Generator set steady state regulation and control method and system based on load change response

Abstract

The invention relates to the technical field of generator set steady-state regulation and control, in particular to a generator set steady-state regulation and control method and system based on load change response, and the method comprises the steps: obtaining a frequency modulation value set, a carbon emission set and an optimal power distribution value set of a generator set, obtaining a comprehensive reward value and an actual output value set, and sequentially extracting actual output values from the actual output value set, calculating the total output value of the correction target, the maximum technical output, the minimum technical output and the maximum allowable climbing rate, if the total output value of the correction target is smaller than the minimum technical output or larger than the maximum technical output, correcting the total output value of the correction target to obtain a corrected total output value, and calculating the corrected total output value to obtain a corrected total output value; and calculating a subsequent period apportioned power value, regulating and controlling the jurisdiction generator set of the extracted actual output value to obtain a normal jurisdiction generator set, and summarizing the normal jurisdiction generator set to obtain a normal jurisdiction generator set. According to the invention, the optimal comprehensive benefit of the system can be realized while the rapid power regulation is ensured.

Description

Technical Field

[0001] This invention relates to the field of generator set steady-state control technology, and in particular to a generator set steady-state control method and system based on load change response. Background Technology

[0002] Load change response refers to the series of adjustments made by the generator set system during operation when the external load (i.e., the demand for electrical energy from electrical equipment, etc.) changes. A generator set is a device that converts other forms of energy (such as mechanical energy, thermal energy, hydropower, etc.) into electrical energy. Steady-state control refers to the use of various control methods and measures during generator set operation to maintain a relatively stable operating state under changing external conditions such as load variations.

[0003] Traditional automatic power generation control systems (AHPs) are centered around conventional thermal and hydropower units, using PID controllers to adjust power based on control deviation signals within the response range. However, PID parameters are fixed, making it difficult to dynamically adapt to changes in system characteristics. When faced with rapid random disturbances caused by new energy sources, they are prone to response lag and excessive overshoot, resulting in slow frequency recovery and insufficient control accuracy. Furthermore, these systems do not adequately consider the differences in response speed, regulation value, and carbon emission intensity among gas turbine, wind power, and photovoltaic units, failing to achieve a comprehensive optimization of regulation costs and carbon emissions. Therefore, the challenge lies in achieving optimal overall system efficiency while ensuring rapid power regulation. Summary of the Invention

[0004] This invention provides a generator set steady-state control method based on load change response and a computer-readable storage medium. Its main purpose is to ensure rapid power regulation while achieving optimal overall system benefits.

[0005] To achieve the above objectives, the present invention provides a generator set steady-state control method based on load change response, comprising: The power grid of the area to be regulated has been identified. The power grid of the area to be regulated includes: multiple intelligent agents and multiple generator sets under their jurisdiction. There is a one-to-one correspondence between the intelligent agents and the generator sets. The intelligent agents include an Actor network. The current control cycle is identified, and based on the current control cycle, multiple intelligent agents, and multiple subordinate generator sets, the frequency regulation value set, carbon emission set, and optimal power allocation value set of the generator sets are obtained. The comprehensive reward value and actual output value set are obtained based on the frequency regulation value set, carbon emission set and the optimal power allocation value set of the unit; The actual output values ​​are extracted sequentially from the set of actual output values, and the total output value of the correction target, the maximum technical output, the minimum technical output, and the maximum allowable climbing rate are calculated based on the extracted actual output values. If the total output value of the target is less than the minimum technical output or greater than the maximum technical output, the total output value of the target is corrected by using the minimum technical output or the maximum technical output to obtain the corrected total output value. The corrected total output value is then calculated based on the corrected total output value and the extracted actual output value. The power allocation value for subsequent cycles is calculated based on the corrected total output value and the maximum allowable climbing rate. The generator sets under the jurisdiction with the extracted actual output value are then adjusted according to the power allocation value for subsequent cycles to obtain the normal generator sets under the jurisdiction. The normal generator sets are summarized to obtain the normal generator set set. The actual output value set, comprehensive reward value and the optimal power allocation value set of the generator sets are stored to obtain the agent training set. Based on the training set of the intelligent agent and the set of normally managed generator sets, the steady-state control of generator sets based on load change response is completed.

[0006] Optionally, the step of obtaining the frequency regulation value set, carbon emission set, and optimal power allocation value set of the units based on the current control cycle, multiple intelligent agents, and multiple subordinate generator sets includes: Real-time monitoring of the power grid in the area to be regulated is performed based on the current control cycle to obtain tie line power deviation and system frequency deviation. The regional control deviation is calculated based on the tie line power deviation, system frequency deviation, and preset regional frequency deviation coefficient. The generator sets under their jurisdiction are extracted sequentially from multiple generator sets under their jurisdiction, and the target agent is identified from multiple agents based on the extracted generator sets under their jurisdiction. Based on the target intelligent agent and the regional control deviation, the value of the extracted generator sets under its jurisdiction is calculated to obtain the frequency regulation value and the optimal power allocation value of the generator sets. The carbon emissions are calculated by multiplying the preset carbon emission intensity coefficient of the unit with the optimal power allocation value of the unit. By summarizing the frequency regulation value, carbon emissions, and optimal power allocation value of the generating units, we can obtain the frequency regulation value set, carbon emission set, and optimal power allocation value set corresponding to multiple generating units under our jurisdiction.

[0007] Optionally, the step of real-time monitoring of the power grid in the area to be regulated according to the current control cycle to obtain tie-line power deviation and system frequency deviation includes: The key node set and inter-regional tie line set of the power grid in the area to be regulated are identified. Frequency is collected for each key node in the key node set according to the current control cycle to obtain the node frequency set. Abnormal frequency elimination operation is performed on the node frequency set to obtain the normal node frequency set. The node weight set is determined based on the key node set, where each node weight corresponds one-to-one with a key node. The system frequency deviation is calculated based on the node weight set and the normal node frequency set. The following operations are performed on each inter-regional tie line in the inter-regional tie line set: Obtain the line voltage and line current of the inter-regional tie lines, and calculate the active power of the tie lines based on the line voltage and line current; The difference between the active power of the tie line and the preset planned power of the tie line is calculated to obtain the power deviation of a single tie line; Summarize the power deviations of individual tie lines to obtain a set of power deviations for each tie line. Then sum the power deviation sets of individual tie lines to obtain the total power deviation of the tie line.

[0008] Optionally, the formula for calculating the system frequency deviation is as follows: in, Indicates the system frequency deviation. This indicates the number of critical nodes in the critical node set. Represents the node weight set of the 1st generation. The node weights of key nodes, Represents the first normal node frequency set Normal node frequency.

[0009] Optionally, the step of calculating the unit value of the extracted generator sets based on the target intelligent agent and regional control deviation to obtain the frequency regulation value and the optimal power allocation value of the units includes: Information is collected from the generator units under its jurisdiction to obtain the generator unit parameters of the power plant; The regional control deviation, system frequency deviation, and power plant unit parameters are input into the Actor network of the target intelligent agent to obtain the optimal power allocation value of the unit; Obtain the historical periodic power allocation value of the extracted generator units, calculate the absolute difference between the historical periodic power allocation value and the optimal power allocation value of the units, and obtain the frequency regulation mileage of the units. The frequency regulation value is calculated based on the preset mileage value, the preset unit regulation performance coefficient, and the unit frequency regulation mileage.

[0010] Optionally, obtaining the comprehensive reward value and actual output value set based on the frequency regulation value set, carbon emission set, and optimal power allocation value set of the unit includes: The carbon emission sets are summed to obtain the total carbon emissions. The total carbon emissions and the preset carbon emission penalty coefficient are then calculated to obtain the carbon emission penalty value. The stability reward value is calculated based on the regional control deviation and the system frequency deviation, and the frequency modulation contribution reward value is calculated based on the frequency modulation value set. The comprehensive reward value is obtained by summing the carbon emission penalty value, the stability reward value, and the frequency regulation contribution reward value. Based on the confirmation of the power plant control system by multiple generator units under its jurisdiction, the optimal power allocation value set of the units is sent to the power plant control system to obtain the actual output value set. The actual output value set includes multiple actual output values, and the actual output values ​​and the optimal power allocation value of the units correspond one-to-one with the generator units under its jurisdiction.

[0011] Optionally, the formula for calculating the stability reward value is as follows: in, This represents the stability reward value. This indicates the preset regional control deviation weight. Indicates regional control deviation. This represents the preset system frequency deviation weight. This indicates the system frequency deviation.

[0012] Optionally, the step of calculating the corrected target total output value based on the extracted actual output value includes: Based on the extracted actual output values, the historical actual output values ​​and the optimal power allocation values ​​for the target units are confirmed; The expected output value is obtained by summing the historical actual output value and the optimal power allocation value of the target unit. The output deviation is obtained by calculating the difference between the expected output value and the extracted actual output value. Obtain the power adjustment value for the next cycle, and calculate and correct the next power adjustment value based on the output deviation and the power adjustment value for the next cycle. The total output value of the correction target is calculated based on the corrected next power adjustment value and the extracted actual output value.

[0013] Optionally, the step of calculating the power distribution value for subsequent cycles based on the corrected total output value and the maximum permissible climbing rate includes: The average ramp rate is calculated based on the corrected total output value and the current control cycle. If the average ramp rate is greater than the maximum allowable ramp rate, the maximum output change value is calculated based on the maximum allowable ramp rate and the current control cycle. The current cycle limit value is calculated based on the maximum output change value and the next power adjustment value. The remaining uncorrected value is calculated based on the current cycle limit value and the next power adjustment value. The remaining uncorrected value is the value obtained by subtracting the current cycle limit value from the next power adjustment value. Calculate the minimum number of cycles required based on the remaining uncorrected values ​​and the maximum output change value; Calculate the power allocation value for subsequent cycles based on the remaining uncorrected values ​​and the minimum number of cycles required.

[0014] To achieve the above objectives, the present invention also provides a generator set steady-state control system based on load change response, comprising: The assigned generator set confirmation module is used to confirm the power grid in the area to be regulated. The power grid in the area to be regulated includes multiple intelligent agents and multiple assigned generator sets. Each intelligent agent corresponds to one generator set. The intelligent agents include an Actor network. The output value calculation module is used to determine the current control cycle, and based on the current control cycle, multiple intelligent agents and multiple subordinate generator sets, obtain the frequency regulation value set, carbon emission set and the optimal power allocation value set of the generator sets, and based on the frequency regulation value set, carbon emission set and the optimal power allocation value set of the generator sets, obtain the comprehensive reward value and the actual output value set. The generator set control module is used to sequentially extract actual output values ​​from the actual output value set, calculate the correction target total output value, maximum technical output, minimum technical output, and maximum allowable ramp rate based on the extracted actual output values, if the correction target total output value is less than the minimum technical output or greater than the maximum technical output, then the correction target total output value is corrected using the minimum or maximum technical output to obtain the corrected total output value, calculate the corrected total output value based on the corrected total output value and the extracted actual output values, calculate the subsequent cycle allocated power value based on the corrected total output value and the maximum allowable ramp rate, and control the generator sets under the extracted actual output values ​​based on the subsequent cycle allocated power value to obtain the normal managed generator sets; The generator set control module is used to summarize the normally managed generator sets to obtain the normally managed generator set set. It stores the actual output value set, comprehensive reward value and the optimal power allocation value set of the generator set to obtain the agent training set. Based on the agent training set and the normally managed generator set set, it completes the steady-state control of the generator set based on the load change response.

[0015] To address the above problems, the present invention also provides an electronic device, the electronic device comprising: Memory, storing at least one instruction; The processor executes the instructions stored in the memory to implement the generator set steady-state control method based on load change response described above.

[0016] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the aforementioned generator set steady-state control method based on load change response.

[0017] To address the problems described in the background, this invention identifies the power grid area to be regulated, which includes multiple intelligent agents and multiple subordinate generator units. Each intelligent agent corresponds one-to-one with a generator unit. Each intelligent agent comprises an Actor network. This invention achieves precise and independent regulation of individual generator units through this one-to-one correspondence between agents and generator units. The Actor network, as the core of reinforcement learning, provides algorithmic support for generating the optimal power allocation scheme, avoiding the delay and insufficient accuracy problems of traditional centralized regulation. The current control cycle is identified, and based on the current control cycle, multiple intelligent agents, and multiple subordinate generator units, the frequency regulation value set, carbon emission set, and the optimal power allocation set for each unit are obtained. This invention utilizes a power allocation value set, dividing the control phases into time-divided phases based on the control cycle to ensure the timeliness and periodicity of control. Based on the frequency regulation value set, carbon emission set, and the optimal power allocation value set of the unit, it obtains a comprehensive reward value and an actual output value set. The comprehensive reward value enables a quantitative evaluation of the power allocation scheme in terms of low carbon emissions, stability, and frequency regulation, providing a unified standard for judging the merits of the scheme. The actual output value set records the actual operating status of the unit, providing feedback on the decision-making execution effect. Actual output values ​​are extracted sequentially from the actual output value set, and based on the extracted actual output values, the correction target total output value, maximum technical output, minimum technical output, and maximum allowable ramp rate are calculated. This invention uses a single actual output value set to... The extraction and correction of force values ​​enables precise control of the unit. The corrected total output value drives the unit's operating state closer to the optimal solution, improving the operating accuracy of a single unit. If the corrected total output value is less than the minimum technical output or greater than the maximum technical output, the minimum or maximum technical output is used to correct the corrected total output value, resulting in a corrected total output value. The corrected total output value is then calculated based on the corrected total output value and the extracted actual output value. This invention, through the constraints of minimum and maximum technical output, prevents the correction target from exceeding the unit's safe operating range, preventing equipment failure or downtime risks. Subsequent cycle intervals are calculated based on the corrected total output value and the maximum allowable ramp rate. The power allocation value is used to adjust the generator sets under its jurisdiction based on the extracted actual output value in subsequent cycles, thus obtaining the normally managed generator sets. This invention solves the problem of generator ramping capability constraints by allocating power in cycles, avoiding overload in single-cycle regulation. The normally managed generator sets are then aggregated to obtain a set of normally managed generator sets. The actual output value set, comprehensive reward value, and optimal power allocation value set of the generator sets are stored to obtain the intelligent agent training set. This invention provides samples for training the intelligent agent model, promoting the model to continuously optimize its decision-making ability through learning, and achieving iterative improvement in regulation effect. Based on the intelligent agent training set and the normally managed generator set set, steady-state regulation of generator sets based on load change response is completed. Therefore, this invention can achieve optimal overall system benefits while ensuring rapid power regulation. Attached Figure Description

[0018] Figure 1This is a flowchart illustrating a generator set steady-state control method based on load change response according to an embodiment of the present invention. Figure 2 A functional block diagram of a generator set steady-state control system based on load change response provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device for implementing the generator set steady-state control method based on load change response, according to an embodiment of the present invention.

[0019] Explanation of reference numerals in the attached figures: 10. Electronic device; 11. Processor; 12. Memory; 13. Bus.

[0020] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0022] This application provides a generator set steady-state control method based on load change response. The executing entity of the generator set steady-state control method based on load change response includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application embodiment: a server, a terminal, etc. In other words, the generator set steady-state control method based on load change response can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0023] Reference Figure 1 The diagram shown is a flowchart illustrating a generator set steady-state control method based on load change response according to an embodiment of the present invention. In this embodiment, the generator set steady-state control method based on load change response includes: S1. Identify the power grid in the area to be regulated, which includes multiple intelligent agents and multiple generator sets under their jurisdiction. Each intelligent agent corresponds to a generator set. The intelligent agents include an Actor network.

[0024] It should be explained that the power grid of the area to be regulated is a sub-grid selected as the target in a large-scale power system with interconnected multiple regions, requiring real-time power generation adjustment. This sub-grid exchanges power with the external power grid through tie lines. The intelligent agent is an intelligent control unit with autonomous perception and decision-making capabilities. Based on real-time acquired global and local system state information, this intelligent control unit can independently calculate the optimal power adjustment recommendation value (optimal power allocation value for the generator set) for the corresponding generator set within the current control cycle. The managed generator sets are all generator sets that can participate in power adjustment within the dispatching jurisdiction of the power grid of the area to be regulated. A generator set is a physical power generation device that completes the conversion of electrical energy. For example, managed generator sets include, but are not limited to: gas turbines, hydroelectric generator sets, coal-fired generator sets, wind turbine generator sets, and photovoltaic power generation units. In this invention, each generator set receives a power adjustment command issued by its corresponding intelligent agent and executes output adjustment through the generator set's local control system, thereby converting intelligent decisions into actual physical power output and ultimately achieving power balance. The Actor network is a neural network model trained offline and fine-tuned online using deep reinforcement learning.

[0025] S2. Confirm the current control cycle, and based on the current control cycle and multiple subordinate generator sets, obtain the frequency regulation value set, carbon emission set, and optimal power allocation value set for the generator sets.

[0026] In detail, the acquisition of frequency regulation value set, carbon emission set, and optimal power allocation value set of units based on the current control cycle, multiple intelligent agents, and multiple subordinate generator units includes: Real-time monitoring of the power grid in the area to be regulated is performed based on the current control cycle to obtain tie line power deviation and system frequency deviation. The regional control deviation is calculated based on the tie line power deviation, system frequency deviation, and preset regional frequency deviation coefficient. The generator sets under their jurisdiction are extracted sequentially from multiple generator sets under their jurisdiction, and the target agent is identified from multiple agents based on the extracted generator sets under their jurisdiction. Based on the target intelligent agent and the regional control deviation, the value of the extracted generator sets under its jurisdiction is calculated to obtain the frequency regulation value and the optimal power allocation value of the generator sets. The carbon emissions are calculated by multiplying the preset carbon emission intensity coefficient of the unit with the optimal power allocation value of the unit. By summarizing the frequency regulation value, carbon emissions, and optimal power allocation value of the generating units, we can obtain the frequency regulation value set, carbon emission set, and optimal power allocation value set corresponding to multiple generating units under our jurisdiction.

[0027] It should be explained that the current control cycle is a fixed and continuous time interval defined to achieve dynamic and precise control of the power grid operation status in the area to be regulated. The regional frequency deviation coefficient is a pre-set constant reflecting the load frequency characteristics of the power grid in the area to be regulated. For example, the regional frequency deviation coefficient is 840MW / 0.1Hz. The calculation formula for the regional control deviation in the step of calculating the regional control deviation based on the tie-line power deviation, system frequency deviation, and the preset regional frequency deviation coefficient is as follows: in, Indicates regional control deviation. Indicates the power deviation of the tie line. Indicates the system frequency deviation. This represents the regional frequency deviation coefficient. The target agent is the unique agent corresponding to the extracted generator set. The generator set carbon emission intensity coefficient is a coefficient characterizing the carbon dioxide emissions generated per unit of power generated by the extracted generator set. Carbon emissions are the product of the generator set carbon emission intensity coefficient and the optimal power allocation value. Carbon emissions represent the expected carbon dioxide emissions from a single generator set due to the execution of the optimal power allocation value. The optimal power allocation value is the power regulation amount recommended for the generator set corresponding to the target agent within the current control cycle, calculated and output by the target agent. The frequency regulation value set is the set of all frequency regulation values. The carbon emission set is the set of all carbon emissions. The optimal power allocation value set is the set of all optimal power allocation values ​​for all generator sets.

[0028] It should be noted that in the above steps of this invention, firstly, by real-time monitoring of the tie-line power deviation and system frequency deviation of the power grid in the area to be regulated within the current control cycle, the frequency regulation demand and operational stability of the power grid in the area to be regulated are determined. Then, the regional control deviation is calculated by combining the regional frequency deviation coefficient. The regional control deviation is an indicator for measuring whether the power grid in the area to be regulated needs frequency regulation. By integrating the tie-line power deviation, system frequency deviation, and regional frequency deviation coefficient, the regulation demand of the power grid in the area to be regulated can be more comprehensively reflected, avoiding the limitations of judging by a single indicator. Afterwards, each subordinate generator unit is extracted sequentially, and the corresponding target agent is identified from multiple agents. The target agent is then used to target... It provides adaptive calculation logic for the characteristics of different generator sets, ensuring the relevance and rationality of subsequent value calculation and power allocation. This avoids the insufficient adaptability of a unified calculation model to different generator sets under its jurisdiction. By combining the adaptability of the target intelligent agent with the needs of regional control deviation, the calculated frequency regulation value can truly reflect the frequency regulation contribution of the generator sets under its jurisdiction. At the same time, it ensures that the optimal power allocation value of the generator sets under its jurisdiction meets the frequency regulation requirements of the power grid and the operating characteristics of the units, thereby achieving optimal resource allocation. Subsequently, by combining the carbon emission intensity coefficient with the optimal power allocation value, the carbon emissions of the units at that power level can be accurately calculated, meeting the current needs of low-carbon regulation and ensuring the scientific and comprehensive nature of the regulation strategy.

[0029] In detail, the real-time monitoring of the power grid in the area to be regulated according to the current control cycle to obtain the tie-line power deviation and system frequency deviation includes: The key node set and inter-regional tie line set of the power grid in the area to be regulated are identified. Frequency is collected for each key node in the key node set according to the current control cycle to obtain the node frequency set. Abnormal frequency elimination operation is performed on the node frequency set to obtain the normal node frequency set. The node weight set is determined based on the key node set, where each node weight corresponds one-to-one with a key node. The system frequency deviation is calculated based on the node weight set and the normal node frequency set. The following operations are performed on each inter-regional tie line in the inter-regional tie line set: Obtain the line voltage and line current of the inter-regional tie lines, and calculate the active power of the tie lines based on the line voltage and line current; The difference between the active power of the tie line and the preset planned power of the tie line is calculated to obtain the power deviation of a single tie line; Summarize the power deviations of individual tie lines to obtain a set of power deviations for each tie line. Then sum the power deviation sets of individual tie lines to obtain the total power deviation of the tie line.

[0030] It should be explained that the critical node set refers to the collection of all critical nodes. Critical nodes include, but are not limited to: electrical centers, hub substations, grid connection points of major power plants, and important load centers in the power grid to be regulated. The inter-regional tie line set is the collection of all inter-regional tie lines. Inter-regional tie lines are transmission lines connecting the power grid to be regulated to adjacent external power grids. The node frequency set is the collection of all node frequencies, where the node frequency is the frequency measurement value of the critical node. The steps for abnormal frequency removal from the node frequency set are: calculating the average and standard deviation of all node frequencies, and removing node frequencies that deviate from the average by more than a preset threshold (e.g., 3 times the standard deviation). The normal node frequency set is the collection of valid node frequencies after the removal operation. The node weight set is the collection of all node weights. Node weight is the frequency measurement value of a critical node, indicating its importance in calculating the overall system frequency. The node weights described in this invention can be set based on factors such as the electrical centrality of the node, the capacity of the connected generator or load, and the accuracy level of the measuring equipment. The calculation steps for system frequency deviation will be given later. The weighted average method is existing technology and will not be elaborated further here. The formula for calculating the active power of the tie line is as follows: in, Indicates the active power of the tie line. Indicates line voltage. Indicates line current. This represents the preset power factor. Represents the cosine function. This indicates the phase angle.

[0031] Understandably, line voltage and line current are the voltage values ​​between two phase lines (live wires) and the current values ​​flowing through a single phase line, respectively, in a three-phase AC transmission system. The power factor is the cosine of the phase angle between voltage and current. It should be noted that the power factor ranges from 0 to 1. The planned power of a tie line is a pre-set power value that should be transmitted through a certain inter-regional tie line within a certain time period. The power deviation of a single tie line is the value obtained by subtracting the planned power of the tie line from the active power of the tie line. If the power deviation of a single tie line is negative, it indicates that the total active power actually flowing into the grid of the region to be regulated from the external grid (through the inter-regional tie line) exceeds the planned power of the tie line, and the grid of the region to be regulated may be under-generating. If the power deviation of a single tie line is positive, it indicates that the actual power transmitted by the inter-regional tie line exceeds the planned power of the tie line, and the grid of the region to be regulated may be over-generating. The set of power deviations for a single tie line is the set of all power deviations for a single tie line. The tie line power deviation is the sum of the sets of power deviations for a single tie line.

[0032] It should be noted that the frequency of key nodes can directly reflect the local operating status of the power grid, and the inter-regional tie lines are the core channels for power exchange in the regional power grid. The monitoring data of both are the core basis for calculating the system frequency deviation and tie line power deviation. Therefore, this invention first identifies the set of key nodes and the set of inter-regional tie lines in the power grid to be regulated. Abnormal frequencies are usually caused by equipment failure or transient interference. If retained, they will affect the accuracy of the system frequency deviation calculation results. Therefore, it is necessary to perform an abnormal frequency removal operation on the collected node frequency set. Since different key nodes have different locations and load proportions in the power grid to be regulated, their impact on the system frequency also varies. Therefore, by setting node weights that correspond one-to-one with key nodes, the subsequent calculation of system frequency deviation can be made to better reflect the actual operation of the power grid and improve the rationality of the calculation results. On this basis, the system frequency deviation is calculated by combining the node weight set and the normal node frequency set. The normal frequencies of each key node can be integrated by weighting, comprehensively reflecting the frequency deviation of the entire power grid to be regulated, and providing a reliable basis for judging the frequency regulation needs of the power grid.

[0033] In detail, the formula for calculating the system frequency deviation is as follows: in, Indicates the system frequency deviation. This indicates the number of critical nodes in the critical node set. Represents the node weight set of the 1st generation. The node weights of key nodes, Represents the first normal node frequency set Normal node frequency.

[0034] It should be explained that the formula for calculating the system frequency deviation uses node weights to weight and integrate the normal node frequencies of each key node, thus more accurately reflecting the overall frequency level of the power grid in the area to be regulated. In the formula, This represents the sum of the products of the node weights of all critical nodes and the frequencies of their corresponding normal nodes. The sum of the node weights of all key nodes is represented by the sum of the node weights. The system frequency deviation is obtained by dividing the sum of the weights and the weights. This calculation method fully considers the impact of the differences in the location and load share of different key nodes in the power grid on the system frequency. It avoids the limitations of simple arithmetic averages that ignore the differences in node importance, and makes the calculation results more consistent with the actual operation of the power grid.

[0035] In detail, the step of calculating the unit value of the extracted generator sets based on the target intelligent agent and regional control deviation to obtain the frequency regulation value and the optimal power allocation value of the units includes: Information is collected from the generator units under its jurisdiction to obtain the generator unit parameters of the power plant; The regional control deviation, system frequency deviation, and power plant unit parameters are input into the Actor network of the target intelligent agent to obtain the optimal power allocation value of the unit; Obtain the historical periodic power allocation value of the extracted generator units, calculate the absolute difference between the historical periodic power allocation value and the optimal power allocation value of the units, and obtain the frequency regulation mileage of the units. The frequency regulation value is calculated based on the preset mileage value, the preset unit regulation performance coefficient, and the unit frequency regulation mileage.

[0036] It should be explained that the power plant unit parameters include: current actual output, upper limit of available adjustable capacity, lower limit of available adjustable capacity, ramp rate, and response time delay. Current actual output is the extracted electrical power actually output by the managed generator units within the current control cycle, reflecting the current operating load status of the managed generator units. The upper limit of available adjustable capacity is the maximum output power that the managed generator units can achieve under safe and stable operation. The lower limit of available adjustable capacity is the minimum output power that the managed generator units can achieve without triggering the minimum operating limits. The upper and lower limits of available adjustable capacity together define the power regulation range of the managed generator units when participating in frequency regulation. Ramp rate is the magnitude of electrical power increase or decrease that the managed generator units can achieve per unit time, reflecting the rate capability of power regulation of the managed generator units. Response time delay is the time interval from receiving the frequency regulation command to starting to adjust the output power of the managed generator units. The optimal power allocation value is the optimal output power that the managed generator units should achieve within the current control cycle after inputting the regional control deviation, system frequency deviation, and power plant unit parameters into the Actor network of the target agent. The historical cycle power allocation value is the power allocation value of the generator units under its jurisdiction in the previous control cycle. The unit frequency regulation mileage is a value calculated by the absolute difference between the historical cycle power allocation value and the optimal power allocation value of the unit, used to visually reflect the power regulation range of the generator units under its jurisdiction in adjacent control cycles. Mileage value is a pre-set value coefficient corresponding to a unit frequency regulation mileage. For example, each 1MW power regulation corresponds to a value of 20. The unit regulation performance coefficient is a pre-set coefficient reflecting the regulation performance (such as regulation accuracy, stability, etc.) of the generator units under its jurisdiction, used to distinguish the differences in frequency regulation value between generator units with different performance characteristics. For example, the unit regulation performance coefficient is 0.8.

[0037] It should be noted that the Actor network, as a core component of the reinforcement learning model, possesses the ability to process multi-dimensional input data and output optimal decisions. Therefore, this invention inputs grid regulation deviations and power plant unit parameters into this network, enabling the output power allocation value to meet both the grid frequency regulation requirements and the operational limitations of the assigned generator units, thereby achieving a precise match between regulation requirements and unit capabilities. Simultaneously, mileage value serves as the benchmark for the value of a unit's regulation amplitude, and the unit regulation performance coefficient can distinguish the differences in regulation quality among different assigned generator units. This invention combines these two factors with the unit's frequency regulation mileage, enabling a comprehensive quantification of the frequency regulation contribution of the assigned generator units.

[0038] S3. Obtain the comprehensive reward value and actual output value set based on the frequency regulation value set, carbon emission set and the optimal power allocation value set of the unit.

[0039] In detail, the process of obtaining the comprehensive reward value and the actual output value set based on the frequency regulation value set, the carbon emission set, and the optimal power allocation value set of the unit includes: The carbon emission sets are summed to obtain the total carbon emissions. The total carbon emissions and the preset carbon emission penalty coefficient are then calculated to obtain the carbon emission penalty value. The stability reward value is calculated based on the regional control deviation and the system frequency deviation, and the frequency modulation contribution reward value is calculated based on the frequency modulation value set. The comprehensive reward value is obtained by summing the carbon emission penalty value, the stability reward value, and the frequency regulation contribution reward value. Based on the confirmation of the power plant control system by multiple generator units under its jurisdiction, the optimal power allocation value set of the units is sent to the power plant control system to obtain the actual output value set. The actual output value set includes multiple actual output values, and the actual output values ​​and the optimal power allocation value of the units correspond one-to-one with the generator units under its jurisdiction.

[0040] It should be explained that the total carbon emissions are the expected total carbon dioxide emissions from the power grid in the controlled area, calculated within a control cycle based on the optimal power allocation values ​​of multiple managed generator units and their respective carbon emission intensity coefficients. The carbon emission penalty coefficient is a pre-set value used to multiply the total carbon emissions to calculate the carbon emission penalty value. For example, a penalty of 0.06 values ​​corresponds to the generation of 1 ton of carbon dioxide per 1 kWh of electricity. The carbon emission penalty value is the product of the total carbon emissions and the carbon emission penalty coefficient, used to quantify the negative impact of total carbon emissions. The stability reward value reflects the contribution of the generator unit power allocation scheme to the stability of the power grid in the controlled area within the current control cycle. The smaller the regional control deviation and system frequency deviation, the larger the stability reward value. The frequency regulation contribution reward value is obtained by summing the frequency regulation value set and then taking the negative of the sum. The comprehensive reward value is the sum of the carbon emission penalty value, the stability reward value, and the frequency regulation contribution reward value. The power plant control system is a system used to receive and execute power allocation commands based on confirmation from multiple managed generator units. The actual output value set is the set of power values ​​actually executed and output by each generator unit after the optimal power allocation value set of the generating units is sent to the power plant control system.

[0041] Importantly, this invention incorporates a carbon emission penalty coefficient to quantify the negative impact of carbon emissions, introducing low-carbon constraints for comprehensive evaluation and avoiding the application of high-carbon emission schemes. Next, it calculates stability reward values ​​and frequency regulation contribution reward values. Regional control deviation and system frequency deviation are core indicators for measuring grid operation stability. The stability reward value calculated accordingly quantifies the contribution of the power allocation scheme to grid stability, while the frequency regulation contribution reward value reflects the value of generator units participating in frequency regulation. Both provide evaluation criteria from the perspectives of grid stability and generator unit frequency regulation contribution, respectively. Subsequently, the carbon emission penalty value, stability reward value, and frequency regulation contribution reward value are summed to obtain a comprehensive reward value. By integrating negative penalties and positive rewards, the comprehensive performance of the power allocation scheme in terms of low carbon emissions, stability, and frequency regulation can be comprehensively evaluated. Finally, based on the power plant control system of multiple managed generator units, the optimal power allocation value set of the units is sent to the power plant control system to obtain the actual output value set. Since the power plant control system is the key link between decision-making and execution, sending the optimal power allocation value to this system ensures that the decision is implemented, enabling the managed generator units to operate according to the optimal scheme, thereby achieving the control objective of balancing low carbon emissions, stability, and frequency regulation effects.

[0042] In detail, the formula for calculating the stability reward value is as follows: in, This represents the stability reward value. This indicates the preset regional control deviation weight. Indicates regional control deviation. This represents the preset system frequency deviation weight. This indicates the system frequency deviation.

[0043] Understandably, the system frequency deviation weight is a pre-set value used to differentiate the degree of impact of system frequency deviation on grid stability. The regional control deviation weight is a pre-set value used to differentiate the degree of impact of regional control deviation on grid stability. It should be noted that the sum of the system frequency deviation weight and the regional control deviation weight is 1. For example, if the regional grid has extremely high requirements for frequency stability (such as including a large number of frequency-sensitive precision manufacturing enterprises and data centers), then the system frequency deviation weight is set to 0.6 to strengthen the impact of system frequency deviation on the stability bonus value, resulting in a system frequency deviation of 0.4.

[0044] It should be explained that the formula for calculating the stability reward value in this invention calculates the stability reward value by taking the weighted sum of the squares of the regional control deviation and the system frequency deviation, and then inverting the sum. The weights for the regional control deviation and the system frequency deviation are used to distinguish the degree of influence of these deviations on grid stability. Taking the squares of the regional control deviation and the system frequency deviation in the formula amplifies the negative impact of larger deviations on the stability reward value, while avoiding the problem of positive and negative deviations canceling each other out. Taking the inversion after the weighted sum ensures that the smaller the deviation (regional control deviation or system frequency deviation) (indicating a more stable grid in the controlled area), the larger the stability reward value; conversely, the larger the deviation, the smaller the stability reward value. The purpose of introducing this formula in this invention is to quantify the contribution of power allocation schemes to grid operation stability. By converting the two indicators, regional control deviation and system frequency deviation, into calculable reward values, it provides an assessment basis for the grid stability dimension of the comprehensive reward value calculation, thereby enabling the assessment of the stability of power allocation schemes and avoiding the problems of relying on experience-based judgments and the inability to accurately quantify results in traditional assessments.

[0045] S4. Extract the actual output values ​​from the actual output value set in sequence, and calculate the total output value of the correction target, the maximum technical output, the minimum technical output, and the maximum allowable climbing rate based on the extracted actual output values.

[0046] Specifically, the calculation of the corrected target total output value based on the extracted actual output value includes: Based on the extracted actual output values, the historical actual output values ​​and the optimal power allocation values ​​for the target units are confirmed; The expected output value is obtained by summing the historical actual output value and the optimal power allocation value of the target unit. The output deviation is obtained by calculating the difference between the expected output value and the extracted actual output value. Obtain the power adjustment value for the next cycle, and calculate and correct the next power adjustment value based on the output deviation and the power adjustment value for the next cycle. The total output value of the correction target is calculated based on the corrected next power adjustment value and the extracted actual output value.

[0047] It should be noted that the maximum technical output and minimum technical output are the upper and lower limits of the output of the generator units under its jurisdiction, respectively, which are physically capable of safe and stable operation. The maximum allowable ramp rate is the maximum rate at which the output of the generator units under its jurisdiction can change per unit time (e.g., per minute), determining the shortest time required to change from the current output to the target output. The historical actual output value is the actual output value of the generator units under its jurisdiction measured at the end of the previous control cycle (cycle k-1). The target unit's optimal power allocation value is the amount of power regulation expected to be completed by the generator units under its jurisdiction in the current control cycle (cycle k), calculated and issued by the agent (Actor network). The expected output value is the sum of the historical actual output value and the target unit's optimal power allocation value. The output deviation is the value obtained by subtracting the extracted actual output value from the expected output value. The next cycle's power regulation value is the power regulation value initially generated by the upper-level Model Predictive Controller (MPC) for the next cycle (cycle k+1). The corrected next power regulation value is the sum of the output deviation and the next cycle's power regulation value. The target total output value is the sum of the next power adjustment value and the extracted actual output value.

[0048] For example, if the output deviation is -2MW and the power adjustment value for the next cycle is +5MW, then the corrected next power adjustment value = 5MW + (-2MW) = 3MW. The current actual output value is 108MW, and the corrected target total output value = 108MW + 3MW = 111MW.

[0049] Importantly, this invention combines output deviation calculation to correct the next power adjustment value. The next cycle power adjustment value is the initial adjustment plan. Introducing output deviation allows for the correction of the initial adjustment plan based on the current actual gap, making the adjustment demand more in line with the actual situation. Finally, the correction target total output value is calculated based on the corrected next power adjustment value and the extracted actual output value. By combining the corrected adjustment demand with the current actual output, a correction target total output value that conforms to both the adjustment plan and the current state of the managed generator units can be obtained. This ensures the feasibility of the correction scheme, promotes the adjustment of the managed generator units towards the optimal operating state, and improves the overall control effect.

[0050] S5. If the total output value of the target is less than the minimum technical output or greater than the maximum technical output, the total output value of the target is corrected by using the minimum technical output or the maximum technical output to obtain the corrected total output value. The corrected total output value is calculated based on the corrected total output value and the extracted actual output value.

[0051] It should be explained that if the total output value of the correction target is less than the minimum technical output, it means that the total output value of the correction target is lower than the minimum output power that the generator set under the jurisdiction can maintain under the premise of safe and stable operation. If it is forced to operate at this value, the generator set may experience problems such as unstable combustion and excessive turbine vibration, or even trigger the protection device to shut down, affecting the reliability of power grid supply. If the total output value of the correction target is greater than the maximum technical output, it means that the total output value of the correction target exceeds the maximum output power that the generator set under the jurisdiction can achieve under the constraints of equipment structure and thermal efficiency. If it is forced to operate at this value, the generator set may experience malfunctions such as accelerated equipment wear and abnormal temperature rise due to overload, which will also threaten the stable operation of the generator set and the power grid of the area to be regulated. Since the total output value of the correction target exceeds the actual operating capacity range of the generator set, it cannot be directly executed. Therefore, this invention needs to use the minimum or maximum technical output to correct the total output value of the correction target, so that the corrected target value falls within the power range for safe operation of the generator set under the jurisdiction, ensuring the safety of the generator set equipment and meeting the regulation requirements as closely as possible. The step of correcting the total output value of the calibration target using the minimum or maximum technical output is an operation that replaces the total output value of the calibration target with the minimum or maximum technical output. The corrected total output value is obtained by subtracting the extracted actual output value from the total output value of the calibration target.

[0052] S6. Calculate the power allocation value for subsequent cycles based on the corrected total output value and the maximum allowable climbing rate. Adjust the generator sets under the jurisdiction based on the extracted actual output value to obtain the normal generator sets under jurisdiction.

[0053] Specifically, the calculation of the power distribution value for subsequent cycles based on the corrected total output value and the maximum permissible climbing rate includes: The average ramp rate is calculated based on the corrected total output value and the current control cycle. If the average ramp rate is greater than the maximum allowable ramp rate, the maximum output change value is calculated based on the maximum allowable ramp rate and the current control cycle. The current cycle limit value is calculated based on the maximum output change value and the next power adjustment value. The remaining uncorrected value is calculated based on the current cycle limit value and the next power adjustment value. The remaining uncorrected value is the value obtained by subtracting the current cycle limit value from the next power adjustment value. Calculate the minimum number of cycles required based on the remaining uncorrected values ​​and the maximum output change value; Calculate the power allocation value for subsequent cycles based on the remaining uncorrected values ​​and the minimum number of cycles required.

[0054] It should be explained that the average ramp rate is obtained by dividing the total corrected output value by the current control cycle. If the average ramp rate is greater than the maximum permissible ramp rate, it means that the adjustment requirement of the entire corrected total output is to be completed in one go within the current control cycle, exceeding the upper limit of the power adjustment speed that the generator unit under its jurisdiction can stably complete in a single cycle. If the average ramp rate is not greater than the maximum permissible ramp rate, it means that the adjustment requirement of the entire corrected total output can be completed in one go. The maximum output change value is the product of the maximum permissible ramp rate and the current control cycle. The current cycle limit value is the minimum value taken from the maximum output change value and the next power adjustment value. This is because if the next power adjustment value is less than or equal to the maximum output change value, it means that the adjustment can be completed in one go within the current cycle; if the next power adjustment value is greater than the maximum output change value, it means that the adjustment can only be carried out according to the maximum output change value within the current cycle, and the remaining part needs to be distributed to subsequent cycles. The formula for calculating the minimum number of cycles required in the step of calculating the minimum number of cycles required based on the remaining uncompleted correction value and the maximum output change value is as follows: in, Indicates the minimum number of cycles required. This indicates the remaining uncorrected values. This represents the maximum output change value. The symbol indicates rounding up.

[0055] Understandably, the power allocation value for subsequent cycles is obtained by dividing the remaining uncompleted correction value by the minimum number of cycles required. This method of calculating the power allocation value for subsequent cycles ensures that the remaining adjustment demand is completed evenly within the minimum number of cycles, avoiding situations where the adjustment amount in a particular cycle is too large or too small, thus guaranteeing the stability of the unit's adjustment process.

[0056] S7. Summarize the normally managed generator sets to obtain the normally managed generator set set. Store the actual output value set, comprehensive reward value and the optimal power allocation value set of the generator sets to obtain the agent training set.

[0057] It should be explained that the normally managed generator set set is a collection of normally managed generator sets. The agent training set is a dataset formed by integrating data such as the actual output value set, the comprehensive reward value, and the optimal power allocation value set of the units. The agent training set provides samples for training agent models (such as reinforcement learning models), enabling the agent model to continuously optimize its decision-making capabilities through learning, thereby achieving continuous optimization of the control effect.

[0058] S8. Based on the training set of the intelligent agent and the set of normally managed generator sets, complete the steady-state control of generator sets based on load change response.

[0059] It should be explained that this invention combines the trained and optimized intelligent agent with samples of safely operating generating units to achieve rapid response and steady-state regulation to load changes, ensuring that the regional power grid can still maintain multiple objectives such as frequency stability, low carbon emissions, and reliable power supply when the load fluctuates, thereby improving the overall operational stability of the power grid.

[0060] To address the problems described in the background, this invention identifies the power grid area to be regulated, which includes multiple intelligent agents and multiple subordinate generator units. Each intelligent agent corresponds one-to-one with a generator unit. Each intelligent agent comprises an Actor network. This invention achieves precise and independent regulation of individual generator units through this one-to-one correspondence between agents and generator units. The Actor network, as the core of reinforcement learning, provides algorithmic support for generating the optimal power allocation scheme, avoiding the delay and insufficient accuracy problems of traditional centralized regulation. The current control cycle is identified, and based on the current control cycle, multiple intelligent agents, and multiple subordinate generator units, the frequency regulation value set, carbon emission set, and the optimal power allocation set for each unit are obtained. This invention utilizes a power allocation value set, dividing the control phases into time-divided phases based on the control cycle to ensure the timeliness and periodicity of control. Based on the frequency regulation value set, carbon emission set, and the optimal power allocation value set of the unit, it obtains a comprehensive reward value and an actual output value set. The comprehensive reward value enables a quantitative evaluation of the power allocation scheme in terms of low carbon emissions, stability, and frequency regulation, providing a unified standard for judging the merits of the scheme. The actual output value set records the actual operating status of the unit, providing feedback on the decision-making execution effect. Actual output values ​​are extracted sequentially from the actual output value set, and based on the extracted actual output values, the correction target total output value, maximum technical output, minimum technical output, and maximum allowable ramp rate are calculated. This invention uses a single actual output value set to... The extraction and correction of force values ​​enables precise control of the unit. The corrected total output value drives the unit's operating state closer to the optimal solution, improving the operating accuracy of a single unit. If the corrected total output value is less than the minimum technical output or greater than the maximum technical output, the minimum or maximum technical output is used to correct the corrected total output value, resulting in a corrected total output value. The corrected total output value is then calculated based on the corrected total output value and the extracted actual output value. This invention, through the constraints of minimum and maximum technical output, prevents the correction target from exceeding the unit's safe operating range, preventing equipment failure or downtime risks. Subsequent cycle intervals are calculated based on the corrected total output value and the maximum allowable ramp rate. The power allocation value is used to adjust the generator sets under its jurisdiction based on the extracted actual output value in subsequent cycles, thus obtaining the normally managed generator sets. This invention solves the problem of generator ramping capability constraints by allocating power in cycles, avoiding overload in single-cycle regulation. The normally managed generator sets are then aggregated to obtain a set of normally managed generator sets. The actual output value set, comprehensive reward value, and optimal power allocation value set of the generator sets are stored to obtain the intelligent agent training set. This invention provides samples for training the intelligent agent model, promoting the model to continuously optimize its decision-making ability through learning, and achieving iterative improvement in regulation effect. Based on the intelligent agent training set and the normally managed generator set set, steady-state regulation of generator sets based on load change response is completed. Therefore, this invention can achieve optimal overall system benefits while ensuring rapid power regulation.

[0061] like Figure 2The diagram shown is a functional block diagram of a generator set steady-state control system based on load change response provided in an embodiment of the present invention.

[0062] The generator set steady-state control system 100 based on load change response described in this invention can be installed in an electronic device. Depending on the functions implemented, the generator set steady-state control system 100 may include a generator set confirmation module 101, an output value calculation module 102, a generator set control module 103, and a generator set control completion module 104. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and which are stored in the memory of the electronic device. The power generation unit confirmation module 101 is used to confirm the power grid of the area to be regulated, wherein the power grid of the area to be regulated includes: multiple intelligent agents and multiple power generation units under their jurisdiction, wherein the intelligent agents correspond one-to-one with the power generation units, and wherein the intelligent agents include an Actor network. The output value calculation module 102 is used to determine the current control cycle, obtain the frequency regulation value set, carbon emission set and the optimal power allocation value set of the generator set based on the current control cycle, multiple intelligent agents and multiple subordinate generator sets, and obtain the comprehensive reward value and actual output value set based on the frequency regulation value set, carbon emission set and the optimal power allocation value set of the generator set. The generator set control module 103 is used to sequentially extract actual output values ​​from the actual output value set, calculate the correction target total output value, maximum technical output, minimum technical output, and maximum allowable ramp rate based on the extracted actual output values, if the correction target total output value is less than the minimum technical output or greater than the maximum technical output, then the correction target total output value is corrected using the minimum technical output or maximum technical output to obtain the corrected correction total output value, calculate the corrected total output value based on the corrected correction total output value and the extracted actual output values, calculate the subsequent cycle allocated power value based on the corrected total output value and the maximum allowable ramp rate, and control the generator sets under the extracted actual output values ​​based on the subsequent cycle allocated power value to obtain the normal managed generator sets; The generator set control completion module 104 is used to summarize the normally managed generator sets to obtain a normally managed generator set set, store the actual output value set, comprehensive reward value and the optimal power allocation value set of the generator set to obtain an intelligent agent training set, and complete the generator set steady-state control based on load change response based on the intelligent agent training set and the normally managed generator set set.

[0063] In detail, the modules in the generator set steady-state control system 100 based on load change response described in this embodiment of the invention adopt the same characteristics as described above during use. Figure 1The method used is the same as the steady-state control method for generator sets based on load change response described in the previous section, and can produce the same technical effect, so it will not be repeated here.

[0064] like Figure 3 The diagram shown is a schematic diagram of an electronic device for implementing a generator set steady-state control method based on load change response, according to an embodiment of the present invention.

[0065] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a generator set steady-state control method program based on load change response.

[0066] The memory 11 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 1. Furthermore, the memory 11 includes both internal storage units and external storage devices of the electronic device 1. The memory 11 can be used not only to store application software and various types of data installed on the electronic device 1, such as the code of a generator steady-state control method program based on load change response, but also to temporarily store data that has been output or will be output.

[0067] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., a generator set steady-state control method program based on load change response) and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0068] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11 and at least one processor 10, etc.

[0069] Figure 3 Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0070] For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0071] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is typically used to establish communication connections between the electronic device 1 and other electronic devices.

[0072] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), and optionally, a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0073] The generator set steady-state control method program based on load change response stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When run in the processor 10, it can achieve the following: The power grid of the area to be regulated has been identified. The power grid of the area to be regulated includes: multiple intelligent agents and multiple generator sets under their jurisdiction. There is a one-to-one correspondence between the intelligent agents and the generator sets. The intelligent agents include an Actor network. The current control cycle is identified, and based on the current control cycle, multiple intelligent agents, and multiple subordinate generator sets, the frequency regulation value set, carbon emission set, and optimal power allocation value set of the generator sets are obtained. The comprehensive reward value and actual output value set are obtained based on the frequency regulation value set, carbon emission set and the optimal power allocation value set of the unit; The actual output values ​​are extracted sequentially from the set of actual output values, and the total output value of the correction target, the maximum technical output, the minimum technical output, and the maximum allowable climbing rate are calculated based on the extracted actual output values. If the total output value of the target is less than the minimum technical output or greater than the maximum technical output, the total output value of the target is corrected by using the minimum technical output or the maximum technical output to obtain the corrected total output value. The corrected total output value is then calculated based on the corrected total output value and the extracted actual output value. The power allocation value for subsequent cycles is calculated based on the corrected total output value and the maximum allowable climbing rate. The generator sets under the jurisdiction with the extracted actual output value are then adjusted according to the power allocation value for subsequent cycles to obtain the normal generator sets under the jurisdiction. The normal generator sets are summarized to obtain the normal generator set set. The actual output value set, comprehensive reward value and the optimal power allocation value set of the generator sets are stored to obtain the agent training set. Based on the training set of the intelligent agent and the set of normally managed generator sets, the steady-state control of generator sets based on load change response is completed.

[0074] Specifically, the processor 10's implementation method for the above instructions can be found in [reference needed]. Figures 1 to 3 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0075] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0076] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following: The power grid of the area to be regulated has been identified. The power grid of the area to be regulated includes: multiple intelligent agents and multiple generator sets under their jurisdiction. There is a one-to-one correspondence between the intelligent agents and the generator sets. The intelligent agents include an Actor network. The current control cycle is identified, and based on the current control cycle, multiple intelligent agents, and multiple subordinate generator sets, the frequency regulation value set, carbon emission set, and optimal power allocation value set of the generator sets are obtained. The comprehensive reward value and actual output value set are obtained based on the frequency regulation value set, carbon emission set and the optimal power allocation value set of the unit; The actual output values ​​are extracted sequentially from the set of actual output values, and the total output value of the correction target, the maximum technical output, the minimum technical output, and the maximum allowable climbing rate are calculated based on the extracted actual output values. If the total output value of the target is less than the minimum technical output or greater than the maximum technical output, the total output value of the target is corrected by using the minimum technical output or the maximum technical output to obtain the corrected total output value. The corrected total output value is then calculated based on the corrected total output value and the extracted actual output value. The power allocation value for subsequent cycles is calculated based on the corrected total output value and the maximum allowable climbing rate. The generator sets under the jurisdiction with the extracted actual output value are then adjusted according to the power allocation value for subsequent cycles to obtain the normal generator sets under the jurisdiction. The normal generator sets are summarized to obtain the normal generator set set. The actual output value set, comprehensive reward value and the optimal power allocation value set of the generator sets are stored to obtain the agent training set. Based on the training set of the intelligent agent and the set of normally managed generator sets, the steady-state control of generator sets based on load change response is completed.

[0077] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and actual implementations may have other classification methods.

[0078] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0079] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0080] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A steady-state control method for generator sets based on load change response, characterized in that, The method includes: The power grid of the area to be regulated has been identified. The power grid of the area to be regulated includes: multiple intelligent agents and multiple generator sets under their jurisdiction. There is a one-to-one correspondence between the intelligent agents and the generator sets. The intelligent agents include an Actor network. The current control cycle is identified, and based on the current control cycle, multiple intelligent agents, and multiple subordinate generator sets, the frequency regulation value set, carbon emission set, and optimal power allocation value set of the generator sets are obtained. The comprehensive reward value and actual output value set are obtained based on the frequency regulation value set, carbon emission set and the optimal power allocation value set of the unit; The actual output values ​​are extracted sequentially from the set of actual output values, and the total output value of the correction target, the maximum technical output, the minimum technical output, and the maximum allowable climbing rate are calculated based on the extracted actual output values. If the total output value of the target is less than the minimum technical output or greater than the maximum technical output, the total output value of the target is corrected by using the minimum technical output or the maximum technical output to obtain the corrected total output value. The corrected total output value is then calculated based on the corrected total output value and the extracted actual output value. The power allocation value for subsequent cycles is calculated based on the corrected total output value and the maximum allowable climbing rate. The generator sets under the jurisdiction with the extracted actual output value are then adjusted according to the power allocation value for subsequent cycles to obtain the normal generator sets under the jurisdiction. The normal generator sets are summarized to obtain the normal generator set set. The actual output value set, comprehensive reward value and the optimal power allocation value set of the generator sets are stored to obtain the agent training set. Based on the training set of the intelligent agent and the set of normally managed generator sets, the steady-state control of generator sets based on load change response is completed.

2. The generator set steady-state control method based on load change response as described in claim 1, characterized in that, The process of obtaining frequency regulation value sets, carbon emission sets, and optimal power allocation value sets for generator units based on the current control cycle, multiple intelligent agents, and multiple subordinate generator units includes: Real-time monitoring of the power grid in the area to be regulated is performed based on the current control cycle to obtain tie line power deviation and system frequency deviation. The regional control deviation is calculated based on the tie line power deviation, system frequency deviation, and preset regional frequency deviation coefficient. The generator sets under their jurisdiction are extracted sequentially from multiple generator sets under their jurisdiction, and the target agent is identified from multiple agents based on the extracted generator sets under their jurisdiction. Based on the target intelligent agent and the regional control deviation, the value of the extracted generator sets under its jurisdiction is calculated to obtain the frequency regulation value and the optimal power allocation value of the generator sets. The carbon emissions are calculated by multiplying the preset carbon emission intensity coefficient of the unit with the optimal power allocation value of the unit. By summarizing the frequency regulation value, carbon emissions, and optimal power allocation value of the generating units, we can obtain the frequency regulation value set, carbon emission set, and optimal power allocation value set corresponding to multiple generating units under our jurisdiction.

3. The generator set steady-state control method based on load change response as described in claim 2, characterized in that, The process of real-time monitoring of the power grid in the area to be regulated according to the current control cycle to obtain tie-line power deviation and system frequency deviation includes: The key node set and inter-regional tie line set of the power grid in the area to be regulated are identified. Frequency is collected for each key node in the key node set according to the current control cycle to obtain the node frequency set. Abnormal frequency elimination operation is performed on the node frequency set to obtain the normal node frequency set. The node weight set is determined based on the key node set, where each node weight corresponds one-to-one with a key node. The system frequency deviation is calculated based on the node weight set and the normal node frequency set. The following operations are performed on each inter-regional tie line in the inter-regional tie line set: Obtain the line voltage and line current of the inter-regional tie lines, and calculate the active power of the tie lines based on the line voltage and line current; The difference between the active power of the tie line and the preset planned power of the tie line is calculated to obtain the power deviation of a single tie line; Summarize the power deviations of individual tie lines to obtain a set of power deviations for each tie line. Then sum the power deviation sets of individual tie lines to obtain the total power deviation of the tie line.

4. The generator set steady-state control method based on load change response as described in claim 3, characterized in that, The formula for calculating the system frequency deviation is as follows: in, Indicates the system frequency deviation. This indicates the number of critical nodes in the critical node set. Represents the node weight set of the 1st generation. The node weights of key nodes, Represents the first normal node frequency set Normal node frequency.

5. The generator set steady-state control method based on load change response as described in claim 4, characterized in that, The step of calculating the unit value of the extracted generator sets based on the target intelligent agent and regional control deviation to obtain the frequency regulation value and the optimal power allocation value of the units includes: Information is collected from the generator units under its jurisdiction to obtain the generator unit parameters of the power plant; The regional control deviation, system frequency deviation, and power plant unit parameters are input into the Actor network of the target intelligent agent to obtain the optimal power allocation value of the unit; Obtain the historical periodic power allocation value of the extracted generator units, calculate the absolute difference between the historical periodic power allocation value and the optimal power allocation value of the units, and obtain the frequency regulation mileage of the units. The frequency regulation value is calculated based on the preset mileage value, the preset unit regulation performance coefficient, and the unit frequency regulation mileage.

6. The generator set steady-state control method based on load change response as described in claim 5, characterized in that, The process of obtaining the comprehensive reward value and actual output value set based on the frequency regulation value set, carbon emission set, and optimal power allocation value set of the unit includes: The carbon emission sets are summed to obtain the total carbon emissions. The total carbon emissions and the preset carbon emission penalty coefficient are then calculated to obtain the carbon emission penalty value. The stability reward value is calculated based on the regional control deviation and the system frequency deviation, and the frequency modulation contribution reward value is calculated based on the frequency modulation value set. The comprehensive reward value is obtained by summing the carbon emission penalty value, the stability reward value, and the frequency regulation contribution reward value. Based on the confirmation of the power plant control system by multiple generator units under its jurisdiction, the optimal power allocation value set of the units is sent to the power plant control system to obtain the actual output value set. The actual output value set includes multiple actual output values, and the actual output values ​​and the optimal power allocation value of the units correspond one-to-one with the generator units under its jurisdiction.

7. The generator set steady-state control method based on load change response as described in claim 6, characterized in that, The formula for calculating the stability reward value is as follows: in, This represents the stability reward value. This indicates the preset regional control deviation weight. Indicates regional control deviation. This represents the preset system frequency deviation weight. This indicates the system frequency deviation.

8. The generator set steady-state control method based on load change response as described in claim 7, characterized in that, The calculation of the corrected target total output value based on the extracted actual output value includes: Based on the extracted actual output values, the historical actual output values ​​and the optimal power allocation values ​​for the target units are confirmed; The expected output value is obtained by summing the historical actual output value and the optimal power allocation value of the target unit. The output deviation is obtained by calculating the difference between the expected output value and the extracted actual output value. Obtain the power adjustment value for the next cycle, and calculate and correct the next power adjustment value based on the output deviation and the power adjustment value for the next cycle. The total output value of the correction target is calculated based on the corrected next power adjustment value and the extracted actual output value.

9. The generator set steady-state control method based on load change response as described in claim 8, characterized in that, The calculation of the power distribution value for subsequent cycles based on the corrected total output value and the maximum allowable climbing rate includes: The average ramp rate is calculated based on the corrected total output value and the current control cycle. If the average ramp rate is greater than the maximum allowable ramp rate, the maximum output change value is calculated based on the maximum allowable ramp rate and the current control cycle. The current cycle limit value is calculated based on the maximum output change value and the next power adjustment value. The remaining uncorrected value is calculated based on the current cycle limit value and the next power adjustment value. The remaining uncorrected value is the value obtained by subtracting the current cycle limit value from the next power adjustment value. Calculate the minimum number of cycles required based on the remaining uncorrected values ​​and the maximum output change value; Calculate the power allocation value for subsequent cycles based on the remaining uncorrected values ​​and the minimum number of cycles required.

10. A generator set steady-state control system based on load change response, characterized in that, The system includes: The assigned generator set confirmation module is used to confirm the power grid in the area to be regulated. The power grid in the area to be regulated includes multiple intelligent agents and multiple assigned generator sets. Each intelligent agent corresponds to one generator set. The intelligent agents include an Actor network. The output value calculation module is used to determine the current control cycle, and based on the current control cycle, multiple intelligent agents and multiple subordinate generator sets, obtain the frequency regulation value set, carbon emission set and the optimal power allocation value set of the generator sets, and based on the frequency regulation value set, carbon emission set and the optimal power allocation value set of the generator sets, obtain the comprehensive reward value and the actual output value set. The generator set control module is used to sequentially extract actual output values ​​from the actual output value set, calculate the correction target total output value, maximum technical output, minimum technical output, and maximum allowable ramp rate based on the extracted actual output values, if the correction target total output value is less than the minimum technical output or greater than the maximum technical output, then the correction target total output value is corrected using the minimum or maximum technical output to obtain the corrected total output value, calculate the corrected total output value based on the corrected total output value and the extracted actual output values, calculate the subsequent cycle allocated power value based on the corrected total output value and the maximum allowable ramp rate, and control the generator sets under the extracted actual output values ​​based on the subsequent cycle allocated power value to obtain the normal managed generator sets; The generator set control module is used to summarize the normally managed generator sets to obtain the normally managed generator set set. It stores the actual output value set, comprehensive reward value and the optimal power allocation value set of the generator set to obtain the agent training set. Based on the agent training set and the normally managed generator set set, it completes the steady-state control of the generator set based on the load change response.

Images