculation; and inputting the working condition data into a state observer to obtain an optimal estimated temperature value. And the deviation between the optimal estimated temperature value and a preset temperature set value is calculated, and a feedback air volume compensation instruction is obtained through calculation of a feedback controller according to the deviation. And fusing the feed-forward air volume instruction and the feedback air volume compensation instruction to obtain a final air volume control instruction, and issuing the final air volume control instruction to a fan frequency converter for execution. And continuously monitoring the numerical value and the change trend of the feedback air volume compensation instruction, taking the feedback air volume compensation instruction as a prediction error signal of the powder box model, and adaptively adjusting key thermal parameters in the powder box model. The method is used in the control process of the ring cooling fan.

Description

Technical Field

[0001] This application relates to the field of ring-cooled fans, and in particular to an automatic energy-saving control method and device for ring-cooled fans based on load feedback. Background Technology

[0002] Traditional ring-cooled fan control schemes often rely on a single front-end control logic. These schemes typically employ feedforward control based on fixed parameters or simple PID feedback regulation strategies. This means that based on data collected from limited measurement points (such as inlet and outlet water temperatures), airflow demand is calculated using static heat balance formulas or empirical formulas, and then the ring-cooled fan is regulated. Because these methods depend on preset models and fixed control structures, there is an inherent time delay between data acquisition, deviation calculation, and command issuance, making it difficult to respond in real-time to dynamic changes in operating conditions. Especially when load fluctuates frequently or external environmental conditions change abruptly, the airflow regulation of the ring-cooled fan lags significantly, failing to accurately match actual heat dissipation needs and easily leading to problems such as "overcooling" (energy waste) or "undercooling" (excessive equipment temperature rise). This not only seriously affects the reliability and lifespan of the water pump but may also trigger system protection mechanisms due to excessive temperature, leading to unplanned shutdowns and restricting overall energy efficiency and operational economy. Summary of the Invention

[0003] This application provides an automatic energy-saving control method and device for a ring-cooled fan based on load feedback, which solves the technical problem of delay in regulation under energy-saving working conditions in the prior art.

[0004] To achieve the above objectives, this application adopts the following technical solution: Firstly, an automatic energy-saving control method for annular cooling fans based on load feedback includes: inputting operating data into a preset toner cartridge model to calculate a feedforward airflow command. The toner cartridge model is a hybrid model that dynamically integrates energy conservation mechanisms, fluid dynamics mechanisms, and data-driven parameters, and the key thermal and hydraulic parameters within the toner cartridge model are dynamically mapped and generated by a machine learning module based on real-time operating conditions. The operating data is input into a state observer to obtain an optimal estimated temperature value. The deviation between the optimal estimated temperature value and the preset temperature setpoint is calculated, and a feedback airflow compensation command is calculated based on the deviation using a feedback controller. The feedforward airflow command and the feedback airflow compensation command are fused to obtain the final airflow control command, which is then sent to the fan inverter for execution. The value and trend of the feedback airflow compensation command are continuously monitored, and the feedback airflow compensation command is used as a prediction error signal for the toner cartridge model. The key thermal parameters in the toner cartridge model are adaptively adjusted to bring the feedback airflow compensation command closer to zero.

[0005] Based on the above technical solution, in the automatic energy-saving control method for ring-cooled fans based on load feedback provided in this application, a powder box model is constructed to obtain feedforward commands through a hybrid model architecture of energy conservation and fluid mechanics mechanism framework and data-driven parameter dynamic mapping. Then, a state observer is constructed by embedding extended Kalman filter or unscented Kalman filter algorithm through a nonlinear state-space model to obtain feedback compensation. This allows the feedforward commands and feedback compensation to be weighted and fused according to historical weights, and the inverter efficiency curve is superimposed to find the optimal value. This achieves the goal of adjusting the efficiency of the ring-cooled fan according to the actual use of the pumping station. This not only avoids energy waste during normal use, but also avoids the impact of control delay on the efficiency of the ring-cooled fan during energy-saving use, and reduces the impact of nonlinear factors in feedforward control on control accuracy.

[0006] In conjunction with the first aspect mentioned above, in one possible implementation, the construction of the powder box model includes: establishing a basic mechanism model based on the thermal balance mechanism and fluid dynamics mechanism of the pumping station, taking the pump operating frequency, inlet flow rate, outlet pressure, inlet water temperature, and ambient temperature from the operating data as input, and the theoretical heat dissipation air volume required by the annular cooling fan as output. The computational unit of the basic mechanism model is... Q req For the theoretical heat dissipation airflow, P loss The total heat loss power of the equipment is calculated using the pump operating frequency, inlet flow rate, and outlet pressure. η is the overall efficiency coefficient of the heat dissipation system, ΔT is the desired cooling temperature difference setpoint, and C is the total heat loss power of the equipment. pair P is the specific heat capacity of air. air K is the density of air. flow This refers to the duct fluid resistance coefficient. A machine learning module is constructed using historical and real-time operating data as training sets. It dynamically maps key dynamic parameters from the basic mechanism model for output. These key dynamic parameters include the calibration coefficient for the total heat loss power of the equipment, the overall efficiency coefficient of the heat dissipation system, and the duct fluid resistance coefficient. Operating data includes inlet water temperature, outlet water temperature, ambient temperature, historical fan operating airflow and corresponding temperature change rate, pump operating frequency, inlet water flow rate, and outlet water pressure. The key dynamic parameters output by the machine learning module are input into the basic mechanism model for real-time correction, forming a powder box model. The theoretical heat dissipation airflow is then output as a feedforward airflow command.

[0007] In conjunction with the first aspect mentioned above, one possible implementation involves constructing a state observer that includes: establishing a nonlinear state-space model based on the thermodynamic mechanism structure of the powder box model, with fan airflow, inlet water temperature, ambient temperature, and equipment thermal inertia from operating data as state variables, and outlet water temperature as the observed output. The nonlinear state-space model is discretized, and a state observer is constructed based on either the extended Kalman filter algorithm or the unscented Kalman filter algorithm. The real-time collected fan airflow, inlet water temperature, and ambient temperature are input to the state observer, and the optimal estimate of the outlet water temperature is obtained through recursive calculation.

[0008] In conjunction with the first aspect mentioned above, in one possible implementation, the feedback controller calculates the feedback airflow compensation command based on the deviation, including: subtracting the optimal estimated temperature value from the preset temperature setpoint to obtain the temperature deviation and the rate of change of the deviation; dynamically selecting a proportional-integral-derivative control algorithm or a fuzzy control algorithm based on the magnitude of the temperature deviation and the rate of change of the deviation to calculate the temperature deviation and the rate of change of the deviation, generating a preliminary airflow compensation command; and performing output limiting and integral anti-saturation processing on the preliminary airflow compensation command to obtain the final feedback airflow compensation command.

[0009] In conjunction with the first aspect mentioned above, in one possible implementation, based on the magnitude of the temperature deviation and its rate of change, a proportional-integral-derivative (PI-DE) control algorithm or a fuzzy control algorithm is dynamically selected to calculate the temperature deviation and its rate of change, generating a preliminary airflow compensation command. This includes: dynamically setting a preset threshold based on the average of the rate of change of the best estimated temperature value in historical data; the preset threshold is updated according to a preset hydropower station maintenance cycle. When the absolute value of the temperature deviation is less than or equal to the preset threshold, the PI-DE control algorithm is selected, and proportional, integral, and derivative operations are performed based on the temperature deviation and its rate of change to generate a preliminary airflow compensation command. When the absolute value of the temperature deviation is greater than the preset threshold, the fuzzy control algorithm is selected, and calculations are performed based on the fuzzy logic rules of the temperature deviation and its rate of change to generate a preliminary airflow compensation command.

[0010] In conjunction with the first aspect mentioned above, in one possible implementation, the preliminary airflow compensation command is subjected to output limiting and integral anti-saturation processing to obtain the final feedback airflow compensation command. This includes comparing the preliminary airflow compensation command with preset upper and lower limits for the airflow command. If the preliminary airflow compensation command exceeds the upper limit, the final feedback airflow compensation command is limited to the upper limit. If the preliminary airflow compensation command is lower than the lower limit, the final feedback airflow compensation command is limited to the lower limit. If the preliminary airflow compensation command is between the upper and lower limits, it is directly used as the final feedback airflow compensation command. When a proportional-integral-derivative (PID) control algorithm is used and the preliminary airflow compensation command remains in a limited state, anti-saturation processing is applied to the integral term to stop integral accumulation or reduce the integral action, preventing excessive accumulation of the integral term from causing control overshoot.

[0011] In conjunction with the first aspect mentioned above, one possible implementation involves fusing the feedforward airflow command and the feedback airflow compensation command to obtain the final airflow control command, which is then sent to the fan inverter for execution. This includes: presetting feedforward weighting coefficients and feedback weighting coefficients based on historical loop-cooled fan control data; weighting the feedforward airflow command according to the feedforward weighting coefficients and the feedback airflow compensation command according to the feedback weighting coefficients; adding the weighted feedforward airflow command and the weighted feedback airflow compensation command to obtain the preliminary fused command; acquiring the current operating efficiency curve of the fan inverter and, with the goal of minimizing overall system energy consumption, performing airflow and energy consumption optimization calculations on the preliminary fused command to obtain the optimal airflow command; and limiting the preliminary optimal airflow command to ensure it does not exceed the maximum safe operating airflow of the fan inverter, then sending the limited command as the final airflow control command to the fan inverter.

[0012] In conjunction with the first aspect mentioned above, one possible implementation involves adaptively adjusting key thermal parameters in the toner cartridge model, including: using the feedback airflow compensation command as the dynamic prediction error signal of the toner cartridge model. Based on a preset adaptive adjustment law, the key dynamic parameters output by the machine learning module in the toner cartridge model are adjusted according to the value and time integral of the dynamic prediction error signal, so that the feedback airflow compensation command approaches zero. The adaptive adjustment law is as follows: θ represents the key dynamic parameters to be adjusted, including the calibration coefficient for the total heat loss power of the equipment, the overall efficiency coefficient of the heat dissipation system, and the fluid resistance coefficient of the air duct. new The adjusted parameter value, θ old The parameters are the values ​​before adjustment, α is the adaptive learning rate, and M is the value before adjustment. fb The air volume compensation command is used for feedback, where t is time.

[0013] In conjunction with the first aspect mentioned above, one possible implementation involves constructing a dual-loop control system. This dual-loop control system includes: a feedforward and feedback control outer loop that rapidly responds to load changes, and a model self-updating inner loop that slowly optimizes the feedforward model parameters. The feedforward and feedback control outer loop generates feedforward airflow commands through the toner cartridge model and generates feedback airflow compensation commands through a state observer and a feedback controller. The feedforward and feedback airflow compensation commands are then merged and the final airflow control command is issued for execution, controlling the airflow of the air cooler. The model self-updating inner loop continuously monitors the feedback airflow compensation commands, using them as dynamic prediction error signals for the toner cartridge model. It dynamically adjusts the key dynamic parameters output by the machine learning module in the toner cartridge model using an adaptive adjustment law, enabling online self-correction and slow optimization of the feedforward model.

[0014] Secondly, an automatic energy-saving control device for annular cooling fans based on load feedback is provided, comprising: a communication unit and a processing unit; the communication unit is used to collect real-time operating data of water pumps and annular cooling fans in water plant pumping stations; the processing unit is used to input the operating data into a preset powder box model and calculate the feedforward air volume command. The powder box model is a hybrid model that dynamically integrates energy conservation mechanisms, fluid dynamics mechanisms, and data-driven parameters, and the key thermal and hydraulic parameters inside the powder box model are dynamically mapped and generated by a machine learning module according to real-time operating conditions. The operating data is input to a state observer to obtain the optimal estimated temperature value. The deviation between the optimal estimated temperature value and the preset temperature setpoint is calculated, and the feedback controller calculates the feedback air volume compensation command based on the deviation. The feedforward air volume command and the feedback air volume compensation command are fused to obtain the final air volume control command, which is then sent to the fan inverter for execution. The numerical value and trend of the feedback air volume compensation command are continuously monitored, and the feedback air volume compensation command is used as the prediction error signal of the toner cartridge model. By adaptively adjusting the key thermal parameters in the toner cartridge model, the feedback air volume compensation command is brought closer to zero.

[0015] Thirdly, this application provides an automatic energy-saving control device for an annular fan based on load feedback, comprising: a processor and a storage medium; the storage medium includes instructions, and the processor is used to execute the instructions to implement the method described in the first aspect and any possible implementation thereof. This automatic energy-saving control device for an annular fan based on load feedback can be an electronic device or a chip within an electronic device.

[0016] Fourthly, this application provides a computer-readable storage medium storing instructions that, when executed on the load-feedback-based automatic energy-saving control device for an annular fan, cause the load-feedback-based automatic energy-saving control device to perform the method described in the first aspect and any possible implementation thereof.

[0017] Fifthly, this application provides a computer program product containing instructions that, when the computer program product is run on the load feedback-based automatic energy-saving control device for an annular fan, causes the load feedback-based automatic energy-saving control device for an annular fan to perform the method described in the first aspect and any possible implementation thereof.

[0018] This application provides an automatic energy-saving control method and device for ring-cooled fans based on load feedback. It constructs a powder box model to obtain feedforward commands through a hybrid model architecture combining energy conservation and fluid mechanics mechanisms with data-driven parameter dynamic mapping. Then, by embedding an extended Kalman filter or unscented Kalman filter algorithm into a nonlinear state-space model, a state observer is built to obtain feedback compensation. This allows the feedforward commands and feedback compensation to be weighted and fused according to historical weights, and then superimposed on the inverter efficiency curve for optimization. This achieves the goal of adjusting the efficiency of the ring-cooled fan according to the actual use of the pumping station. This not only avoids energy waste during normal use but also prevents control delays from affecting the ring-cooled fan's efficiency during energy-saving periods, reducing the impact of nonlinear factors in feedforward control on control accuracy.

[0019] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this application do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments. Attached Figure Description

[0020] Figure 1 A system architecture diagram of an automatic energy-saving control system for a ring-cooled fan based on load feedback is provided in an embodiment of this application; Figure 2A flowchart illustrating an automatic energy-saving control method for a ring-cooled fan based on load feedback, provided in an embodiment of this application; Figure 3 A flowchart illustrating another automatic energy-saving control method for a ring-cooled fan based on load feedback provided in an embodiment of this application; Figure 4 A flowchart illustrating another automatic energy-saving control method for a ring-cooled fan based on load feedback provided in an embodiment of this application; Figure 5 A flowchart illustrating another automatic energy-saving control method for a ring-cooled fan based on load feedback provided in an embodiment of this application; Figure 6 This is a schematic diagram of an automatic energy-saving control device for a ring-cooled fan based on load feedback, provided as an embodiment of this application. Detailed Implementation

[0021] In the description of this application, unless otherwise stated, " / " means "or," for example, A / B can mean A or B. The "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one" means one or more, and "multiple" means two or more. The terms "first," "second," etc., do not limit the quantity or order of execution, and "first," "second," etc., do not necessarily imply differences.

[0022] It should be noted that, in this application, the terms "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0023] To address the limitations of existing technologies where control logic is simplistic and relies on fixed parameters, many systems employ feedforward control based on fixed parameters or simple PID feedback regulation strategies. These systems adjust the airflow of the ring-cooled fan based solely on data collected from limited measurement points (such as inlet and outlet water temperatures) using static heat balance formulas or empirical formulas to calculate airflow demand. This reliance on pre-defined models and fixed control structures results in an inherent time delay between data acquisition, deviation calculation, and command issuance, making it difficult to respond in real-time to dynamic changes in operating conditions. Furthermore, in scenarios with frequent load fluctuations or sudden changes in external environmental conditions, the airflow adjustment of the ring-cooled fan lags significantly, failing to accurately match actual heat dissipation needs. This leads to two main problems: overcooling and undercooling. Overcooling wastes energy, while undercooling causes excessively high equipment temperatures, severely impacting pump reliability and lifespan, and potentially triggering system protection mechanisms due to excessive temperature, resulting in unplanned shutdowns. Finally, traditional control methods do not consider nonlinear factors in the system, such as changes in duct fluid resistance and fluctuations in heat dissipation system efficiency. They rely solely on static parameter calculations and cannot dynamically adapt to these nonlinear changes, further reducing control accuracy and exacerbating the deviation between airflow regulation and actual demand. This makes it difficult to achieve precise and timely regulation under energy-saving operating conditions. This application provides an automatic energy-saving control method for ring-cooled fans based on load feedback. This method constructs a powder box model to obtain feedforward commands through a hybrid model architecture that combines the mechanism framework of energy conservation and fluid mechanics with dynamic mapping of data-driven parameters. Then, it constructs a state observer by embedding an extended Kalman filter or unscented Kalman filter algorithm through a nonlinear state-space model to obtain feedback compensation. This allows the feedforward commands and feedback compensation to be weighted and fused according to historical weights, and the inverter efficiency curve is superimposed for optimization. This achieves the goal of adjusting the efficiency of the ring-cooled fan according to the actual use of the pumping station. This not only avoids energy waste during normal use but also avoids control delays that affect the efficiency of the ring-cooled fan during energy-saving use, reducing the impact of nonlinear factors in feedforward control on control accuracy.

[0024] like Figure 1 As shown in the embodiment of this application, an automatic energy-saving control method for a ring-cooled fan based on load feedback is provided, including: Step 101: Collect real-time operating data of water pumps and cooling fans in the water plant's pumping station.

[0025] Among them, operating condition data refers to real-time parameters of the equipment during operation, including pump operating frequency, inlet water flow, outlet water pressure, inlet water temperature, ambient temperature, outlet water temperature, historical operating air volume of the fan and corresponding temperature change rate, etc., which are used to reflect the current working status of the system.

[0026] In some implementations, a frequency sensor is used to collect the operating frequency of the water pump, a flow meter measures the inlet water flow, a pressure sensor monitors the outlet water pressure, and a temperature sensor acquires the inlet water temperature, ambient temperature, and outlet water temperature. At the same time, an air volume meter records the historical operating air volume of the fan and the corresponding temperature change rate. These sensors transmit real-time data to the processing unit through a communication unit to ensure that the data is continuous and synchronized.

[0027] Step 102: Input the operating data into the preset toner cartridge model to calculate the feedforward air volume command. The toner cartridge model is a hybrid model that dynamically integrates energy conservation mechanisms, fluid dynamics mechanisms, and data-driven parameters. The key thermal and hydraulic parameters inside the toner cartridge model are dynamically mapped and generated by the machine learning module based on real-time operating conditions.

[0028] Among them, the feedforward air volume command is the theoretical heat dissipation air volume output value of the ring-cooled fan calculated based on the toner cartridge model, which is used to predict load change requirements.

[0029] In some implementations, real-time collected operating data (including pump operating frequency, inlet flow rate, outlet pressure, inlet water temperature, and ambient temperature) is input into a preset powder cartridge model. The basic mechanism module of the powder cartridge model dynamically calculates the operating parameters from the pump using the energy conservation formula. Simultaneously, the machine learning module uses historical and real-time operating data (such as inlet and outlet water temperatures, historical fan airflow, and temperature change rate) as a training set to dynamically map and generate key dynamic parameters that need to be calibrated in the basic mechanism model (including the calibration coefficient of the total heat loss power of the equipment, the comprehensive efficiency coefficient η of the heat dissipation system, and the fluid resistance coefficient K of the air duct). flow Then, the dynamic parameters output by the machine learning module are used to correct the basic mechanism model in real time, forming a dynamically updated powder box model, and finally outputting the theoretical heat dissipation air volume as the feedforward air volume command.

[0030] Step 103: Input the operating data into the status observer to obtain the optimal estimated temperature value. Calculate the deviation between the optimal estimated temperature value and the preset temperature setpoint, and use the feedback controller to calculate the feedback airflow compensation command based on the deviation.

[0031] The state observer is a nonlinear state-space model constructed based on the thermodynamic mechanism of the powder box model. It is discretized using either an extended Kalman filter or an unscented Kalman filter to estimate the optimal outlet water temperature. The optimal estimated temperature is the best estimated outlet water temperature output by the state observer. The preset temperature setpoint is the target temperature value required by the process. The deviation is the difference between the optimal estimated temperature and the preset temperature setpoint. The feedback controller is a device that dynamically selects a proportional-integral-derivative control algorithm or a fuzzy control algorithm based on the deviation to generate airflow compensation commands. The feedback airflow compensation command is the airflow adjustment value calculated by the feedback controller, used to compensate for errors in the feedforward model.

[0032] In some implementations, real-time collected operating data (including fan airflow, inlet water temperature, ambient temperature, and equipment thermal inertia) is input into a state observer. The state observer, based on the thermodynamic mechanism of the powder box model, establishes a nonlinear state-space model with fan airflow, inlet water temperature, ambient temperature, and equipment thermal inertia as state variables and outlet water temperature as the observed output. After discretizing the model, an extended Kalman filter or an unscented Kalman filter is used for recursive calculation. The state observer reduces measurement noise and hysteresis through filtering, outputting the optimal estimated outlet water temperature. Then, the deviation and rate of change of the deviation between the optimal estimated temperature and the preset temperature setpoint are calculated. Based on the magnitude of the deviation, a control algorithm is dynamically selected to generate a preliminary airflow compensation command. Finally, the preliminary airflow compensation command is output-limited (compared to preset upper and lower limits; if it exceeds the upper limit, it is limited to the upper limit; if it is below the lower limit, it is limited to the lower limit). During continuous limiting, anti-saturation processing is applied to the integral term (stopping integral accumulation or reducing the integral effect), outputting the final feedback airflow compensation command.

[0033] Step 104: Merge the feedforward air volume command and the feedback air volume compensation command to obtain the final air volume control command, and send it to the fan inverter for execution.

[0034] The feedforward airflow command and the feedback airflow compensation command are combined to obtain the final airflow control command, which is then sent to the fan inverter for execution, including: The feedforward weighting coefficient and feedback weighting coefficient are preset based on historical control data of the air cooler.

[0035] The feedforward air volume command is weighted according to the feedforward weighting coefficient, and the feedback air volume compensation command is weighted according to the feedback weighting coefficient.

[0036] The weighted feedforward air volume command and the weighted feedback air volume compensation command are added together to obtain the preliminary fusion command.

[0037] Obtain the current operating efficiency curve of the wind turbine frequency converter, and with the goal of minimizing the overall energy consumption of the system, perform optimization calculations on the air volume and energy consumption of the preliminary fusion command to obtain the optimal air volume command.

[0038] The initial optimal air volume command is limited to prevent it from exceeding the maximum safe operating air volume of the fan inverter. The limited command is then sent to the fan inverter as the final air volume control command.

[0039] The airflow control command is the final control signal after fusion. The fan inverter is the actuator that adjusts the fan speed. The feedforward weighting coefficient and feedback weighting coefficient are weighted proportional coefficients preset based on historical loop-cooled fan control data. The preliminary fusion command is the command value after weighted summation. The current operating efficiency curve of the fan inverter is the energy consumption efficiency relationship curve of the inverter under different airflow rates. Minimizing the overall system energy consumption is the optimization goal, requiring the minimization of total power consumption. The airflow and energy consumption optimization calculation is an optimization operation of the airflow command with the goal of minimizing energy consumption. The optimal airflow command is the command value after optimization calculation. Limiting processing is the process of limiting the command value to not exceed the maximum safe operating airflow. The maximum safe operating airflow is the highest safe airflow threshold allowed by the fan inverter.

[0040] In some implementations, feedforward and feedback weighting coefficients are first preset based on historical loop-cooled fan control data to reflect the relative importance of feedforward and feedback commands in the fusion process. Then, the feedforward airflow command is weighted according to the feedforward weighting coefficient, and the feedback airflow compensation command is weighted according to the feedback weighting coefficient. The weighted feedforward airflow command and the weighted feedback airflow compensation command are then added to obtain the preliminary fusion command. Next, the current operating efficiency curve of the fan inverter is obtained to illustrate the relationship between airflow and energy consumption. With the goal of minimizing overall system energy consumption, the preliminary fusion command undergoes airflow and energy consumption optimization calculations. The airflow value corresponding to the lowest energy consumption point can be solved iteratively or by algorithm, outputting the optimal airflow command. Finally, the preliminary optimal airflow command is limited. The command value is compared with the preset maximum safe operating airflow limit. If it exceeds the limit, it is limited to the limit value; otherwise, the original value is retained, and the limited command is sent to the fan inverter as the final airflow control command for execution.

[0041] Step 105: Continuously monitor the value and trend of the feedback air volume compensation command, and use the feedback air volume compensation command as the prediction error signal of the toner cartridge model. Adaptively adjust the key thermal parameters in the toner cartridge model to make the feedback air volume compensation command approach zero.

[0042] By adaptively adjusting key thermal parameters in the powder box model, including: The feedback air volume compensation command is used as the dynamic prediction error signal for the toner cartridge model.

[0043] Based on the preset adaptive adjustment law, the key dynamic parameters output by the machine learning module in the toner cartridge model are adjusted according to the value of the dynamic prediction error signal and the time integral, so as to make the feedback air volume compensation command approach zero.

[0044] The adaptive adjustment law is θ represents the key dynamic parameters to be adjusted, including the calibration coefficient for the total heat loss power of the equipment, the overall efficiency coefficient of the heat dissipation system, and the fluid resistance coefficient of the air duct. new The adjusted parameter value, θ old The parameters are the values ​​before adjustment, α is the adaptive learning rate, and M is the value before adjustment. fb The air volume compensation command is used for feedback, where t is time.

[0045] Among them, the dynamic prediction error signal is a quantitative indicator of the prediction deviation of the toner cartridge model, which uses the feedback air volume compensation command as the basis. The adaptive adjustment law is a preset parameter update formula.

[0046] In some implementations, the real-time values ​​and trends of the feedback airflow compensation commands (such as continuous positive or negative compensation states) are continuously monitored. The feedback airflow compensation commands are then input as dynamic prediction error signals to the toner cartridge model and fed into the parameter adjustment module. Based on a preset adaptive adjustment law formula, the calibration coefficients of three key parameters dynamically generated by the machine learning module in the toner cartridge model (total heat loss power of the equipment, overall efficiency coefficient of the heat dissipation system η, and air duct fluid resistance coefficient K) are adjusted. flow The parameters are iteratively updated; finally, the updated parameters are backfilled into the toner cartridge model. By dynamically correcting the model parameters, the feedforward prediction is made closer to the actual needs, and the feedback air volume compensation command is driven to approach zero.

[0047] Based on the above technical solution, a powder box model is constructed using a hybrid model architecture that combines a mechanistic framework (energy conservation + fluid dynamics) with data-driven parameter dynamic mapping to obtain feedforward commands. Then, a state observer is constructed by embedding an extended Kalman filter or unscented Kalman filter algorithm through a nonlinear state-space model to obtain feedback compensation. This allows the feedforward commands and feedback compensation to be weighted and fused according to historical weights, and then superimposed on the inverter efficiency curve for optimization. This enables the efficiency of the ring-cooled fan to be adjusted according to the actual use of the pumping station. This not only avoids energy waste during normal use, but also avoids control delays that affect the efficiency of the ring-cooled fan during energy-saving use, reducing the impact of nonlinear factors in feedforward control on control accuracy.

[0048] In one possible implementation of the embodiments of this application, combined with Figure 1 ,like Figure 2 As shown, the powder box model can be constructed through the following steps 201 to 203, which are explained in detail below: Step 201: Based on the heat balance mechanism and fluid mechanics mechanism of the pumping station, establish a basic mechanism model with the pump operating frequency, inlet flow rate, outlet pressure, inlet water temperature, and ambient temperature from the operating data as input, and the theoretical heat dissipation air volume required by the annular cooling fan as output. The calculation unit of the basic mechanism model is... Q reqFor the theoretical heat dissipation airflow, P loss The total heat loss power of the equipment is calculated using the pump operating frequency, inlet flow rate, and outlet pressure. η is the overall efficiency coefficient of the heat dissipation system, ΔT is the desired cooling temperature difference setpoint, and C is the total heat loss power of the equipment. pair P is the specific heat capacity of air. air K is the density of air. flow This is the fluid resistance coefficient of the air duct.

[0049] The pump station heat balance mechanism is based on the principle of energy conservation, calculating the balance between heat generation and dissipation in the pump station. The fluid mechanics mechanism involves the physical principles of fluid flow, including pressure, flow rate, and resistance. The theoretical cooling air volume required by the annular cooling fan is calculated based on the heat balance and represents the air volume the fan needs to provide to dissipate heat.

[0050] In some implementations, based on the heat balance mechanism and fluid mechanics mechanism of the pumping station, the pump operating frequency, inlet flow rate, and outlet pressure are first used as input parameters. The total heat loss power of the equipment is calculated using fluid mechanics formulas to reflect the energy loss during pump operation. Then, combined with the inlet water temperature and ambient temperature, the desired cooling temperature difference setpoint is determined, and preset comprehensive efficiency coefficient of the heat dissipation system, air specific heat capacity, air density, and duct fluid resistance coefficient are introduced. Finally, based on the heat balance equation, the total heat loss power of the equipment, the cooling temperature difference setpoint, the efficiency coefficient, air characteristic parameters, and the resistance coefficient are comprehensively calculated to obtain the theoretical heat dissipation air volume required by the annular cooling fan as the output, thus completing the construction of the basic mechanism model.

[0051] It should be noted that preset parameters such as the overall efficiency coefficient of the heat dissipation system, specific heat capacity of air, air density, and air duct fluid resistance coefficient can be calibrated through experiments or historical data to ensure accuracy.

[0052] Step 202: Construct a machine learning module using historical and real-time operating condition data as training sets, and output key dynamic parameters in the basic mechanism model through dynamic mapping. Key dynamic parameters include the calibration coefficient of the total heat loss power of the equipment, the comprehensive efficiency coefficient of the heat dissipation system, and the fluid resistance coefficient of the air duct. Operating condition data include inlet water temperature, outlet water temperature, ambient temperature, historical operating air volume of the fan and the corresponding temperature change rate, water pump operating frequency, inlet water flow rate, and outlet water pressure.

[0053] Historical operating data refers to the set of parameters recorded during the system's past operation, while real-time operating data refers to the set of real-time parameters collected during the system's current operation. The machine learning module is a program unit that learns data patterns through algorithms. Dynamic mapping refers to the process of generating output parameters in real time based on input data. The calibration coefficient for the total heat loss power of the equipment is a proportional factor used to correct the heat loss value calculated by the mechanistic model. The comprehensive efficiency coefficient of the heat dissipation system is a variable characterizing the overall energy conversion efficiency of the heat dissipation system. The duct fluid resistance coefficient is a parameter reflecting the airflow resistance characteristics within the duct. The inlet water temperature and outlet water temperature are the water temperature values ​​at the pump inlet and outlet, respectively.

[0054] In some implementations, historical operating condition data (such as previously recorded inlet water temperature, outlet water temperature, ambient temperature, fan air volume and temperature change rate, pump frequency, inlet water flow rate, and outlet water pressure) and real-time collected data of similar operating conditions are used as training sets and input into the machine learning module. The machine learning module analyzes the inherent correlation of the input data through supervised learning algorithms (such as neural networks or random forests), dynamically maps and outputs three key dynamic parameters in the basic mechanism model. Among the three key dynamic parameters, a calibration coefficient for the total heat loss power of the equipment is first generated to correct the heat loss value calculated by the mechanism model. Then, the comprehensive efficiency coefficient of the heat dissipation system is output to characterize the actual efficiency level of the current heat dissipation system. Finally, the air duct fluid resistance coefficient is generated to reflect the real-time resistance characteristics of the air duct structure. Furthermore, the three key dynamic parameters are continuously updated according to changes in real-time operating conditions.

[0055] Step 203: Input the key dynamic parameters output by the machine learning module into the basic mechanism model, perform real-time correction on the basic mechanism model, form the powder box model, and output the theoretical heat dissipation air volume as the feedforward air volume command.

[0056] The powder box model is a hybrid model formed by dynamically correcting the basic mechanism model. The theoretical heat dissipation airflow is the theoretical airflow requirement of the annular fan calculated by the powder box model. The feedforward airflow command is a control signal used to pre-adjust the airflow of the annular fan.

[0057] In some implementations, the calibration coefficient of the total heat loss power of the equipment, the comprehensive efficiency coefficient of the heat dissipation system, and the fluid resistance coefficient of the air duct are input into the basic mechanism model in real time. Then, the calibration coefficient is used to correct the calculation result of the total heat loss power of the equipment. At the same time, the fixed efficiency coefficient preset by the original model is replaced with the dynamically generated efficiency coefficient, and the air duct resistance parameter is updated with the dynamic resistance coefficient. After the parameter replacement is completed, the basic mechanism model is corrected into a powder box model in real time. Finally, the powder box model can run the original heat balance calculation formula based on the current operating data (pump frequency, flow rate, pressure, etc.), output the theoretical heat dissipation air volume, and send it directly to the control system as a feedforward air volume command.

[0058] Based on the above technical solution, a hybrid model structure is constructed using a mechanistic model as the underlying computational framework and a machine learning module as the parameter generation layer. This allows for the integration of data-driven flexibility while preserving physical interpretability, avoiding the rigidity of pure mechanistic model parameters and the lack of physical constraints in pure data models. Secondly, multi-source operating condition data is input into the machine learning module to generate calibration coefficients, efficiency coefficients, and drag coefficients in real time for dynamic parameter mapping, enabling adaptive correction of heat loss calculations, heat dissipation efficiency, and airflow resistance. Finally, by directly replacing the static parameters in the basic model with dynamic parameters, the output of the powder box model is updated instantly for online real-time correction, thereby improving the accuracy of feedforward instructions and reducing reliance on feedback compensation.

[0059] In one possible implementation of this application embodiment, the state observer can be constructed through the following steps 301 to 303, which are described in detail below: Step 301: Based on the thermodynamic mechanism structure of the powder box model, establish a nonlinear state-space model with the fan air volume, inlet water temperature, ambient temperature and equipment thermal inertia in the working condition data as state variables and the outlet water temperature as the observed output. Thermodynamic mechanism structure refers to the model framework based on the principles of energy conservation and thermal balance. Nonlinear state-space model is a mathematical model that uses nonlinear state equations and output equations to describe the dynamic behavior of a system.

[0060] In some implementations, based on the thermodynamics of the powder box model and the mechanism structure of the principle of energy conservation and fluid mechanics, the fan air volume, inlet water temperature, ambient temperature and equipment thermal inertia in the operating data are selected as state variables to jointly characterize the thermal dynamic characteristics of the system. Then, the outlet water temperature is used as the observed output to reflect the cooling effect. Finally, a nonlinear state-space model is established to describe the relationship between the state variables and the observed output through nonlinear equations to accurately capture the complex dynamic behavior of the system.

[0061] Step 302: Discretize the nonlinear state-space model and construct a state observer based on the extended Kalman filter algorithm or the unscented Kalman filter algorithm.

[0062] Discretization is the process of converting a continuous-time model into a discrete-time model to adapt to digital control. The Extended Kalman Filter (EKF) algorithm is a state estimation algorithm for nonlinear systems that uses linearization. The Unscented Kalman Filter (UKF) algorithm utilizes the unscented transform to process the state estimation of nonlinear systems. A state observer is a device constructed based on filtering algorithms to estimate the internal state of a system, such as the outlet water temperature.

[0063] In some implementations, a nonlinear state-space model is established based on the thermodynamic mechanism structure of the powder box model. The system's thermal dynamics are described by nonlinear equations using operating data such as fan airflow, inlet water temperature, ambient temperature, and equipment thermal inertia as state variables, and outlet water temperature as the observed output. This nonlinear state-space model is then discretized from continuous-time to discrete-time form, facilitating real-time calculations in a digital control system. Next, a state observer is constructed based on either the Extended Kalman Filter (EKF) or Unscented Kalman Filter (UKF) algorithm. The EKF performs state estimation by linearizing the nonlinear model, while the UKF utilizes unscented transformations to avoid linearization errors. Finally, the real-time collected fan airflow, inlet water temperature, and ambient temperature are injected into the state observer. The state variables are updated progressively through recursive calculations, ultimately outputting the optimal estimate of the outlet water temperature.

[0064] Step 303: Input the real-time collected fan operating air volume, inlet water temperature, and ambient temperature into the state observer, and obtain the optimal estimate of the outlet water temperature through recursive calculation.

[0065] The optimal estimate is the accurate state estimate obtained by reducing the influence of noise through a filtering algorithm.

[0066] In some implementations, the fan's operating air volume, inlet water temperature, and ambient temperature are collected in real time as input signals and transmitted to the control system via sensors. These input signals are then injected into a state observer, which converts them into digital form through discretization. The state observer then uses an extended Kalman filter or an unscented Kalman filter to perform recursive calculations, iteratively updating state variables, such as equipment thermal inertia and outlet water temperature, with time steps. Finally, the recursive calculations progressively optimize the state estimate, outputting the optimal estimate of the outlet water temperature while reducing measurement noise and hysteresis effects, for subsequent feedback control.

[0067] Based on the above technical solution, by deeply integrating the energy conservation mechanism with data-driven parameters, thermodynamic dynamic behavior can be captured through nonlinear equations, thereby achieving high-precision feedforward control and reducing model prediction errors. Simultaneously, the powder box model dynamically adjusts its parameters through a machine learning module, improving its adaptability under varying operating conditions (such as changes in material properties or environmental fluctuations), thus enhancing the accuracy of feedforward commands. Reduced control latency allows the system to respond quickly to load changes, avoiding energy waste. Furthermore, discretization and Kalman filtering algorithms effectively reduce measurement noise and system hysteresis.

[0068] In one possible implementation of this application embodiment, the feedback air volume compensation command calculated by the feedback controller based on the deviation can be achieved through the following steps 401 to 403, which are described in detail below: Step 401: Subtract the optimal estimated temperature value from the preset temperature setting value to obtain the temperature deviation and the rate of change of deviation.

[0069] The preset temperature setpoint is the target temperature value required by the process, which is preset by the system. The deviation change rate is the rate of change of temperature deviation over time, representing the speed at which the deviation changes.

[0070] In some implementations, the optimal estimated temperature value output from the state observer and the preset temperature setting value set by the user are obtained in real time; then the optimal estimated temperature value is subtracted from the preset temperature setting value to obtain the real-time temperature deviation; then, by monitoring the change of temperature deviation over time, the deviation change rate is calculated using a difference method, for example, taking the difference between the temperature deviation at the current moment and the temperature deviation at the previous moment, and then dividing it by the time interval to obtain the deviation change rate.

[0071] Step 402: Based on the magnitude of the temperature deviation and the rate of change of the deviation, dynamically select the proportional-integral-derivative control algorithm or the fuzzy control algorithm to calculate the temperature deviation and the rate of change of the deviation, and generate a preliminary air volume compensation command.

[0072] Based on the magnitude of the temperature deviation and its rate of change, a proportional-integral-derivative (PID) control algorithm or a fuzzy control algorithm is dynamically selected to calculate the temperature deviation and its rate of change, generating preliminary airflow compensation commands, including: The preset threshold is dynamically set based on the mean of the rate of change of the best estimated temperature value in historical data, and the preset threshold is updated according to the preset maintenance cycle of the hydropower station.

[0073] When the absolute value of the temperature deviation is less than or equal to the preset threshold, the proportional-integral-derivative control algorithm is selected. Based on the temperature deviation and the rate of change of the deviation, proportional, integral, and derivative operations are performed to generate a preliminary air volume compensation command.

[0074] When the absolute value of the temperature deviation is greater than the preset threshold, the fuzzy control algorithm is selected, and calculations are performed according to the fuzzy logic rules of the temperature deviation and the rate of change of the deviation to generate a preliminary air volume compensation command.

[0075] Among them, the Proportional-Integral-Derivative (PID) control algorithm is a control method based on deviation and the rate of change of deviation, performing proportional, integral, and derivative operations. It is suitable for linear or small-deviation scenarios. The fuzzy control algorithm is a control method based on fuzzy logic rules to calculate deviation and the rate of change of deviation, suitable for nonlinear or large-deviation scenarios. The initial airflow compensation command is the initial command value generated by the controller to compensate for the airflow of the circulating air cooler.

[0076] Step 403: Perform output limiting and integral anti-saturation processing on the preliminary air volume compensation command to obtain the final feedback air volume compensation command.

[0077] The initial airflow compensation command is subjected to output limiting and integral anti-saturation processing to obtain the final feedback airflow compensation command, which includes: The initial airflow compensation command is compared with the preset upper and lower limits of the airflow command.

[0078] If the initial airflow compensation command exceeds the upper limit, the final feedback airflow compensation command will be limited to the upper limit.

[0079] If the initial airflow compensation command is lower than the lower limit, the final feedback airflow compensation command will be limited to the lower limit.

[0080] If the initial airflow compensation command is between the upper and lower limits, it will be directly used as the final feedback airflow compensation command.

[0081] When a proportional-integral-derivative control algorithm is used and the initial air volume compensation command is continuously in a limited state, anti-saturation processing is performed on the integral term to stop integral accumulation or reduce integral action, in order to prevent excessive accumulation of the integral term from causing control overshoot.

[0082] The output limiting mechanism compares the initial command with preset upper and lower limits for the airflow command. If the command exceeds the upper limit, it is limited to the upper limit; if it falls below the lower limit, it is limited to the lower limit; if it is within the range, it is used directly, ensuring the command remains within a safe range. Integral anti-saturation processing stops integral accumulation or reduces integral action when a proportional-integral-derivative control algorithm is used and the command remains in a limited state, preventing excessive accumulation of the integral term and control overshoot. The final feedback airflow compensation command is the compensation value after output limiting and integral anti-saturation processing, used to generate the final control command with the fused feedforward command.

[0083] Based on the above technical solution, the optimal estimated temperature value is output by the state observer, and the difference between this value and the preset temperature setpoint is used to obtain the temperature deviation and the rate of change of deviation. Next, based on a preset threshold, a PID algorithm or a fuzzy control algorithm is dynamically selected to generate an initial airflow compensation command. Finally, through output limiting and integral anti-saturation processing, the final feedback airflow compensation command is output. This avoids the problems of low control accuracy and lag caused by the inability to dynamically handle nonlinear relationships, which are inherent in traditional single PID or fuzzy control algorithms. Simultaneously, it avoids the situation where the integral term easily accumulates excessively under continuous limiting conditions, leading to system oscillation or equipment damage, as is common in traditional PID control.

[0084] In one possible implementation of this application embodiment, a dual-loop control system is constructed. The dual-loop control system includes: a feedforward and feedback control outer loop that quickly responds to load changes and a model self-updating inner loop that slowly optimizes feedforward model parameters. This can be achieved through the following steps 501 to 403, which are described in detail below: Step 501: The feedforward and feedback control outer loop, which responds quickly to load changes, generates a feedforward airflow command through the powder box model and a feedback airflow compensation command through the state observer and feedback controller. After fusing the feedforward airflow command and the feedback airflow compensation command, the final airflow control command is issued for execution to control the airflow of the ring-cooled fan.

[0085] In some implementations, the outer-loop toner cartridge model first calculates and generates feedforward airflow commands based on real-time detected load changes, allowing it to predict the required airflow adjustment to offset the main impact of load disturbances. Then, a state observer uses measurable input / output signals (such as fan speed, pressure, or flow rate) to estimate indirectly measurable state variables within the system (such as actual airflow and temperature distribution), providing more comprehensive system operating information. Simultaneously, the feedback controller receives the state estimates from the state observer and compares them with the system's desired target values. Based on the generated error signal (such as airflow deviation), a feedback airflow compensation command is generated according to a preset control algorithm (such as PID regulation) to correct the deficiencies of the feedforward command and cope with unmodeled disturbances. Next, the feedforward airflow command and the feedback airflow compensation command are fused to ensure smooth command transition and overall control stability. Finally, the fused final airflow control command is sent to the actuator (such as a frequency converter or damper driver) to directly adjust the operating parameters of the ring-cooled fan (such as motor frequency or blade angle), thereby achieving fast and accurate airflow control to cope with load changes and maintain system stability.

[0086] Step 502: The self-updating inner loop of the slow-optimization feedforward model continuously monitors and feeds back the air volume compensation command. The feedback air volume compensation command is used as the dynamic prediction error signal of the toner cartridge model. The key dynamic parameters output by the machine learning module in the toner cartridge model are dynamically adjusted through the adaptive adjustment law for online self-correction and slow optimization of the feedforward model.

[0087] In some implementations, feedback airflow compensation commands are continuously monitored to reflect the deviation between the feedforward model output and actual demand. These commands are then converted into a dynamic prediction error signal for the toner cartridge model, quantifying the model's prediction inaccuracy under current parameters. An adaptive adjustment law is then activated, generating parameter adjustments based on the magnitude and direction of the dynamic prediction error signal, ensuring the adjustment process follows gradient descent or Lyapunov stability principles to guarantee stability and convergence. Simultaneously, a machine learning module, as part of the toner cartridge model, continuously runs, learning from historical data and outputting key dynamic parameters (such as nonlinear mapping coefficients or time constants) to improve the model's prediction accuracy. The adaptive adjustment law dynamically adjusts these key dynamic parameters, gradually reducing the dynamic prediction error signal through fine-tuning. Finally, the adjusted parameters are applied to the feedforward model, enabling online self-correction and slow optimization, thereby improving the model's accuracy and adaptability over long-term operation.

[0088] Based on the above technical solution, the outer loop integrates the toner cartridge model (feedforward airflow command) with the commands generated by the state observer and feedback controller (feedback airflow compensation command), forming a composite control command. This composite command is responsible for real-time airflow control and improves the speed of forward response and the robustness of feedback. Meanwhile, the feedback airflow compensation command is directly treated as the dynamic prediction error signal of the toner cartridge model to self-update the inner loop, allowing it to focus on the continuous optimization of the feedforward model. This synchronization between the outer and inner loops enables decoupling and parallel processing of control and optimization, achieving a balance between the control system's response speed and optimization accuracy. It can quickly respond to disturbances and continuously improve itself, enhancing long-term control accuracy.

[0089] The above primarily describes the solutions of the embodiments of this application from the perspective of device implementation. It is understood that each device, for example, an automatic energy-saving control device for a ring-cooled fan based on load feedback, includes at least one of the hardware structures and software modules corresponding to each function in order to achieve the above-mentioned functions. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0090] This application embodiment can divide the automatic energy-saving control device for a ring-cooled fan based on load feedback into functional units according to the above method example. For example, each function can be divided into separate functional units, or two or more functions can be integrated into the same processing unit. The integrated unit can be implemented in hardware or as a software functional unit. It should be noted that the unit division in this application embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0091] When using integrated units, Figure 6 A possible structural schematic diagram of an automatic energy-saving control device for a ring-cooled fan based on load feedback (referred to as an automatic energy-saving control device for a ring-cooled fan based on load feedback 60) involved in the above embodiments is shown. The automatic energy-saving control device for a ring-cooled fan based on load feedback 60 includes a processing unit 601 and a communication unit 602, and may also include a storage unit 603. Figure 6 The structural diagram shown can be used to illustrate the structure of the automatic energy-saving control device for the ring-cooled fan based on load feedback involved in the above embodiments.

[0092] when Figure 6 The schematic diagram shown illustrates the structure of the load feedback-based automatic energy-saving control device for a ring-cooled fan involved in the above embodiments. The processing unit 601 is used to control and manage the operation of the load feedback-based automatic energy-saving control device for a ring-cooled fan. The communication unit 602 is used for the load feedback-based automatic energy-saving control device for a ring-cooled fan to communicate with other devices. The storage unit 603 is used to store the program code and data of the load feedback-based automatic energy-saving control device for a ring-cooled fan.

[0093] For example, the communication unit 602 is used to collect real-time operating data of water pumps and cooling fans in the water plant pumping station; The processing unit 601 is used to input operating condition data into a preset toner cartridge model and calculate the feedforward airflow command. The toner cartridge model is a hybrid model that dynamically integrates energy conservation mechanisms, fluid dynamics mechanisms, and data-driven parameters. Key thermal and hydraulic parameters within the toner cartridge model are dynamically mapped and generated by a machine learning module based on real-time operating conditions. Operating condition data is input to a state observer to obtain the optimal estimated temperature value. The deviation between the optimal estimated temperature value and the preset temperature setpoint is calculated, and a feedback airflow compensation command is calculated based on this deviation using a feedback controller. The feedforward airflow command and the feedback airflow compensation command are fused to obtain the final airflow control command, which is then sent to the fan inverter for execution. The value and trend of the feedback airflow compensation command are continuously monitored, and the feedback airflow compensation command is used as a prediction error signal for the toner cartridge model. Key thermal parameters in the toner cartridge model are adaptively adjusted to bring the feedback airflow compensation command closer to zero.

[0094] In one possible implementation, the processing unit 601 is also used to construct the powder box model, including: based on the thermal balance mechanism and fluid dynamics mechanism of the pumping station, establishing a basic mechanism model with the pump operating frequency, inlet flow rate, outlet pressure, inlet water temperature, and ambient temperature from the operating data as input, and the theoretical heat dissipation air volume required by the ring-cooled fan as output. The calculation unit of the basic mechanism model is... Q req For the theoretical heat dissipation airflow, P loss The total heat loss power of the equipment is calculated using the pump operating frequency, inlet flow rate, and outlet pressure. η is the overall efficiency coefficient of the heat dissipation system, ΔT is the desired cooling temperature difference setpoint, and C is the total heat loss power of the equipment. pair P is the specific heat capacity of air. air K is the density of air. flow This refers to the duct fluid resistance coefficient. A machine learning module is constructed using historical and real-time operating data as training sets. It dynamically maps key dynamic parameters from the basic mechanism model for output. These key dynamic parameters include the calibration coefficient for the total heat loss power of the equipment, the overall efficiency coefficient of the heat dissipation system, and the duct fluid resistance coefficient. Operating data includes inlet water temperature, outlet water temperature, ambient temperature, historical fan operating airflow and corresponding temperature change rate, pump operating frequency, inlet water flow rate, and outlet water pressure. The key dynamic parameters output by the machine learning module are input into the basic mechanism model for real-time correction, forming a powder box model. The theoretical heat dissipation airflow is then output as a feedforward airflow command.

[0095] In one possible implementation, the processing unit 601 is further used to construct the state observer, including: establishing a nonlinear state-space model based on the thermodynamic mechanism structure of the powder box model, with fan airflow, inlet water temperature, ambient temperature, and equipment thermal inertia from operating data as state variables, and outlet water temperature as the observed output; discretizing the nonlinear state-space model; and constructing the state observer based on the extended Kalman filter algorithm or the unscented Kalman filter algorithm. The real-time collected fan operating airflow, inlet water temperature, and ambient temperature are input to the state observer, and the optimal estimate of the outlet water temperature is obtained through recursive calculation.

[0096] In one possible implementation, the processing unit 601 is further configured to calculate a feedback airflow compensation command based on the deviation using a feedback controller, including: subtracting the optimal estimated temperature value from a preset temperature setpoint to obtain the temperature deviation and the rate of change of the deviation; dynamically selecting a proportional-integral-derivative control algorithm or a fuzzy control algorithm based on the magnitude of the temperature deviation and the rate of change of the deviation to calculate the temperature deviation and the rate of change of the deviation, generating a preliminary airflow compensation command; and performing output limiting and integral anti-saturation processing on the preliminary airflow compensation command to obtain the final feedback airflow compensation command.

[0097] In one possible implementation, the processing unit 601 is further configured to dynamically select a proportional-integral-derivative (PID) control algorithm or a fuzzy control algorithm to calculate the temperature deviation and its rate of change based on the magnitude of the temperature deviation and the rate of change of the deviation, generating a preliminary airflow compensation command. This includes: dynamically setting a preset threshold based on the average of the rate of change of the best estimated temperature value in historical data, with the preset threshold being updated according to a preset hydropower station maintenance cycle. When the absolute value of the temperature deviation is less than or equal to the preset threshold, the PID control algorithm is selected, and proportional, integral, and derivative operations are performed based on the temperature deviation and its rate of change of the deviation to generate a preliminary airflow compensation command. When the absolute value of the temperature deviation is greater than the preset threshold, the fuzzy control algorithm is selected, and calculations are performed based on the fuzzy logic rules of the temperature deviation and its rate of change of the deviation to generate a preliminary airflow compensation command.

[0098] In one possible implementation, the processing unit 601 is further configured to perform output limiting and integral anti-saturation processing on the preliminary airflow compensation command to obtain the final feedback airflow compensation command, including: comparing the preliminary airflow compensation command with preset upper and lower limits of the airflow command. If the preliminary airflow compensation command exceeds the upper limit, the final feedback airflow compensation command is limited to the upper limit. If the preliminary airflow compensation command is lower than the lower limit, the final feedback airflow compensation command is limited to the lower limit. If the preliminary airflow compensation command is between the upper and lower limits, it is directly used as the final feedback airflow compensation command. When a proportional-integral-derivative control algorithm is used and the preliminary airflow compensation command remains in a limited state, anti-saturation processing is performed on the integral term to stop integral accumulation or reduce the integral action, in order to prevent excessive accumulation of the integral term from causing control overshoot.

[0099] In one possible implementation, the processing unit 601 is further configured to fuse the feedforward airflow command and the feedback airflow compensation command to obtain the final airflow control command, and send it to the fan inverter for execution. This includes: presetting feedforward weighting coefficients and feedback weighting coefficients based on historical loop-cooled fan control data; weighting the feedforward airflow command according to the feedforward weighting coefficients, and weighting the feedback airflow compensation command according to the feedback weighting coefficients; adding the weighted feedforward airflow command and the weighted feedback airflow compensation command to obtain a preliminary fused command; acquiring the current operating efficiency curve of the fan inverter, and performing airflow and energy consumption optimization calculations on the preliminary fused command with the goal of minimizing overall system energy consumption, to obtain the optimal airflow command; and limiting the preliminary optimal airflow command to ensure that the preliminary fused command does not exceed the maximum safe operating airflow of the fan inverter, and sending the limited command as the final airflow control command to the fan inverter.

[0100] In one possible implementation, the processing unit 601 is further configured to adaptively adjust key thermal parameters in the toner cartridge model, including: using the feedback airflow compensation command as a dynamic prediction error signal for the toner cartridge model. Based on a preset adaptive adjustment law, and according to the value of the dynamic prediction error signal and its time integral, the key dynamic parameters output by the machine learning module in the toner cartridge model are adjusted to bring the feedback airflow compensation command close to zero. The adaptive adjustment law is... θ represents the key dynamic parameters to be adjusted, including the calibration coefficient for the total heat loss power of the equipment, the overall efficiency coefficient of the heat dissipation system, and the fluid resistance coefficient of the air duct. new The adjusted parameter value, θ old The parameters are the values ​​before adjustment, α is the adaptive learning rate, and M is the value before adjustment. fb The air volume compensation command is used for feedback, where t is time.

[0101] In one possible implementation, the processing unit 601 is further configured to construct a dual-loop control system, comprising: a feedforward and feedback control outer loop for rapid response to load changes, and a model self-updating inner loop for slow optimization of feedforward model parameters. The feedforward and feedback control outer loop generates feedforward airflow commands through the toner cartridge model and generates feedback airflow compensation commands through a state observer and a feedback controller. The feedforward and feedback airflow compensation commands are then merged and the final airflow control command is issued for execution, controlling the airflow of the air cooler. The model self-updating inner loop continuously monitors the feedback airflow compensation commands, using them as dynamic prediction error signals for the toner cartridge model. It dynamically adjusts the key dynamic parameters output by the machine learning module in the toner cartridge model using an adaptive adjustment law, enabling online self-correction and slow optimization of the feedforward model.

[0102] The processing unit 601 can be a processor or a controller, and the communication unit 602 can be a communication interface, transceiver, transceiver circuit, transceiver device, etc. The term "communication interface" is a general term and may include one or more interfaces. The storage unit 603 can be a memory. When the load feedback-based automatic energy-saving control device 60 for the air-cooled fan is a chip, the processing unit 601 can be a processor or a controller, and the communication unit 602 can be an input interface and / or an output interface, pins, or circuits, etc. The storage unit 603 can be a storage unit within the chip (e.g., a register, cache, etc.) or a storage unit located outside the chip (e.g., read-only memory (ROM), random access memory (RAM, etc.)).

[0103] The communication unit can also be called a transceiver unit. The antenna and control circuit with transceiver functions in the load feedback-based automatic energy-saving control device 60 for the ring-cooled fan can be considered as the communication unit 602 of the load feedback-based automatic energy-saving control device 60, and the processor with processing functions can be considered as the processing unit 601 of the load feedback-based automatic energy-saving control device 60. Optionally, the device in the communication unit 602 used to implement the receiving function can be considered as the communication unit, which is used to execute the receiving steps in the embodiments of this application. The communication unit can be a receiver, a receiver circuit, etc. The device in the communication unit 602 used to implement the transmitting function can be considered as the transmitting unit, which is used to execute the transmitting steps in the embodiments of this application. The transmitting unit can be a transmitter, a transmitter, a transmitting circuit, etc.

[0104] Figure 6If the integrated units in the process are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, 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.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. Storage media for storing computer software products include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0105] In implementation, each step of the method provided in this embodiment can be completed by integrated logic circuits in the processor hardware or by instructions in software. The steps of the method disclosed in the embodiments of this application can be directly manifested as being executed by a hardware processor, or being executed by a combination of hardware and software modules in the processor.

[0106] The processor in this application may include, but is not limited to, at least one of the following: a central processing unit (CPU), a microprocessor, a digital signal processor (DSP), a microcontroller unit (MCU), or an artificial intelligence processor, etc., and other computing devices that run software. Each computing device may include one or more cores for executing software instructions to perform calculations or processing. The processor may be a standalone semiconductor chip or integrated with other circuits into a single semiconductor chip. For example, it may form a System-on-a-Chip (SoC) with other circuits (such as encoding / decoding circuits, hardware acceleration circuits, or various bus and interface circuits), or it may be integrated as a built-in processor within an ASIC. The ASIC of this integrated processor may be packaged separately or together with other circuits. In addition to the cores for executing software instructions to perform calculations or processing, the processor may further include necessary hardware accelerators, such as field-programmable gate arrays (FPGAs), programmable logic devices (PLDs), or logic circuits that implement dedicated logic operations.

[0107] The memory in the embodiments of this application may include at least one of the following types: read-only memory (ROM) or other types of static storage devices capable of storing static information and instructions; random access memory (RAM) or other types of dynamic storage devices capable of storing information and instructions; or electrically erasable programmable-only memory (EEPROM). In some scenarios, the memory may also be compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media, or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0108] This application also provides a computer-readable storage medium including instructions that, when run on a computer, cause the computer to perform any of the methods described above.

[0109] This application also provides a computer program product containing instructions that, when run on a computer, cause the computer to perform any of the methods described above.

[0110] This application also provides a chip including a processor and an interface circuit. The interface circuit is coupled to the processor. The processor is used to run computer programs or instructions to implement the above-described method. The interface circuit is used to communicate with other modules outside the chip.

[0111] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software programs, implementation can be, in whole or in part, in the form of a computer program product. This computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device containing one or more servers, data centers, etc., that can be integrated with the medium. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

[0112] Although this application has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings, the disclosure, and the appended claims, will understand and implement other variations of the disclosed embodiments in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude multiple components. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce good results.

[0113] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely illustrative descriptions of the application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from the spirit and scope of this application. Thus, if such modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is also intended to include such modifications and variations.

Claims

1. An automatic energy-saving control method for a ring-cooled fan based on load feedback, characterized in that, include: Real-time acquisition of operating data of water pumps and cooling fans in water plant pumping stations; The operating condition data is input into a preset toner cartridge model to calculate the feedforward air volume command; The powder box model is a hybrid model that dynamically integrates energy conservation mechanism, fluid dynamics mechanism and data-driven parameters, and the key thermal and hydraulic parameters inside the powder box model are dynamically mapped and generated by the machine learning module according to real-time operating conditions. The operating condition data is input into the condition observer to obtain the optimal estimated temperature value; The deviation between the optimal estimated temperature value and the preset temperature setting value is calculated, and a feedback air volume compensation command is calculated based on the deviation by the feedback controller. The feedforward air volume command and the feedback air volume compensation command are merged to obtain the final air volume control command, which is then sent to the fan inverter for execution. The value and trend of the feedback air volume compensation command are continuously monitored, and the feedback air volume compensation command is used as the prediction error signal of the toner cartridge model. The key thermal parameters in the toner cartridge model are adaptively adjusted to make the feedback air volume compensation command approach zero.

2. The automatic energy-saving control method for a ring-cooled fan based on load feedback according to claim 1, characterized in that, The construction of the powder box model includes: Based on the heat balance mechanism and fluid mechanics mechanism of the pumping station, a basic mechanism model is established. This model takes the pump operating frequency, inlet flow rate, outlet pressure, inlet water temperature, and ambient temperature from the aforementioned operating data as input, and the theoretical heat dissipation air volume required by the annular cooling fan as output. The computational unit of this basic mechanism model is... Q req For the theoretical heat dissipation airflow, P loss The total heat loss power of the equipment is calculated using the pump operating frequency, inlet flow rate, and outlet pressure. η is the overall efficiency coefficient of the heat dissipation system, ΔT is the desired cooling temperature difference setpoint, and C is the total heat loss power of the equipment. pair P is the specific heat capacity of air. air K is the density of air. flow This refers to the fluid resistance coefficient of the air duct. A machine learning module is constructed using historical and real-time operating condition data as training sets. It dynamically maps key dynamic parameters in the basic mechanism model and outputs them. The key dynamic parameters include the calibration coefficient of the total heat loss power of the equipment, the comprehensive efficiency coefficient of the heat dissipation system, and the fluid resistance coefficient of the air duct. The operating condition data includes inlet water temperature, outlet water temperature, ambient temperature, historical operating air volume of the fan and the corresponding temperature change rate, water pump operating frequency, inlet water flow rate, and outlet water pressure. The key dynamic parameters output by the machine learning module are input into the basic mechanism model to perform real-time correction on the basic mechanism model, forming the powder box model, and the theoretical heat dissipation air volume is output as a feedforward air volume command.

3. The automatic energy-saving control method for a ring-cooled fan based on load feedback according to claim 2, characterized in that, The construction of the state observer includes: Based on the thermodynamic mechanism structure of the powder box model, a nonlinear state-space model is established with the fan air volume, inlet water temperature, ambient temperature and equipment thermal inertia in the operating data as state variables and the outlet water temperature as the observed output. The nonlinear state-space model is discretized, and a state observer is constructed based on the extended Kalman filter algorithm or the unscented Kalman filter algorithm. The real-time collected data on the fan's operating air volume, inlet water temperature, and ambient temperature are input into the state observer, and the optimal estimate of the outlet water temperature is obtained through recursive calculation.

4. The automatic energy-saving control method for a ring-cooled fan based on load feedback according to claim 3, characterized in that, The step of obtaining the feedback airflow compensation command by the feedback controller based on the deviation includes: The temperature deviation and the rate of change of deviation are obtained by subtracting the optimal estimated temperature value from the preset temperature setting value. Based on the magnitude of the temperature deviation and the rate of change of the deviation, a proportional-integral-derivative control algorithm or a fuzzy control algorithm is dynamically selected to calculate the temperature deviation and the rate of change of the deviation, and a preliminary air volume compensation command is generated. The initial airflow compensation command is subjected to output limiting and integral anti-saturation processing to obtain the final feedback airflow compensation command.

5. The automatic energy-saving control method for a ring-cooled fan based on load feedback according to claim 4, characterized in that, The step of dynamically selecting a proportional-integral-derivative (PI-DI) control algorithm or a fuzzy control algorithm based on the magnitude of the temperature deviation and its rate of change to calculate the temperature deviation and its rate of change, and generating a preliminary airflow compensation command, includes: A preset threshold is dynamically set based on the mean of the rate of change of the best estimated temperature value in historical data, and the preset threshold is updated according to the preset hydropower station maintenance cycle. When the absolute value of the temperature deviation is less than or equal to the preset threshold, the proportional-integral-derivative control algorithm is selected, and proportional, integral, and derivative operations are performed based on the temperature deviation and the rate of change of the deviation to generate a preliminary air volume compensation command. When the absolute value of the temperature deviation is greater than the preset threshold, the fuzzy control algorithm is selected, and calculations are performed according to the fuzzy logic rules of the temperature deviation and the rate of change of the deviation to generate a preliminary air volume compensation command.

6. The automatic energy-saving control method for a ring-cooled fan based on load feedback according to claim 5, characterized in that, The process of output limiting and integral anti-saturation processing on the initial airflow compensation command to obtain the final feedback airflow compensation command includes: The initial airflow compensation command is compared with the preset upper and lower limits of the airflow command. If the initial airflow compensation command exceeds the upper limit value, the final feedback airflow compensation command will be limited to the upper limit value. If the initial airflow compensation command is lower than the lower limit value, the final feedback airflow compensation command will be limited to the lower limit value. If the initial airflow compensation command is between the upper and lower limits, it will be directly used as the final feedback airflow compensation command. When a proportional-integral-derivative control algorithm is used and the initial air volume compensation command is continuously in a limited state, anti-saturation processing is performed on the integral term to stop integral accumulation or reduce integral action, in order to prevent excessive accumulation of the integral term from causing control overshoot.

7. The automatic energy-saving control method for a ring-cooled fan based on load feedback according to claim 6, characterized in that, The step of fusing the feedforward airflow command and the feedback airflow compensation command to obtain the final airflow control command, and then sending it to the fan inverter for execution, includes: Preset feedforward and feedback weight coefficients based on historical control data of the air cooler; The feedforward air volume command is weighted according to the feedforward weighting coefficient, and the feedback air volume compensation command is weighted according to the feedback weighting coefficient. The weighted feedforward air volume command and the weighted feedback air volume compensation command are added together to obtain the preliminary fusion command; Obtain the current operating efficiency curve of the wind turbine frequency converter, and with the minimum overall system energy consumption as the optimization goal, perform air volume and energy consumption optimization calculations on the preliminary fusion command to obtain the optimal air volume command; The initial optimal airflow command is subjected to amplitude limiting processing to restrict the initial fusion command from exceeding the maximum safe operating airflow of the fan inverter, and the amplitude-limited command is sent to the fan inverter as the final airflow control command.

8. The automatic energy-saving control method for a ring-cooled fan based on load feedback according to claim 7, characterized in that, The adaptive adjustment of key thermal parameters in the powder box model includes: The feedback air volume compensation command is used as the dynamic prediction error signal of the toner cartridge model; Based on the preset adaptive adjustment law, the key dynamic parameters output by the machine learning module in the powder box model are adjusted according to the value and time integral of the dynamic prediction error signal, so as to make the feedback air volume compensation command approach zero. The adaptive adjustment law is: θ represents the key dynamic parameters to be adjusted, including the calibration coefficient of the total heat loss power of the equipment, the overall efficiency coefficient of the heat dissipation system, and the fluid resistance coefficient of the air duct. new The adjusted parameter value, θ old The parameters are the values ​​before adjustment, α is the adaptive learning rate, and M is the value before adjustment. fb The air volume compensation command is used for feedback, where t is time.

9. The automatic energy-saving control method for a ring-cooled fan based on load feedback according to claim 8, characterized in that, A dual-closed-loop control system is constructed, comprising: an outer loop of feedforward and feedback control that responds quickly to load changes and an inner loop of model self-updating that optimizes feedforward model parameters slowly. The feedforward and feedback control outer loop that responds quickly to load changes generates a feedforward airflow command through the powder box model and a feedback airflow compensation command through the state observer and feedback controller. After fusing the feedforward airflow command and the feedback airflow compensation command, the final airflow control command is issued for execution to control the airflow of the ring-cooled fan. The self-updating inner loop of the slow-optimized feedforward model parameters continuously monitors the feedback airflow compensation command, uses the feedback airflow compensation command as the dynamic prediction error signal of the toner cartridge model, and dynamically adjusts the key dynamic parameters output by the machine learning module in the toner cartridge model through an adaptive adjustment law, for online self-correction and slow optimization of the feedforward model.

10. An automatic energy-saving control device for a ring-cooled fan based on load feedback, characterized in that, The device includes: a communication unit and a processing unit; The communication unit is used to collect real-time operating data of water pumps and cooling fans in the water plant's pumping station; The processing unit is used to input the operating condition data into a preset powder box model and calculate the feedforward air volume command. The powder box model is a hybrid model that dynamically integrates the energy conservation mechanism, the fluid dynamics mechanism and the data-driven parameters. The key thermal and hydraulic parameters inside the powder box model are dynamically mapped and generated by the machine learning module according to the real-time operating conditions. The operating data is input into the status observer to obtain the optimal estimated temperature value; the deviation between the optimal estimated temperature value and the preset temperature setpoint is calculated, and the feedback controller calculates the feedback air volume compensation command based on the deviation. The feedforward air volume command and the feedback air volume compensation command are merged to obtain the final air volume control command, which is then sent to the fan inverter for execution. The value and trend of the feedback air volume compensation command are continuously monitored, and the feedback air volume compensation command is used as the prediction error signal of the toner cartridge model. The key thermal parameters in the toner cartridge model are adaptively adjusted to make the feedback air volume compensation command approach zero.

Images