On-line evaluation method and system for grid-connected capacity margin and trend of Sagomean new energy base

Abstract

The invention discloses an online evaluation method and system for the grid-connected capacity margin and trend of a Sagomean new energy base. According to the method, on the basis of meteorological data, power grid operation data and equipment state data which are collected in real time, the real-time grid-connected capacity margin of the new energy base is calculated through multi-source data fusion and dynamic modeling, and the short-term change trend of the new energy base is predicted through a time sequence analysis and machine learning algorithm. The system comprises a data acquisition layer, a fusion analysis layer, an evaluation prediction layer and a visual interaction layer, and full-process automation from data to decision is realized. According to the invention, the new energy consumption capability of the Sagomean region can be improved, the safety and economy of power grid operation are enhanced, and a scientific basis is provided for scheduling decision making.

Description

Technical Field

[0001] This invention belongs to the field of new energy grid connection technology, specifically involving an online assessment technology for the grid-connectable capacity margin of new energy bases and its changing trend. Background Technology

[0002] Against the backdrop of the global energy transition, the Gobi Desert region, with its vast land resources and abundant solar and wind energy, has become a strategic location for the construction of centralized photovoltaic and wind power projects. my country has planned large-scale wind and photovoltaic power bases in the Gobi Desert with a capacity of hundreds of millions of kilowatts. However, the grid connection and operation of these bases face a series of world-class technical challenges. The core contradiction lies in the conflict between the strong uncertainty and volatility of renewable energy output and the limited and rigid carrying capacity of the power grid itself. Traditional methods for assessing grid-connectable capacity have proven severely inadequate in such scenarios.

[0003] First, from the perspective of resource and grid characteristics, the new energy base in the desert and Gobi region exhibits distinct "two highs and two weaknesses": The "two highs" refer to high resource volatility and high equipment concentration. The region's complex meteorological conditions make it susceptible to sandstorms, high temperatures, strong radiation, and drastic temperature differences, causing photovoltaic module output power to vary non-linearly with sunlight, temperature, and dust cover. Wind turbine output is significantly affected by turbulent winds and extreme wind speed shear, with minute-level and hourly power spikes and drops becoming commonplace. Simultaneously, the dense deployment of thousands of power generation devices amplifies the volatility due to their aggregation effect, creating significant uncertainty on the "source" side. The "two weaknesses" refer to a weak grid structure and weak regulation capacity. The desert and Gobi region is typically located at the end of the grid, with a sparse grid structure, limited transmission channels, low short-circuit capacity, and poor voltage support. Furthermore, local load levels are low, lacking sufficient conventional spinning reserve and rapid regulation resources (such as gas turbines and hydropower), resulting in severely insufficient flexibility on both the "load" and "storage" sides of the grid. This coupling of a "strong fluctuation source" and a "weak load-bearing network" makes the real-time safe operation boundary of the power grid extremely narrow and dynamically time-varying.

[0004] Secondly, existing assessment methods have significant limitations, mainly in three aspects: staticity, offline calculation, and one-sidedness. Firstly, traditional methods often rely on offline calculations based on typical days, typical scenarios, or worst-case conditions to determine a fixed "upper limit of installed capacity" or "guaranteed output curve." This method completely fails to reflect the dynamic changes in actual renewable energy output at the minute and hourly levels, as well as real-time changes in grid operating status (such as a line being shut down for maintenance or random load fluctuations). The results are either overly conservative, leading to significant wind and solar power curtailment and reduced economic benefits, or overly aggressive, creating potential safety hazards such as grid over-limitation or even grid collapse. Secondly, relying on manual periodic calculations (e.g., annually, monthly, daily) results in a response speed measured in "days," which cannot meet the decision-making needs of real-time grid dispatch (measured in "minutes" or "seconds"). When encountering sudden weather events (such as a rapid sandstorm causing a sharp drop in photovoltaic output), offline plans become completely ineffective, and dispatchers can only rely on experience for emergency response, which carries extremely high risks. Thirdly, traditional assessments often focus only on a single stability limit, such as thermal stability limits or static voltage safety constraints. However, the grid-connected capacity margin of the Shagohuang renewable energy base is a complex indicator influenced by a combination of factors across multiple dimensions and time scales. It requires comprehensive consideration of at least: a) static safety constraints (line / transformer thermal stability, node voltage upper and lower limits); b) transient stability constraints (the impact of large-scale renewable energy disconnection on system frequency and voltage transients); c) power quality constraints (such as harmonics, flicker, and especially the risk of broadband oscillations caused by the connection of numerous power electronic devices); d) equipment operation constraints (inverter over-temperature derating, wind turbine overspeed protection, energy storage SOC limits); and e) environmental factors (temperature-induced correction of line current carrying capacity, dust-induced degradation of photovoltaic panel efficiency, and wind speed-induced control of wind turbine start-up and shutdown). Existing methods lack a unified, dynamic model to integrate all of these factors.

[0005] Therefore, power grid operators have long faced the dilemma of being "unable to see, unable to estimate accurately, and unable to control": they cannot see in real time how much renewable energy the grid can currently absorb; they cannot accurately predict how this absorption capacity will change in the future; thus, they find it difficult to precisely control the output plans and energy storage operations of renewable energy plants. This directly leads to high curtailment rates under conservative dispatching or safety risks under risky operation, severely restricting the efficiency of the Shagohuang renewable energy base and the achievement of the national carbon emission reduction strategy. To solve the above problems, it is urgent to invent a technical solution that can online, dynamically, and comprehensively assess the grid-connectable capacity margin of the Shagohuang renewable energy base and predict its trend, transforming the grid's "implicit" safety boundary into "explicit," quantifiable, and trend-predictive decision indicators, providing core data-driven solutions for intelligent dispatching, market trading, and energy storage synergy. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide an online assessment method and system for the grid-connectable capacity margin and trend of the Shagohuang new energy base, so as to achieve dynamic, accurate and visualized assessment and prediction.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: an online assessment method for the grid-connectable capacity margin and trend of the Shagohuang new energy base, step 1, multi-source data acquisition, including meteorological monitoring data, power grid monitoring data, new energy station monitoring data, energy storage power station monitoring data, and management data; Step 2, data fusion processing: the multi-source data from Step 1 is spatiotemporally aligned, the data is cleaned, outliers are identified and removed, missing data is filled, key features are extracted, and normalization processing is implemented to form a unified data pool. Step 3, Dynamic Modeling and Evaluation: Establish an online, real-time updated model. Using the current grid operating point as a benchmark, with the total output increment ΔP of the new energy base as the optimization variable, and max ΔP as the objective function, the constraints include: Power flow equality constraints: f(P, Q, V, θ) = 0; P is the dielectric injected active power, Q is the node injected reactive power, V is the node voltage amplitude, θ is the node voltage phase angle, and f is the power flow equation vector function; Inequality safety constraints: V min ≤ V ≤ V max , |P line | ≤ P line-max ; V min It is the minimum allowable voltage amplitude at the node. V max It is the maximum allowable voltage amplitude at the node. |P line | This refers to the actual active power transmitted through the transmission line / channel. P line-max This is the maximum permissible active power transmitted by the line; equipment capacity constraint: 0 ≤ P _new_i ≤P _avail_i ;P _new_i Let P be the actual grid-connected active power of the i-th new energy device. _avail_i Let P be the current maximum available active power of the i-th renewable energy device; stability boundary constraint: ΔP ≤ P _critical ;P _critical The maximum active power disturbance threshold that the system can withstand to maintain stability is determined; then the theoretical maximum grid-connected capacity at the current moment is calculated, and the current real-time grid-connected capacity margin is determined. Step 4, Trend Forecasting Analysis, combines data-driven direct forecasting with model-driven rolling forecasting. Data-driven direct forecasting is based on historical margin time series, combined with predicted illumination and wind speed data, and input into a neural network for calculation to obtain the predicted capacity margin value and its confidence interval for every 15 minutes in the next 2 hours. Model-driven rolling forecasting uses predicted weather and load data for each future time as input, and performs rolling calculations on the dynamic modeling evaluation in Step 3 to obtain the predicted capacity margin change trend for the next 15 minutes to 24 hours. Step 5, Visualization and Alarms: Display graphs related to capacity margin changes through a web interface and issue alarms when the preset threshold is exceeded.

[0008] Specifically, the meteorological monitoring data mentioned in step 1 includes the collection of wind speed, wind direction, light intensity, ambient temperature, humidity, air pressure, dust concentration, satellite cloud image data, radar data, and numerical weather prediction microscale correction data; the power grid monitoring data includes the voltage, phase angle, active / reactive power flow of lines / transformers, switch status, protection signals, and system frequency of each node in the power grid; the new energy power station monitoring data includes real-time active / reactive power output, generator terminal voltage, current, equipment internal temperature, fault codes, available capacity, and control mode; the energy storage power station monitoring data includes the real-time charging and discharging power, SOC, health status, maximum charging and discharging capacity, voltage, and current of the energy storage power station and each battery cluster; and the management data includes planned maintenance information, load forecast curves, and market transaction plans.

[0009] Preferably, in step 2, the spatiotemporal alignment uses the same timestamp of the PTP synchronous clock to implement data unification with the same spatial reference system.

[0010] Preferably, in step 2, the data cleaning adopts a state estimation residual analysis model or an isolated forest algorithm; the missing data is filled by interpolation, regression imputation based on correlation, or deep learning generation.

[0011] Preferably, the key features in step 2 include line load rate, voltage deviation rate, instantaneous value of new energy penetration rate, and conversion factor of meteorological factors to power output.

[0012] Preferably, the method for calculating the current margin in step 3 is linear programming, quadratic programming, or interior point method.

[0013] The graphs related to capacity margin changes in step 5 include real-time capacity margin values, historical and predicted trend curves, heat maps of power flow and margin distribution at key grid sections, and meteorological information overlay maps. Alarm methods include audible and visual alarms, system pop-ups, and SMS messages.

[0014] An online assessment system for the grid-connectable capacity margin and trend of the Shagohuang new energy base, implementing an online assessment method for the grid-connectable capacity margin and trend of the Shagohuang new energy base, includes a data acquisition layer, a data fusion and processing layer, an assessment and prediction layer, and an application and display layer, wherein the data acquisition layer collects multi-source data; The data fusion and processing layer cleans, aligns, and merges multi-source data to unify the data. The assessment and prediction layer includes a dynamic assessment module and a trend prediction module. The dynamic assessment module calculates the current capacity margin based on an online real-time updated model, and the trend prediction module predicts the trend of capacity margin changes. The application and presentation layer provides a visual representation of capacity margins and trends, and issues alerts for exceeding threshold ranges.

[0015] Furthermore, the application and presentation layer includes a report generation module and an external transmission interface. The report generation module automatically generates an assessment report, and the external transmission interface is used to push the margin assessment and prediction results to the upper-level dispatch system, automatic generation control system, energy storage coordination control system, and power trading platform.

[0016] The advantages of this invention are: overcoming the shortcomings of static, offline, and one-sided evaluation in existing technologies, realizing real-time evaluation and trend prediction of the grid-connectable capacity of the Shagohuang new energy base, improving the new energy absorption capacity, reducing wind and solar curtailment, enhancing grid operation safety, preventing equipment overload and voltage exceedance, and providing data support for grid dispatching, energy storage control, and trading plans. It features real-time online operation, dynamic self-adaptation, and multi-dimensional integration. Attached Figure Description

[0017] Figure 1 This describes the visual interface and the display of the operating results of the system of the present invention; Figure 2 This is the power grid constraint regulation and state diagram of the system of the present invention; Figure 3 This refers to the grid-connected capacity margin displayed by the system of this invention at a certain moment; Figure 4 This is the evaluation parameter adjustment page of the system of this invention; Figure 5 It is the change in margin and curtailment rate curves of historical and predicted trends. Detailed Implementation

[0018] An online assessment method for the grid-connectable capacity margin and trend of the Shagohuang new energy base, step 1: multi-source data collection, including meteorological monitoring data, power grid monitoring data, new energy station monitoring data, energy storage power station monitoring data, and management data; Step 2, data fusion processing: the multi-source data from Step 1 is spatiotemporally aligned, the data is cleaned, outliers are identified and removed, missing data is filled, key features are extracted, and normalization processing is implemented to form a unified data pool. Step 3, Dynamic Modeling and Evaluation: Establish an online, real-time updated model. Using the current grid operating point as a benchmark, the total output increment ΔP of the new energy base is used as the optimization variable, and max ΔP is the objective function. Constraints include: power flow equality constraint: f(P, Q, V, θ) = 0; inequality safety constraints: V min ≤ V ≤ V max , |P line | ≤ P line-max Equipment capacity constraint: 0 ≤ P _new_i ≤P _avail_i Stability boundary constraint: ΔP≤P _critical Then, calculate the theoretical maximum grid-connected capacity at the current moment, and calculate the current real-time grid-connected capacity margin. Step 4, Trend Forecasting Analysis, combines data-driven direct forecasting with model-driven rolling forecasting. Data-driven direct forecasting is based on historical margin time series, combined with predicted illumination and wind speed data, and input into a neural network for calculation to obtain the predicted capacity margin value and its confidence interval for every 15 minutes in the next 2 hours. Model-driven rolling forecasting uses predicted weather and load data for each future time as input, and performs rolling calculations on the dynamic modeling evaluation in Step 3 to obtain the predicted capacity margin change trend for the next 15 minutes to 24 hours. Step 5, Visualization and Alarms: Display graphs related to capacity margin changes through a web interface and issue alarms when the preset threshold is exceeded.

[0019] An online assessment system for the grid-connectable capacity margin and trend of a new energy base in the desert region includes a data acquisition layer, a data fusion and processing layer, an assessment and prediction layer, and an application and display layer.

[0020] 1. Data Acquisition Layer: Acting as the system's "sensory nerves," it is responsible for high-frequency (second- to minute-level) acquisition of multi-source heterogeneous data. This is achieved through deployment or integration with the following systems: (1) Meteorological monitoring network: high-precision meteorological stations inside and around the base (collecting wind speed, wind direction, light intensity, ambient temperature, humidity, air pressure, dust concentration), satellite cloud image data, radar data and numerical weather prediction (NWP) microscale correction data.

[0021] (2) Power grid monitoring system: real-time power grid data is obtained from the dispatch center EMS / SCADA system through standard interfaces (such as IEC 61850, 104 protocol), including voltage of each node, phase angle, active / reactive power flow of lines / transformers, switch status, protection signals, system frequency, etc.

[0022] (3) New energy power station monitoring system: Obtain single-unit or combiner box level operation data of wind turbines and photovoltaic inverters from the monitoring system of each new energy power station, including real-time active / reactive power output, terminal voltage, current, internal temperature of equipment, fault codes, available capacity, control mode (PQ, PV, VF, etc.).

[0023] (4) Energy storage power station monitoring system: to obtain real-time charging and discharging power, SOC, state of health (SOH), maximum charging and discharging capacity, voltage and current of energy storage power station and each battery cluster.

[0024] (5) Management data: including planned maintenance information, load forecast curves, market transaction plans, etc.

[0025] 2. Data Fusion and Processing Layer: As the "information hub" of the system, it performs in-depth processing on the massive, asynchronous, multi-scale data reported by the data acquisition layer, which may contain noise and missing data.

[0026] (1) Spatiotemporal alignment: unify all data to the same timestamp (such as PTP synchronization clock) and the same spatial reference system (such as mapping device data to specific nodes or branches of the power grid model).

[0027] (2) Data cleaning: Apply algorithms based on statistics (such as the 3σ principle), models (such as state estimation residual analysis) or machine learning (such as isolated forest) to identify and remove outliers and bad data.

[0028] (3) Missing data imputation: Interpolation, regression imputation based on correlation, or deep learning generation methods are used to reasonably repair short-term missing data.

[0029] (4) Feature engineering: Extract key features from the raw data, such as calculating line load rate, voltage deviation rate, instantaneous value of new energy penetration rate, and conversion factor of meteorological factors on power output. Finally, a clean, consistent "feature dataset" with spatiotemporal labels is formed for subsequent evaluation and prediction.

[0030] 3. Evaluation and Prediction Layer: As the "decision brain" of the system, it contains two core modules.

[0031] (1) Dynamic Evaluation Module: Its core is an online, rolling optimization model. The model uses the current grid operating point as a benchmark and the total output increment ΔP of the new energy base as the optimization variable. The objective function is max ΔP. The constraint set C is updated in real time, including: ① Power flow equation constraint: f(P, Q, V, θ) = 0, where P is the dielectric injected active power, Q is the node injected reactive power, V is the node voltage magnitude, and θ is the node voltage phase angle. f is the power flow equation vector function.

[0032] ② Inequality safety constraints: V min ≤ V ≤ V max , |P line | ≤ P line-max ,in, V min It is the minimum allowable voltage amplitude at the node. V max It is the maximum allowable voltage amplitude at the node. |P line | This refers to the actual active power transmitted through the transmission line / channel. P line-max It is the maximum permissible active power transmitted by the line.

[0033] ③ Equipment capacity constraint: 0≤P _new_i ≤P _avail_i , where P _avail_i It is determined by both the real-time status of the equipment and meteorological data. Among them, P _new_i Let P be the actual grid-connected active power of the i-th new energy device. _avail_i Let be the current maximum available active power of the i-th new energy device.

[0034] ④ Stability boundary constraint: ΔP ≤ P _critical P _critical It can be obtained from a pre-stored stability limit library or from online simplified simulations. Among them, P... _critical The maximum active power disturbance threshold that the system can withstand to maintain stability is not a measured value, but a core indicator obtained through power system stability simulation.

[0035] This optimization problem can be solved quickly using efficient algorithms such as linear programming (LP), quadratic programming (QP), or interior-point methods, outputting the theoretical maximum grid-connectable capacity P at the current moment. _max_theoretical and real-time grid-connectable capacity margin ΔP _available ΔP _available = P _max_theoretical- P _current P _current It has contributed to the current actual grid connection.

[0036] (2) Trend prediction module: Receives the historical margin time series {ΔP} output by the dynamic evaluation module. _available (t)}, and future weather forecast sequences, load forecast sequences, and known equipment commissioning and decommissioning plans. A hybrid forecasting strategy is adopted: ① Data-driven direct prediction: {ΔP} _available (t)} is used as the main sequence (representing the set of discrete time series consisting of the available active power margin of the evaluation object at different historical time points, which is the historical observation data output by the evaluation module), combined with relevant variables (such as predicted illumination and wind speed), and input into the trained LSTM neural network for end-to-end multi-step prediction, directly outputting the capacity margin prediction value and its confidence interval every 15 minutes for the next 2 hours.

[0037] ② Model-driven rolling forecast: Using meteorological forecasts and load data at various future times as input, the dynamic assessment module performs rolling calculations to generate a "predicted capacity margin trajectory". This method has a clearer physical meaning and is especially suitable for scenarios with major planned operations (such as line maintenance).

[0038] The prediction results from the two methods can be weighted and fused to obtain the final trend prediction curve.

[0039] 4. Application and Presentation Layer: This layer serves as the system's "interactive interface" and is provided in the form of a web service or client.

[0040] (1) Visualization panel: Displays real-time capacity margin values, historical and forecast trend curves (attached) Figure 5 Heat maps of power flow and capacity margin distribution at key power grid sections, and overlay maps of meteorological information, etc.

[0041] (2) Alarm Center: Set multiple threshold levels (such as warning, critical, emergency). When the real-time capacity margin is lower than the threshold or the predicted margin will fall below the threshold within a specified time, trigger alarms through multiple channels such as sound and light, system pop-ups, and SMS.

[0042] (3) Reports and interfaces: Automatically generate assessment reports and push the margin assessment and prediction results to the upper-level dispatch system, automatic generation control (AGC) system, energy storage coordination control system and power trading platform through standardized APIs (such as RESTful) to form a closed-loop application.

[0043] This invention views the power grid as a dynamically evolving organism. Through high-frequency, multi-source synchronous sensing, it constructs a digital twin dynamic evaluation model to solve online for the maximum increase in renewable energy power (i.e., grid-connected capacity margin) at the current moment while satisfying all safety, quality, and equipment constraints. It also uses data-driven algorithms to proactively predict the trajectory of this increase. This invention shortens the evaluation cycle from "days / hours" to "minutes / seconds," achieving online rolling calculations in sync with the power grid energy management system (EMS). By incorporating fine-grained data neglected by traditional models, such as real-time meteorological micro-forecasts and equipment status monitoring, the calculation accuracy for renewable energy output and grid carrying capacity is significantly improved. For example, it can account for real-time derating of photovoltaic inverters due to afternoon high temperatures or pollution losses on component surfaces caused by sandstorms, thus providing margin values ​​that more closely reflect actual physical processes. The dynamic evaluation model constructed in this invention is a multi-objective, multi-constraint optimization problem. Its objective function is to maximize the total grid-connected power increment of renewable energy bases while ensuring absolute grid security. Its constraints are a large, dynamically updated set. This invention transforms complex assessment and prediction results into intuitive graphics, charts, map overlays, and early warning information through a visual interaction layer. When the margin falls below a preset safety threshold, or when the predicted trend indicates a rapid contraction, the system can automatically trigger multi-level alarms and generate corresponding scheduling suggestions. This transforms dispatchers from passive "event responders" to proactive "trend managers," significantly improving the economy and safety of power grid operation and providing core technical support tools for the Shagohuang New Energy Base to achieve "full power generation and efficient consumption."

Claims

1. An online assessment method for the grid-connectable capacity margin and trend of a new energy base in the desert region, characterized in that, Step 1: Multi-source data acquisition, including meteorological monitoring data, power grid monitoring data, new energy power station monitoring data, energy storage power station monitoring data, and management data; Step 2, data fusion processing: the multi-source data from Step 1 is spatiotemporally aligned, the data is cleaned, outliers are identified and removed, missing data is filled, key features are extracted, and normalization processing is implemented to form a unified data pool. Step 3, Dynamic Modeling and Evaluation: Establish an online, real-time updated model. Using the current grid operating point as a benchmark, with the total output increment ΔP of the new energy base as the optimization variable, and max ΔP as the objective function, the constraints include: Power flow equality constraints: f(P, Q, V, θ) = 0; P is the dielectric injected active power, Q is the node injected reactive power, V is the node voltage amplitude, θ is the node voltage phase angle, and f is the power flow equation vector function; Inequality safety constraints: V min ≤ V ≤ V max , |P line | ≤ P line-max ; V min It is the minimum allowable voltage amplitude at the node. V max It is the maximum allowable voltage amplitude at the node. |P line | This refers to the actual active power transmitted through the transmission line / channel. P line-max This is the maximum permissible active power transmitted by the line; equipment capacity constraint: 0 ≤ P _new_i ≤P _avail_i ;P _new_i Let P be the actual grid-connected active power of the i-th new energy device. _avail_i Let P be the current maximum available active power of the i-th renewable energy device; stability boundary constraint: ΔP ≤ P _critical ;P _critical The maximum active power disturbance threshold that the system can withstand to maintain stability is determined; then the theoretical maximum grid-connected capacity at the current moment is calculated, and the current real-time grid-connected capacity margin is determined. Step 4, Trend Forecasting Analysis, combines data-driven direct forecasting with model-driven rolling forecasting. Data-driven direct forecasting is based on historical margin time series, combined with predicted illumination and wind speed data, and input into a neural network for calculation to obtain the predicted capacity margin value and its confidence interval for every 15 minutes in the next 2 hours. Model-driven rolling forecasting uses predicted weather and load data for each future time as input, and performs rolling calculations on the dynamic modeling evaluation in Step 3 to obtain the predicted capacity margin change trend for the next 15 minutes to 24 hours. Step 5, Visualization and Alarms: Display graphs related to capacity margin changes through a web interface and issue alarms when the preset threshold is exceeded.

2. The online assessment method for the grid-connectable capacity margin and trend of the Shagohuang new energy base according to claim 1, characterized in that, The meteorological monitoring data mentioned in step 1 includes the collection of wind speed, wind direction, light intensity, ambient temperature, humidity, air pressure, dust concentration, satellite cloud image data, radar data, and numerical weather prediction microscale correction data; The power grid monitoring data includes voltage, phase angle, active / reactive power flow of lines / transformers, switch status, protection signals, and system frequency at each node of the power grid; the new energy power plant monitoring data includes real-time active / reactive power output, generator terminal voltage, current, equipment internal temperature, fault codes, available capacity, and control mode; the energy storage power station monitoring data includes real-time charging and discharging power, SOC, health status, maximum charging and discharging capacity, voltage, and current of the energy storage power station and each battery cluster; and the management data includes planned maintenance information, load forecast curves, and market transaction plans.

3. The online assessment method for the grid-connectable capacity margin and trend of the Shagohuang new energy base according to claim 1, characterized in that, In step 2, the spatiotemporal alignment uses the same timestamp of the PTP synchronous clock to implement data unification with the same spatial reference frame.

4. The online assessment method for the grid-connectable capacity margin and trend of the Shagohuang new energy base according to claim 1, characterized in that, In step 2, the data cleaning uses a state estimation residual analysis model or an isolated forest algorithm; missing data is filled using interpolation, regression-based imputation, or deep learning generation.

5. The online assessment method for the grid-connectable capacity margin and trend of the Shagohuang new energy base according to claim 1, characterized in that, Key features in step 2 include line load rate, voltage deviation rate, instantaneous value of new energy penetration rate, and conversion factor of meteorological factors to power output.

6. The online assessment method for the grid-connectable capacity margin and trend of the Shagohuang new energy base according to claim 1, characterized in that, In step 3, the current margin is calculated using linear programming, quadratic programming, or interior point methods.

7. The online assessment method for the grid-connectable capacity margin and trend of the Shagohuang new energy base according to claim 1, characterized in that, The graphs related to capacity margin changes in step 5 include real-time capacity margin values, historical and predicted trend curves, heat maps of power flow and margin distribution at key grid sections, and meteorological information overlay maps. Alarm methods include audible and visual alarms, system pop-ups, and SMS messages.

8. An online evaluation system for the grid-connectable capacity margin and trend of a new energy base in the desert region, characterized in that, The online assessment method for the grid-connectable capacity margin and trend of the Shagohuang new energy base as described in any one of claims 1-7 includes a data acquisition layer, a data fusion and processing layer, an assessment and prediction layer, and an application and display layer, wherein the data acquisition layer acquires multi-source data; The data fusion and processing layer cleans, aligns, and merges multi-source data to unify the data. The assessment and prediction layer includes a dynamic assessment module and a trend prediction module. The dynamic assessment module calculates the current capacity margin based on an online real-time updated model, and the trend prediction module predicts the trend of capacity margin changes. The application and presentation layer provides a visual representation of capacity margins and trends, and issues alerts for exceeding threshold ranges.

9. The online assessment system for the grid-connectable capacity margin and trend of the Shagohuang new energy base according to claim 8, characterized in that, The application and presentation layer includes a report generation module and an external transmission interface. The report generation module automatically generates assessment reports, and the external transmission interface is used to push the margin assessment and forecast results to the upper-level dispatch system, automatic generation control system, energy storage coordination control system, and power trading platform.

Images