Multi-unit gas power station intelligent grid-connected control method and system

Abstract

The invention provides an intelligent grid-connected control method and system for a multi-unit gas power station, and relates to the technical field of power system automation control, and the method comprises the steps: dividing a generator set group in a target power station into three levels, namely a unit level, a group level and a power station level, on an electrical structure; based on the three levels, setting three intelligent grid-connected control modes: a grouping independent mode, a group parallel mode and a high-voltage side grid-connected mode; performing cost optimization calculation on the three control modes through a genetic algorithm, and dynamically switching to an optimal control mode according to real-time operation data; according to the method, the huge generator set group is logically decomposed and physically grouped, so that efficient and reliable global grid-connected control is realized.

Description

Technical Field

[0001] This invention relates to the field of power system automation control technology, specifically to an intelligent grid-connected control method and system for multi-unit gas-fired power plants. Background Technology

[0002] With the transformation of the energy structure and the increasing demand for peak-shaving capacity in the power system, large-scale gas-fired power plants are playing an increasingly important role in modern power grids due to their advantages such as rapid start-up and shutdown, flexible regulation, and clean and efficient operation. However, the continuous expansion of power plant scale, especially the popularization of multi-unit group coordinated operation mode, poses a severe challenge to traditional grid-connected control systems.

[0003] Currently, in the field of power system automation control, traditional large-scale power plant grid-connected systems generally adopt a centralized control architecture. Under this architecture, the operating data of all units must be uploaded to a central control unit, which performs unified calculations and decisions before sending control commands downwards. This mode can operate effectively when the number of units is small, but when the power plant expands and the number of units exceeds 32, system bottlenecks become increasingly prominent: First, the centralized transmission and processing of massive amounts of data leads to significant communication delays, greatly reducing the real-time performance of the control system and making it difficult to meet the stringent requirements of the power grid for rapid grid connection; second, the load on the central processing unit increases dramatically, easily causing computational bottlenecks, resulting in decreased control accuracy and deterioration of system stability, and a failure at a single control point may even paralyze the entire grid-connected system.

[0004] Furthermore, existing grid-connected control technologies fall short when faced with the complex operational demands of large gas-fired power plants with multiple generating units. Fluctuations in grid load, maintenance schedules for generating units, and intermittent access from different energy sources all require power plants to flexibly group operating units and implement differentiated control strategies. However, traditional rigid and fixed control systems struggle to achieve this dynamic configuration management and cannot intelligently divide dozens of units into multiple independent grid-connected units and implement dynamic grid-connected control based on real-time task requirements. This results in suboptimal resource allocation, slow response times, and difficulty in further improving overall operational efficiency.

[0005] Therefore, existing intelligent grid-connected control technologies for multi-unit gas-fired power plants cannot overcome the inherent defects of centralized architecture, cannot meet the requirements in terms of real-time performance and reliability, and cannot support flexible group management and dynamic control. Summary of the Invention

[0006] To address the aforementioned problems, this invention proposes an intelligent grid-connected control method and system for multi-unit gas-fired power plants. By logically decomposing and physically grouping a large group of generator units, efficient and reliable global grid-connected control is achieved.

[0007] According to some embodiments, the present invention adopts the following technical solution: A method for intelligent grid-connected control of a multi-unit gas-fired power plant includes: The generator group in the target power station is divided into three levels in terms of electrical structure: unit level, group level and power station level; Based on the three levels, three intelligent grid connection control modes are set: group independent mode, group parallel mode, and high-voltage side grid connection mode. The genetic algorithm is used to perform cost optimization calculations for the three control modes and dynamically switch to the optimal control mode based on real-time operating data.

[0008] According to some embodiments, the present invention adopts the following technical solution: A smart grid-connected control system for a multi-unit gas-fired power plant includes: The structural partitioning module is configured to divide the generator group in the target power station into three levels in terms of electrical structure: unit level, group level and power station level; The control setting module is configured to set three intelligent grid connection control modes based on the three levels: group independent mode, group parallel mode and high-voltage side grid connection mode; The mode optimization module is configured to perform cost optimization calculations on three control modes using a genetic algorithm, and dynamically switch to the optimal control mode based on real-time operating data.

[0009] According to some embodiments, the present invention adopts the following technical solution: A computer program product includes a computer program that, when executed by a processor, implements the aforementioned intelligent grid-connected control method for a multi-unit gas-fired power plant.

[0010] According to some embodiments, the present invention adopts the following technical solution: A non-transitory computer-readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the aforementioned intelligent grid-connected control method for a multi-unit gas-fired power plant.

[0011] According to some embodiments, the present invention adopts the following technical solution: An electronic device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the aforementioned intelligent grid-connected control method for a multi-unit gas-fired power plant.

[0012] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention proposes a three-level distributed control architecture, which decomposes the complex task of controlling several generating units into the coordination problem of multiple small-scale groups through a hierarchical aggregation method of "unit level → group level → power plant level". This structure effectively breaks through the capacity limitations of traditional parallel systems in terms of physical layout and control logic, and has good scalability, flexibility and reliability, making it suitable for grid-connected operation control of power generation systems with a large number of generating units.

[0013] This invention designs a unique hybrid communication scheme combining CAN bus and optical fiber, solving the problem of long-distance communication in large power plants; it realizes dynamic reconfigurability of power plant operation modes and supports switching between three typical operation modes based on genetic algorithms. Attached Figure Description

[0014] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0015] Figure 1 This is a diagram of the distributed three-level control architecture of Example 1. Figure 2 This is the logic diagram for adjusting the standard deviation σ in Example 1. Figure 3 This is a schematic diagram of CAN communication between the group controller and the unit controller in Example 1. Figure 4 This is a schematic diagram of CAN communication between the group controller, bus tie controller, and grid-connected controller in Example 1. Detailed Implementation

[0016] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0017] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0018] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0019] Example 1 One embodiment of the present invention provides a smart grid-connected control method for a multi-unit gas-fired power plant, comprising: The generator group in the target power station is divided into three levels in terms of electrical structure: unit level, group level and power station level; Based on the three levels, three intelligent grid connection control modes are set: group independent mode, group parallel mode, and high-voltage side grid connection mode. The genetic algorithm is used to perform cost optimization calculations for the three control modes and dynamically switch to the optimal control mode based on real-time operating data.

[0020] As one embodiment, the present invention provides an intelligent grid-connected control method for multi-unit gas-fired power plants. By logically decomposing and physically grouping a large group of generator units, it achieves efficient and reliable global grid-connected control. The specific implementation process is as follows: The distributed three-level control architecture proposed in this embodiment is as follows: Figure 1 As shown, this architecture achieves efficient and reliable global grid-connected control by logically decomposing and physically grouping a large number of generator sets. Its core is a three-level coordinated control system at the generator set level, group level, and power station grid-connected level.

[0021] Figure 1 In the diagram, IG-NTC-GC is the unit controller, IG1000-BC1-5 is the group controller, IG1000-BTB1-4 is the bus tie controller, IM1010-MCB1-3 is the grid connection controller, GCB is the unit circuit breaker, BC1-5 is the group circuit breaker, BTB1-4 is the bus tie circuit breaker, MCB1-3 is the grid connection circuit breaker, BANK1-5 is group 1-5, and G is the gas generator set.

[0022] 1. Overall Architecture The system is divided into three levels in terms of electrical structure: unit level (level 1), group level (level 2), and power plant level (level 3), as detailed below: Level 1: Unit Level Main components: It includes five common busbars, each of which connects 10, 15, 15, 13, and 13 generator sets respectively, for a total of 66 units; each unit is equipped with a unit circuit breaker GCB and a unit controller (IG-NTC-GC); each busbar is equipped with a group circuit breaker BC and a group controller G1000-BC.

[0023] Connection structure: Five generating busbars, each equipped with a unit controller to enable automatic parallel operation and load balancing among units on the same busbar. The generated power from each unit is collected at its corresponding common busbar and then output to supply loads or transmit to higher levels.

[0024] Features: Enables coordinated control of multiple units on the same bus, ensuring power balance and stable operation among units on the same bus.

[0025] Level 2: Group Level Main components: including 5 group circuit breakers BC1-5, 4 bus tie circuit breakers BTB1-4, 5 group controllers IG1000-BC1-5 and 4 bus tie controllers IG1000-BTB1-4, as well as the connecting busbars between the various devices.

[0026] Connection Relationship: All 66 generating units are divided into 5 groups. Each group coordinates the total output power of its internal units through a group controller and realizes power distribution between groups. The bus tie controller is responsible for the synchronous closing and sectional control of the bus tie circuit breakers between adjacent busbars, and supports electrical connection and isolation between groups.

[0027] Features: Each group can be dispatched as an independent unit. The groups can be flexibly combined with the group circuit breaker and the bus tie circuit breaker to realize modular operation of the system and multiple power output modes.

[0028] Level 3: Power Plant Level Main components: Includes 3-channel grid-connected switchgear and 3 power station grid-connected controllers.

[0029] Connection relationship: As the interface between the power plant and the external power grid, the grid-connected controller integrates and schedules the electrical energy output from each group to realize grid connection / off-grid operation with the upper-level power grid.

[0030] Features: It enables unified management of the entire station's power supply, and while meeting its own power needs, it achieves safe grid connection and power interaction with higher-level power grids.

[0031] 2. Functions of each controller Unit controller: Each unit is independently configured and is responsible for the unit's start-up, shutdown, speed regulation, voltage regulation, parallel operation and load distribution within the same unit.

[0032] Group Controller: There is one group controller on each common bus. All units on the same common bus constitute a group BANK, with a total of 5 groups. The group controller is responsible for coordinating the total output power of all units in the group and participating in load sharing between groups. For example, group controller IG1000-BC1 is responsible for setting the total output power of 10 units in BANK1. Group controllers IG1000-BC1 and IG1000-BC2 participate in load sharing between groups BANK1 and BANK2.

[0033] Bus tie controller: Used to control the bus tie circuit breaker between two bus sections to achieve synchronous grid connection and electrical isolation between groups.

[0034] Power plant grid connection controller: As the top-level control unit, it is responsible for the overall regulation of the power output of the entire power plant, realizing the safe grid connection and power dispatch of the power plant with the external power grid.

[0035] Based on the above architecture and controller functions, a specific implementation example is provided, taking a foreign power plant project (66 gas generator units) as an example. It includes 5 groups, totaling 66 gas generator units, supplying power to the production and living loads of the oilfield. The unit parameters are: rated power 2000KW, rated voltage 10.5kV, rated frequency 50HZ, and rated power factor 0.8. The distributed three-level control architecture is shown in the appendix. Figure 1 In the figure, BC1, BC2, BC3, MCB1, MCB2, MCB3, BTB1, BTB2, BTB3, BTB4, and GCB all represent circuit breakers. This embodiment uses a three-level distributed control architecture for a large number of units (>32 units), which breaks through the traditional limitation on the number of parallel control modules and greatly improves control efficiency.

[0036] The following section uses mathematical modeling of a scalable communication model within a distributed control architecture to innovatively quantify and analyze the relationship between the number of generating units and control efficiency: Assuming the communication delay of each control layer (GC / BC / MCB, i.e., unit / group / power station) is a fixed value t0, the number of units is N (in this embodiment, N=66), and the number of control layer levels is L=3, then the total communication delay is: (1) In calculating total communication latency, the latency growth of a distributed system is typically not linear; therefore, logarithm (log2) is used. N It can perform quick and easy estimations, making it more reasonable to reflect the growth trend of communication latency in large-scale systems and fully embody the parallelism and hierarchical characteristics of distributed architecture.

[0037] When N=66, substituting into equation (1) yields:

[0038] This demonstrates the difference in latency compared to a centralized control architecture.

[0039] The communication latency of the three-tier distributed architecture is reduced by approximately It significantly improves communication efficiency when multiple units are operating in parallel.

[0040] 3. Control Strategy Logical grouping of generating units is achieved by configuring the IG1000-BC group controller, and flexible grid connection between groups is achieved by using the IG1000-BTB bus tie controller. It supports three operating modes: "group independent mode," "group parallel mode," and "high-voltage side grid connection mode," which are described below: (1) Group independent mode operation Disconnect 4 BTB1-4 bus tie circuit breakers and 3 MCB1-3 grid connection circuit breakers. The 5 groups operate in a grouped independent mode, with a maximum distance of 2000 meters between the 5 groups, each carrying its own bus section load.

[0041] Each group executes a dynamic load balancing strategy and power management mode. Taking the first group, BANK1, as an example, the IG1000-BC1 group controller monitors the total load in real time and communicates with the IG-NTC-GC unit controllers of the 10 units via CAN. Its main functions are: ① Dynamic load balancing strategy The IG1000-BC1 group controller monitors the total output power of the BANK1 group in real time, i.e. the total load of the BANK1 group, by collecting the current signal of the circuit where the BC1 circuit breaker is located. It then broadcasts this information to the IG-NTC-GC unit controller of the operating unit via the CAN bus.

[0042] The target power of a single generating unit is equal to the total load divided by the number of generating units in operation, calculated as follows:

[0043] Among them, P 单target For the target power of a single unit, P 总load n represents the total load of the group. 运行 This refers to the number of units online.

[0044] The IG-NTC-GC unit controller for each unit is responsible for performing closed-loop power regulation to ensure the actual output power P 单actual It can quickly and accurately track the target power P sent by the IG1000-BC1 controller. 单target The specific control process is as follows: The IG-NTC-GC unit controller acquires the voltage and current signals of the unit in real time and calculates the actual power P of a single unit. 单actual and with P 单target Comparison, generating power deviation ΔP:

[0045] The above formula can be used to calculate both active power deviation and reactive power deviation.

[0046] a. Speed ​​control (frequency / power regulation): If the active power ΔP>0, the active power load demand increases. The IG-NTC-GC unit controller sends an acceleration signal to the engine governor to increase the gas intake, thereby increasing the unit's output active power.

[0047] If the active power ΔP < 0, the active power load demand decreases. The IG-NTC-GC unit controller sends a speed reduction signal to the engine governor to reduce the gas intake, thereby reducing the unit's output active power.

[0048] b. Voltage regulation control (reactive power / voltage regulation): If the active power ΔP>0, the reactive power load demand increases. The IG-NTC-GC unit controller sends a boost signal to the generator voltage regulator AVR board to increase the excitation current, thereby increasing the unit's output reactive power.

[0049] If the active power ΔP < 0, the reactive power load demand decreases. The IG-NTC-GC unit controller sends a step-down signal to the generator voltage regulator AVR board to reduce the excitation current, thereby reducing the unit's output reactive power.

[0050] ② Power Management Mode Power management is based on total active load. According to changes in total active load demand, it automatically starts and stops a corresponding number of generating units. Specifically, when the total active load exceeds 80% (adjustable) of the total capacity of currently operating units, the next unit is started, and load distribution is controlled among all operating units. When the total active load is lower than 70% (adjustable) of the total capacity after subtracting one unit from the current number of operating units, the unit that has been running the longest is shut down to reduce energy consumption. Detailed control strategies are as follows: a. Activation conditions (taking an 80% threshold as an example) All online units are pre-configured with startup sequences 1 to n. 总 n 总 The total number of units in the group is n. In group BANK1, n 总 =10; when P 总有功load >0.8×P 单rated × n 运行 (P) 单rated n is the rated power of a single unit. 运行 When the total active load exceeds 80% of the total rated power of the currently operating units (the number of currently operating units), the BC group controller sends a start command to the nth unit via CAN communication. 运行 +1 IG-NTC-GC unit controller, nth 运行 +1 IG-NTC-GC unit controller controls the unit's automatic start-up, speed increase to rated speed, automatic excitation generation, and automatic synchronous closing on the nth day. 运行 +1 GCB unit circuit breaker; nth 运行+1 unit is added to the operating group, increasing the number of operating units by one; the group will redistribute the load according to the dynamic load averaging strategy described above ①. b. Shutdown conditions (taking a 70% threshold as an example) When P 总有功load <0.7×P 单rated ×(n 运行 When the total active load is less than 70% of the total rated power of the currently operating units, the BC controller sends a shutdown command to the controller of the IG-NTC-GC unit with the longest operating time via CAN communication. The IG-NTC-GC unit controller controls the unit to transfer the load to other operating units. When the load it carries is close to the circuit breaker tripping value, the output circuit breaker GCB of that unit is disconnected. The unit is then shut down after cooling down, and the number of currently operating units is reduced by one. The unit group then redistributes the load according to the dynamic load averaging strategy described in ① above. (2) Group parallel operation Taking the joint operation of the first group BANK1 and the second group BANK2 as an example, disconnect the three bus tie BTB circuit breakers (BTB2 / 3 / 4) and the three MCB grid-connected circuit breakers. The BTB1 bus tie circuit breaker remains controllable. When the load of BANK1 group is at its peak, and all units in BANK1 are operating at their maximum sustainable power, but a power deficit still exists, the parallel connection between groups is triggered. The detailed control strategy is as follows: ① Power deficit determination The IG1000-BC1 group controller of group BANK1 detected P 总load When the maximum sustainable power of the operating unit is greater than ∑(the maximum sustainable power of the unit), a parallel connection request is sent to the IG1000-BTB1 bus tie controller via CAN communication.

[0051] ②BTB1 Synchronous Verification and Adjustment After receiving the parallel connection request, the IG1000-BTB1 bus tie controller compares the voltage difference, frequency difference, and phase difference of the two bus sections of group BANK1 and group BANK2 in real time to see if they are within the allowable synchronization conditions.

[0052] If the synchronization is not within the range, the bus tie controller IG1000-BTB1 communicates with the unit controllers IG-NTC-GC in groups BANK1 and BANK2. IG-NTC-GC adjusts the speed and voltage values ​​of their respective units until the two bus sections meet the synchronization conditions.

[0053] a. Frequency adjustment: If the frequency of BANK1 is higher than that of BANK2, the generator controller IG-NTC-GC in BANK1 sends a speed reduction signal to the engine governor to reduce the gas intake and lower the generator frequency. In BANK2, the IG-NTC-GC sends a speed increase signal to the engine governor to increase the gas intake and raise the generator frequency. If the frequency of BANK2 is higher than that of BANK1, the adjustment process is the opposite.

[0054] b. Voltage regulation: If the voltage in BANK1 is higher than that in BANK2, the unit controller IG-NTC-GC in BANK1 sends a step-down signal to the generator voltage regulator AVR board to reduce the excitation current and lower the generator voltage. Conversely, if the voltage in BANK2 is higher than that in BANK1, the regulation process is the opposite.

[0055] Once the synchronization conditions are met, the IG1000-BTB1 controller sends a closing command to BTB1, and BANK1 and BANK2 operate in parallel.

[0056] ③BANK1 and BANK2 operate together with load and load transfer a. Adjustment of BANK2 unit output According to the preset control strategy (average distribution / fixed total load), the output of BANK2 unit is increased to share the load shortfall with BANK1.

[0057] Average distribution: The total load is evenly distributed among the units in group BANK1 and BANK2.

[0058] Fixed total load: BANK2 maintains a constant total output power and gradually transfers the fixed load of BANK1 to BANK2.

[0059] b. Load decrease and transfer in BANK1 Slowly reduce the output of BANK1 unit and gradually transfer the load to BANK2 unit.

[0060] (3) High-voltage side grid connection mode (taking 35kV as an example) The 35KV high-voltage side is connected to the grid, which can realize dual power supply to the load side to improve power supply reliability and load distribution flexibility. On the other hand, when the load in the station is low, the surplus power can also be connected to the grid.

[0061] Taking the grid-connected operation of group BANK1 and No. 1 10.5 / 35KV transformer as an example, the bus tie circuit breaker BTB1 is disconnected and the group circuit breaker BC1 is closed, and BANK1 operates as an independent group. After receiving the synchronization grid connection signal sent by the human or the background monitoring system, the grid-connected circuit breaker MCB1 automatically detects the consistency of the electrical signals on both sides of MCB1, namely the group side and the low-voltage side of the transformer, including voltage, frequency and phase, and whether the synchronization conditions are met.

[0062] If the conditions are not met, the grid-connected controller IM1010-MCB1 communicates with the group controller IG1000-BC1, which then sends a signal to the unit controllers IG-NTC-GC within the group. IG-NTC-GC adjusts the speed and voltage values ​​of its respective units in a manner similar to the synchronization verification and adjustment strategy of ②BTB1 in parallel operation of the group; however, the adjustment targets only include all units within group BANK1, and the 35KV side of the main grid cannot be adjusted. Until the synchronization conditions are met on the group side and the transformer side of the grid-connected circuit breaker MCB1, the grid-connected controller IM1010-MCB1 sends a closing command to the grid-connected circuit breaker MCB1, and group BANK1 and the No. 1 10.5 / 35KV transformer are connected to the grid for operation.

[0063] By setting the parameters of the grid-connected controller IM1010-MCB1 and the group controller IG1000-BC1, various load-bearing modes for the unit group and the mains power can be achieved, including: Mode 1, Group BANK1 outputs fixed load: After deducting local load usage, surplus electricity is fed into the grid; its characteristics are stable load, which is beneficial to unit operation, and full utilization of natural gas feedstock; Mode 2, 35KV side mains power + BANK1 supplement: The mains power supply provides local loads at a fixed power, and BANK1 is used for peak shaving; the feature is that the backup power supply has sufficient reserves and fast response, which is particularly suitable for situations with large load fluctuations.

[0064] Mode 3, BANK1 fixed internet access, means that in addition to supplying local loads, it provides fixed power to the internet; its feature is that it can not only meet the needs of free loads, but also guarantee revenue from electricity sales.

[0065] 4. Mode Optimization To minimize long-term operating costs and maximize profits, this embodiment employs economic optimization based on a dynamically reconfigurable model using a genetic algorithm for mode switching. The following is an innovative functional modeling of mode switching costs: First, define the mode switching cost C, including the unit start-up and shutdown cost C. start / stop Load transfer cost C load Electricity sales revenue R sellThe formula is as follows (this embodiment simplifies the formula using only 3 basic influencing factors; in reality, more parameters are involved): C=α·C start / stop +β·C load γ·R sell (2) The parameters α, β, and γ can be optimized using a genetic algorithm to minimize long-term operating costs. The specific implementation process is as follows: First, N sets of random initial individuals need to be generated. Before generating them, the parameter range needs to be set to simplify the calculation: α∈[α_min, α_max], β∈[β_min, β_max], γ∈[γ_min, γ_max], let min=1 and max=3, then α∈[1, 3], β∈[1, 3], γ∈[1, 3].

[0066] Next, within the above range, N individuals of α, β, and γ are randomly generated, each individual being (α... i ,β i γ i ), where i = 1, 2, ..., N: Individual 1: (α1, β1, γ1) = (0.5, 1.2, 0.8); Individual 2: (α2, β2, γ2) = (0.3, 0.9, 1.1); ... Individual N: (α) N ,β N γ N ).

[0067] Next, these individuals are substituted into formula (3) to calculate the fitness of each individual in order to measure their merits. The process is as follows: For each individual (α) i ,β i γ i ), calculate the long-term operating cost C i After calculation, substitute it into the fitness function: f i =1 / C i (3) As can be seen from equation (3), the lower the cost, the higher the fitness.

[0068] Assume start-up and shutdown costs C start / stop Load transfer cost C load Electricity sales revenue R sell When both are 1, calculate C. i Substituting into formula (3), we get: C1=2.5, f1=0.4; C² = 2.3, f² = 0.435; C N f n .

[0069] The next step is to select individuals with high fitness as "parents" to reproduce the next generation. Here, a "roulette wheel selection" method is used to increase the probability of selecting individuals with high fitness, such as C1 and C2.

[0070] The next step is to randomly select a crossover point to exchange some of the parents' genes. This embodiment uses a "single-point crossover" method, as follows: Individual 1: (α1, β1, γ1) = (0.5, 1.2, 0.8), parents; Individual 2: (α2, β2, γ2) = (0.3, 0.9, 1.1), parents; Assuming the intersection occurs at the second point, the offspring would be: Offspring 1: (0.5, 0.9, 1.1); Offspring 2: (0.3, 1.2, 0.8); The next step is to randomly alter the genes of the offspring to increase population diversity. Here, a Gaussian mutation method is used to add random noise with a Gaussian distribution to a certain parameter. Offspring 1: (0.5, 0.9, 1.1) → after mutation becomes (0.5, 0.9+δ, 1.1); Offspring 2: (0.3, 1.2, 0.8) → after mutation becomes (0.3, 1.2, 0.8 + δ); Where δ~N(0, σ 2 ), where δ is a random variable, and ~ indicates that it follows a normal distribution (i.e., Gaussian distribution); the mean μ=0 indicates that the center point of the noise is symmetrical with positive and negative disturbances; σ is the standard deviation; the variance σ² indicates the amplitude of noise fluctuation, and the smaller σ² is, the more concentrated the noise is.

[0071] During each mutation, δ is independently sampled from the normal distribution N(0, σ²) to ensure the randomness of each perturbation. Its function is to explore the neighborhood of the parameter space by adding a small amount of random noise, so as to avoid the algorithm getting stuck in local optima.

[0072] In this embodiment, the standard deviation σ is determined using a trial-and-error method in parameter tuning. The principle is to use σ as a hyperparameter of the genetic algorithm and observe the impact of different σ values ​​on the algorithm's convergence speed and the quality of the optimal individual through multiple trial-and-error experiments. If σ is too large (e.g., σ=1), it may cause the parameter variation to exceed the reasonable range (e.g., α mutates from 0.5 to 3.5, exceeding the range of [1, 3]), thus destroying superior genes; If σ is too small (e.g., σ=0.01), the mutation effect is weak, the population diversity is insufficient, and it is easy to get trapped in local optima; Therefore, this embodiment started testing with σ=0.5 (initially using a larger σ to enhance global search capabilities, and later reducing σ=0.1 for fine-tuning), gradually adjusting until the algorithm performed optimally.

[0073] The steps for determining σ are as follows: First, fix the value of σ (e.g., σ=0.5, 0.4, 0.3....), run the genetic algorithm multiple times, record the average number of generations of convergence, the long-term running cost corresponding to the best individual, and whether the parameters frequently exceed the limits (e.g., α<1 or α>3). After comprehensive judgment, select the σ with the best performance.

[0074] Secondly, analyze the sensitivity of parameters. If the electricity sales revenue (γ) has the greatest impact on the cost, a smaller σ (such as 0.1) can be used for it to avoid excessive disturbance; if the start-up and shutdown costs (α) are more robust, a larger σ (such as 0.3) can be used.

[0075] Secondly, adopting such Figure 2 The logic shown is to adjust the size of σ as follows: when the fitness of the population has not improved for several consecutive generations, increase σ to escape the local optimum; when the fitness has significantly improved, decrease σ to achieve stable convergence.

[0076] like Figure 2 As shown, a variable named `sigma` is first defined and initialized to 0.5. Then, the value of `sigma` is adjusted based on the number of iterations (`generation`) and whether the optimal fitness has improved: if the number of iterations exceeds 100 and the optimal fitness has not improved in the last 20 iterations, `sigma` is increased by 10% to increase search diversity; if the number of iterations exceeds 50 and the optimal fitness has improved, `sigma` is decreased by 10% for a more refined search. Finally, each gene of the individual is traversed, and for each gene, Gaussian mutation is performed with a certain probability (`mutation_rate`). A random value drawn from a Gaussian distribution with a mean of 0 and a standard deviation of `sigma` is added to the mutated gene value. The `np.clip` function is used to ensure that the mutated gene value remains within the range [1, 3] to prevent the gene value from exceeding the preset range.

[0077] Finally, convergence curves were plotted for different σ values ​​to observe changes in cost and fitness.

[0078] Using the above method, when the preset number of iterations is reached, the fitness reaches the threshold, and the population fitness no longer increases significantly, the individual with the highest fitness is output, which is the optimal parameter combination (α*, β*, γ*).

[0079] When start-up and shutdown costs Cstart / stop Or load transfer cost C load Or revenue from electricity sales R sell When significant changes occur, the system based on the parameter combination (α*, β*, γ*) determines mode switching through the following steps: Calculate the cost of the current mode: C current =α * ·C start / stop +β * ·C load γ * ·R sell Predicting the cost of the new model: C new =α * ·C start / stop′ +β * ·C load′ γ * ·R sell′ If C new <C current If so, switch to the new mode; If C new ≥C current If so, the current mode will be maintained.

[0080] The core of the dynamic reconfigurable mode is to adjust the system configuration based on real-time economics and operating conditions. When α* has a high weight, the system pays more attention to start-up and shutdown costs and tends to reduce the number of unit start-ups and shutdowns; when β* has a high weight, the system pays more attention to start-up and shutdown costs and tends to reduce the number of unit start-ups and shutdowns. * When the weight is high, the system pays more attention to load transfer costs and tends to smooth load distribution; when γ * When the weighting is high, the system pays more attention to electricity sales revenue and tends to increase output when electricity prices are high. The optimal parameter combination (α) is found through a genetic algorithm. * ,β * γ * The system can switch operating modes in a timely manner to minimize long-term operating costs while maximizing profits.

[0081] The simplified switching steps are as follows: a. Collect data on start-up and shutdown costs, load transfer costs, and electricity sales revenue in real time; b. Use a genetic algorithm to calculate the optimal parameter combination (α) for each of the three operating modes. * ,β * γ * ); c. If costs or benefits change significantly, use formula (2) to calculate the costs of the current model and the new model; d. Compare the C modes of the three models. currentand C new Decide whether to switch modes; e. such as C new <C current Then the execution mode switches to the optimal mode among the three modes, such as C. new ≥C current Then maintain the current mode.

[0082] 5. Communication Scheme 1) CAN bus communication method The group controllers IG1000-BC1 / BC2 / BC3, grid-connected controllers IM1010-MCB1 / MCB2 / MCB3, bus tie controllers IG1000-BTB1 / BTB2 / BTB3 / BTB4, and unit controllers IG-NTC-GC all communicate via CAN bus. A communication diagram is shown below. Figure 3 and Figure 4 2) Hybrid networking of CAN bus and fiber optic cable Due to the large number of generating units and the extensive layout of the power station, a hybrid networking approach using CAN bus and fiber optic cables was adopted. The theoretical maximum transmission distance of the CAN bus at a rate of 250kbps is 200 meters, but the actual distance may be shorter due to factors such as cable quality and interference. By using a CAN-to-fiber optic module, the communication distance can be extended to several kilometers. Typical transmission distances for multimode fiber are 2km, while single-mode fiber can reach over 20km.

[0083] This embodiment, through a unique hybrid communication scheme combining CAN bus and fiber optic, not only solves the long-distance communication problem in large power plants but also greatly improves communication reliability. The following is a mathematical model of the reliability of this hybrid communication scheme: Assume the CAN bus failure rate is p CAN The fiber optic failure rate is p OF The reliability calculation formula for the hybrid communication scheme is: R hybrid =1 (p CAN ·p OF (4) Typically, the failure rate of a CAN bus is about one in a thousand (10). -3 The failure rate of fiber optic communication is generally about one in ten thousand (10). -4 Substituting this into formula (4), we can see that: R hybrid ≈0.9999999 It is evident that the reliability of the hybrid communication scheme using CAN bus and fiber optic cable is improved by approximately three orders of magnitude.

[0084] Example 2 One embodiment of the present invention provides an intelligent grid-connected control system for a multi-unit gas-fired power plant, comprising: The structure partitioning controller is configured to divide the generator group in the target power station into three levels in terms of electrical structure: unit level, group level and power station level; The control settings controller is configured to set three intelligent grid connection control modes based on the three levels: group independent mode, group parallel mode, and high-voltage side grid connection mode; The mode optimization controller is configured to perform cost optimization calculations on three control modes using a genetic algorithm, and dynamically switch to the optimal control mode based on real-time operating data.

[0085] Example 3 One embodiment of the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned intelligent grid-connected control method for a multi-unit gas-fired power plant.

[0086] Example 4 In one embodiment of the present invention, a non-transitory computer-readable storage medium is provided for storing computer instructions. When the computer instructions are executed by a processor, they implement the aforementioned intelligent grid-connected control method for a multi-unit gas-fired power plant.

[0087] Example 5 One embodiment of the present invention provides an electronic device, including: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the aforementioned intelligent grid-connected control method for a multi-unit gas-fired power plant.

[0088] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0089] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0090] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for intelligent grid-connected control of a multi-unit gas-fired power plant, characterized in that, include: The generator group in the target power station is divided into three levels in terms of electrical structure: unit level, group level and power station level; Based on the three levels, three intelligent grid connection control modes are set: group independent mode, group parallel mode, and high-voltage side grid connection mode. The genetic algorithm is used to perform cost optimization calculations for the three control modes and dynamically switch to the optimal control mode based on real-time operating data.

2. The intelligent grid-connected control method for a multi-unit gas-fired power plant as described in claim 1, characterized in that, The unit level includes several common busbars, with several generator sets connected to each busbar, and each generator set is equipped with a unit controller; The unit controller is responsible for starting, stopping, speed regulation, voltage regulation, parallel operation and load distribution within the same unit.

3. The intelligent grid-connected control method for a multi-unit gas-fired power plant as described in claim 1, characterized in that, The group level includes several group switch cabinets, several bus tie switch cabinets, several group controllers and several bus tie controllers, and the devices are connected by busbars. The group controller is provided for each bus and is responsible for coordinating the total output power of all units in the group and participating in load distribution between groups; the bus tie controller is used to control the bus tie circuit breaker between two bus sections to achieve synchronous grid connection and electrical isolation between groups.

4. The intelligent grid-connected control method for a multi-unit gas-fired power plant as described in claim 1, characterized in that, The power station level includes several grid-connected switchgear and several power station grid-connected controllers; The power station grid-connected controller, as the top-level control unit, is responsible for the overall regulation of the power output of the entire station, realizing the safe grid connection and power dispatch of the power station with the external power grid.

5. The intelligent grid-connected control method for a multi-unit gas-fired power plant as described in claim 1, characterized in that, The group-independent mode uses a single group as the scheduling unit to perform dynamic load averaging or proportional distribution and power management among several generator sets within the target group.

6. The intelligent grid-connected control method for a multi-unit gas-fired power plant as described in claim 1, characterized in that, The group parallel mode is a flexible combination of group circuit breakers and bus tie circuit breakers between groups to achieve system-controlled operation and multiple power output modes.

7. The intelligent grid-connected control method for a multi-unit gas-fired power plant as described in claim 1, characterized in that, The high-voltage side grid connection mode is a process in which the power output of the group is integrated and scheduled by the grid connection controller to realize grid connection and off-grid operation with the external power grid.

8. A smart grid-connected control system for a multi-unit gas-fired power plant, characterized in that, include: The structure partitioning controller is configured to divide the generator group in the target power station into three levels in terms of electrical structure: unit level, group level and power station level; The control settings controller is configured to set three intelligent grid connection control modes based on the three levels: group independent mode, group parallel mode, and high-voltage side grid connection mode; The mode optimization controller is configured to perform cost optimization calculations on three control modes using a genetic algorithm, and dynamically switch to the optimal control mode based on real-time operating data.

9. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the intelligent grid-connected control method for a multi-unit gas-fired power plant as described in any one of claims 1-7.

10. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the intelligent grid-connected control method for a multi-unit gas-fired power plant as described in any one of claims 1-7.

Images