Power distribution network voltage and frequency adaptive regulation and control system based on multi-source cooperation

Abstract

The invention provides a power distribution network voltage and frequency adaptive regulation and control system based on multi-source cooperation, and belongs to the technical field of intelligent power grid and power system automation, and the system comprises an equipment layer which comprises a distributed power supply, a hybrid energy storage system, an on-load voltage regulation transformer, a switchable capacitor bank and an intelligent load; the edge control layer is an edge controller deployed at each power supply node and is used for calculating a local voltage out-of-limit index and a frequency out-of-limit index in real time; the centralized optimization layer is used for predicting a controlled central server by taking a real-time calculation result as the input of a new energy output prediction model, generating a global optimization instruction and issuing the global optimization instruction to the equipment layer for control execution; and the communication layer is used for realizing data interaction between the edge control layer and the centralized optimization layer based on 5G communication technology and optical fiber hybrid networking. Dynamic balance control of voltage and frequency is realized, and network loss and equipment action cost are reduced.

Description

Technical Field

[0001] The present invention relates to the field of smart grid and power system automation technology, and in particular to a distribution network voltage and frequency adaptive control system based on multi-source collaboration. Background Art

[0002] With the large-scale access of distributed energy to the distribution network, its volatility and uncertainty have caused severe challenges to the operation of the power grid.

[0003] Voltage and frequency coupling: Power fluctuations from distributed power sources (such as photovoltaic inverters) simultaneously cause voltage and frequency deviations, making it difficult to address both with traditional single-step regulation methods (such as voltage or frequency regulation). Existing technologies lack coordination between energy storage systems and on-load tap changers (OLTCs), potentially exacerbating grid losses due to opposing regulation directions (for example, when energy storage discharge raises voltage, the OLTC lowers the tap).

[0004] Therefore, the present invention proposes a distribution network voltage and frequency adaptive control system based on multi-source collaboration. Summary of the Invention

[0005] The present invention provides a distribution network voltage and frequency adaptive control system based on multi-source collaboration, which is used to achieve dynamic balanced control of voltage and frequency while reducing network losses and equipment operation costs.

[0006] Equipment layer: including distributed power supply, hybrid energy storage system, on-load tap-changing transformer, switchable capacitor bank and intelligent load; Edge control layer: An edge controller deployed at each power node is used to calculate local voltage and frequency over-limit indicators in real time. When the edge controller detects that the voltage has dropped to the product of a preset coefficient and a rated value and the frequency has not yet exceeded the limit, it switches all switchable capacitor banks and sends a reactive power priority instruction to the hybrid energy storage system. At this time, if the locally regulated voltage is still calculated to be undervoltage, the centralized optimization layer is reported as an emergency voltage over-limit and the current device status. Centralized optimization layer: The real-time calculation results are used as the input of the new energy output prediction model to predict the central server of the control, generate global optimization instructions, and send them to the equipment layer for control execution. In combination with the whole network power flow calculation, if the voltage at the end of the feeder in the corresponding area drops synchronously, it is judged that the line impedance is too large, and a global optimization instruction is generated. The global optimization instruction is transmitted to the edge controller and executed in the order of capacitor bank switching, on-load tap-changing transformer, and intelligent load control. After each adjustment step is completed, the real-time voltage data is fed back to the centralized optimization layer until the voltage returns to the specified value, entering the secondary adjustment stage; Communication layer: Based on 5G communication technology and optical fiber hybrid networking, data interaction between the edge control layer and the centralized optimization layer is realized.

[0007] Preferably, the edge controller uses Kalman filtering to pre-process the electrical quantity of the corresponding power supply node, and uploads it to the centralized optimization layer based on the communication layer.

[0008] Preferably, the centralized optimization layer: dynamically adjusts the optimization weights of voltage and frequency based on the uploaded pre-processed data and through fuzzy logic; The new energy processing prediction model is updated according to the optimization weight, wherein the objective function of the new energy processing prediction model is:

[0009] Where: α(t) and β(t) are the optimization weights at time t, which are adaptively adjusted according to the new energy penetration rate ρ(t). Ploss is the total active power loss of the distribution network. is the energy storage operation cost, T represents the total number of time t; N is the total number of data sampling moments; is the voltage deviation collected for the i-th time at time t; is the frequency deviation of the ith acquisition at time t; are adjustment coefficients respectively; min is the minimum value symbol.

[0010] Preferably, the constraints of the objective function are: Voltage safety limit: ; Frequency Deviation: ; Energy storage charging and discharging power: ; in, is the voltage collected for the i-th time at time t; is the rated voltage; is the power variable; is the minimum power; is the maximum power.

[0011] Preferably, the edge controller includes: A determining unit: determining a characteristic set presented by the electrical quantity of each power node, wherein the characteristic set includes a plurality of first characteristics; A first sorting unit: performing a first sorting according to the feature weight of the first characteristic to construct a first Kalman filter model; A second sorting unit: performing a second sorting according to the occurrence frequency of each first characteristic in the characteristic set in all power supply nodes, and constructing a second Kalman filter model; Estimation unit: Inputs the electrical quantity data monitored in real time by the corresponding power node into different Kalman filter models to obtain the corresponding state estimation results; The first increasing unit: when the fluctuation type is aggravated fluctuation and the state estimation result is qualified, the process noise covariance Q is calculated according to A first increase is performed, wherein is the data variance of the corresponding electrical quantity data; Setting thresholds for qualified performance and increased fluctuation type; The first reduction unit: when the fluctuation type is stable fluctuation and the state estimation result is qualified, the measurement noise covariance R is calculated according to performing a first reduction; The second increasing unit: when the fluctuation type is aggravated fluctuation and the state estimation result is unqualified, the process noise covariance Q is calculated according to Perform a second enlargement; The second reduction unit: when the fluctuation type is stable fluctuation and the state estimation result is qualified, the measurement noise covariance R is calculated according to performing a second reduction; Preprocessing data is obtained based on the acquisition results of the first Kalman filter model and combined with the corresponding change results.

[0012] Preferably, the centralized optimization layer includes: Set construction unit: constructs voltage deviation rate sets and frequency deviation rate sets at different times based on preprocessed data; Weight optimization unit: The voltage deviation rate variable and frequency deviation rate variable under each deviation rate centralized comparison number are input into the fuzzy logic system respectively, and the latest voltage optimization weight and the latest frequency optimization weight under the corresponding number of times are obtained to realize dynamic adjustment.

[0013] Preferably, the centralized optimization layer further includes: A server determination unit: determines each central server to be controlled, which is regarded as a first server; Priority determination unit: determines the instruction priority of each optimization instruction according to the optimization instruction and service function of each first server, and performs priority sorting to obtain a global optimization instruction.

[0014] Preferably, the priority determination unit includes: According to the detailed description and applicable scenario of the optimization instruction of each first server, the first coefficient is matched with the functional description and functional scenario of the service function of each first server respectively;

[0015] in, Indicates a detailed description based on the corresponding optimization instruction Functional description of the corresponding service function The similarity function of , the value range is (0, 1); Indicates the applicable scenarios based on the corresponding optimization instructions Functional scenarios corresponding to service functions The similarity function of , the value range is (0, 1); Indicates the minimum value symbol; ln indicates the logarithmic function symbol; a first coefficient representing a matching relationship between the corresponding optimization instruction and the j-th service function in the first server; determining a priority value based on all first coefficients under the corresponding optimization instruction;

[0016] in, Indicates the priority value of the corresponding optimization instruction; represents the maximum coefficient among all first coefficients related to the j-th service function in all first servers; Indicates the total number of service functions corresponding to the first server; Match the instruction priority of each priority value from the value-priority comparison table; All optimization instructions are sorted in order of priority to obtain global optimization instructions.

[0017] Compared with the prior art, the present invention has the following advantages: Achieve dynamic balanced control of voltage and frequency while reducing network losses and equipment operation costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 This is a structural diagram of a distribution network voltage and frequency adaptive control system based on multi-source collaboration provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0021] The present invention provides a distribution network voltage and frequency adaptive control system based on multi-source collaboration, such as Figure 1 As shown, including: Equipment layer: including distributed power supply, hybrid energy storage system, on-load tap-changing transformer, switchable capacitor bank and intelligent load; Edge control layer: An edge controller deployed at each power node is used to calculate local voltage and frequency over-limit indicators in real time. When the edge controller detects that the voltage has dropped to the product of a preset coefficient and a rated value and the frequency has not yet exceeded the limit, it switches all switchable capacitor banks and sends a reactive power priority instruction to the hybrid energy storage system. At this time, if the locally regulated voltage is still calculated to be undervoltage, the centralized optimization layer is reported as an emergency voltage over-limit and the current device status. Centralized optimization layer: The real-time calculation results are used as the input of the new energy output prediction model to predict the central server of the control, generate global optimization instructions, and send them to the equipment layer for control execution. In combination with the whole network power flow calculation, if the voltage at the end of the feeder in the corresponding area drops synchronously, it is judged that the line impedance is too large, and a global optimization instruction is generated. The global optimization instruction is transmitted to the edge controller and executed in the order of capacitor bank switching, on-load tap-changing transformer, and intelligent load control. After each adjustment step is completed, the real-time voltage data is fed back to the centralized optimization layer until the voltage returns to the specified value, entering the secondary adjustment stage; Communication layer: Based on 5G communication technology and optical fiber hybrid networking, data interaction between the edge control layer and the centralized optimization layer is realized.

[0022] Preferably, the edge controller uses Kalman filtering to pre-process the electrical quantity of the corresponding power supply node, and uploads it to the centralized optimization layer based on the communication layer.

[0023] Preferably, the centralized optimization layer: dynamically adjusts the optimization weights of voltage and frequency based on the uploaded pre-processed data and through fuzzy logic; The new energy processing prediction model is updated according to the optimization weight, wherein the objective function of the new energy processing prediction model is:

[0024] Where: α(t) and β(t) are the optimization weights at time t, which are adaptively adjusted according to the new energy penetration rate ρ(t). Ploss is the total active power loss of the distribution network. is the energy storage operation cost, T represents the total number of time t; N is the total number of data sampling moments; is the voltage deviation collected for the i-th time at time t; is the frequency deviation of the ith acquisition at time t; are adjustment coefficients respectively; min is the minimum value symbol.

[0025] Preferably, the constraints of the objective function are: Voltage safety limit: ; Frequency Deviation: ; Energy storage charging and discharging power: ; in, is the voltage collected for the i-th time at time t; is the rated voltage; is the power variable; is the minimum power; is the maximum power.

[0026] In this embodiment, a multi-timescale coordinated control strategy is adopted, where supercapacitors perform second-level response, OLTC and capacitor banks are regulated at the minute level, and demand-side response is initiated at the hour level; The data-physics fusion prediction method combines LSTM and physical models to improve the accuracy of renewable energy output prediction.

[0027] In this embodiment, distributed power sources (photovoltaic, wind power), energy storage systems (lithium battery + supercapacitor hybrid energy storage), OLTC, switchable capacitor banks, smart loads, etc.

[0028] In this embodiment, the central server (deployed at the main station of the distribution network) generates rolling optimization instructions based on the global topology, load forecast, and weather data (sunlight / wind speed).

[0029] In this embodiment, the distributed power source is such as a solar photovoltaic panel, a small wind turbine, etc.; Hybrid energy storage systems such as battery energy storage (lithium-ion batteries, lead-acid batteries, etc.) and supercapacitors; On-load tap-changing transformers can adjust the voltage ratio of the transformer under load; The switchable capacitor bank consists of multiple capacitors, and some capacitors can be switched on or off as needed to adjust reactive power. Smart loads are load devices with intelligent control functions that can adjust their power consumption behavior according to the grid status and their own needs.

[0030] In this embodiment, the edge controller is deployed near the power node and is a device that processes and controls local data in real time.

[0031] In this embodiment, the renewable energy output prediction model uses historical data and real-time monitoring data to predict the power generation of renewable energy sources (such as solar and wind energy). Taking a wind farm as an example, the model combines real-time meteorological data such as wind speed, wind direction, and temperature with historical wind turbine power generation data to predict the wind turbine's power generation over a period of time.

[0032] In this embodiment, the central server collects data such as voltage over-limit, frequency over-limit, and renewable energy power generation forecasts uploaded by each edge controller, and generates global optimization instructions after comprehensive analysis, such as adjusting the tap position of the on-load tap-changing transformer and controlling the switching of the switchable capacitor bank.

[0033] In this embodiment, when the 5G network coverage is good and the optical fiber is working normally, the data transmission delay can be controlled at the millisecond level and the packet loss rate is less than 0.1%.

[0034] In this embodiment, the rated voltage is 10 kV and the rated frequency is 50 Hz.

[0035] The beneficial effect of this technical solution is that, through a layered architecture of "edge rapid response + centralized global optimization," combined with dynamic weight adjustment and multi-timescale coordination, it addresses the voltage and frequency instability issues caused by a high proportion of renewable energy access. The system supports local autonomy during communication outages, reducing network losses by over 20% and significantly improving the economic efficiency and safety of distribution network operations.

[0036] The present invention provides a distribution network voltage and frequency adaptive control system based on multi-source collaboration, wherein the edge controller includes: A determining unit: determining a characteristic set presented by the electrical quantity of each power node, wherein the characteristic set includes a plurality of first characteristics; A first sorting unit: performing a first sorting according to the feature weight of the first characteristic to construct a first Kalman filter model; A second sorting unit: performing a second sorting according to the occurrence frequency of each first characteristic in the characteristic set in all power supply nodes, and constructing a second Kalman filter model; Estimation unit: Inputs the electrical quantity data monitored in real time by the corresponding power node into different Kalman filter models to obtain the corresponding state estimation results; The first increasing unit: when the fluctuation type is aggravated fluctuation and the state estimation result is qualified, the process noise covariance Q is calculated according to A first increase is performed, wherein is the data variance of the corresponding electrical quantity data; Setting thresholds for qualified performance and increased fluctuation type; The first reduction unit: when the fluctuation type is stable fluctuation and the state estimation result is qualified, the measurement noise covariance R is calculated according to performing a first reduction; The second increasing unit: when the fluctuation type is aggravated fluctuation and the state estimation result is unqualified, the process noise covariance Q is calculated according to Perform a second enlargement; The second reduction unit: when the fluctuation type is stable fluctuation and the state estimation result is qualified, the measurement noise covariance R is calculated according to performing a second reduction; Preprocessing data is obtained based on the acquisition results of the first Kalman filter model and combined with the corresponding change results.

[0037] In this embodiment, the characteristic set refers to a set of a series of characteristics presented by the electrical quantity of the power node. These characteristics can reflect the changing laws and stability of the electrical quantity. For example, for a certain power node, the characteristic set of its electrical quantity may include characteristics such as the voltage fluctuation amplitude, the current change trend, and the stability of the power factor, and the first characteristic is a specific characteristic item in the characteristic set. Taking voltage as an example, the voltage fluctuation amplitude is a first characteristic. Suppose that at a power node in a microgrid, after a period of monitoring, it is found that its voltage fluctuation amplitude is between ±5%. This fluctuation amplitude is a first characteristic in the electrical quantity characteristic set of the power node. During verification, long-term monitoring data can be used to observe whether the characteristic exists stably. If the voltage fluctuation amplitudes obtained from multiple measurements are all around ±5%, it means that the determination of the characteristic is accurate.

[0038] In this embodiment, the sum of the feature weights is 1, and the weights of different features are different, but are all set in advance.

[0039] In this embodiment, the first Kalman filter model is a model for processing electrical quantity data, constructed by sorting the first characteristics according to their characteristic weights. For example, at a power node containing three first characteristics (voltage, current, and power factor), these three characteristics are sorted according to their characteristic weights, with the voltage characteristic (with the highest weight) being placed first in the model input, followed by current and power factor. During verification, known electrical quantity data is input and the deviation between the state estimation result output by the model and the actual situation is observed. If the deviation is within an acceptable range, such as the deviation between the estimated voltage value and the actual value is within ±1%, it indicates that the model is reasonably constructed and effective.

[0040] In this embodiment, the occurrence frequency refers to the number of times each first characteristic in the characteristic set appears across all power nodes. For example, in a power grid supplied by multiple cells, monitoring of each power node reveals that the first characteristic "large voltage fluctuation amplitude" appears in 6 out of 10 power nodes. Therefore, the occurrence frequency is 6.

[0041] In this embodiment, the second Kalman filter model is constructed by ranking the frequency of occurrence of each first characteristic across all power nodes. In the aforementioned residential grid example, the characteristic of "large voltage fluctuations," which occurs frequently, is given higher priority in model construction to construct the second Kalman filter model. During verification, actual electrical quantity data is also input, and the model output is compared with the actual situation. If the model can effectively handle electrical quantity characteristics common to most power nodes, such as accurately estimating the voltage status for multiple power nodes experiencing large voltage fluctuations, the model demonstrates its practicality.

[0042] In this embodiment, real-time voltage and current data for a power node are input into the first Kalman filter model. The model outputs an estimated voltage of 220.5V and an estimated current of 5A for the power node at the current moment. This is the state estimation result. During verification, the state estimation result output by the model can be compared with the actual electrical quantity data measured by a high-precision measuring instrument. If the deviation between the two is small over a period of time, such as within ±0.5V for voltage and ±0.1A for current, the estimation unit is functioning properly and the model estimation is accurate.

[0043] In this embodiment, increasing fluctuations means that the amplitude of the change in the electrical quantity gradually increases, for example, the voltage of the power node gradually changes from a fluctuation of ±5% to ±10% over a period of time. Stable fluctuations mean that the amplitude of the fluctuation of the electrical quantity is relatively stable, for example, the voltage always fluctuates around ±3%.

[0044] In this embodiment, when the fluctuation type is exacerbated and performance is acceptable, the process noise covariance Q is first increased according to the formula, where is the data variance of the corresponding electrical quantity data, and is the threshold value for acceptable performance and exacerbated fluctuation. For example, if is set to 0.05, and the variance of the voltage data at a power node is 0.1, and the fluctuation is exacerbated and performance is acceptable, the process noise covariance Q is increased according to the formula to enhance the Kalman filter's ability to track system state changes. During verification, the Kalman filter's processing effect on the electrical quantity data after increasing the Q value is observed. If the Kalman filter can more accurately track the changing trend of the electrical quantity, the operation of the first increase unit is effective.

[0045] In this embodiment, when the power node electrical quantity fluctuations are stable and performance is acceptable, the measurement noise covariance R is reduced, making the filter more reliant on the measured data and improving estimation accuracy. During verification, the model's estimation accuracy of the electrical quantity data is compared before and after reducing R. If the deviation between the estimated and actual values ​​decreases, such as when the voltage estimation deviation decreases from ±1V to ±0.5V, the first reduction unit operation is reasonable and effective.

[0046] In this embodiment, if the voltage is less than 210V or greater than 230V, and the frequency is less than 49.5Hz or greater than 50.5Hz, performance is considered unsatisfactory. When the fluctuation type is exacerbated and performance is unsatisfactory, the process noise covariance Q is increased in a second step. This is similar to the first step, but the magnitude or method of increase may differ to accommodate more complex situations. During verification, the Kalman filter is observed to better handle electrical quantity data with unsatisfactory performance and exacerbated fluctuations after increasing the Q value in this case. If the estimated result more closely matches the actual variation trend, the second step is operating correctly.

[0047] The beneficial effect of this technical solution is that the first Kalman filter model adjusts the process noise covariance Q and measurement noise covariance R based on different electrical quantity characteristics, fluctuation types, and performance conditions. Ultimately, the model outputs processed electrical quantity data, known as preprocessed data, which ensures data reliability and enables more accurate analysis and control of subsequent power systems.

[0048] The present invention provides a distribution network voltage and frequency adaptive control system based on multi-source collaboration, wherein the centralized optimization layer includes: Set construction unit: constructs voltage deviation rate sets and frequency deviation rate sets at different times based on preprocessed data; Weight optimization unit: The voltage deviation rate variable and frequency deviation rate variable under each deviation rate centralized comparison number are input into the fuzzy logic system respectively, and the latest voltage optimization weight and the latest frequency optimization weight under the corresponding number of times are obtained to realize dynamic adjustment.

[0049] In this embodiment, the voltage deviation rate set is a collection of voltage deviation rates in the power system at different times. The voltage deviation rate refers to the difference between the actual voltage and the rated voltage as a percentage of the rated voltage. For example, if the actual voltage at a point in the power system is 225V and the rated voltage is 220V at a certain moment, the voltage deviation rate at that moment is (225-220) ÷ 220 × 100% ≈ 2.27%. Collecting these voltage deviation rates at multiple different times constitutes a voltage deviation rate set.

[0050] In this embodiment, the frequency deviation rate set is a collection of frequency deviation rates in the power system at different times. The frequency deviation rate refers to the difference between the actual frequency and the rated frequency as a percentage of the rated frequency. For example, if the rated frequency of my country's power system is 50 Hz and the actual measured frequency at a certain moment is 50.5 Hz, then the frequency deviation rate at that moment is (50.5 - 50) ÷ 50 × 100% = 1%. Multiple frequency deviation rates at different times constitute a frequency deviation rate set.

[0051] The number of centralized deviation rate comparisons is the number of times the voltage deviation rate and frequency deviation rate are observed and recorded within a statistical period. For example, if the voltage deviation rate and frequency deviation rate are recorded every 10 minutes within an hour, there will be six comparisons within that hour.

[0052] The voltage deviation rate variable and frequency deviation rate variable refer to the specific voltage deviation rate value and frequency deviation rate value for each comparison in the voltage deviation rate set and frequency deviation rate set, respectively. For example, in the above example, if the voltage deviation rate of the first record is 2.27%, then this 2.27% is a voltage deviation rate variable; if the frequency deviation rate of the first record is 1%, then this 1% is a frequency deviation rate variable.

[0053] A fuzzy logic system is a system based on fuzzy set theory and fuzzy logic reasoning, capable of processing uncertain and imprecise information. In a power system, it can derive corresponding optimization weights based on the input voltage deviation rate and frequency deviation rate variables through a series of fuzzy rules and reasoning. For example, when the voltage deviation rate is large and the frequency deviation rate is small, the fuzzy logic system may, based on preset rules, assign a larger voltage optimization weight and a smaller frequency optimization weight to adjust the power system's control strategy.

[0054] The latest voltage optimization weight and the latest frequency optimization weight are weight coefficients calculated by the fuzzy logic system based on the input deviation rate variable, used to adjust the voltage and frequency control in the power system. These weights are dynamically adjusted according to the changes in the deviation rate at different times. For example, at a certain moment, based on the voltage deviation rate and frequency deviation rate at that time, the fuzzy logic system calculates the latest voltage optimization weight to be 0.6 and the latest frequency optimization weight to be 0.4. This means that in the control of the power system at that moment, voltage control is relatively important, accounting for 60%, while frequency control accounts for 40%. By continuously and dynamically adjusting these weights, the power system can operate more stably and efficiently.

[0055] Verification result example: Assume that the above solution is verified in a small power system. Over a period of time, the following data is recorded: Initially, the voltage deviation rate set is {1.5%, 2.0%, 1.8%}, and the frequency deviation rate set is {0.5%, 0.8%, 0.6%}.

[0056] After calculation by the fuzzy logic system, the initial latest voltage optimization weight is 0.55, and the latest frequency optimization weight is 0.45.

[0057] Over time, the power system's operating state changes, with the newly recorded voltage deviation rate set becoming {2.5%, 2.2%, 2.0%} and the frequency deviation rate set becoming {0.3%, 0.4%, 0.5%}. These values ​​are then fed back into the fuzzy logic system for calculation, yielding the latest voltage optimization weights of 0.65 and frequency optimization weights of 0.35.

[0058] The beneficial effect of the above technical solution is that it is combined with a fuzzy logic system based on a set of voltage and frequency deviation rates to achieve dynamic adjustment of weights.

[0059] The present invention provides a distribution network voltage and frequency adaptive control system based on multi-source collaboration, wherein the centralized optimization layer further includes: A server determination unit: determines each central server to be controlled, which is regarded as a first server; Priority determination unit: determines the instruction priority of each optimization instruction according to the optimization instruction and service function of each first server, and performs priority sorting to obtain a global optimization instruction.

[0060] Preferably, the priority determination unit includes: According to the detailed description and applicable scenario of the optimization instruction of each first server, the first coefficient is matched with the functional description and functional scenario of the service function of each first server respectively;

[0061] in, Indicates a detailed description based on the corresponding optimization instruction Functional description of the corresponding service function The similarity function of , the value range is (0, 1); Indicates the applicable scenarios based on the corresponding optimization instructions Functional scenarios corresponding to service functions The similarity function of , the value range is (0, 1); Indicates the minimum value symbol; ln indicates the logarithmic function symbol; a first coefficient representing a matching relationship between the corresponding optimization instruction and the j-th service function in the first server; determining a priority value based on all first coefficients under the corresponding optimization instruction;

[0062] in, Indicates the priority value of the corresponding optimization instruction; represents the maximum coefficient among all first coefficients related to the j-th service function in all first servers; Indicates the total number of service functions corresponding to the first server; Match the instruction priority of each priority value from the value-priority comparison table; All optimization instructions are sorted in order of priority to obtain global optimization instructions.

[0063] In this embodiment, "shorten the response time of the product information query interface to less than 1 second during peak system access times" is an optimization instruction. The detailed description includes the optimization goal (shortening the response time) and means, and is applicable to the peak system access scenario.

[0064] In this embodiment, the functional description is a detailed description of the business logic and operational procedures implemented by the function; the functional scenario is the business scenario in which the function operates normally or takes effect. For example, the functional description of the "product information query" function of the product information management server will involve database query logic, etc., and the functional scenario may be when a user searches for products on an e-commerce platform.

[0065] In this embodiment, the value-priority comparison table is a pre-set table that corresponds priority values ​​to instruction priorities. Through this comparison table, the calculated priority value can be converted into a specific instruction priority, such as high, medium, low, etc.

[0066] In this embodiment, if after executing instruction 1, the performance of product information query in high-concurrency scenarios is significantly improved, such as the response time is shortened from the original 3 seconds to less than 1 second; then after executing instruction 2, the order processing process is smoother in daily operations and the processing efficiency is improved, it means that the scheme can effectively determine the priority of the optimization instructions, obtain a reasonable global optimization instruction sequence, and achieve the optimization goals of the system.

[0067] The beneficial effect of the above technical solution is: the instruction sequence obtained by sorting all optimization instructions according to the instruction priority determines the execution order of each optimization instruction in the system, which helps the system to perform optimization operations in an orderly manner and improve overall performance and efficiency.

[0068] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A distribution network voltage and frequency adaptive control system based on multi-source collaboration, characterized in that: include: Equipment layer: including distributed power supply, hybrid energy storage system, on-load tap-changing transformer, switchable capacitor bank and intelligent load; Edge control layer: An edge controller deployed at each power node is used to calculate local voltage and frequency over-limit indicators in real time. When the edge controller detects that the voltage has dropped to the product of a preset coefficient and a rated value and the frequency has not yet exceeded the limit, it switches all switchable capacitor banks and sends a reactive power priority instruction to the hybrid energy storage system. At this time, if the locally regulated voltage is still calculated to be undervoltage, the centralized optimization layer is reported as an emergency voltage over-limit and the current device status. Centralized optimization layer: The real-time calculation results are used as the input of the new energy output prediction model to predict the central server of the control, generate global optimization instructions, and send them to the equipment layer for control execution. In combination with the whole network power flow calculation, if the voltage at the end of the feeder in the corresponding area drops synchronously, it is judged that the line impedance is too large, and a global optimization instruction is generated. The global optimization instruction is transmitted to the edge controller and executed in the order of capacitor bank switching, on-load tap-changing transformer, and intelligent load control. After each adjustment step is completed, the real-time voltage data is fed back to the centralized optimization layer until the voltage returns to the specified value, entering the secondary adjustment stage; Communication layer: Based on 5G communication technology and optical fiber hybrid networking, data interaction between the edge control layer and the centralized optimization layer is realized.

2. The distribution network voltage and frequency adaptive control system based on multi-source collaboration according to claim 1 is characterized in that: The edge controller uses Kalman filtering to pre-process the electrical quantity of the corresponding power supply node and uploads it to the centralized optimization layer based on the communication layer.

3. The distribution network voltage and frequency adaptive control system based on multi-source collaboration according to claim 2 is characterized in that: The centralized optimization layer dynamically adjusts the optimization weights of voltage and frequency based on the uploaded pre-processed data and through fuzzy logic; The new energy processing prediction model is updated according to the optimization weight, wherein the objective function of the new energy processing prediction model is: Where: α(t) and β(t) are the optimization weights at time t, which are adaptively adjusted according to the new energy penetration rate ρ(t). Ploss is the total active power loss of the distribution network. is the energy storage operation cost, T represents the total number of time t; N is the total number of data sampling moments; is the voltage deviation collected for the i-th time at time t; is the frequency deviation of the ith acquisition at time t; are adjustment coefficients respectively; min is the minimum value symbol.

4. The distribution network voltage and frequency adaptive control system based on multi-source collaboration according to claim 3 is characterized in that: The constraints of the objective function are: Voltage safety limit: ; Frequency deviation: ; Energy storage charging and discharging power: ; in, is the voltage collected for the i-th time at time t; is the rated voltage; is the power variable; is the minimum power; is the maximum power.

5. The distribution network voltage and frequency adaptive control system based on multi-source collaboration according to claim 2 is characterized in that: The edge controller includes: A determining unit: determining a characteristic set presented by the electrical quantity of each power node, wherein the characteristic set includes a plurality of first characteristics; A first sorting unit: performing a first sorting according to the feature weight of the first characteristic to construct a first Kalman filter model; A second sorting unit: performing a second sorting according to the occurrence frequency of each first characteristic in the characteristic set in all power supply nodes, and constructing a second Kalman filter model; Estimation unit: Inputs the electrical quantity data monitored in real time by the corresponding power node into different Kalman filter models to obtain the corresponding state estimation results; The first increasing unit: when the fluctuation type is aggravated fluctuation and the state estimation result is qualified, the process noise covariance Q is calculated according to A first increase is performed, wherein is the data variance of the corresponding electrical quantity data; Setting thresholds for qualified performance and increased fluctuation type; The first reduction unit: when the fluctuation type is stable fluctuation and the state estimation result is qualified, the measurement noise covariance R is calculated according to performing a first reduction; The second increasing unit: when the fluctuation type is aggravated fluctuation and the state estimation result is unqualified, the process noise covariance Q is calculated according to Perform a second enlargement; The second reduction unit: when the fluctuation type is stable fluctuation and the state estimation result is qualified, the measurement noise covariance R is calculated according to performing a second reduction; Preprocessing data is obtained based on the acquisition results of the first Kalman filter model and combined with the corresponding change results.

6. The distribution network voltage and frequency adaptive control system based on multi-source collaboration according to claim 3 is characterized in that: The centralized optimization layer includes: Set construction unit: constructs voltage deviation rate sets and frequency deviation rate sets at different times based on preprocessed data; Weight optimization unit: The voltage deviation rate variable and frequency deviation rate variable under each deviation rate centralized comparison number are input into the fuzzy logic system respectively, and the latest voltage optimization weight and the latest frequency optimization weight under the corresponding number of times are obtained to realize dynamic adjustment.

7. The distribution network voltage and frequency adaptive control system based on multi-source collaboration according to claim 1 is characterized in that: The centralized optimization layer further includes: A server determination unit: determines each central server to be controlled, which is regarded as a first server; Priority determination unit: determines the instruction priority of each optimization instruction according to the optimization instruction and service function of each first server, and performs priority sorting to obtain a global optimization instruction.

8. The distribution network voltage and frequency adaptive control system based on multi-source collaboration according to claim 7 is characterized in that: The priority determination unit includes: According to the detailed description and applicable scenario of the optimization instruction of each first server, the first coefficient is matched with the functional description and functional scenario of the service function of each first server respectively; in, Indicates a detailed description based on the corresponding optimization instruction Functional description of the corresponding service function The similarity function of , the value range is (0, 1); Indicates the applicable scenarios based on the corresponding optimization instructions Functional scenarios corresponding to service functions The similarity function of , the value range is (0, 1); Indicates the minimum value symbol; ln indicates the logarithmic function symbol; a first coefficient representing a matching relationship between the corresponding optimization instruction and the j-th service function in the first server; determining a priority value based on all first coefficients under the corresponding optimization instruction; in, Indicates the priority value of the corresponding optimization instruction; represents the maximum coefficient among all first coefficients related to the j-th service function in all first servers; Indicates the total number of service functions corresponding to the first server; Match the instruction priority of each priority value from the value-priority comparison table; All optimization instructions are sorted in order of priority to obtain global optimization instructions.

Images