A hierarchical collaborative control method and system for isolated network operation of a same power level double-steam turbine generator unit

Abstract

This invention relates to the field of dual-turbine generator control, specifically to a hierarchical collaborative control method and system for islanded operation of dual-turbine generator sets of the same power rating. The method includes performing synchronous calibration of the first and second units of the dual-turbine generator sets under the same operating conditions to obtain an initial control parameter set; calculating the incremental base power command for the first and second units; calculating the target power and correcting deviations using dynamic impedance matching logic; combining real-time frequency deviation correction and amplitude limiting calibration to output a feedforward compensation signal and the final target power control command. This invention, through dynamic impedance matching logic, calculates the equivalent power response impedance difference based on the real-time operating conditions of the two units, and constructs a weighted correction coefficient based on this difference. This eliminates load distribution deviations, suppresses abnormal operating conditions such as load grabbing and reverse regulation, and ensures the power response synchronization of the two units throughout the entire operating range, providing stable and balanced power support for the islanded grid system.

Description

Technical Field

[0001] This invention relates to the field of dual-turbine generator control, and more specifically to a hierarchical collaborative control method and system for isolated grid operation of dual-turbine generator sets of the same power level. Background Technology

[0002] Steam turbine generator sets typically operate in grid-connected mode, where the primary control objective is to regulate active power according to dispatch instructions, with the system frequency maintained by the main power grid. However, when the unit operates in islanded mode, the control objective fundamentally changes: the unit must share the entire system load and autonomously establish and maintain stable system frequency and voltage. In this situation, traditional grid-connected control strategies are no longer applicable.

[0003] For an isolated grid system consisting of two steam turbine generator sets of the same power, the main challenges are: dynamic response issues; the small inertia of isolated grid systems makes them prone to large frequency fluctuations or even instability when the load changes abruptly, requiring the units to have a fast and coordinated power response capability; and power distribution issues, requiring automatic and precise distribution of active load between the two units to avoid load grabbing or load shirking, while ensuring that the distribution ratio meets the preset requirements.

[0004] Existing islanded grid control systems mostly employ simple frequency droop control, which can achieve basic power distribution but has inherent drawbacks: droop control is a differential regulation, and load changes inevitably lead to frequency deviations; response speed depends on the droop coefficient, and there is a contradiction between speed and stability; for the same type of unit, even slight differences in the characteristics of the speed control system can easily lead to static errors and dynamic oscillations in power distribution. Summary of the Invention

[0005] This invention addresses the technical problems existing in the prior art by providing a hierarchical collaborative control method and system for isolated grid operation of dual steam turbine generator sets of the same power level.

[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: A hierarchical collaborative control method for isolated grid operation of dual turbine generator sets of the same power level, comprising the following steps: S1, performing synchronous calibration of the first and second units of the dual turbine generator sets under the same operating conditions, obtaining the rated parameter set and safe operation boundary threshold of the units, configuring unified control parameters and setting system safety boundaries for the first and second units, obtaining an initial control parameter set, aligning the control references of the first and second units, and using the maximum allowable time delay difference threshold, performing time delay measurement and synchronous calibration of all control channels of the first and second units to obtain control channel time delay parameters; S2. Generate a synchronization frequency deviation signal shared by the first unit and the second unit based on the unit's rated parameter set and control channel delay parameters, calculate the basic power command increment of the first unit and the second unit, and perform amplitude limit verification on the basic power command increment of the first unit and the second unit in combination with the safe operation boundary threshold. S3. Based on the initial control parameter set and the rated parameter set of the unit, combined with the actual output power of the first and second units collected in real time, the target power is calculated and the deviation is corrected by dynamic impedance matching logic. Then, a deviation compensation mechanism is introduced to complete the real-time observation of external disturbances and the pre-compensation of periodic power deviation. Then, the power deviation is predicted in advance through interval prediction, and the power correction commands of the first and second units are generated. After synchronization verification, the coordinated correction signal is output. S4. Based on the real-time collected total load data of the isolated grid system, the initial feedforward compensation amount of the first unit and the second unit is generated according to the preset target allocation coefficient. After the correction and amplitude limiting calibration are completed by combining the real-time frequency deviation, the feedforward compensation signal is output. After setting a fixed time delay compensation value, the final target power control command is output. The final target power control command is converted into a turbine valve opening command and driven to execute the hydraulic servo mechanism.

[0007] In a preferred embodiment, step S1 performs synchronous calibration of the rated power, rated frequency, rated speed, and turbine flow characteristic curves of the first and second units of the dual-turbine generator set under the same operating conditions. It should be noted that all calibration operations must be completed under completely identical rated operating conditions of the two units to ensure uniform calibration benchmarks. The main steam pressure, main steam temperature, and condenser vacuum of the first and second units are all adjusted to the rated operating conditions. Starting from the fully closed state, the turbine valves are gradually opened to the fully open state in units of valve characteristic calibration opening increments. For each adjustment of the opening point, the opening is kept stable for no less than the single-point opening stability time of the valve characteristic calibration. After the standby unit output power is completely stable, record the current valve opening values ​​of the first unit and the second unit, as well as the actual output active power values ​​of each unit under the corresponding valve opening values. It should be noted that the valve characteristic calibration opening step size refers to the single step amplitude of the valve opening adjustment, which is determined by the test results of the unit's valve adjustment accuracy. Its constraint is that the step amplitude must cover the linear adjustment range of the entire valve stroke. The valve characteristic calibration single-point opening stabilization time refers to the shortest holding time required for power stabilization under single-point opening, which is determined by the test results of the unit's power response lag characteristics. Its constraint is that the holding time must be greater than the stabilization time of the unit's power closed-loop regulation. The valve opening values ​​of the first and second units and the actual output active power values ​​are organized to form a one-to-one corresponding valve flow characteristic curve. Since the turbine valve gradually opens to the fully open state during the above process, the curve completely covers all operating ranges from fully closed to fully open valves. The corresponding power value can be queried by any opening value, and the corresponding opening value can be queried by any power value. Based on the calibrated valve flow characteristic curves, the output power deviation between the first unit and the second unit is ensured to be within the maximum allowable power deviation threshold of the valve calibration for the first unit and the second unit, thus obtaining the set of rated parameters of the units after collaborative calibration.

[0008] The maximum permissible power deviation threshold for the regulating valves of the first and second generating units refers to the maximum permissible power deviation between the first and second generating units at the same opening degree during the calibration process. It is controlled by the manufacturing tolerance standard of the same type of equipment of the generating units, and the constraint is that the deviation value must meet the basic synchronization requirements of the parallel operation of the generating units. In summary, it can be seen that the rated parameter set of the generating units after collaborative calibration includes the rated power, rated frequency, rated speed, and regulating valve flow characteristic curves of the two generating units after unified calibration.

[0009] In a preferred embodiment, step S1 sets identical frequency droop coefficient reference values ​​for the first and second generating units based on the unit's rated parameter set. In some other specific embodiments, the frequency droop coefficient reference value is a frequency-power droop characteristic coefficient synchronously set for the first and second generating units, used to represent the magnitude of the change in unit power command corresponding to a change in system frequency. In the preceding static characteristic test of the speed control system, the test results of the static characteristic test of the speed control system under the rated operating conditions of the units are used as the data source to configure the target allocation coefficients for the first and second generating units. Specifically, the configuration setting is such that the sum of the target allocation coefficients of the two generating units equals 1, and in the even distribution scenario, the distribution is based on the generating units. The load is evenly distributed by quantity. In a custom proportional scenario, the load is converted according to the operating requirements and then fixed into the control program. For example, if the workload of the first unit is 0.6, the preset workload of the second unit is 0.4. Thus, the target allocation coefficient refers to the proportion of the total active load of the system borne by a single unit. The allocation can be customized according to the unit's operating conditions and load distribution requirements. Then, according to the actual operating conditions, the system frequency allowable fluctuation boundary, the unit power operating lower limit threshold, the unit power operating upper limit threshold, the unit maximum allowable power change rate threshold, and the maximum allowable time delay difference threshold of the control channels of the first and second units are set in sequence to form the initialization control parameter set and the system safe operation boundary threshold.

[0010] In a preferred embodiment, S1 further includes: using the unified clock of the central controller as a reference, synchronously sending trigger signals to the frequency acquisition, power acquisition, and digital electro-hydraulic control command sending channels, recording the transmission and reception time of each channel, and calculating the one-way transmission delay; Based on the maximum transmission delay of all channels, a delay compensation value is configured for channels whose transmission delay is less than the maximum transmission delay of the channel. The compensation value is the difference between the maximum delay and the actual delay of the channel, so as to achieve consistency of the total transmission delay of all channels. After calibration, the sampling and transmission delay difference of the control signals of the first and second units is verified to obtain the control channel delay parameters after synchronous calibration.

[0011] The control channel delay parameters specifically include the control channel delay parameters after synchronous calibration, including the delay compensation value of each channel, the total delay after calibration, and the verification result of the delay difference between the first unit and the second unit.

[0012] In a preferred embodiment, S2 uses the rated frequency calibrated by the unit's rated parameters as the frequency deviation of the common bus of the islanded network system, and uses the difference between the rated frequency and the actual operating frequency as the frequency deviation signal of the first unit and the second unit as the only common control signal. After synchronous calibration, the signal is synchronously sent to the droop control module of the first unit and the second unit through the equal time delay channel to obtain the real-time system frequency deviation and common frequency deviation signal. The ratio of frequency deviation to the reference value of frequency droop coefficient is inverted by multiplying the product of the rated power of the unit to obtain the basic power command increment of a single unit. Since the two units use completely identical rated parameters, droop coefficients and frequency deviation signals, their basic power command increments are exactly the same, and the synchronous basic power command increments of the first unit and the synchronous basic power command increments of the second unit are obtained. The incremental synchronous base power command of the first unit and the incremental synchronous base power command of the second unit are respectively superimposed with the actual output power of the corresponding previous cycle to obtain the superimposed power command. The superimposed power command is compared with the upper and lower limit thresholds of the unit power operation. If it exceeds the boundary, the superimposed power command is limited to the boundary value to obtain the verified incremental synchronous base power command of the first unit and the incremental synchronous base power command of the second unit.

[0013] In a preferred embodiment, S3 sums the actual output active power of the first unit and the actual output active power of the second unit based on the initialization control parameter set and the unit rated parameter set to obtain the current total active power demand of the system. The total active power demand is then multiplied by the target allocation coefficients corresponding to the first unit and the second unit to obtain the real-time target power of the first unit and the second unit. Then extract the power value corresponding to the current valve opening and the two power values ​​corresponding to the previous and next openings adjacent to the current opening. Use the three openings as the horizontal axis and the three power values ​​as the vertical axis, calculate the average slope between the three points, obtain the valve opening change corresponding to the unit power change under the current opening, and use it as the equivalent power response impedance difference between the first unit and the second unit, and the maximum allowable impedance difference of the equivalent power response between the first unit and the second unit. The original power deviation of the first and second units is obtained by using the difference between the real-time target power and the corresponding actual output active power of the first and second units. Then, the equivalent power response impedance difference of the first and second units is introduced, and the weight correction coefficient is calculated to obtain the corrected power deviation. Specifically, after determining the maximum allowable impedance difference of the equivalent power response of the first and second units, that is, the maximum difference of the equivalent power response impedance corresponding to all operating points of the two units in the entire operating range from fully closed to fully open, the current impedance difference is divided by the maximum allowable impedance difference, and 1 is added to obtain the deviation correction weight of a single unit. The original power deviation of a single unit is multiplied by the corresponding correction weight to finally obtain the corrected power deviation.

[0014] In a preferred embodiment, S3 further includes: The theoretical valve opening is obtained by converting the basic power command increment as the input signal of the governor, and then the theoretical output power is obtained after the steam filling delay is compensated for the steam volume lag of the steam turbine. This process is the whole physical process from command to power. The theoretical output power is subtracted from the actual output power collected by the first and second units to obtain the total disturbance observation value, the disturbance compensation value of the first unit, and the disturbance compensation value of the second unit. The total disturbance observation value is then subjected to a first-order low-pass filter to obtain a smoothed disturbance compensation value. The corrected historical power deviation data for continuous operating cycles is collected. After removing DC trends and random fluctuations, a discrete Fourier transform is performed to convert the time-domain data of the historical power deviation data into frequency-domain data. The fixed frequency component with the largest amplitude is selected from the frequency-domain data. This component is the periodic fluctuation component. Periodic deviation pre-compensation values ​​for the first unit and the second unit, which are equal in magnitude and opposite in direction to the largest fixed frequency component, are generated, thus completing the introduction of the deviation compensation mechanism.

[0015] In a preferred embodiment, step S3 is based on the current corrected power deviation, the disturbance compensation value of the first unit and the disturbance compensation value of the second unit, as well as the periodic deviation pre-compensation value of the first unit and the periodic deviation pre-compensation value of the second unit, and extracts the power deviation change rate of the past multiple consecutive control cycles, calculates the average of the power deviation change rate of the past multiple consecutive control cycles as the prediction slope, and then, starting from the current corrected power deviation, predicts the power deviation prediction value cycle by cycle according to the prediction slope. Furthermore, this application calculates the deviation of the predicted value based on the linear relationship. If the dual-turbine generator has a nonlinear relationship, a nonlinear prediction algorithm, such as mode decomposition or support vector machine, is used to obtain the predicted power deviation value. However, considering that most of the predictions of existing dual-turbine generators have obvious linear characteristics, this application chooses the above scheme, which will not be elaborated here. Based on the safe operation boundary threshold, the allowable range of prediction deviation is set. If the predicted power deviation exceeds the allowable range, an advance correction value is generated. If it does not exceed the allowable range, a steady correction value is generated. The allowable range of prediction deviation is obtained by adding or subtracting the safe operation boundary threshold, the real-time power of the two units, and the current corrected power deviation. Then, the corrected power deviation, the pre-compensation value of the periodic deviation of the first unit, the pre-compensation value of the periodic deviation of the second unit, and the advance correction value are superimposed to form the final power deviation correction signal of the first unit and the second unit. The transmission delay of the final power deviation correction signal of the first and second units is checked to ensure that it does not exceed the maximum allowable delay difference threshold of the dual-unit control channel. If it exceeds the tolerance, the delay parameter is compensated and adjusted to obtain the power correction command of the first unit and the power correction command of the second unit after synchronous verification.

[0016] In a preferred embodiment, step S4 collects the total active load value on the common bus of the isolated network in real time, and obtains the load change in the current cycle by subtracting the total active load value of the current cycle from the total active load value of the previous cycle. The effective load change is multiplied by the target allocation coefficients of the first unit and the second unit respectively to obtain the initial feedforward compensation amount of the first unit and the second unit. The initial feedforward compensation amount of the first unit and the second unit is corrected according to the feedforward static gain coefficient to obtain the corrected initial feedforward compensation amount of the first unit and the second unit. It should be noted that the feedforward static gain coefficient is the ratio of the actual output power change of the unit to the power command change. The initial feedforward compensation is multiplied by this gain coefficient to correct the compensation deviation caused by the nonlinearity of the turbine control valve and the power response lag, thus obtaining the corrected initial feedforward compensation.

[0017] Based on the upper and lower limits of unit power operation in the safe operation boundary threshold, the current adjustable upper and lower limits of the first and second units are calculated. The upper limit is the unit power operation upper limit threshold minus the current actual output power of the unit, and the lower limit is the current actual output power of the unit minus the unit power operation lower limit threshold. The feedforward compensation amount must not exceed the range of the upper and lower limits. The corrected initial feedforward compensation amount of the first and second units is locked within the maximum allowable limit boundary, a feedforward compensation signal is generated, and a fixed time delay compensation value is set on the feedforward compensation signal according to the synchronously calibrated control channel time delay parameter, and the final target power control command is output.

[0018] This invention also provides a hierarchical collaborative control system for islanded operation of dual steam turbine generator sets of the same power rating, comprising: Initialization module; used to establish control benchmarks, complete unified calibration through synchronous calibration under the same operating conditions, set core control benchmarks such as frequency droop coefficient and load target allocation coefficient, and set safe operating boundaries; Synchronization control module; used to provide underlying frequency support for isolated grid systems, calculate the frequency deviation of the isolated grid common bus in real time, and synchronously send it to the two units through the equal time delay channel after synchronization calibration. At the same time, it completes the preliminary safety limit verification of basic commands based on safety boundary thresholds, and outputs the verified dual-machine synchronization basic power command increment and real-time system frequency deviation. Deviation Co-correction Module: It is used to calculate the real-time target power of the two machines and complete the weight correction of the power deviation through dynamic impedance matching logic. Then, it extracts and compensates for periodic power fluctuations through repetitive control to eliminate the dynamic deviation caused by disturbances. Finally, it generates the lead correction component and outputs the dual-machine power co-correction command after synchronization verification. Feedforward compensation module: used to collect the total active load of the isolated network system in real time and generate the initial feedforward compensation amount, improve the compensation accuracy through static gain correction, and output the final feedforward compensation command.

[0019] The beneficial effects of this invention are as follows: This invention uses dynamic impedance matching logic to calculate the equivalent power response impedance difference based on the real-time operating conditions of the two turbines, and constructs a weight correction coefficient based on this. This completes the differential correction of power deviation, which can adaptively adapt to the nonlinear characteristics of the turbine control valve throughout its entire stroke. It eliminates the load distribution deviation caused by the inherent characteristic differences between the two turbines, suppresses abnormal operating conditions such as load grabbing and reverse adjustment from the root, ensures the power response synchronization of the two turbines throughout the entire operating range, and provides stable and balanced power support for the isolated grid system. By modeling the entire physical process from command to power, the difference between theoretical and actual output power is used as the total disturbance observation value. This enables real-time observation and smooth compensation of non-periodic external disturbances, achieving rapid suppression of disturbances such as load abrupt changes and steam parameter fluctuations. Furthermore, frequency domain transformation extracts the dominant periodic fluctuation component from the power deviation, generating a reverse pre-compensation value to actively eliminate inherent periodic fluctuations during unit regulation. Compared to traditional lag-based closed-loop regulation, this scheme upgrades passive deviation elimination to active disturbance compensation, significantly reducing the impact of various disturbances on the frequency and power of the isolated grid system, avoiding system oscillations caused by continuous disturbance amplification, and significantly improving the frequency stability of the isolated grid system. Attached Figure Description

[0020] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] As attached Figure 1 As shown, this embodiment provides a hierarchical collaborative control method for islanded operation of dual steam turbine generator sets of the same power rating, including the following steps: S1. Perform synchronous calibration of the first and second units of the dual-turbine generator set under the same operating conditions. After obtaining the rated parameter set and safe operating boundary threshold of the units, configure the unified control parameters of the first and second units and set the system safety boundary to obtain the initial control parameter set. Align the control reference of the first and second units. And use the maximum allowable time delay difference threshold to measure and synchronously calibrate the time delay of all control channels of the first and second units to obtain the control channel time delay parameters. S1 performs synchronous calibration of the rated power, rated frequency, rated speed, and turbine control valve flow characteristic curves of the first and second units of the dual-turbine generator set under the same operating conditions. It should be noted that all calibration operations must be completed under completely identical rated operating conditions of the two units to ensure that the calibration benchmark is consistent. The main steam pressure, main steam temperature, and condenser vacuum of the first and second units are all adjusted to the rated operating conditions. Starting from the fully closed state, the turbine control valves are gradually opened to the fully open state in units of the control valve characteristic calibration opening size. After adjusting each opening size point, the opening size is kept stable for no less than the single-point opening stability time of the control valve characteristic calibration. After the standby unit output power is completely stable, record the current valve opening values ​​of the first unit and the second unit, as well as the actual output active power values ​​of each unit under the corresponding valve opening values. It should be noted that the valve characteristic calibration opening step size refers to the single step amplitude of the valve opening adjustment, which is determined by the test results of the unit's valve adjustment accuracy. Its constraint is that the step amplitude must cover the linear adjustment range of the entire valve stroke. The valve characteristic calibration single-point opening stabilization time refers to the shortest holding time required for power stabilization under single-point opening, which is determined by the test results of the unit's power response lag characteristics. Its constraint is that the holding time must be greater than the stabilization time of the unit's power closed-loop regulation. The valve opening values ​​of the first and second units and the actual output active power values ​​are organized to form a one-to-one corresponding valve flow characteristic curve. Since the turbine valve gradually opens to the fully open state during the above process, the curve completely covers all operating ranges from fully closed to fully open valves. The corresponding power value can be queried by any opening value, and the corresponding opening value can be queried by any power value. Based on the calibrated valve flow characteristic curves, the output power deviation between the first unit and the second unit is ensured to be within the maximum allowable power deviation threshold of the valve calibration for the first unit and the second unit, thus obtaining the set of rated parameters of the units after collaborative calibration.

[0023] The maximum permissible power deviation threshold for the regulating valves of the first and second generating units refers to the maximum permissible power deviation between the first and second generating units at the same opening degree during the calibration process. It is controlled by the manufacturing tolerance standard of the same type of equipment of the generating units, and the constraint is that the deviation value must meet the basic synchronization requirements of the parallel operation of the generating units. In summary, it can be seen that the rated parameter set of the generating units after collaborative calibration includes the rated power, rated frequency, rated speed, and regulating valve flow characteristic curves of the two generating units after unified calibration.

[0024] S1 sets identical frequency droop coefficient benchmark values ​​for the first and second generating units based on the unit's rated parameter set. In some other specific implementations, the frequency droop coefficient benchmark value is a frequency-power droop characteristic coefficient synchronously set for the first and second generating units, used to represent the magnitude of the change in unit power command corresponding to a change in system frequency. In the preceding static characteristic test of the speed control system, the test results of the static characteristic test of the speed control system under the rated operating conditions of the units are used as the data source to configure the target allocation coefficients for the first and second generating units. Specifically, the configuration settings are such that the sum of the target allocation coefficients of the two generating units equals 1, and the load is evenly allocated according to the number of generating units in the equal distribution scenario. The load is calculated according to the operational requirements under the custom proportional scenario and then fixed into the control program. For example, the workload undertaken by the first unit is 0.6, and the preset workload of the second unit is 0.4. Thus, the target allocation coefficient refers to the proportion of the total active load of the system undertaken by a single unit. It can be customized according to the unit's operating conditions and load allocation requirements. Then, according to the actual operating conditions, the system frequency allowable fluctuation boundary, the lower limit threshold of unit power operation, the upper limit threshold of unit power operation, the maximum allowable power change rate threshold of the unit, and the maximum allowable time delay difference threshold of the control channels of the first and second units are set in sequence. These are summarized to form the initialization control parameter set and the system safe operation boundary threshold. To further understand the relevant settings disclosed in this application, the system frequency allowable fluctuation boundary is the allowable frequency deviation range for normal operation of the isolated grid system, determined based on the power quality standards for isolated grid operation; the lower limit threshold for unit power operation refers to the minimum output power of the unit for stable operation, which can be derived based on the results of unit low-load stable operation tests or past operating experience; the upper limit threshold for unit power operation is the maximum allowable output power of the unit in a short time, determined by the power when the unit is overloaded; the maximum allowable time delay difference threshold for the control channels of the first and second units is the maximum allowable time difference in the transmission of control signals between the two units.

[0025] Based on the unified clock of the central controller, trigger signals are synchronously sent to the frequency acquisition, power acquisition, and digital electro-hydraulic control command transmission channels, and the transmission and reception times of each channel are recorded and the one-way transmission delay is calculated. Based on the maximum transmission delay of all channels, a delay compensation value is configured for channels whose transmission delay is less than the maximum transmission delay of the channel. The compensation value is the difference between the maximum delay and the actual delay of the channel, so as to achieve consistency of the total transmission delay of all channels. After calibration, the sampling and transmission delay difference of the control signals of the first and second units is verified to obtain the control channel delay parameters after synchronous calibration.

[0026] The control channel delay parameters specifically include the control channel delay parameters after synchronous calibration, including the delay compensation value of each channel, the total delay after calibration, and the verification result of the delay difference between the first unit and the second unit.

[0027] S2. Generate a synchronization frequency deviation signal shared by the first unit and the second unit based on the unit's rated parameter set and control channel delay parameters, calculate the basic power command increment of the first unit and the second unit, and perform amplitude limit verification on the basic power command increment of the first unit and the second unit in combination with the safe operation boundary threshold. Based on the rated frequency calibrated by the unit's rated parameters, the difference between the rated frequency and the actual operating frequency is used as the frequency deviation of the common bus of the islanded grid system. The frequency deviation signal is used as the only common control signal of the first and second units. After synchronous calibration, the signal is synchronously sent to the droop control modules of the first and second units through the equal time delay channel to obtain the real-time system frequency deviation and common frequency deviation signal. The ratio of frequency deviation to the reference value of frequency droop coefficient is inverted by multiplying the product of the rated power of the unit to obtain the basic power command increment of a single unit. Since the two units use completely identical rated parameters, droop coefficients and frequency deviation signals, their basic power command increments are exactly the same, and the synchronous basic power command increments of the first unit and the synchronous basic power command increments of the second unit are obtained. The incremental synchronous base power command of the first unit and the incremental synchronous base power command of the second unit are respectively superimposed with the actual output power of the corresponding previous cycle to obtain the superimposed power command. The superimposed power command is compared with the upper and lower limit thresholds of the unit power operation. If it exceeds the boundary, the superimposed power command is limited to the boundary value to obtain the verified incremental synchronous base power command of the first unit and the incremental synchronous base power command of the second unit.

[0028] S3. Based on the initial control parameter set and the rated parameter set of the unit, combined with the actual output power of the first and second units collected in real time, the target power is calculated and the deviation is corrected by dynamic impedance matching logic. Then, a deviation compensation mechanism is introduced to complete the real-time observation of external disturbances and the pre-compensation of periodic power deviation. Then, the power deviation is predicted in advance through interval prediction, and the power correction commands of the first and second units are generated. After synchronization verification, the coordinated correction signal is output. Based on the initialization control parameter set and the unit rated parameter set, the actual output active power of the first unit and the actual output active power of the second unit are summed to obtain the current total active power demand of the system. The real-time target power of the first unit and the second unit is obtained by multiplying the total active power demand by the target allocation coefficients corresponding to the first unit and the second unit respectively. Then extract the power value corresponding to the current valve opening and the two power values ​​corresponding to the previous and next openings adjacent to the current opening. Use the three openings as the horizontal axis and the three power values ​​as the vertical axis, calculate the average slope between the three points, obtain the valve opening change corresponding to the unit power change under the current opening, and use it as the equivalent power response impedance difference between the first unit and the second unit, and the maximum allowable impedance difference of the equivalent power response between the first unit and the second unit. The original power deviation of the first and second units is obtained by using the difference between the real-time target power and the corresponding actual output active power of the first and second units. Then, the equivalent power response impedance difference of the first and second units is introduced, and the weight correction coefficient is calculated to obtain the corrected power deviation. Specifically, after determining the maximum allowable impedance difference of the equivalent power response of the first and second units, that is, the maximum difference of the equivalent power response impedance of the two units at all operating points in the full operating range from fully closed to fully open, the current impedance difference is divided by the maximum allowable impedance difference, and 1 is added to obtain the deviation correction weight of a single unit. The original power deviation of a single unit is multiplied by the corresponding correction weight to finally obtain the corrected power deviation. To further understand the technical solutions disclosed above in this application, it should be noted that the equivalent power response impedance refers to the change in valve opening corresponding to a unit power change at the current opening degree, which is the absolute value of the difference between the equivalent power response impedances of the two units under the current operating conditions, and the impedance deviation must be within the maximum allowable range. The difference in equivalent power response impedance between the first unit and the second unit refers to the maximum difference in impedance between the two units within the full opening degree range, which is determined by the stroke calibration results.

[0029] The theoretical valve opening is obtained by converting the basic power command increment as the input signal of the governor, and then the theoretical output power is obtained after the steam filling delay is compensated for the steam volume lag of the steam turbine. This process is the whole physical process from command to power. The theoretical output power is subtracted from the actual output power collected by the first and second units to obtain the total disturbance observation value, the disturbance compensation value of the first unit, and the disturbance compensation value of the second unit. The total disturbance observation value is then subjected to a first-order low-pass filter to obtain a smoothed disturbance compensation value. The corrected historical power deviation data for continuous operating cycles is collected. After removing DC trends and random fluctuations, a discrete Fourier transform is performed to convert the time-domain data of the historical power deviation data into frequency-domain data. The fixed frequency component with the largest amplitude is selected from the frequency-domain data. This component is the periodic fluctuation component. Periodic deviation pre-compensation values ​​for the first unit and the second unit, which are equal in magnitude and opposite in direction to the largest fixed frequency component, are generated, thus completing the introduction of the deviation compensation mechanism.

[0030] In some other specific implementations, the complete filtering steps include initializing the initial value of the filter output, multiplying the current disturbance observation value by the filter coefficient, multiplying the previous filter output value by the complement of the filter coefficient, adding the two products to obtain the current filter output, updating the previous filter output value for the next calculation, and filtering out high-frequency sampling noise to obtain a smoothed disturbance compensation value.

[0031] Based on the current corrected power deviation, the disturbance compensation value of the first unit and the disturbance compensation value of the second unit, as well as the periodic deviation pre-compensation value of the first unit and the periodic deviation pre-compensation value of the second unit, and extracting the power deviation change rate of multiple consecutive control cycles in the past, the average value of the power deviation change rate of multiple consecutive control cycles in the past is calculated as the prediction slope, and then the current corrected power deviation is used as the starting point to predict the power deviation prediction value cycle by cycle according to the prediction slope. Furthermore, this application calculates the deviation of the predicted value based on the linear relationship. If the dual-turbine generator has a nonlinear relationship, a nonlinear prediction algorithm, such as mode decomposition or support vector machine, is used to obtain the predicted power deviation value. However, considering that most of the predictions of existing dual-turbine generators have obvious linear characteristics, this application chooses the above scheme, which will not be elaborated here. Based on the safe operation boundary threshold, the allowable range of prediction deviation is set. If the predicted power deviation exceeds the allowable range, an advance correction value is generated. If it does not exceed the allowable range, a steady correction value is generated. The allowable range of prediction deviation is obtained by adding or subtracting the safe operation boundary threshold, the real-time power of the two units, and the current corrected power deviation. Then, the corrected power deviation, the pre-compensation value of the periodic deviation of the first unit, the pre-compensation value of the periodic deviation of the second unit, and the advance correction value are superimposed to form the final power deviation correction signal of the first unit and the second unit. The transmission delay of the final power deviation correction signal of the first and second units is checked to ensure that it does not exceed the maximum allowable delay difference threshold of the dual-unit control channel. If it exceeds the tolerance, the delay parameter is compensated and adjusted to obtain the power correction command of the first unit and the power correction command of the second unit after synchronous verification.

[0032] S4. Based on the real-time collected total load data of the isolated grid system, the initial feedforward compensation amount of the first unit and the second unit is generated according to the preset target allocation coefficient. After the correction and amplitude limiting calibration are completed by combining the real-time frequency deviation, the feedforward compensation signal is output. After setting a fixed time delay compensation value, the final target power control command is output. The final target power control command is converted into a turbine valve opening command and driven to execute the hydraulic servo mechanism.

[0033] S4 collects the total active load value on the common bus of the isolated network in real time, and obtains the load change in the current cycle by subtracting the total active load value of the current cycle from the total active load value of the previous cycle. The effective load change is multiplied by the target allocation coefficient of the first unit and the second unit respectively to obtain the initial feedforward compensation amount of the first unit and the second unit. The initial feedforward compensation amount of the first unit and the second unit is corrected according to the feedforward static gain coefficient to obtain the corrected initial feedforward compensation amount of the first unit and the second unit. It should be noted that the feedforward static gain coefficient is the ratio of the actual output power change of the unit to the power command change. The initial feedforward compensation is multiplied by this gain coefficient to correct the compensation deviation caused by the nonlinearity of the turbine control valve and the power response lag, thus obtaining the corrected initial feedforward compensation.

[0034] Based on the upper and lower limits of unit power operation in the safe operation boundary threshold, the current adjustable upper and lower limits of the first and second units are calculated. The upper limit is the unit power operation upper limit threshold minus the current actual output power of the unit, and the lower limit is the current actual output power of the unit minus the unit power operation lower limit threshold. The feedforward compensation amount must not exceed the range of the upper and lower limits. The corrected initial feedforward compensation amount of the first and second units is locked within the maximum allowable limit boundary, a feedforward compensation signal is generated, and a fixed time delay compensation value is set on the feedforward compensation signal according to the synchronously calibrated control channel time delay parameter, and the final target power control command is output.

[0035] This invention also provides a hierarchical collaborative control system for islanded operation of dual steam turbine generator sets of the same power rating, comprising: Initialization module; used to establish control benchmarks, complete unified calibration through synchronous calibration under the same operating conditions, set core control benchmarks such as frequency droop coefficient and load target allocation coefficient, and set safe operating boundaries; Synchronization control module; used to provide underlying frequency support for isolated grid systems, calculate the frequency deviation of the isolated grid common bus in real time, and synchronously send it to the two units through the equal time delay channel after synchronization calibration. At the same time, it completes the preliminary safety limit verification of basic commands based on safety boundary thresholds, and outputs the verified dual-machine synchronization basic power command increment and real-time system frequency deviation. Deviation Co-correction Module: It is used to calculate the real-time target power of the two machines and complete the weight correction of the power deviation through dynamic impedance matching logic. Then, it extracts and compensates for periodic power fluctuations through repetitive control to eliminate the dynamic deviation caused by disturbances. Finally, it generates the lead correction component and outputs the dual-machine power co-correction command after synchronization verification. Feedforward compensation module: used to collect the total active load of the isolated network system in real time and generate the initial feedforward compensation amount, improve the compensation accuracy through static gain correction, and output the final feedforward compensation command.

Claims

1. A hierarchical collaborative control method for isolated grid operation of dual steam turbine generator sets of the same power rating, characterized in that, Includes the following steps: S1. Perform synchronous calibration of the first and second units of the dual-turbine generator set under the same operating conditions. After obtaining the rated parameter set and safe operating boundary threshold of the units, configure the unified control parameters of the first and second units and set the system safety boundary to obtain the initial control parameter set. Align the control reference of the first and second units. And use the maximum allowable time delay difference threshold to measure and synchronously calibrate the time delay of all control channels of the first and second units to obtain the control channel time delay parameters. S2. Generate a synchronization frequency deviation signal shared by the first unit and the second unit based on the unit's rated parameter set and control channel delay parameters, calculate the basic power command increment of the first unit and the second unit, and perform amplitude limit verification on the basic power command increment of the first unit and the second unit in combination with the safe operation boundary threshold. S3. Based on the initial control parameter set and the rated parameter set of the unit, combined with the actual output power of the first and second units collected in real time, the target power is calculated and the deviation is corrected by dynamic impedance matching logic. Then, a deviation compensation mechanism is introduced to complete the real-time observation of external disturbances and the pre-compensation of periodic power deviation. Then, the power deviation is predicted in advance through interval prediction, and the power correction commands of the first and second units are generated. After synchronization verification, the coordinated correction signal is output. S4. Based on the real-time collected total load data of the isolated grid system, the initial feedforward compensation amount of the first unit and the second unit is generated according to the preset target allocation coefficient. After the correction and amplitude limiting calibration are completed by combining the real-time frequency deviation, the feedforward compensation signal is output. After setting a fixed time delay compensation value, the final target power control command is output. The final target power control command is converted into a turbine valve opening command and driven to execute the hydraulic servo mechanism.

2. The hierarchical collaborative control method for isolated grid operation of dual steam turbine generator sets of the same power rating according to claim 1, characterized in that, S1 performs synchronous calibration of the rated power, rated frequency, rated speed, and turbine flow characteristic curves of the first and second units of the dual-turbine generator set under the same operating conditions. The main steam pressure, main steam temperature, and condenser vacuum of the first and second units are all adjusted to the rated operating conditions. Starting from the fully closed state, the turbine valves are gradually opened to the fully open state in units of valve characteristic calibration opening size. For each adjustment of the opening size, the opening size is kept stable for no less than the single-point opening stability time of the valve characteristic calibration. After the standby unit output power is completely stable, record the current valve opening values ​​of the first unit and the second unit, as well as the actual output active power values ​​of each unit under the corresponding valve opening values. The valve opening values ​​of the first and second units and the actual output active power values ​​are organized to form a one-to-one corresponding valve flow characteristic curve. Based on the calibrated valve flow characteristic curves, the output power deviation between the first unit and the second unit is ensured to be within the maximum allowable power deviation threshold of the valve calibration for the first unit and the second unit, thus obtaining the set of rated parameters of the units after collaborative calibration.

3. The hierarchical collaborative control method for isolated grid operation of dual steam turbine generator sets of the same power rating according to claim 2, characterized in that, S1 sets identical frequency droop coefficient benchmark values ​​for the first and second units based on the unit's rated parameter set, configures the target allocation coefficients for the first and second units, and then sequentially sets the system frequency allowable fluctuation boundary, unit power operation lower limit threshold, unit power operation upper limit threshold, unit maximum allowable power change rate threshold, and the maximum allowable time delay difference threshold of the control channels of the first and second units according to the actual operating conditions. These are then summarized to form the initialization control parameter set and the system safe operation boundary threshold.

4. The hierarchical collaborative control method for isolated grid operation of dual steam turbine generator sets of the same power rating according to claim 1, characterized in that, S1 further includes: using the unified clock of the central controller as a reference, synchronously sending trigger signals to the frequency acquisition, power acquisition, and digital electro-hydraulic control command sending channels, recording the transmission and reception time of each channel and calculating the one-way transmission delay; Based on the maximum transmission delay of all channels, a delay compensation value is configured for channels whose transmission delay is less than the maximum transmission delay of the channel. The compensation value is the difference between the maximum delay and the actual delay of the channel, so as to achieve consistency of the total transmission delay of all channels. After calibration, the sampling and transmission delay difference of the control signals of the first and second units is verified to obtain the control channel delay parameters after synchronous calibration.

5. The hierarchical collaborative control method for isolated grid operation of dual steam turbine generator sets of the same power rating according to claim 1, characterized in that, S2 is based on the rated frequency calibrated by the unit's rated parameters. The difference between the rated frequency and the actual operating frequency is used as the frequency deviation of the common bus of the islanded network system. The frequency deviation signal is used as the only common control signal of the first unit and the second unit. After synchronous calibration, it is synchronously sent to the first unit and the second unit through the equal time delay channel to obtain the real-time system frequency deviation and common frequency deviation signal. The ratio of frequency deviation to the reference value of frequency droop coefficient is inverted by multiplying the product of the rated power of the unit to obtain the basic power command increment of a single unit, and the synchronous basic power command increment of the first unit and the synchronous basic power command increment of the second unit are obtained. The incremental synchronous base power command of the first unit and the incremental synchronous base power command of the second unit are respectively superimposed with the actual output power of the corresponding previous cycle to obtain the superimposed power command. The superimposed power command is compared with the upper and lower limit thresholds of the unit power operation. If it exceeds the boundary, the superimposed power command is limited to the boundary value to obtain the verified incremental synchronous base power command of the first unit and the incremental synchronous base power command of the second unit.

6. The hierarchical collaborative control method for isolated grid operation of dual steam turbine generator sets of the same power rating according to claim 1, characterized in that, S3, based on the initialization control parameter set and the unit rated parameter set, sums the actual output active power of the first unit and the actual output active power of the second unit to obtain the current total active power demand of the system. Then, by multiplying the total active power demand by the target allocation coefficients corresponding to the first unit and the second unit respectively, the real-time target power of the first unit and the second unit is obtained. Then extract the power value corresponding to the current valve opening and the two power values ​​corresponding to the previous and next openings adjacent to the current opening. Use the three openings as the horizontal axis and the three power values ​​as the vertical axis, calculate the average slope between the three points, obtain the valve opening change corresponding to the unit power change under the current opening, and use it as the equivalent power response impedance difference between the first unit and the second unit, and the maximum allowable impedance difference of the equivalent power response between the first unit and the second unit. The original power deviation between the first and second units is obtained by using the difference between the real-time target power and the corresponding actual output active power of the first and second units. Then, the equivalent power response impedance difference between the first and second units is introduced, and the weighting correction coefficient is calculated to obtain the corrected power deviation.

7. The hierarchical collaborative control method for isolated grid operation of dual steam turbine generator sets of the same power rating according to claim 6, characterized in that, S3 further includes: The theoretical valve opening is obtained by converting the basic power command increment as the input signal of the governor, and then the theoretical output power is obtained after the compensation steam filling delay for the steam volume lag of the steam turbine. The theoretical output power is subtracted from the actual output power collected by the first and second units to obtain the total disturbance observation value, the disturbance compensation value of the first unit, and the disturbance compensation value of the second unit. The total disturbance observation value is then subjected to a first-order low-pass filter to obtain a smoothed disturbance compensation value. The corrected historical power deviation data for continuous operating cycles is collected. After removing DC trends and random fluctuations, a discrete Fourier transform is performed to convert the time-domain data of the historical power deviation data into frequency-domain data. The fixed frequency component with the largest amplitude is selected from the frequency-domain data, and the periodic deviation pre-compensation value of the first unit and the periodic deviation pre-compensation value of the second unit, which are equal in magnitude and opposite in direction to the largest fixed frequency component, are generated, thus completing the introduction of the deviation compensation mechanism.

8. The hierarchical collaborative control method for isolated grid operation of dual steam turbine generator sets of the same power rating according to claim 1, characterized in that, S3 is based on the current corrected power deviation, the disturbance compensation value of the first unit and the disturbance compensation value of the second unit, as well as the periodic deviation pre-compensation value of the first unit and the periodic deviation pre-compensation value of the second unit. It also extracts the power deviation change rate of the past multiple consecutive control cycles, calculates the average of the power deviation change rate of the past multiple consecutive control cycles as the prediction slope, and then uses the current corrected power deviation as the starting point to predict the power deviation prediction value cycle by cycle according to the prediction slope. Based on the safety operation boundary threshold, the allowable range of prediction deviation is set. If the predicted power deviation exceeds the allowable range, an advance correction value is generated. If it does not exceed the allowable range, a steady correction value is generated. The corrected power deviation, the pre-compensation value of the periodic deviation of the first unit and the pre-compensation value of the periodic deviation of the second unit, and the advance correction value are then superimposed to form the final power deviation correction signal of the first unit and the second unit. The transmission delay of the final power deviation correction signal of the first unit and the second unit is checked to obtain the power correction command of the first unit and the power correction command of the second unit after synchronous verification.

9. A hierarchical collaborative control method for isolated grid operation of dual steam turbine generator sets of the same power rating according to claim 8, characterized in that, S4 collects the total active load value on the common bus of the isolated network in real time, and obtains the load change in the current period by subtracting the total active load value of the current period from the total active load value of the previous period. The effective load change is multiplied by the target allocation coefficient of the first unit and the second unit respectively to obtain the initial feedforward compensation amount of the first unit and the second unit. The initial feedforward compensation amount of the first unit and the second unit is corrected according to the feedforward static gain coefficient to obtain the corrected initial feedforward compensation amount of the first unit and the second unit. Based on the upper and lower limits of unit power operation in the safe operation boundary threshold, the current adjustable upper and lower limits of the first and second units are calculated. The initial feedforward compensation amount of the first and second units after correction is locked within the maximum allowable limit boundary, a feedforward compensation signal is generated, and a fixed time delay compensation value is set on the feedforward compensation signal according to the control channel time delay parameter after synchronous calibration, and the final target power control command is output.

10. A hierarchical collaborative control system for isolated grid operation of dual steam turbine generator sets of the same power rating, characterized in that, include: Initialize the module; Used to establish control benchmarks, complete unified calibration through synchronous calibration under the same operating conditions, set core control benchmarks such as frequency droop coefficient and load target allocation coefficient, and set safe operating boundaries; Synchronization control module; It is used to provide underlying frequency support for isolated grid systems, calculate the frequency deviation of the isolated grid common bus in real time, and synchronously send it to the two units through the equal time delay channel after synchronous calibration. At the same time, it completes the preliminary safety limit verification of basic commands based on safety boundary thresholds, and outputs the verified dual-machine synchronous basic power command increment and real-time system frequency deviation. Deviation Co-correction Module: It is used to calculate the real-time target power of the two machines and complete the weight correction of the power deviation through dynamic impedance matching logic. Then, it extracts and compensates for periodic power fluctuations through repetitive control to eliminate the dynamic deviation caused by disturbances. Finally, it generates the lead correction component and outputs the dual-machine power co-correction command after synchronization verification. Feedforward compensation module; It is used to collect the total active load of the isolated network system in real time and generate the initial feedforward compensation amount, and improve the compensation accuracy through static gain correction, and output the final feedforward compensation command.

Images