Condensate pump frequency conversion energy-saving optimization method and system

Abstract

This invention relates to the field of frequency conversion energy saving and process automation control technology for condensate systems in thermal power plants, and particularly to a method and system for optimizing the frequency conversion energy saving of condensate pumps. The method includes acquiring real-time data such as back pressure of the direct air-cooled island, ambient wind speed, outdoor temperature, condensate header pressure, deaerator tank water level, and feedwater pump operating status; calculating the heat transfer coefficient attenuation slope using meteorological parameters and converting it into a pressure compensation feedforward value; upon receiving a start-up command from a standby feedwater pump, freezing pressure and water level measuring points to determine a steady-state baseline, and calculating the feedwater shortage value accordingly, converting it into a condensate pump acceleration predictive command; finally, superimposing the basic feedback quantity and the feedforward compensation component to generate a comprehensive frequency control signal to adjust the motor frequency. This invention solves the problems of frequency conversion regulation lag, system pressure oscillation, and energy efficiency imbalance caused by extreme back pressure fluctuations and sudden changes in feedwater pump flow.

Description

Technical Field

[0001] This invention relates to the field of frequency conversion energy saving and process automation control technology for condensate systems in thermal power plants, and particularly to a method and system for optimizing energy saving of condensate pumps using frequency conversion. Background Technology

[0002] In existing direct air-cooled thermal power units, condensate pumps are responsible for transporting the condensate formed after the exhaust steam from the low-pressure cylinder of the turbine is condensed in the air-cooled island from the condenser hot well to the subsequent regenerative system. The core of variable frequency drive (VFD) energy-saving technology in this scenario lies in using a VFD to change the operating frequency of the condensate pump drive motor, thereby smoothly adjusting the motor speed and pump output flow. The system automatically generates control signals based on changes in the condenser hot well water level or unit load. After receiving the signals, the VFD adjusts its output frequency to precisely match the motor speed with the real-time demand of the condensate system.

[0003] The existing variable frequency energy-saving optimization technology for condensate pumps in direct air-cooled thermal power units has the following technical pain points: In the application scenario of direct air-cooled thermal power units, the vacuum changes drastically with the ambient wind speed and temperature, and the feedwater system performs high dynamic transient processes such as parallel switching of feedwater pumps. The drastic fluctuation of the back pressure of the direct air-cooling island will change the outlet back pressure reference of the condensate pump. Coupled with the massive feedwater demand generated by the standby pump speed-up and grid connection during the one-button start-up of the feedwater pump, the system will generate significant pressure drop and flow deficit in a very short time. Conventional variable frequency regulation algorithms are based on feedback deviation for lag compensation and cannot match the dynamic gain offset caused by the inertia of the physical process in real time. This induces pressure fluctuations in the condensate header and large swings in the deaerator water level, forming a nonlinear oscillation cycle. For example, when a strong wind suddenly hits the environment, causing a sharp drop in the condensation capacity of the air-cooled island and resulting in an increase in back pressure, the feedwater pump set triggers an automatic paralleling program. The condensate system needs to fill the water supply gap by increasing the speed through frequency conversion while resisting the change in back pressure. If the frequency adjustment cannot predict the peak energy demand, the deaerator water level will quickly drop to the low limit, triggering a safety alarm and forcing the operators to abandon the frequency conversion energy-saving mode and switch to the power frequency high-energy consumption mode to maintain system stability. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a variable frequency energy-saving optimization method and system for condensate pumps. This invention solves the dynamic control pain point of nonlinear oscillations in condensate system pressure and deaerator water level caused by the lag in variable frequency regulation due to extreme back pressure fluctuations in direct air-cooled systems and sudden flow changes during "one-click start" of feedwater pumps, making it difficult to achieve optimal energy efficiency. "One-click start" refers to the automatic opening of the standby feedwater pump's outlet valve, pump speed increase, pressure matching, and automatic grid connection during the transient process of switching from single-pump operation to dual-pump parallel operation in a thermal power plant feedwater system, through a pre-set logical sequence in the centralized control system (DCS). This process generates a significant step demand for feedwater flow in a very short time, instantly disrupting the original flow balance of the condensate system.

[0005] To solve the above-mentioned technical problems, the specific contents of the present invention are as follows: In a first aspect, the condensate pump frequency conversion energy-saving optimization method provided by the present invention includes: Step 1: Obtain the real-time back pressure, ambient wind speed, outdoor temperature, condensate header pressure, deaerator tank water level, main feed water pump output flow rate, standby feed water pump speed increase rate, and standby feed water pump start command of the direct air-cooled island. Step 2: Calculate the heat transfer coefficient decay slope using the obtained real-time back pressure, ambient wind speed, and outdoor temperature. Calculate the back pressure offset using the heat transfer coefficient decay slope and the real-time back pressure. Then, convert the back pressure offset into a pressure compensation feedforward value. Step 3: Receive the start command of the backup water supply pump, freeze the obtained pressure of the condensate header and the water level of the deaerator tank, and obtain the steady-state reference pressure and steady-state reference water level; Step 4: Calculate the current supply flow rate based on the steady-state reference pressure and the steady-state reference water level; calculate the total water demand based on the obtained rate of increase of the standby water pump speed and the output flow rate of the main water pump; calculate the water supply gap value by subtracting the total water demand from the current supply flow rate; and convert the water supply gap value into a condensate acceleration predictive command. Step 5: Based on the obtained basic feedback quantity of the deoxygenated water tank level extraction, the basic feedback quantity, the pressure compensation feedforward value, and the condensate acceleration predictive command are superimposed to generate a comprehensive frequency control signal; the comprehensive frequency control signal is sent to the frequency converter execution terminal to instruct the motor to change its operating frequency according to the comprehensive frequency control signal.

[0006] Furthermore, in the condensate pump frequency conversion energy-saving optimization method of the present invention, step 2 includes: Extract transient wind direction data and gust wind speed data from the ambient wind speed, and extract spatial temperature gradient data from the outdoor temperature; The flow field vector synthesis operation is performed based on the transient wind direction data and the gust wind speed data, and the forced convection heat transfer coefficient decay slope is calculated by combining the spatial temperature gradient data. The forced convection heat transfer coefficient decay slope is used as the heat transfer coefficient decay slope. Substituting the heat transfer coefficient decay slope into the preset cold-end heat exchange mapping matrix, the a priori trend signal of the decrease in exhaust steam condensation capacity is calculated. The prior trend signal and the real-time back pressure are subjected to nonlinear fitting to obtain the back pressure offset, and the pressure compensation feedforward value is calculated based on the back pressure offset.

[0007] Furthermore, in the condensate pump frequency conversion energy-saving optimization method of the present invention, step 2 further includes: The vacuum decrease rate in the exhaust channel is calculated based on the prior trend signal; the vacuum decrease rate and the real-time back pressure are used to calculate the peak value of the back pressure of the direct air-cooled island and the time when the peak value occurs, and the peak value is used as the back pressure offset. The back pressure offset and the peak occurrence time are spatiotemporally combined to generate a spatiotemporal prediction trajectory for back pressure rise. Extract the peak data from the back pressure rise spatiotemporal prediction trajectory and convert the peak data into a back pressure pulse perturbation feedforward quantity; The preset pipeline resistance coefficient is retrieved, and the pipeline resistance coefficient is multiplied with the back pressure pulse disturbance feedforward to obtain the additional dynamic head. The additional dynamic head is used as the pressure compensation feedforward value.

[0008] Furthermore, in the condensate pump frequency conversion energy-saving optimization method of the present invention, step 4 includes: Obtain the preset outlet valve opening time axis; The speed increase rate of the standby water pump is combined with the time axis of the outlet valve opening to generate a speed opening curve matrix; The data corresponding to the speed opening curve matrix are added to the output flow of the main feedwater pump to calculate the total water demand at the moment of grid connection. The difference between the total water demand and the current supply flow is used to calculate the water shortage value; Retrieve the preset mapping matrix between volumetric efficiency and motor frequency; The water supply gap value is numerically converted by the motor frequency mapping correlation matrix to obtain an additional applied frequency parameter, which is then used as the predictive command for condensate acceleration.

[0009] Furthermore, in the condensate pump frequency conversion energy-saving optimization method of the present invention, step 4 further includes: The back pressure offset is obtained and set as an environmental correction factor; The environmental correction coefficient is injected into the motor frequency mapping correlation matrix and matrix multiplication is performed to generate an updated motor frequency mapping correlation matrix. The water supply gap value is input into the updated motor frequency mapping correlation matrix for inverse calculation to obtain the feedforward target frequency increment. The feedforward target frequency increment is output as the additional applied frequency parameter, and the combined parameters are used to generate the condensate acceleration predictive command.

[0010] Furthermore, in the condensate pump frequency conversion energy-saving optimization method of the present invention, step 5 includes: The basic feedback value, the pressure compensation feedforward value, and the condensate acceleration predictive command are added together to obtain the synthetic frequency drive command. A preset filtering algorithm is invoked to perform smoothing filtering calculations on the synthesized frequency driving command, thereby obtaining a smoothed frequency command. The smoothing frequency command is compared with a preset rising slope threshold, and a rate of change limiting operation is performed to obtain a limiting frequency command. The limiting frequency command is output as the integrated frequency control signal.

[0011] Furthermore, in the condensate pump frequency conversion energy-saving optimization method of the present invention, step 5 further includes: Retrieve the preset motor stator frequency and pump shaft mechanical work transfer function; By using the stator frequency of the motor and the mechanical work transfer function of the pump shaft to perform reverse analytical calculation on the limiting frequency command, the reference speed increase increment of the motor is obtained; A preset phase advance compensation operator is retrieved, and the motor reference speed increase increment is concatenated with the phase advance compensation operator to generate a frequency advance injection sequence with a time gradient. The acceleration ramp curve is redrawn according to the data nodes corresponding to the frequency advance injection sequence, and the redrawn acceleration ramp curve data is output as the comprehensive frequency control signal.

[0012] Furthermore, in the condensate pump frequency conversion energy-saving optimization method of the present invention, step 5 further includes: The real-time operating frequency of the motor and the fluid pressure feedback value after transient adjustment of the pipeline network are obtained. The difference between the integrated frequency control signal and the real-time operating frequency is calculated and compared to obtain the first residual value. The difference between the steady-state reference pressure and the fluid pressure feedback value is calculated and compared to obtain the second residual value. Calculate the proportional-integral update coefficients based on the first residual value and the second residual value; adjust the proportional constant and integral constant in the feedforward adjustment logic according to the proportional-integral update coefficients.

[0013] Furthermore, in the condensate pump frequency conversion energy-saving optimization method of the present invention, step 5 further includes: Continue to extract the absolute value of the second residual; Determine whether the absolute value of the second residual is less than a preset convergence threshold; When the judgment result is yes, the superposition weight of the pressure compensation feedforward value and the condensate acceleration predictive command in the comprehensive frequency control signal is reduced cyclically according to the preset attenuation step size. The weight of the product of the basic feedback quantity and the data in the integrated frequency control signal is increased synchronously until the data corresponding to the integrated frequency control signal is equal to the data of the basic feedback quantity.

[0014] Secondly, the condensate pump frequency conversion energy-saving optimization system provided by the present invention is applied to the condensate pump frequency conversion energy-saving optimization method as described above, including: The data acquisition module is used to acquire real-time back pressure, ambient wind speed, outdoor temperature, condensate header pressure, deaerator water tank level, main feed water pump output flow rate, standby feed water pump speed increase rate, and standby feed water pump start command of the direct air-cooled island. The feedforward calculation module is used to calculate the heat transfer coefficient decay slope using the acquired real-time back pressure, the ambient wind speed and the outdoor temperature, and to calculate the back pressure offset using the heat transfer coefficient decay slope and the real-time back pressure, and to convert the back pressure offset into a pressure compensation feedforward value. The reference freezing module is used to receive the start command of the standby water supply pump, freeze the obtained pressure of the condensate header and the water level of the deaerator tank, and obtain the steady-state reference pressure and steady-state reference water level. The predictive command generation module is used to calculate the current supply flow rate based on the steady-state reference pressure and the steady-state reference water level; calculate the total water demand based on the obtained rate of increase of the standby water pump speed and the output flow rate of the main water pump; calculate the water supply gap value by subtracting the total water demand from the current supply flow rate; and convert the water supply gap value into a condensate acceleration predictive command. The integrated control module is used to extract basic feedback based on the obtained water level of the deoxygenated water tank, and superimpose the basic feedback, the pressure compensation feedforward value, and the condensate acceleration predictive command to generate a comprehensive frequency control signal; the comprehensive frequency control signal is sent to the frequency converter execution terminal to instruct the motor to change its operating frequency according to the comprehensive frequency control signal.

[0015] Beneficial effects of this invention: The technical solution of this invention deeply integrates the evolution of external environmental meteorological patterns with the operational logic of internal electromechanical equipment. It uses ambient wind speed and outdoor temperature to calculate the attenuation slope of the heat transfer coefficient and converts it into a pressure compensation feedforward value to cope with vacuum fluctuations. Simultaneously, upon intercepting the start command of the standby water pump, it establishes a static reference coordinate system including steady-state reference pressure and steady-state reference water level. It accurately calculates the water supply gap caused by the transient network cavitation effect and converts it into a condensate acceleration prediction command. The pressure compensation feedforward value, representing advanced sensing capability, and the condensate acceleration prediction command are superimposed on the basic feedback quantity to form a comprehensive frequency control signal, fundamentally breaking through the conventional closed-loop water level regulation. The mechanism's inherent time lag barrier in the face of highly dynamic transient processes enables the variable frequency drive to operate in advance before the actual pressure drop and severe flow deficit occur. Relying on pre-established fluid kinetic energy, it precisely offsets the abnormal rise in pipeline resistance caused by extreme back pressure fluctuations and the massive water replenishment demand caused by the instantaneous parallel operation of feedwater pumps. It cuts off the nonlinear oscillation cycle formed by repeated over-adjustment of condensate header pressure and deaerator tank water level. Under the premise of maintaining the fully automatic dynamic balance of the thermal power plant's thermal system under complex and harsh operating conditions, it completely eliminates the ineffective high-frequency speed fluctuations caused by blindly chasing feedback deviations, and achieves a smooth reconstruction of the optimal energy efficiency state of the condensate pump variable frequency system. Attached Figure Description

[0016] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on the drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the variable frequency energy-saving optimization method for condensate pumps according to the present invention. Detailed Implementation

[0018] To make the technical solution of the present invention clearer, the present invention will be clearly and completely described below with reference to specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. The present invention provided by various embodiments will be described in detail below with reference to the accompanying drawings. To better understand the purpose of the present invention, the present invention will be described in further detail below.

[0019] In a first aspect, the condensate pump frequency conversion energy-saving optimization method provided by the present invention includes: Step 1: Obtain the real-time back pressure, ambient wind speed, outdoor temperature, condensate header pressure, deaerator tank water level, main feed water pump output flow rate, standby feed water pump speed increase rate, and standby feed water pump start command of the direct air-cooled island. Step 2: Calculate the heat transfer coefficient decay slope using the obtained real-time back pressure, ambient wind speed, and outdoor temperature. Calculate the back pressure offset using the heat transfer coefficient decay slope and the real-time back pressure. Then, convert the back pressure offset into a pressure compensation feedforward value. Step 3: Receive the start command of the backup water supply pump, freeze the obtained pressure of the condensate header and the water level of the deaerator tank, and obtain the steady-state reference pressure and steady-state reference water level; Step 4: Calculate the current supply flow rate based on the steady-state reference pressure and the steady-state reference water level; calculate the total water demand based on the obtained rate of increase of the standby water pump speed and the output flow rate of the main water pump; calculate the water supply gap value by subtracting the total water demand from the current supply flow rate; and convert the water supply gap value into a condensate acceleration predictive command. Step 5: Based on the obtained basic feedback quantity of the deoxygenated water tank level extraction, the basic feedback quantity, the pressure compensation feedforward value, and the condensate acceleration predictive command are superimposed to generate a comprehensive frequency control signal; the comprehensive frequency control signal is sent to the frequency converter execution terminal to instruct the motor to change its operating frequency according to the comprehensive frequency control signal.

[0020] Direct air-cooled units in thermal power plants face the dual challenges of drastic environmental fluctuations and frequent switching of internal operating conditions during actual operation. When constructing predictive models, researchers comprehensively collected real-time back pressure, ambient wind speed, and outdoor temperature data from the direct air-cooling island using a sensor network deployed on-site. The control system simultaneously connected to the thermal pipeline system to extract condensate header pressure and deaerator tank water level. To anticipate transient hydrodynamic changes in the feedwater system, the control system concurrently retrieved the main feedwater pump output flow rate, the standby feedwater pump speed increase rate, and the standby feedwater pump start-up command. The control system then uniformly aggregated the multi-dimensional operational data into the underlying in-memory database for periodic updates.

[0021] Drastic fluctuations in environmental meteorological parameters directly deteriorate the condensation physical environment of the air-cooled island. The control logic performs thermodynamic deduction based on extracted ambient wind speed and outdoor temperature, calculating the heat transfer coefficient decay slope, representing the degree of heat dissipation capacity degradation. The control logic then performs a cross-product operation using the heat transfer coefficient decay slope and real-time back pressure to derive the potential back pressure offset within the future time window of the direct air-cooled island. An increase in back pressure inevitably raises the water flow resistance at the end of the condensate pipe network. The control logic converts the back pressure offset into a pressure compensation feedforward value according to the pipe network resistance characteristics. The control system allocates the pressure compensation feedforward value to the internal feedforward adjustment channel for standby buffering.

[0022] The operation of internal system electromechanical equipment can cause transient hydraulic shocks to the condensate pipe network. When the water supply system performs parallel operation of the standby pump, the original fluid balance is momentarily disrupted. The control system triggers a core data locking mechanism the instant it receives the start command for the standby water supply pump. This data locking mechanism forcibly freezes the condensate header pressure and deaerator tank level at a specific moment. The freezing action physically eliminates the disturbance fluctuations generated during the pump parallel operation. The control system records the values ​​after freezing to obtain the steady-state reference pressure and steady-state reference water level. The control system stores the steady-state reference pressure and steady-state reference water level as static reference coordinate coefficient values ​​in the logic unit.

[0023] The control system quantifies the supply-demand imbalance of the water system based on a static reference coordinate system. The computational unit calculates the current supply flow rate based on the steady-state reference pressure and steady-state reference water level. The computational unit retrieves the standby feedwater pump's speed increase rate and the main feedwater pump's output flow rate to perform summation simulations, predicting the total feedwater demand faced by the two pumps operating in parallel. The computational unit subtracts the total feedwater demand from the current supply flow rate to obtain the feedwater gap value reflecting the transient water replenishment vacuum zone. The feedwater gap value is a purely fluid dynamics variable. The control system retrieves the variable frequency electromechanical mapping correlation matrix to map and convert the feedwater gap value into a condensate acceleration predictive command. The control system synchronously pushes the condensate acceleration predictive command to the command synthesis node of the computational architecture.

[0024] Dynamic balancing of the pipeline network under complex operating conditions relies not only on feedforward prediction mechanisms but also on traditional feedback closed-loop systems. The control system extracts basic feedback quantities based on the difference between the deaerator water tank level and the set target water level. At the command synthesis node, the control system logically superimposes the basic feedback quantities, pressure compensation feedforward values, and condensate acceleration prediction commands. After superposition and calculation of multi-source control components, a comprehensive frequency control signal covering steady-state regulation and transient compensation is generated. The control system sends the comprehensive frequency control signal to the frequency converter actuator via an industrial communication bus. The inverter component inside the frequency converter actuator analyzes the received control signal and instructs the motor to change its operating frequency in real time according to the comprehensive frequency control signal.

[0025] The control logic extracts data from the sensor array deployed on the windward side of the direct air-cooled island, separating transient wind direction and gust wind speed data from the ambient wind speed, and simultaneously extracting spatial temperature gradient data from the outdoor temperature. The wind field exhibits three-dimensional fluid vector characteristics in physical space. The control system performs flow field vector synthesis calculations based on the transient wind direction and gust wind speed data. The flow field vector synthesis calculation outputs a comprehensive aerodynamic disturbance factor. The control system combines the spatial temperature gradient data to extrapolate and calculate the comprehensive aerodynamic disturbance factor, deriving the forced convection heat transfer coefficient decay slope, representing the deterioration of the local cooling flow field on the heat dissipation surface. The control logic inputs the forced convection heat transfer coefficient decay slope as the heat transfer coefficient decay slope into the underlying computational architecture. The control system substitutes the heat transfer coefficient decay slope into a preset cold-end heat exchange mapping matrix for matrix transformation and solution. The matrix transformation and solution directly outputs a priori trend signal indicating a decrease in exhaust steam condensation capacity. The control logic reads the priori trend signal and performs differential derivation to calculate the vacuum decrease rate within the exhaust channel. The control system uses the vacuum decrease rate and real-time back pressure to perform forward algebraic calculations to derive the peak rise of the direct air-cooled island back pressure and the timing of its occurrence within a future physical time window. The control system imports this peak rise as a back pressure offset into the predictive model database. The control logic combines the back pressure offset and the peak occurrence timing in a spatiotemporal dimension, generating a spatiotemporal prediction trajectory of the back pressure rise in memory, characterizing the dynamic evolution of the pressure. The control system extracts the peak data from the spatiotemporal prediction trajectory of the back pressure rise and converts it into a back pressure pulse disturbance feedforward. This feedforward represents the maximum physical resistance intensity that the pipeline fluid will encounter. The control logic retrieves a preset pipeline resistance coefficient and multiplies it with the back pressure pulse disturbance feedforward to derive the additional dynamic pressure head required to overcome the abnormal rise in pipeline resistance. The control system allocates this additional dynamic pressure head as a pressure compensation feedforward value to the feedforward adjustment channel for standby buffering.

[0026] Addressing the abrupt changes in hydrodynamics caused by parallel pump switching requires establishing a low-level predictive control architecture. The control system acquires a preset outlet valve opening time axis. The control logic cross-merges the standby feedwater pump's speed increase rate with the outlet valve opening time axis to generate a speed-opening curve matrix depicting the standby pump's dynamic water output capacity during startup. The control system extracts the data corresponding to the speed-opening curve matrix, adds the extracted data to the main feedwater pump's output flow rate, and calculates the total feedwater demand at the moment of grid connection between the two electromechanical devices. The control logic extracts the frozen current supply flow rate from memory, subtracts the total feedwater demand from the current supply flow rate, and calculates the feedwater gap value reflecting the local evacuation state of the condensate network. The control system retrieves a preset volumetric efficiency-motor frequency mapping correlation matrix. Deterioration in environmental thermodynamic conditions alters the inverter's mechanical work efficiency benchmark. The control system acquires the back pressure offset calculated at the front end and sets it as an environmental correction factor. The control system injects environmental correction coefficients into the motor frequency mapping correlation matrix and performs matrix multiplication to generate an updated motor frequency mapping correlation matrix that can adapt to extreme back pressure conditions. The control system then inputs the feedwater shortage value into the updated motor frequency mapping correlation matrix for inverse calculation to obtain the feedforward target frequency increment required to fill the transient water shortage. The control logic outputs the feedforward target frequency increment as an additional applied frequency parameter as a digital quantity, which is then logically reorganized to generate a condensate acceleration predictive command and pushed to the command convergence node.

[0027] The coordinated convergence of multi-source control commands needs to balance the system response rate and the electrical tolerance limits of the frequency converter actuators. At the command convergence node, the control logic adds the basic feedback quantity, the pressure compensation feedforward value, and the condensate acceleration prediction command to obtain the synthetic frequency drive command. The originally generated synthetic frequency drive command includes physical step characteristics and high-frequency glitches. The control system retrieves a preset filtering algorithm to perform smoothing filtering calculations on the synthetic frequency drive command, resulting in a smoothed frequency command that removes high-frequency oscillation components. The control logic retrieves a preset rising slope threshold, cross-compares the smoothed frequency command with the rising slope threshold, and forcibly performs a rate-of-change limiting calculation to obtain a limited frequency command that complies with the frequency converter hardware safety specifications. The control system needs to overcome the fluid mass inertia and motor rotor mechanical hysteresis unique to long-distance water pipeline networks. The control logic retrieves preset motor stator frequency and pump shaft mechanical work transfer function. The control system uses the motor stator frequency and pump shaft mechanical work transfer function to perform reverse analytical calculations on the limited frequency command to obtain the motor reference acceleration increment that drives the pump impeller to start ahead of time. The control logic retrieves a preset phase lead compensation operator, concatenates the motor's base speed increase increment with the phase lead compensation operator, and generates a frequency lead injection sequence with a time gradient and pre-intervention waveform. The control system redraws the acceleration ramp curve of the frequency converter according to the data nodes corresponding to the frequency lead injection sequence. The control unit sends the redrawn acceleration ramp curve data as a comprehensive frequency control signal to the underlying communication bus.

[0028] The closed-loop feedback mechanism is used to smooth out the physical execution deviations generated by the feedforward prediction model and reconstruct the system steady state. The control system acquires the real-time operating frequency of the motor and the fluid pressure feedback value after transient regulation of the pipeline network through field instruments. The control logic performs a difference comparison between the integrated frequency control signal and the real-time operating frequency to obtain a first residual value reflecting the degree of electrical tracking response hysteresis. Simultaneously, the control logic performs a difference comparison between the steady-state reference pressure and the fluid pressure feedback value to obtain a second residual value reflecting the true hydraulic recovery state of the pipeline network. The control unit calculates the proportional-integral update coefficients by substituting the first and second residual values ​​into the algebraic equation. The control system dynamically adjusts the proportional and integral constants in the feedforward regulation logic according to the proportional-integral update coefficients. As the unit's thermodynamic conditions gradually pass through the high-dynamic transient process, the dominant control needs to be smoothly transferred to the conventional water level regulation loop. The control logic continuously extracts the absolute value of the second residual value and inputs it into the internal logic comparator. The control system determines whether the absolute value of the second residual value is less than a preset convergence threshold. When the judgment result is yes, the control logic triggers the feedforward exit mechanism, cyclically reducing the superposition weight of the pressure compensation feedforward value and the condensate acceleration prediction command in the integrated frequency control signal according to a preset decay step size. Simultaneously, the control system increases the proportion of the data product of the basic feedback quantity in the integrated frequency control signal while reducing the feedforward prediction weight. The weight handover process continues to execute within the control cycle until the data corresponding to the integrated frequency control signal is equal to the data of the basic feedback quantity.

[0029] The control logic extracts data from a sensor array deployed on the windward side of the direct air-cooled island, separating transient wind direction and gust wind speed data from the ambient wind speed, and simultaneously extracting spatial temperature gradient data from the outdoor temperature. The control system performs flow field vector synthesis calculations based on the transient wind direction and gust wind speed data, outputting a comprehensive aerodynamic disturbance factor. The specific flow field vector synthesis calculation is implemented using a vector dot product formula: In the formula, Represents the comprehensive aerodynamic disturbance factor. Represents gust wind speed data. This represents the angle between the transient wind direction data and the normal direction of the windward surface of the direct air-cooled island. The control system, combined with spatial temperature gradient data, calculates the comprehensive aerodynamic disturbance factor to derive the forced convection heat transfer coefficient decay slope. The specific calculation formula is as follows: In the formula, This represents the slope of the forced convection heat transfer coefficient decay. Represents the wind field disturbance weighting constant. Represents the comprehensive aerodynamic disturbance factor. Represents the temperature gradient weighting constant. This represents the spatial temperature gradient data. The control logic uses the forced convection heat transfer coefficient decay slope as the input to the underlying computational architecture. The control system substitutes the heat transfer coefficient decay slope into a preset cold-end heat exchange mapping matrix for matrix transformation and solution, directly outputting a priori trend signal indicating a decrease in exhaust steam condensation capacity. The computation process is as follows: In the formula, Represents a priori trend signal, Represents the cold-end heat exchange mapping matrix. This represents the slope of the heat transfer coefficient decay. The control logic reads the prior trend signal and performs differential derivation to calculate the vacuum decrease rate in the exhaust channel. The control system uses the vacuum decrease rate and real-time back pressure to perform forward algebraic calculations to obtain the peak rise of the back pressure in the direct air-cooled island. The specific calculation formula is as follows: In the formula, Represents the peak of the rise. Represents real-time back pressure. Represents the system's airtightness conversion factor. Represents the rate of decrease in vacuum level. This represents the continuous span of the future physical time window. The control system imports the rising peak value as the back pressure offset into the prediction model database. The control logic combines the back pressure offset with the peak occurrence time in a spatiotemporal dimension to generate the spatiotemporal prediction trajectory of the back pressure rise. The control system extracts the peak data from the spatiotemporal prediction trajectory of the back pressure rise and converts the peak data into a back pressure pulse disturbance feedforward. The control logic retrieves the preset pipeline resistance coefficient and multiplies it with the back pressure pulse disturbance feedforward to obtain the additional dynamic head. The specific multiplication operation is as follows: In the formula, This represents the additional dynamic pressure head. Represents the pipeline resistance coefficient. This represents the back pressure pulse disturbance feedforward. The control system allocates the additional dynamic pressure head as a pressure compensation feedforward value to the feedforward regulation channel for standby buffering. The control system extracts the data corresponding to the speed opening curve matrix, adds the extracted data to the main feedwater pump output flow rate, and calculates the total feedwater demand. The control logic extracts the current supply flow rate calculated from the steady-state reference pressure and steady-state reference water level, and performs a difference operation between the total feedwater demand and the current supply flow rate to obtain the feedwater deficit value. The specific formula for the difference operation is: In the formula, Represents the water supply deficit value. Represents the total water demand. This represents the current supply flow rate. The control system retrieves the preset volumetric efficiency-motor frequency mapping correlation matrix. The control system obtains the back pressure offset calculated by the front end and sets it as an environmental correction factor. The control system injects the environmental correction factor into the motor frequency mapping correlation matrix and performs matrix multiplication to generate an updated motor frequency mapping correlation matrix. The matrix update formula is: In the formula, This represents the updated motor frequency mapping correlation matrix. Represents the environmental correction factor. This represents the motor frequency mapping correlation matrix. The control system inputs the water supply gap value into the updated motor frequency mapping correlation matrix for inverse calculation to obtain the feedforward target frequency increment. The inverse calculation process is as follows: In the formula, This represents the feedforward target frequency increment. This represents the inverse of the updated motor frequency mapping correlation matrix. This represents the feedwater shortage value. The control logic converts the feedforward target frequency increment into a digital output as an additional applied frequency parameter, which is then converted into a condensate acceleration predictive command. The coordinated convergence of multi-source control commands needs to consider both the system response rate and the electrical tolerance limits of the variable frequency actuators. At the command convergence node, the control logic adds the basic feedback quantity, the pressure compensation feedforward value, and the condensate acceleration predictive command to obtain the synthesized frequency drive command. The addition logic is as follows: In the formula, Represents the synthesized frequency drive instruction. Represents the basic feedback volume. This represents the pressure compensation feedforward value. This represents a predictive command to accelerate condensate flow. The control system retrieves a preset filtering algorithm to perform smoothing filtering on the synthesized frequency drive command, resulting in a smoothed frequency command. The control logic extracts the integrated frequency control signal and compares it with the real-time operating frequency to obtain the first residual value. The comparison formula is: In the formula, Represents the first residual value. Represents the integrated frequency control signal. This represents the real-time operating frequency. The control logic synchronously calculates the difference between the steady-state reference pressure and the fluid pressure feedback value to obtain the second residual value. The formula for the difference comparison is: In the formula, Represents the second residual value. Represents steady-state baseline pressure. This represents the fluid pressure feedback value. The control unit calculates the proportional-integral update coefficients by substituting the first and second residual values ​​into the algebraic equation. The specific algebraic equation is as follows: In the formula, Representative proportional integral update coefficient, Represents the frequency tracking error weighting constant. Represents the first residual value. Represents the pipeline pressure error weighting constant. This represents the second residual value. The control system dynamically adjusts the proportional and integral constants in the feedforward control logic according to the proportional-integral update coefficients.

[0030] Secondly, the condensate pump frequency conversion energy-saving optimization system provided by the present invention is applied to the condensate pump frequency conversion energy-saving optimization method as described above, including: The data acquisition module is used to acquire real-time back pressure, ambient wind speed, outdoor temperature, condensate header pressure, deaerator water tank level, main feed water pump output flow rate, standby feed water pump speed increase rate, and standby feed water pump start command of the direct air-cooled island. The feedforward calculation module is used to calculate the heat transfer coefficient decay slope using the acquired real-time back pressure, the ambient wind speed and the outdoor temperature, and to calculate the back pressure offset using the heat transfer coefficient decay slope and the real-time back pressure, and to convert the back pressure offset into a pressure compensation feedforward value. The reference freezing module is used to receive the start command of the standby water supply pump, freeze the obtained pressure of the condensate header and the water level of the deaerator tank, and obtain the steady-state reference pressure and steady-state reference water level. The predictive command generation module is used to calculate the current supply flow rate based on the steady-state reference pressure and the steady-state reference water level; calculate the total water demand based on the obtained rate of increase of the standby water pump speed and the output flow rate of the main water pump; calculate the water supply gap value by subtracting the total water demand from the current supply flow rate; and convert the water supply gap value into a condensate acceleration predictive command. The integrated control module is used to extract basic feedback based on the obtained water level of the deoxygenated water tank, and superimpose the basic feedback, the pressure compensation feedforward value, and the condensate acceleration predictive command to generate a comprehensive frequency control signal; the comprehensive frequency control signal is sent to the frequency converter execution terminal to instruct the motor to change its operating frequency according to the comprehensive frequency control signal.

[0031] In the daily operation environment of direct air-cooled units in thermal power plants, the system frequently faces extreme conditions such as sudden changes in environmental meteorological conditions and the superposition of parallel operation of internal feedwater pumps. Drastic fluctuations in ambient wind speed and outdoor temperature can cause significant oscillations in the real-time back pressure of the direct air-cooled island. The control unit comprehensively collects real-time back pressure, ambient wind speed, outdoor temperature, condensate header pressure, and deaerator tank level of the direct air-cooled island based on a sensor network deployed on-site. To proactively detect dynamic flow changes within the hydraulic network, the control unit simultaneously connects to the centralized control system to extract the main feedwater pump output flow, the standby feedwater pump speed increase rate, and the standby feedwater pump start-up command.

[0032] The deterioration of external environmental meteorological conditions directly weakens the condensation physical capacity of the direct air-cooled island. The computational logic extracts transient wind direction and gust wind speed data from the ambient wind speed and spatial temperature gradient data from the outdoor temperature. Based on the transient wind direction and gust wind speed data, the logic performs flow field vector synthesis calculations and, combined with the spatial temperature gradient data, calculates the forced convection heat transfer coefficient decay slope. The system uses this forced convection heat transfer coefficient decay slope as the heat transfer coefficient decay slope, substitutes it into a preset cold-end heat exchange mapping matrix for matrix transformation, and calculates the a priori trend signal characterizing the decline in exhaust steam condensation capacity. Based on the a priori trend signal, the computational logic calculates the vacuum decrease rate within the exhaust channel and uses the vacuum decrease rate and real-time back pressure to deduce the peak rise of the direct air-cooled island back pressure and the time of its occurrence. The system sets the peak rise as the back pressure offset and performs a spatiotemporal combination of the back pressure offset and the peak occurrence time to generate a spatiotemporal prediction trajectory for the back pressure rise. The arithmetic unit extracts the peak data from the spatiotemporal prediction trajectory of the back pressure rise and converts the peak data into a back pressure pulse disturbance feedforward. The arithmetic unit retrieves the preset pipeline resistance coefficient and multiplies the pipeline resistance coefficient with the back pressure pulse disturbance feedforward to obtain the additional dynamic head. The system converts the additional dynamic head into a pressure compensation feedforward value to cope with vacuum fluctuations.

[0033] The parallel operation of the standby pump in the water supply system will disrupt the inherent flow balance of the condensate system in a very short time. Upon receiving the start command of the standby water supply pump, the control system triggers a freezing mechanism for the underlying data. The system forcibly freezes the condensate header pressure and the deaerator tank level, recording the frozen values ​​to obtain the steady-state reference pressure and level. The calculation unit calculates the current supply flow rate based on the steady-state reference pressure and level. The system obtains the preset outlet valve opening time axis and merges the standby water supply pump speed increase rate with the outlet valve opening time axis to generate a speed opening curve matrix. The control logic extracts the data corresponding to the speed opening curve matrix and performs a summation calculation with the main water supply pump output flow rate to predict the total water demand at the moment of grid connection. The calculation unit calculates the difference between the total water demand and the current supply flow rate to obtain the water supply gap value. The system obtains the calculated back pressure offset and sets it as an environmental correction coefficient. This environmental correction coefficient is then injected into a preset motor frequency mapping correlation matrix to perform matrix multiplication, generating an updated motor frequency mapping correlation matrix. The computational logic inputs the water supply gap value into the updated motor frequency mapping correlation matrix for reverse calculation, obtains the feedforward target frequency increment required to fill the transient water supply vacuum, outputs the feedforward target frequency increment as an additional applied frequency parameter, and finally converts it into a predictive command for condensate acceleration.

[0034] The system not only relies on feedforward prediction but also incorporates conventional water level closed-loop regulation. The computational logic extracts basic feedback based on the deviation of the deaerator tank water level from the target level. The control node adds the basic feedback, pressure compensation feedforward value, and condensate acceleration prediction command to obtain the synthetic frequency drive command. The computational unit retrieves a preset filtering algorithm to perform smoothing filtering on the synthetic frequency drive command, resulting in a smoothed frequency command. The system cross-compares the smoothed frequency command with a preset rising slope threshold and forces a rate-of-change limiting operation to obtain a limited frequency command. The computational logic retrieves preset motor stator frequency and pump shaft mechanical work transfer function, and uses these functions to perform reverse analytical calculations on the limited frequency command to obtain the motor reference acceleration increment. The system retrieves a preset phase advance compensation operator, concatenates the motor reference acceleration increment with the phase advance compensation operator, and generates a frequency advance injection sequence with a time gradient. The control unit redraws the frequency conversion acceleration ramp curve according to the data nodes corresponding to the frequency advance injection sequence, and sends the redrawn acceleration ramp curve data as a comprehensive frequency control signal to the frequency conversion execution terminal.

[0035] The inverter component inside the variable frequency drive commands the motor to change its operating frequency in real time according to the integrated frequency control signal. Field monitoring instruments collect the real-time operating frequency of the motor and the fluid pressure feedback value after transient adjustment of the pipeline network. The control system calculates the first residual value by comparing the integrated frequency control signal with the real-time operating frequency. Simultaneously, the system calculates the second residual value by comparing the steady-state reference pressure with the fluid pressure feedback value. The calculation logic calculates the proportional-integral update coefficient based on the first and second residual values, and dynamically adjusts the proportional and integral constants in the feedforward control logic according to the proportional-integral update coefficient. After the unit gradually passes through the high-dynamic transient process, the system continuously extracts the absolute value of the second residual value and determines whether the absolute value of the second residual value is less than a preset convergence threshold. When the determination result is yes, the system cyclically reduces the superposition weight of the pressure compensation feedforward value and the condensate acceleration predictive command in the integrated frequency control signal according to a preset decay step size. While reducing the predictive weight, the control system simultaneously increases the proportion of the data product of the basic feedback quantity in the integrated frequency control signal.

[0036] Embodiment 1 of this invention: Under extremely cold winter weather conditions, direct air-cooled units in thermal power plants face a severe operating environment with sudden strong winds. An anemometer network captures gust wind speed data and transient wind direction data within the ambient wind speed. Temperature sensors simultaneously extract spatial temperature gradient data from the outdoor temperature. The computing unit performs flow field vector synthesis calculations based on the transient wind direction data and gust wind speed data, and calculates the forced convection heat transfer coefficient decay slope by combining it with the spatial temperature gradient data. The control system uses the forced convection heat transfer coefficient decay slope as the heat transfer coefficient decay slope, and substitutes it into a preset cold-end heat exchange mapping matrix for matrix transformation. The matrix transformation calculation yields a priori trend signal characterizing the decrease in exhaust steam condensation capacity. The computing logic calculates the vacuum decrease rate within the exhaust steam channel based on the priori trend signal. The system uses the vacuum decrease rate and real-time back pressure to deduce the peak rise of the back pressure in the direct air-cooled island and the time of its occurrence. The control logic sets the peak rise as the back pressure offset, and combines the back pressure offset with the peak occurrence time in a spatiotemporal manner to generate a spatiotemporal prediction trajectory for the back pressure rise. The arithmetic unit extracts the peak data from the spatiotemporal prediction trajectory of the back pressure rise and converts the peak data into a back pressure pulse disturbance feedforward. The arithmetic unit retrieves the preset pipeline resistance coefficient and multiplies the pipeline resistance coefficient with the back pressure pulse disturbance feedforward to obtain the additional dynamic head.

[0037] Embodiment 2 of this invention: During the summer peak demand period, the generating unit not only faces the challenge of high-temperature environment but also needs to respond to the feedwater pump operation caused by deep grid peak shaving. The control unit intercepts the standby feedwater pump start command. At the instant of receiving the standby feedwater pump start command, the control unit freezes the condensate header pressure and deaerator tank water level, and records the frozen values ​​to obtain the steady-state reference pressure and steady-state reference water level. The calculation logic calculates the current supply flow rate based on the steady-state reference pressure and steady-state reference water level. The system obtains the preset outlet valve opening time axis and merges the standby feedwater pump speed increase rate with the outlet valve opening time axis to generate a speed opening curve matrix. The control node extracts the data corresponding to the speed opening curve matrix and performs summation calculation with the main feedwater pump output flow rate to obtain the total feedwater demand at the grid connection transient moment. The calculation unit performs a difference calculation between the total feedwater demand and the current supply flow rate to obtain the feedwater gap value. The control system obtains the back pressure offset calculated by the front end and sets it as the environmental correction coefficient, injects the environmental correction coefficient into the preset motor frequency mapping correlation matrix, and performs matrix multiplication operation. Matrix multiplication generates an updated motor frequency mapping correlation matrix. The computational logic inputs the feedwater shortage value into the updated motor frequency mapping correlation matrix for inverse calculation, deriving the feedforward target frequency increment required to fill the transient makeup water vacuum. The system outputs the feedforward target frequency increment as an additional applied frequency parameter, ultimately converting it into a predictive command for condensate acceleration.

[0038] Embodiment 3 of this invention: The collaborative convergence of multi-source control commands requires smoothing out the physical execution deviations generated by the feedforward prediction model and reconstructing the system steady state. The computational logic extracts the basic feedback quantity based on the deviation of the deaerator tank water level from the target water level. The control node adds the basic feedback quantity, the pressure compensation feedforward value, and the condensate acceleration prediction command to obtain the synthetic frequency drive command. The computational unit retrieves a preset filtering algorithm to perform smoothing filtering calculations on the synthetic frequency drive command to obtain a smoothed frequency command. The system cross-compares the smoothed frequency command with a preset rising slope threshold and forcibly performs a rate-of-change limiting calculation to obtain a limited frequency command. The computational logic retrieves a preset motor stator frequency and pump shaft mechanical work transfer function, and uses the motor stator frequency and pump shaft mechanical work transfer function to perform reverse analytical calculations on the limited frequency command to obtain the motor reference acceleration increment. The system retrieves a preset phase advance compensation operator, concatenates the motor reference acceleration increment with the phase advance compensation operator, and generates a frequency advance injection sequence with a time gradient. The control unit redraws the frequency conversion acceleration ramp curve according to the data nodes corresponding to the frequency advance injection sequence. The control unit sends the redrawn acceleration ramp curve data as a comprehensive frequency control signal to the frequency converter actuator. The frequency converter actuator changes its operating frequency according to the comprehensive frequency control signal. Field monitoring instruments collect the real-time operating frequency of the motor and the fluid pressure feedback value after transient adjustment of the pipeline network. The control system calculates the first residual value by comparing the comprehensive frequency control signal with the real-time operating frequency. Simultaneously, the system calculates the second residual value by comparing the steady-state reference pressure with the fluid pressure feedback value. The calculation logic calculates the proportional-integral update coefficient based on the first and second residual values, and dynamically adjusts the proportional and integral constants in the feedforward control logic according to the proportional-integral update coefficient. The system continuously extracts the absolute value of the second residual value and determines whether it is less than a preset convergence threshold. If the determination result meets the condition, the system cyclically reduces the superposition weight of the pressure compensation feedforward value and the condensate acceleration predictive command in the comprehensive frequency control signal according to a preset decay step size.

[0039] The control logic extracts data from a sensor array deployed on the windward side of the direct air-cooled island, separating transient wind direction and gust wind speed data from the ambient wind speed, and simultaneously extracting spatial temperature gradient data from the outdoor temperature. The control system performs flow field vector synthesis calculations based on the transient wind direction and gust wind speed data, outputting a comprehensive aerodynamic disturbance factor. The specific flow field vector synthesis calculation is implemented using a vector dot product formula: In the formula, Represents the comprehensive aerodynamic disturbance factor. Represents gust wind speed data. This represents the angle between the transient wind direction data and the normal direction of the windward surface of the direct air-cooled island. The control system, combined with spatial temperature gradient data, calculates the comprehensive aerodynamic disturbance factor to derive the forced convection heat transfer coefficient decay slope. The specific calculation formula is as follows: In the formula, This represents the slope of the forced convection heat transfer coefficient decay. Represents the wind field disturbance weighting constant. Represents the comprehensive aerodynamic disturbance factor. Represents the temperature gradient weighting constant. This represents the spatial temperature gradient data. The control logic uses the forced convection heat transfer coefficient decay slope as the input to the underlying computational architecture. The control system substitutes the heat transfer coefficient decay slope into a preset cold-end heat exchange mapping matrix for matrix transformation and solution, directly outputting a priori trend signal indicating a decrease in exhaust steam condensation capacity. The computation process is as follows: In the formula, Represents a priori trend signal, Represents the cold-end heat exchange mapping matrix. This represents the slope of the heat transfer coefficient decay.

[0040] The control logic reads the prior trend signal and performs differential derivation to calculate the vacuum decrease rate in the exhaust channel. The control system then uses the vacuum decrease rate and real-time back pressure to perform forward algebraic calculations to obtain the peak rise of the back pressure in the direct air-cooled island. The specific calculation formula is as follows: In the formula, Represents the peak of the rise. Represents real-time back pressure. Represents the system's airtightness conversion factor. Represents the rate of decrease in vacuum level. This represents the continuous span of the future physical time window. The control system imports the rising peak value as the back pressure offset into the prediction model database. The control system extracts the peak data from the spatiotemporal prediction trajectory of the back pressure rise and converts the peak data into a back pressure pulse disturbance feedforward. The control logic retrieves the preset pipeline resistance coefficient and multiplies it with the back pressure pulse disturbance feedforward to obtain the additional dynamic head. The specific multiplication operation is as follows: In the formula, This represents the additional dynamic pressure head. Represents the pipeline resistance coefficient. This represents the back pressure pulse disturbance feedforward. The control system allocates the additional dynamic pressure head as a pressure compensation feedforward value to the feedforward adjustment channel.

[0041] The control system calculates the current supply flow rate based on the steady-state reference pressure and steady-state reference water level, and then calculates the difference between the total water demand and the current supply flow rate to obtain the water supply deficit value. The specific formula for the difference calculation is as follows: In the formula, Represents the water supply deficit value. Represents the total water demand. This represents the current supply flow rate. The control system obtains the back pressure offset calculated by the front end and sets it as an environmental correction factor. The control system injects the environmental correction factor into the motor frequency mapping correlation matrix and performs matrix multiplication to generate an updated motor frequency mapping correlation matrix. The matrix update formula is: In the formula, This represents the updated motor frequency mapping correlation matrix. Represents the environmental correction factor. This represents the motor frequency mapping correlation matrix. The control system inputs the water supply gap value into the updated motor frequency mapping correlation matrix for inverse calculation to obtain the feedforward target frequency increment. The inverse calculation process is as follows: In the formula, This represents the feedforward target frequency increment. This represents the inverse of the updated motor frequency mapping correlation matrix. This represents the water supply shortfall value.

[0042] The coordinated convergence of multi-source control commands needs to balance the system response rate and the electrical tolerance limits of the variable frequency drive components. At the command convergence node, the control logic adds the basic feedback quantity, the pressure compensation feedforward value, and the condensate acceleration prediction command to obtain the synthesized frequency drive command. The addition logic is as follows: In the formula, Represents the synthesized frequency drive instruction. Represents the basic feedback volume. This represents the pressure compensation feedforward value. This represents a predictive command to accelerate condensate flow. The control logic extracts the integrated frequency control signal and compares it with the real-time operating frequency to obtain the first residual value. The comparison formula is: In the formula, Represents the first residual value. Represents the integrated frequency control signal. This represents the real-time operating frequency. The control logic synchronously calculates the difference between the steady-state reference pressure and the fluid pressure feedback value to obtain the second residual value. The formula for the difference comparison is: In the formula, Represents the second residual value. Represents steady-state baseline pressure. This represents the fluid pressure feedback value. The control unit calculates the proportional-integral update coefficients by substituting the first and second residual values ​​into the algebraic equation. The specific algebraic equation is as follows: In the formula, Representative proportional integral update coefficient, Represents the frequency tracking error weighting constant. Represents the first residual value. Represents the pipeline pressure error weighting constant. This represents the value of the second residual.

[0043] In the specific implementation environment of this invention, the heat transfer coefficient decay slope essentially reflects the dynamic weakening of exhaust steam condensation capacity due to the deterioration of the airflow field state outside the direct air-cooled island tube bundle. From a lower-level technical perspective, the heat transfer coefficient decay slope is not merely a simple numerical rate of change, but a physical quantity derived based on a three-dimensional fluid dynamics and heat transfer model. It is derived by extracting the three-dimensional vector characteristics of gust wind speed and the temperature gradient data of the ambient space, using specific empirical correlations for convective heat transfer. The heat transfer coefficient decay slope characterizes the evolution trend of forced convection heat transfer efficiency deviating from the steady-state design value over a short timescale, providing a quantitative benchmark for the system to detect thermodynamic disturbances on the environmental side in advance.

[0044] Pressure compensation feedforward refers to the forward-looking adjustment component pre-calculated and superimposed on the variable frequency control loop to overcome the increase in resistance at the end of the pipeline network caused by extreme back pressure fluctuations in the direct air-cooled system. In practical implementation, the system transforms the transient peak value of the back pressure and its spatiotemporal trajectory into resistance boundary conditions in the pipeline hydraulic equations. Combining this with the pipeline's unique resistance coefficient, the corresponding additional dynamic head demand is calculated. This additional dynamic head is then equivalently converted into the required increase in the operating frequency of the pump motor. The introduction of pressure compensation feedforward allows the variable frequency actuator to generate mechanical torque before the actual pressure drop occurs, effectively mitigating nonlinear disturbances caused by vacuum fluctuations.

[0045] Steady-state reference pressure and steady-state reference water level are snapshots of the fluid operating state forcibly frozen and saved by the system at the moment it captures the command to start the standby feedwater pumps in parallel, via a hardware interrupt or a high-priority task. Under complex operating conditions, the one-button start-up process of the feedwater pumps can trigger severe water hammer effects and flow oscillations within the pipeline network. If real-time feedback values ​​are directly used for adjustment, it is easy to trigger a vicious cycle of divergence in the system. Therefore, by locking the stable measurement point data of the last control cycle before the start command is issued as the reference coordinate system, the false interference fluctuations caused by transient operations are eliminated, providing a reliable static reference anchor point for subsequent accurate calculation of the actual supply and demand imbalance difference of the water system.

[0046] The condensate acceleration predictive command represents a digitally driven credential for proactively accelerating water flow in response to an impending flow deficit. The underlying logic of this command lies in the system's ability to dynamically deduce the total feedwater demand at the moment of grid connection by pre-integrating the mechanical acceleration characteristics of the standby feedwater pump with the opening stroke time axis of the outlet valve. The system then calculates the absolute value of the feedwater deficit by subtracting the total demand from the existing supply flow. Subsequently, the system utilizes a built-in volumetric efficiency model to convert the physical flow deficit into the required advance response frequency parameters for the pump motor, thus eliminating the deep hysteresis inherent in traditional control systems at its source.

[0047] The motor frequency mapping correlation matrix is ​​a multi-dimensional nonlinear data structure or polynomial function library embedded in the control unit's memory. Its core function is to establish the physical mapping relationship between the pump's volumetric flow rate, head requirement, and the motor stator power supply frequency. Considering the drift in mechanical work efficiency of direct air-cooled units under different environmental back pressures, the system synchronously introduces the real-time back pressure offset as an environmental correction coefficient when calling the mapping matrix, and dynamically refreshes the weights of the mapping nodes within the matrix. This ensures that even under deteriorating server-side operating conditions, the system can still accurately reverse-calculate the calculated flow gap into the most suitable feedforward target frequency increment.

[0048] The frequency advance injection sequence is a dynamic output waveform generated by reshaping the time axis of the base frequency command to overcome the massive fluid mass inertia within long-distance condensate pipe networks and the mechanical hysteresis of the motor rotor itself. At the lower-level execution level, the system uses a preset phase advance compensation operator to perform mathematical concatenation operations on the calculated baseline acceleration increment, causing the final output frequency command signal to have an artificial advance offset in time phase. This results in a stepped or specific slope-like advance rise characteristic on the inverter's acceleration ramp curve, thereby precisely offsetting the subsequent flow consumption dips by relying on the pre-accumulated fluid kinetic energy.

[0049] The integrated frequency control signal is a global-level final execution command that comprehensively considers system steady-state maintenance, environmental disturbance resistance, and compensation for internal operating condition changes. In the signal synthesis architecture, the system uses the basic feedback deviation adjustment of the deaerator tank water level as the main backbone, and superimposes the pressure compensation feedforward value representing environmental compensation and the condensate acceleration predictive command representing internal system action compensation in parallel. To ensure the electrical safety and stability of the frequency converter actuators, the superimposed signal also undergoes a built-in smoothing filter algorithm to remove high-frequency glitches, and a strict rate of change limiting calculation is performed by the rising slope threshold limiting module. The final output is an electrical signal that combines fast response capability and hardware operation safety, which is then executed by the underlying driver level.

Claims

1. A method for optimizing energy saving of condensate pumps by frequency conversion, characterized in that, include: Step 1: Obtain the real-time back pressure, ambient wind speed, outdoor temperature, condensate header pressure, deaerator tank water level, main feed water pump output flow rate, standby feed water pump speed increase rate, and standby feed water pump start command of the direct air-cooled island. Step 2: Calculate the heat transfer coefficient decay slope using the obtained real-time back pressure, ambient wind speed, and outdoor temperature. Calculate the back pressure offset using the heat transfer coefficient decay slope and the real-time back pressure. Then, convert the back pressure offset into a pressure compensation feedforward value. Step 3: Receive the start command of the backup water supply pump, freeze the obtained pressure of the condensate header and the water level of the deaerator tank, and obtain the steady-state reference pressure and steady-state reference water level; Step 4: Calculate the current supply flow rate based on the steady-state reference pressure and the steady-state reference water level; calculate the total water demand based on the obtained rate of increase of the standby water pump speed and the output flow rate of the main water pump; calculate the water supply gap value by subtracting the total water demand from the current supply flow rate; and convert the water supply gap value into a condensate acceleration predictive command. Step 5: Based on the obtained basic feedback quantity of the deoxygenated water tank level extraction, the basic feedback quantity, the pressure compensation feedforward value, and the condensate acceleration predictive command are superimposed to generate a comprehensive frequency control signal; the comprehensive frequency control signal is sent to the frequency converter execution terminal to instruct the motor to change its operating frequency according to the comprehensive frequency control signal.

2. According to the condensate pump frequency conversion energy-saving optimization method as described in claim 1, step 2 includes: Extract transient wind direction data and gust wind speed data from the ambient wind speed, and extract spatial temperature gradient data from the outdoor temperature; The flow field vector synthesis operation is performed based on the transient wind direction data and the gust wind speed data, and the forced convection heat transfer coefficient decay slope is calculated by combining the spatial temperature gradient data. The forced convection heat transfer coefficient decay slope is used as the heat transfer coefficient decay slope. Substituting the heat transfer coefficient decay slope into the preset cold-end heat exchange mapping matrix, the a priori trend signal of the decrease in exhaust steam condensation capacity is calculated. The prior trend signal and the real-time back pressure are subjected to nonlinear fitting to obtain the back pressure offset, and the pressure compensation feedforward value is calculated based on the back pressure offset.

3. The condensate pump frequency conversion energy-saving optimization method according to claim 2, characterized in that, Step 2 also includes: The vacuum decrease rate in the exhaust channel is calculated based on the prior trend signal; the vacuum decrease rate and the real-time back pressure are used to calculate the peak value of the back pressure of the direct air-cooled island and the time when the peak value occurs, and the peak value is used as the back pressure offset. The back pressure offset and the peak occurrence time are spatiotemporally combined to generate a spatiotemporal prediction trajectory for back pressure rise. Extract the peak data from the back pressure rise spatiotemporal prediction trajectory and convert the peak data into a back pressure pulse perturbation feedforward quantity; The preset pipeline resistance coefficient is retrieved, and the pipeline resistance coefficient is multiplied with the back pressure pulse disturbance feedforward to obtain the additional dynamic head. The additional dynamic head is used as the pressure compensation feedforward value.

4. The condensate pump frequency conversion energy-saving optimization method according to claim 3, characterized in that, Step 4 includes: Obtain the preset outlet valve opening time axis; The speed increase rate of the standby water pump is combined with the time axis of the outlet valve opening to generate a speed opening curve matrix; The data corresponding to the speed opening curve matrix are added to the output flow of the main feedwater pump to calculate the total water demand at the moment of grid connection. The difference between the total water demand and the current supply flow is used to calculate the water shortage value; Retrieve the preset mapping matrix between volumetric efficiency and motor frequency; The water supply gap value is numerically converted by the motor frequency mapping correlation matrix to obtain an additional applied frequency parameter, which is then used as the predictive command for condensate acceleration.

5. The condensate pump frequency conversion energy-saving optimization method according to claim 4, characterized in that, Step 4 also includes: The back pressure offset is obtained and set as an environmental correction factor; The environmental correction coefficient is injected into the motor frequency mapping correlation matrix and matrix multiplication is performed to generate an updated motor frequency mapping correlation matrix. The water supply gap value is input into the updated motor frequency mapping correlation matrix for inverse calculation to obtain the feedforward target frequency increment. The feedforward target frequency increment is output as the additional applied frequency parameter, and the combined parameters are used to generate the condensate acceleration predictive command.

6. The condensate pump frequency conversion energy-saving optimization method according to claim 5, characterized in that, Step 5 includes: The basic feedback value, the pressure compensation feedforward value, and the condensate acceleration predictive command are added together to obtain the synthetic frequency drive command. A preset filtering algorithm is invoked to perform smoothing filtering calculations on the synthesized frequency driving command, thereby obtaining a smoothed frequency command. The smoothing frequency command is compared with a preset rising slope threshold, and a rate of change limiting operation is performed to obtain a limiting frequency command. The limiting frequency command is output as the integrated frequency control signal.

7. The condensate pump frequency conversion energy-saving optimization method according to claim 6, characterized in that, Step 5 further includes: Retrieve the preset motor stator frequency and pump shaft mechanical work transfer function; By using the stator frequency of the motor and the mechanical work transfer function of the pump shaft to perform reverse analytical calculation on the limiting frequency command, the reference speed increase increment of the motor is obtained; A preset phase advance compensation operator is retrieved, and the motor reference speed increase increment is concatenated with the phase advance compensation operator to generate a frequency advance injection sequence with a time gradient. The acceleration ramp curve is redrawn according to the data nodes corresponding to the frequency advance injection sequence, and the redrawn acceleration ramp curve data is output as the comprehensive frequency control signal.

8. The condensate pump frequency conversion energy-saving optimization method according to claim 7, characterized in that, Step 5 further includes: The real-time operating frequency of the motor and the fluid pressure feedback value after transient adjustment of the pipeline network are obtained. The difference between the integrated frequency control signal and the real-time operating frequency is calculated and compared to obtain the first residual value. The difference between the steady-state reference pressure and the fluid pressure feedback value is calculated and compared to obtain the second residual value. Calculate the proportional-integral update coefficients based on the first residual value and the second residual value; adjust the proportional constant and integral constant in the feedforward adjustment logic according to the proportional-integral update coefficients.

9. The condensate pump frequency conversion energy-saving optimization method according to claim 8, characterized in that, Step 5 further includes: Continue to extract the absolute value of the second residual; Determine whether the absolute value of the second residual is less than a preset convergence threshold; When the judgment result is yes, the superposition weight of the pressure compensation feedforward value and the condensate acceleration predictive command in the comprehensive frequency control signal is reduced cyclically according to the preset attenuation step size. The weight of the product of the basic feedback quantity and the data in the integrated frequency control signal is increased synchronously until the data corresponding to the integrated frequency control signal is equal to the data of the basic feedback quantity.

10. A condensate pump frequency conversion energy-saving optimization system, applied to the condensate pump frequency conversion energy-saving optimization method as described in any one of claims 1 to 9, characterized in that, include: The data acquisition module is used to acquire real-time back pressure, ambient wind speed, outdoor temperature, condensate header pressure, deaerator water tank level, main feed water pump output flow rate, standby feed water pump speed increase rate, and standby feed water pump start command of the direct air-cooled island. The feedforward calculation module is used to calculate the heat transfer coefficient decay slope using the acquired real-time back pressure, the ambient wind speed and the outdoor temperature, and to calculate the back pressure offset using the heat transfer coefficient decay slope and the real-time back pressure, and to convert the back pressure offset into a pressure compensation feedforward value. The reference freezing module is used to receive the start command of the standby water supply pump, freeze the obtained pressure of the condensate header and the water level of the deaerator tank, and obtain the steady-state reference pressure and steady-state reference water level. The predictive instruction generation module is used to calculate the current supply flow rate based on the steady-state reference pressure and the steady-state reference water level; The total water demand is calculated based on the obtained rate of increase of the standby water pump speed and the output flow of the main water pump. The difference between the total water demand and the existing supply flow is used to obtain the water supply gap value. The water supply gap value is then converted into a condensate acceleration predictive command. The integrated control module is used to extract basic feedback based on the obtained water level of the deoxygenated water tank, and superimpose the basic feedback, the pressure compensation feedforward value, and the condensate acceleration predictive command to generate a comprehensive frequency control signal; the comprehensive frequency control signal is sent to the frequency converter execution terminal to instruct the motor to change its operating frequency according to the comprehensive frequency control signal.

Images