Port energy load and wind and light output prediction and early warning method, equipment and medium

Abstract

The invention discloses a port energy load and wind and light output prediction and early warning method and device and a medium, and relates to the technical field of port energy management, and the method comprises the steps: obtaining ship automatic recognition system data, loading and unloading operation plan data, weather forecast data and weather correction data of a port region, the method comprises the steps of determining the punctual berthing probability of a target ship based on ship automatic identification system data, determining plan demand power and a port operation busy index based on loading and unloading operation plan data, determining weather conditions based on weather forecast data, and further completing load prediction by using a dynamic Bayesian network model. The wind speed and the illumination irradiance in the weather forecast data are corrected based on the weather correction data, the wind power generation power is further obtained through calculation based on the corrected wind speed, the photovoltaic power generation power is obtained through calculation based on the corrected illumination irradiance, and wind and light output prediction is completed. The port energy load and wind and light output prediction precision can be improved, and subsequent early warning and regulation are facilitated.

Description

Technical Field

[0001] This application relates to the field of port energy management technology, and in particular to a method, equipment and medium for predicting and early warning of port energy load and wind and solar power output. Background Technology

[0002] As a comprehensive energy hub, a port's load composition is complex, including both continuous base loads and sudden, impactful loads caused by the berthing and departure of large ships (shore power access) and the start-up and shutdown of large loading and unloading equipment. Simultaneously, the widely deployed distributed power sources such as wind and solar power in ports are highly intermittent and volatile due to the influence of local coastal microclimates (such as sea and land breezes and sea surface reflections). These sudden, impactful loads, along with the intermittent and volatile nature of wind and solar power output, pose significant power imbalance risks to the port's energy system, placing extremely high demands on the accuracy of load and wind / solar power output forecasting. This is because accurate forecasting of load and wind / solar power output allows sufficient time for early warning and the implementation of control measures to mitigate power imbalance risks. Therefore, accurate forecasting of load and wind / solar power output is a prerequisite for achieving multi-energy complementary optimal scheduling and local balance control.

[0003] However, the commonly used method for forecasting load is port load forecasting based on historical data. This method mainly performs time series analysis on historical port load data (such as using ARIMA (Autoregressive Integrated Moving Average) or LSTM (Long Short-Term Memory)) to predict port load. However, its essential drawback is that it cannot perceive and predict the step-like and impactful loads caused by the berthing and departure of large ships and the start-up and shutdown of large loading and unloading equipment. This results in a large prediction error during critical periods such as concentrated ship arrivals, with the prediction error usually exceeding 40%. When predicting wind and solar power output, the commonly used method is the general wind and solar power output prediction method. This method usually directly applies the meteorological forecast data of the port area to drive the physical model or statistical model to predict the wind and solar power generation power. However, its essential flaw is that it does not consider the acceleration effect of the special topography of the port on the wind field and the light enhancement effect of the sea surface reflection of the port on photovoltaics. It directly applies macro-level meteorological forecast data to the micro-level port scenario, resulting in a systematic deviation between the predicted and actual wind and solar power output values. The prediction error is usually higher than 20%. Summary of the Invention

[0004] The purpose of this application is to provide a method, equipment, and medium for predicting and warning of port energy load and wind and solar power output, which can improve the accuracy of port energy load and wind and solar power output prediction.

[0005] To achieve the above objectives, this application provides the following solution.

[0006] In a first aspect, this application provides a method for predicting port energy load and wind and solar power output, the method comprising: Acquire data from the Automatic Identification System (AIS) of vessels in the port area, including loading and unloading operation plan data, weather forecast data, and weather correction data. The AIS data includes the planned berthing time and vessel identification code of the vessel. The loading and unloading operation plan data includes the planned start time, planned stop time, and rated power of each loading and unloading equipment. The weather correction data includes the coastline slope, coastline direction, and sea surface reflectivity. Vessels whose planned berthing time falls within the predicted time period are selected as target vessels. Multiple historical data sets are extracted based on the target vessel's vessel identification code. The on-time berthing probability of the target vessel is calculated based on all historical data sets. The historical data sets include historical planned berthing time and historical actual berthing time. Based on the loading and unloading operation plan data, the planned power demand and port operation busy index for each forecast time in the forecast period are calculated. Based on meteorological forecast data, determine the weather conditions at each forecast time within the forecast period; Using the on-time berthing probability of the target vessel, the planned power demand, port operation busy index, and weather conditions at each forecast time in the forecast period as inputs, a dynamic Bayesian network model is used to determine the load power mutation probability and load power mutation amplitude at each forecast time in the forecast period, and the load forecast result is obtained. The wind speed in the weather forecast data is corrected by using the slope and orientation of the coastline to obtain the corrected wind speed at each forecast time in the forecast period. The irradiance in the weather forecast data is corrected by using the sea surface reflectance to obtain the corrected irradiance at each forecast time in the forecast period. Based on the corrected wind speed at each forecast moment within the forecast period, the wind power generation at each forecast moment within the forecast period is calculated. Based on the corrected solar irradiance at each forecast moment within the forecast period, the photovoltaic power generation at each forecast moment within the forecast period is calculated, thus obtaining the wind and solar power output forecast results. The load forecast results and the wind and solar power output forecast results can be used for early warning and regulation of the port energy system.

[0007] Secondly, this application provides a method for early warning of port energy load and wind and solar power output, the method comprising: Obtain the load power mutation probability, load power mutation amplitude, wind power generation and photovoltaic power generation at each prediction moment in the prediction period determined by the above-mentioned port energy load and wind and solar power output prediction methods. Based on the wind power generation and photovoltaic power generation at each prediction time within the prediction period, the rate of change of wind power generation and photovoltaic power generation at each prediction time within the prediction period are calculated. Based on the probability of load power change, the magnitude of load power change, the rate of change of wind power generation, and the rate of change of photovoltaic power generation at each prediction time within the prediction period, it is determined whether an early warning should be triggered, and an early warning trigger signal is generated when an early warning is triggered.

[0008] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the above-described port energy load and wind and solar power output prediction method or the above-described port energy load and wind and solar power output early warning method.

[0009] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned port energy load and wind and solar power output prediction method or the aforementioned port energy load and wind and solar power output early warning method.

[0010] According to the specific embodiments provided in this application, this application has the following technical effects.

[0011] This application provides a method, equipment, and medium for predicting and warning of port energy load and wind and solar power output. It acquires data from the Automatic Identification System (AIS) of vessels, loading and unloading operation plans, weather forecasts, and weather corrections for the port area. When predicting load, based on AIS data, vessels with planned berthing times within the prediction period are selected as target vessels. Multiple historical data sets are extracted based on the target vessels' identification codes, and the on-time berthing probability of the target vessels is calculated based on all historical data sets. Based on the loading and unloading operation plan data, the planned power demand and port operation busy index for each prediction moment within the prediction period are calculated. Based on weather forecast data, the weather conditions for each prediction moment within the prediction period are determined. The on-time berthing probability of the target vessel, along with the planned power demand, port operation busy index, and weather conditions at each forecast moment within the forecast period, are used as inputs. A dynamic Bayesian network model is then used to determine the probability and magnitude of load power mutations at each forecast moment within the forecast period, thus obtaining the load forecast result and completing the load forecast. By incorporating Automatic Identification System (AIS) data and loading / unloading operation plan data into the load forecast process, non-periodic load mutation events driven by AIS data and loading / unloading operation plan data can be considered. After prediction using the dynamic Bayesian network model, it is possible to perceive and predict step-like and impactful loads caused by the berthing and departure of large vessels and the start-up and shutdown of large loading / unloading equipment, thereby improving the accuracy of load forecasting. When predicting wind and solar power output, based on meteorological correction data, the wind speed in the meteorological forecast data is corrected using the coastline slope and coastline orientation to obtain the corrected wind speed at each forecast time within the forecast period. The solar irradiance in the meteorological forecast data is corrected using sea surface reflectivity to obtain the corrected solar irradiance at each forecast time within the forecast period. Based on the corrected wind speed at each forecast time within the forecast period, the wind power generation at each forecast time within the forecast period is calculated. Based on the corrected solar irradiance at each forecast time within the forecast period, the photovoltaic power generation at each forecast time within the forecast period is calculated, thus obtaining the wind and solar power output prediction results and completing the wind and solar power output prediction. Because meteorological correction data is introduced into the wind and solar power output prediction process, and the wind speed is corrected using the coastline slope and coastline orientation, and the solar irradiance is corrected using sea surface reflectivity, the acceleration effect of the special topography of the port on the wind field and the light enhancement effect of the port's sea surface reflection on photovoltaics can be considered, thereby improving the accuracy of wind and solar power output prediction. In summary, this application can improve the accuracy of port energy load and wind and solar power output forecasts, and the load forecast results and wind and solar power output forecast results can be used for early warning and regulation of the port energy system. On the premise of improving the accuracy of port energy load and wind and solar power output forecasts, it is beneficial to conduct better early warning and regulation of the port energy system in the future. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is an application environment diagram of a port energy load and wind and solar power output prediction method provided in Embodiment 1 of this application.

[0014] Figure 2 This is a flowchart illustrating a method for predicting port energy load and wind and solar power output, as provided in Embodiment 1 of this application.

[0015] Figure 3 This is a schematic diagram of the load forecasting process based on a dynamic Bayesian network model provided in Embodiment 1 of this application.

[0016] Figure 4 This is a schematic diagram of the structure of the Bayesian network model provided in Embodiment 1 of this application.

[0017] Figure 5 This is a flowchart illustrating a method for early warning of port energy load and wind and solar power output provided in Embodiment 2 of this application.

[0018] Figure 6 This is a schematic diagram of the prediction-control linkage timing provided in Embodiment 2 of this application.

[0019] Figure 7 This is a schematic diagram of a port energy load and wind and solar power output prediction and early warning system provided in Embodiment 3 of this application.

[0020] Figure 8 This is a schematic diagram of the structure of a computer device provided in Embodiment 4 of this application. Detailed Implementation

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

[0022] Example 1.

[0023] The port energy load and wind and solar power output prediction method provided in this application embodiment can be applied to, for example... Figure 1The application environment shown is as follows. The terminal communicates with the server via a network. The data storage system stores the data the server needs to process. The data storage system can be set up independently, integrated into the server, or placed in the cloud or on another server. The terminal can send a forecast request to be processed to the server. After receiving the forecast request, the server obtains data from the port area's Automatic Identification System (AIS), loading and unloading operation plans, weather forecasts, and weather corrections. It selects vessels whose planned berthing time falls within the forecast period as target vessels, extracts multiple historical data sets based on the target vessel's identification code, and calculates the on-time berthing probability of the target vessel based on all historical data sets. Based on the loading and unloading operation plan data, it calculates the planned power demand and port operation busy index for each forecast moment within the forecast period. Based on the weather forecast data, it determines the weather conditions for each forecast moment within the forecast period. The on-time berthing probability of the target vessel, along with the planned power demand, port operation busy index, and weather conditions for each forecast moment within the forecast period, are used as input data. The system uses a dynamic Bayesian network model to determine the probability and magnitude of load power abrupt changes at each forecast time within the forecast period, thus obtaining the load forecast result. It then uses the coastline slope and orientation to correct the wind speed in the weather forecast data, obtaining the corrected wind speed at each forecast time within the forecast period. Similarly, it uses sea surface reflectivity to correct the solar irradiance in the weather forecast data, obtaining the corrected solar irradiance at each forecast time within the forecast period. Based on the corrected wind speed at each forecast time within the forecast period, the system calculates the wind power generation at each forecast time within the forecast period. Finally, based on the corrected solar irradiance at each forecast time within the forecast period, the system calculates the photovoltaic power generation at each forecast time within the forecast period, thus obtaining the wind and solar power output forecast result. The load forecast result and the wind and solar power output forecast result can be used for early warning and regulation of the port energy system. The server can feed back the obtained load forecast results (i.e., the probability and magnitude of load power mutation at each forecast time in the forecast period) and wind and solar power output forecast results (i.e., wind power generation and photovoltaic power generation at each forecast time in the forecast period) to the terminal.

[0024] In addition, in some embodiments, the port energy load and wind and solar power output prediction methods can also be implemented by a server or a terminal. For example, the terminal can directly process the prediction requests to be processed, or the server can obtain the prediction requests to be processed from the data storage system and process them.

[0025] In one exemplary embodiment, such as Figure 2As shown, a method for predicting port energy load and wind and solar power output is provided. This method is executed by computer equipment, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 The following steps are used as an example of a server in the example.

[0026] Step S1: Obtain Automatic Identification System (AIS) data, loading and unloading operation plan data, weather forecast data, and weather correction data for the port area. AIS data includes the planned berthing time and vessel identification code of the vessel. Loading and unloading operation plan data includes the planned start time, planned stop time, and rated power of each loading and unloading device. Weather correction data includes the coastline slope, coastline orientation, and sea surface reflectivity.

[0027] Step S2: Select vessels whose planned berthing time falls within the predicted time period as target vessels. Extract multiple historical data sets based on the target vessel's vessel identification code, and calculate the on-time berthing probability of the target vessel based on all historical data sets. The historical data sets include historical planned berthing time and historical actual berthing time.

[0028] Step S3: Based on the loading and unloading operation plan data, calculate the planned power demand and port operation busy index for each predicted moment in the predicted time period.

[0029] Step S4: Based on meteorological forecast data, determine the weather conditions at each forecast time within the forecast period.

[0030] Step S5: Using the on-time berthing probability of the target vessel, the planned power demand, port operation busy index, and weather conditions at each forecast time in the forecast period as inputs, the dynamic Bayesian network model is used to determine the load power mutation probability and load power mutation amplitude at each forecast time in the forecast period, and the load forecast result is obtained.

[0031] Step S6: Correct the wind speed in the weather forecast data using the coastline slope and coastline orientation to obtain the corrected wind speed at each forecast time within the forecast period. Correct the irradiance in the weather forecast data using the sea surface reflectance to obtain the corrected irradiance at each forecast time within the forecast period.

[0032] Step S7: Based on the corrected wind speed at each prediction moment in the prediction period, calculate the wind power generation at each prediction moment in the prediction period; based on the corrected solar irradiance at each prediction moment in the prediction period, calculate the photovoltaic power generation at each prediction moment in the prediction period, and obtain the wind and solar power output prediction results; the load prediction results and the wind and solar power output prediction results can be used for early warning and regulation of the port energy system.

[0033] By implementing steps S1 to S7 above, this embodiment can improve the accuracy of port energy load and wind and solar power output prediction, which is beneficial for better early warning and regulation of the port energy system in the future.

[0034] The port energy load and wind and solar power output prediction method provided in this embodiment is a port energy load and wind and solar power output prediction method based on multi-source heterogeneous data fusion. Specifically, it is a port energy load and wind and solar power output prediction method that integrates ship automatic identification system data, loading and unloading operation plan data, weather forecast data, and weather correction data (mainly terrain correction and reflection correction). It can complete ultra-short-term prediction of load and wind and solar power output (prediction time period less than or equal to 15 minutes). Of course, it can not only complete ultra-short-term prediction, but also predict load and wind and solar power output for any prediction time period, providing high-precision prediction input for the local balance control of multi-energy fusion system.

[0035] The following is a detailed description of the port energy load and wind and solar power output prediction method used in this embodiment, including the following steps.

[0036] (a) Real-time acquisition and preprocessing of multi-source heterogeneous data.

[0037] This embodiment can receive real-time Automatic Identification System (AIS) data from the port area (generally referred to as the port or port area) via an API (Application Programming Interface). The AIS is a technical system for automatic ship identification and maritime communication. It utilizes the Global Positioning System (GPS) and communication technology to enable ships to automatically transmit information such as their position, heading, and speed, and exchange this information in real time with surrounding ships and shore-based equipment. AIS data is transmitted via Very High Frequency (VHF) radio waves and encapsulated in the NMEA 0183 protocol format. It includes static information (such as ship name and MMSI (Maritime Mobile Service Identification) number), dynamic information (such as position, heading, and speed), and safety-related information (such as turning intention). The MMSI number is a unique identifier for the ship within the AIS system and cannot be arbitrarily changed. Subsequent parsing of the AIS data allows the extraction of the ship's planned berthing time. Ship identification number (MMSI number) and shore power rated power .

[0038] This embodiment can obtain the port area's loading and unloading operation plan data (also known as a refined loading and unloading operation plan table) from the port production management system. Based on the loading and unloading operation plan data, the planned start time of each loading and unloading equipment within the planning period can be extracted. Planned stop time and rated power .

[0039] This embodiment can collect meteorological station data in and around the port area in real time to obtain weather forecast data for the port area, including wind speed at each forecast time within the forecast period. Wind direction and solar irradiance Cloud cover, precipitation, visibility, ambient temperature, etc.

[0040] This embodiment can retrieve digital elevation model data of the port area from the geographic information system. Based on the digital elevation model data, the slope and orientation of the coastline of the port area can be extracted. The coastline slope is the sea-land boundary benchmark with the elevation value corresponding to the mean sea level. The coastline is the degree of inclination relative to the horizontal plane when it extends seaward (including the surface and nearshore shallow water area). The orientation of the coastline is the overall extension direction of the coastline. At the same time, the sea surface reflectance of the port area can also be obtained to obtain the meteorological correction data of the port area.

[0041] Through the above process, real-time acquisition and preprocessing of multi-source heterogeneous data can be completed. The multi-source heterogeneous data (i.e., ship automatic identification system data, loading and unloading operation plan data, weather forecast data, and weather correction data) includes data with different spatial and temporal resolutions. Ship automatic identification system data and weather correction data are high spatial resolution data, loading and unloading operation plan data and weather forecast data are low spatial resolution data, ship automatic identification system data and loading and unloading operation plan data are high temporal resolution data, and weather forecast data and weather correction data are low temporal resolution data. Data from different sources and with different spatiotemporal characteristics are integrated into a unified analysis framework for subsequent load and wind and solar power output prediction.

[0042] It should be noted that high spatial resolution data, low spatial resolution data, high temporal resolution data, and low temporal resolution data are all relative terms. They only indicate that the spatial resolution of high spatial resolution data is higher than that of low spatial resolution data, and the temporal resolution of high temporal resolution data is higher than that of low temporal resolution data.

[0043] In this embodiment, the system acquires Automatic Identification System (AIS) data, loading and unloading operation plan data, weather forecast data, and weather correction data for the port area. The AIS data includes the planned berthing time and the ship's identification code. The loading and unloading operation plan data includes the planned start time, planned stop time, and rated power of each loading and unloading device. The weather forecast data includes wind speed, wind direction, solar irradiance, cloud cover, precipitation, visibility, and ambient temperature at each forecast time within the forecast period. The weather correction data includes the coastline slope, coastline orientation, and sea surface reflectivity.

[0044] (ii) Load forecasting based on dynamic Bayesian network model.

[0045] Load forecasting based on a dynamic Bayesian network model is the core of this embodiment. It aims to combine the uncertainty of ship dynamic behavior with the planning of port operations to quantitatively calculate the probability and magnitude of sudden load changes in the near future, providing a forward-looking decision-making basis for the control system. Relying on load forecasting based on a dynamic Bayesian network model, this embodiment aims to address the shortcomings of commonly used load forecasting methods and provide an ultra-high-precision port load forecasting method. Specifically, it addresses the following technical problems: how to accurately predict sudden and impactful loads caused by ship arrivals and loading / unloading operations, overcoming the insensitivity of commonly used load forecasting methods to such events; that is, how to overcome the limitation of commonly used load forecasting methods that rely on historical load data, and accurately predict planned sudden loads by integrating real-time dynamic data (Automatic Identification System data) and planned data (loading / unloading operation plan data). This provides forward-looking information input for optimized scheduling, reliable decision-making basis for collaborative strategies, and especially provides precise time windows and power requirements for ship energy consumption.

[0046] like Figure 3 As shown, the load forecasting based on the dynamic Bayesian network model specifically includes: extracting planned berthing times from Automatic Identification System (AIS) data; selecting vessels whose planned berthing times fall within the forecast period as target vessels; extracting multiple historical data sets based on the target vessels' vessel identification codes, including historical planned berthing times and historical actual berthing times; calculating the deviation between historical planned berthing times and historical actual berthing times based on all historical data sets of the target vessels; and further calculating the on-time berthing probability of the target vessels. Based on the loading and unloading operation plan data, the planned start and stop times (including planned start and stop times) and rated power of each loading and unloading equipment are extracted. A parallel operation power superposition model for each loading and unloading equipment is constructed to calculate the power at each prediction time in the prediction period. Planned power demand Simultaneously, based on the equipment status of the loading and unloading equipment (including operating status and shutdown status), each prediction moment in the prediction time period is determined. The port operation busy index. Based on meteorological forecast data, determine each forecast time within the forecast period. Weather conditions. The probability of the target vessel berthing on time. and each prediction moment within the prediction time period Planned power demand The port operation busy index and weather conditions are input into the Dynamic Bayesian Network (DBN) model. The DBN model infers and outputs each prediction moment in the prediction period. load power sudden change probability and load power fluctuation amplitude .

[0047] Among them, the probability of the target ship berthing on time is calculated. Specifically, this includes: analyzing data from the Automatic Identification System (AIS) to determine the planned berthing time of a vessel. Based on the vessel identification code, select vessels whose planned berthing time falls within the predicted time period as target vessels, and use the target vessel's vessel identification code as an index to query the target vessel's past... Next (such as) =20 (the historical planned berthing time and historical actual berthing time can be set according to user needs) to obtain There are several historical data sets. For each historical data set, the difference between the historical planned berthing time and the historical actual berthing time is calculated to obtain the deviation of the historical data set. The deviations of all historical data sets are then combined into a deviation dataset. }, For the first The bias of a historical data set. Assume the biased dataset follows a normal distribution. Based on the bias dataset, the mean is calculated using the maximum likelihood estimation method. and standard deviation On-time is defined as the actual berthing time of the target vessel within [timeframe missing]. Within the range, An acceptable threshold, such as 15 minutes, can also be set according to user needs. The target vessel's planned berthing time is calculated using the cumulative distribution function. front and back The probability of berthing within a given time period is the probability of berthing on time. .

[0048] On-time berthing probability The calculation formula is: ; in, To increase the probability of berthing on time; It is the cumulative distribution function; It follows a normal distribution. The mean, Standard deviation; For the planned berthing time; This is an acceptable threshold.

[0049] Specifically, a parallel operation power superposition model for each loading and unloading device is constructed to calculate the power at each prediction moment within the prediction time period. Planned power demand Specifically, this includes: extracting the data from each loading and unloading equipment based on the loading and unloading operation plan data. Plan launch time Planned stop time and rated power .

[0050] Define device start / stop status functions Characterizing loading and unloading equipment At the predicted time Device status: ; Based on this, each loading and unloading device can be identified. At the predicted time If the equipment status is "working", then the loading and unloading equipment can be recorded as "working loading and unloading equipment".

[0051] Therefore, at each prediction time within the prediction time period Port's planned power requirements It refers to all loading and unloading equipment that is in operation. The formula for calculating the sum of the rated power is: ; in, For loading and unloading equipment i at the predicted time The device status.

[0052] Each prediction time point within the prediction time period Planned power demand This reflects the theoretical load of the port based on loading and unloading operation plan data if there are no ship delays.

[0053] Among them, each prediction moment in the prediction time period is determined. The port operation busy index specifically includes: for each forecast time within the forecast period. If at the predicted time If the number of loading and unloading equipment in operation (i.e., working loading and unloading equipment) is greater than the first preset value, then the predicted time... The port's busyness index is busy at that predicted time. If the number of loading and unloading equipment in operation is less than or equal to a first preset value, but greater than or equal to a second preset value, then the predicted time is... The port operation busy index is normal at this predicted time. If the number of loading and unloading equipment in operation is less than the second preset value, then the predicted time is... The port operation busy index is idle. The first preset value is greater than the second preset value, and both the first and second preset values ​​can be determined by the user based on experience, for example, based on the user's experience of how the number of arriving ships affects the number of loading and unloading equipment.

[0054] Among them, each prediction moment in the prediction time period is determined. The specific weather conditions include: meteorological forecast data including every forecast time within the forecast period. Wind speed, wind direction, solar irradiance, cloud cover, precipitation, visibility, and ambient temperature—these weather data all influence actual ship berthing. Therefore, weather conditions are determined based on this data and input as evidence into a dynamic Bayesian network model. This influences other network nodes within the model, ultimately contributing to inference. For example, the impact of weather data on actual ship berthing can include: wind speed and visibility are crucial evidence. When wind speed exceeds the safe berthing threshold (too high) or visibility falls below the visibility threshold (too low), the berthing time is affected. Weather conditions Determine the prediction time The ship actually berthed The probability of being yes This will be significantly reduced; wind direction and precipitation are supplementary evidence used to fine-tune the actual berthing of ships. This refers to the probability of a ship actually berthing. In other words, by using key meteorological physical quantities (i.e., weather forecast data) that strongly influence ship berthing, weather conditions are determined. By considering the probability of severe weather causing ship delays, the actual berthing time of a ship can be accurately determined. The probability is [not specified]. At this point, weather conditions are categorized as good, fair, and severe, and for each prediction moment within the prediction period... Based on the predicted time The wind speed, wind direction, solar irradiance, cloud cover, precipitation, visibility, and ambient temperature are all factors that users determine the forecast time based on experience. The weather conditions.

[0055] For example, when determining weather conditions, wind speed is used. Irradiance and visibility As indicators, a threshold is set for each indicator to distinguish between normal and abnormal.

[0056] Good: Wind speed and visibility meet the conditions for safe berthing, and solar irradiance is stable. At this point, the following conditions must be met. , ,and , This is the wind speed threshold, which is the maximum wind speed (e.g., 15 m / s) specified by the port for safe berthing of ships. The visibility threshold is the minimum visibility (e.g., 1 km) required for safe berthing of ships as stipulated by the port. For light irradiance at Changes over a period of time Users can customize the settings according to their needs. This represents the threshold for changes in light irradiance.

[0057] Generally: Wind speed and visibility meet the conditions for safe berthing, and there are some fluctuations in solar irradiance. In this case, the following conditions must be met: , ,and .

[0058] Severe conditions: Wind speed or visibility does not meet the safe berthing requirements; in this case, the following conditions must be met. ,or .

[0059] The dynamic Bayesian network model infers and outputs each prediction moment within the prediction time period. load power sudden change probability and load power fluctuation amplitude Specifically, this includes: increasing the probability of the target vessel berthing on time. and each prediction moment within the prediction time period Planned power demand Port operation busyness index and weather conditions (affecting berthing) are used as known evidence and input into the constructed dynamic Bayesian network model. The dynamic Bayesian network model performs forward inference based on the conditional probability table between network nodes. Its core is that it is known that ships may be on time or delayed (with a probability of 1 / 2). ), then at each predicted time in the future Whether the ship has actually berthed, the probability and magnitude of a sudden change in load from the current value, and output for each prediction time within the prediction period. load power sudden change probability and load power fluctuation amplitude Optionally, this embodiment may also use port operation priority as known evidence.

[0060] The dynamic Bayesian network model includes the Bayesian network model at each prediction time, as shown in Table 1 below. The key network nodes of the Bayesian network model include: (1) Observed variables (i.e., input nodes): the probability of the target ship berthing on time. Planned power demand (1) Port operation busy index and weather conditions; (2) Latent variables (i.e., intermediate nodes): actual berthing of ships; (3) Output variables (i.e., output nodes): probability of sudden change in load power. and load power fluctuation amplitude .

[0061] Table 1 Key network nodes of the Bayesian network model

[0062] load power sudden change probability Indicates the time of prediction The probability of a load surge and the magnitude of the load power surge. Indicates if at the predicted time A sudden load change occurs, and the expected increase in load power is significant. If a load change occurs, it indicates a ship is berthing. If a ship berths, then shore power load will be introduced into the port. The value of is approximately equal to It is the change in power after comprehensively considering the possibility of mutation and the planned power demand.

[0063] After identifying the network nodes, the connection relationships between them are further clarified. Since the dynamic Bayesian network model is a multi-time-slice model, there are temporal dependencies and causal relationships between network nodes.

[0064] like Figure 4 As shown, temporal dependencies are dependencies across time slices, including: Weather conditions at time -1 Weather conditions at any time Weather conditions are continuous. Represents a connection; Port Operations Busyness Index at -1 Moment Port Operations Busyness Index at Any Time The level of activity at ports has inertia. The ship's actual berthing time at -1 The actual berthing time of the ship The berthing status of a ship is continuous.

[0065] like Figure 4As shown, the causal relationship is a dependency within the same time slice, including: the probability of the target ship berthing on time. The actual berthing time of the ship Historical on-time performance of vessels affects their actual berthing rate; Weather conditions at any time The actual berthing time of the ship Current weather conditions are affecting the actual berthing of ships; Port Operations Busyness Index at Any Time Power requirements at any given time The port's operational busyness index determines the planned demand power. The actual berthing time of the ship Probability of sudden change in load power at time step Actual berthing events of ships directly lead to sudden changes in load; The actual berthing time of the ship Amplitude of load power change at time The actual berthing of a ship determines whether a sudden change in load power occurs; Power requirements at any given time Amplitude of load power change at time The planned power demand determines the magnitude of sudden changes in load power.

[0066] This embodiment further trains network parameters using historical data to learn the conditional probability relationships between network nodes, resulting in a conditional probability table, for example: (1) The conditional probability of a ship actually berthing is calculated based on the on-time berthing probability of different target ships, weather conditions, and the ship's actual berthing at the previous moment. (2) :when At that time, Approaching 1.0, when Otherwise, A value close to 0.0 allows for small probability values ​​to account for other unknown disturbances; (3) : When the actual berthing of the known ship and planned power demand Under these conditions, the amplitude of sudden load change power The probability distribution is determined by using a Gaussian distribution. To perform modeling, that is The method for determining the mean is as follows: when At that time, the mean was set to This means that if the ship is confirmed to berth, the expected load surge power magnitude is the planned demand power. Otherwise, the mean is set to 0, meaning that if the ship does not berth, no load surge is expected due to this event. The variance is determined by learning from historical data to quantify the degree to which the actual load surge power amplitude fluctuates around the mean. For example, even if the ship berths, the actual shore power load may differ from the planned demand due to differences in ship equipment condition and operation. There are slight deviations.

[0067] At this point, the inference process of the dynamic Bayesian network model includes: (1) Evidence input: As evidence of network node SP, As a network node Evidence using weather conditions as network node data Evidence using the port operation busy index as a network node Evidence; (2) Probabilistic reasoning: Use the forward-backward algorithm or particle filter algorithm for temporal reasoning and calculate the posterior probability distribution. This leads to the derivation of the output distribution. and ; (3) Output generation: Take the output distribution Expected value As the probability of sudden change in load power , Take the output distribution Expected value As the amplitude of load power change .

[0068] Through the above process, this embodiment integrates the uncertainty of ship dynamic behavior (the probability of the target ship berthing on time quantified by the Automatic Identification System data) and the planning of port operations (the planned demand power quantified by the loading and unloading operation plan data) under a unified probabilistic graphical model framework. By capturing the inertial effects of weather conditions and operational status (i.e., the port operation busy index) through time-series modeling, it is more suitable for dynamic port scenarios than static models.

[0069] In this embodiment, vessels whose planned berthing time falls within the predicted time period are selected as target vessels. Multiple historical data sets are extracted based on the target vessel's identification code. The on-time berthing probability of the target vessel is calculated based on all historical data sets, including historical planned berthing times and historical actual berthing times. Based on loading and unloading operation plan data, the planned power demand and port operation busy index for each predicted moment within the predicted time period are calculated. Based on weather forecast data, the weather conditions for each predicted moment within the predicted time period are determined. Using the on-time berthing probability of the target vessel, the planned power demand, port operation busy index, and weather conditions for each predicted moment within the predicted time period as inputs, a dynamic Bayesian network model is used to determine the load power mutation probability and load power mutation amplitude for each predicted moment within the predicted time period, thus obtaining the load prediction result and completing the load prediction.

[0070] Specifically, the on-time berthing probability of the target vessel is calculated based on all historical data sets. This includes: for each historical data set, calculating the difference between the historical planned berthing time and the historical actual berthing time in the historical data set to obtain the deviation of the historical data set; forming a deviation dataset from all historical data sets, and using the maximum likelihood estimation method to calculate the mean and standard deviation of the deviation dataset, which follows a normal distribution; and calculating the on-time berthing probability of the target vessel based on the mean and standard deviation.

[0071] Specifically, based on loading and unloading operation plan data, the planned power demand and port operation busy index for each forecast time in the forecast period are calculated. This includes: for each forecast time in the forecast period, the following steps are performed: for each loading and unloading equipment, it is determined whether the forecast time is between the planned start time and the planned stop time of the loading and unloading equipment. If so, the loading and unloading equipment is recorded as a working loading and unloading equipment; the sum of the rated power of all working loading and unloading equipment is calculated to obtain the planned power demand for the forecast time; and the port operation busy index for the forecast time is determined based on the number of working loading and unloading equipment.

[0072] The dynamic Bayesian network model includes a Bayesian network model for each prediction time in the prediction period. The Bayesian network model includes a first input node, a second input node, a third input node, a fourth input node, an intermediate node, a first output node, and a second output node.

[0073] The input terminal of the first input node is used to receive the on-time berthing probability of the target vessel, and the output terminal of the first input node is connected to the first input terminal of the intermediate node.

[0074] The first input terminal of the second input node is used to receive the planned power demand at the predicted time, and the output terminal of the second input node is connected to the first input terminal of the second output node.

[0075] The first input of the third input node is used to receive the port operation busy index at the prediction time. The second input of the third input node is connected to the output of the third input node of the Bayesian network model at the previous prediction time. The output of the third input node is connected to the second input of the second input node.

[0076] The first input of the fourth input node is used to receive the weather conditions at the prediction time. The second input of the fourth input node is connected to the output of the fourth input node of the Bayesian network model at the previous prediction time. The output of the fourth input node is connected to the second input of the intermediate node.

[0077] The third input of the intermediate node is connected to the output of the intermediate node of the Bayesian network model at the previous prediction time. The first output of the intermediate node is connected to the input of the first output node. The second output of the intermediate node is connected to the second input of the second output node. The output of the first output node is used to output the probability of load power change at the prediction time. The output of the second output node is used to output the magnitude of load power change at the prediction time.

[0078] For the first prediction time, the outputs of the third input node, the fourth input node, and the intermediate nodes of the Bayesian network model at the previous prediction time are all empty (set to 0).

[0079] (iii) Prediction of wind and solar power output based on terrain perception.

[0080] The core innovation of this embodiment, based on terrain perception-based wind and solar power output forecasting, is tailored to the unique geographical environment of ports (coastal location with complex terrain). It aims to localize and refine meteorological forecast data, significantly improving the accuracy of wind and solar power output forecasting in port areas and providing more reliable input for energy dispatch. Relying on terrain perception-based wind and solar power output forecasting, this embodiment aims to address the shortcomings of commonly used wind and solar power output forecasting methods, providing an ultra-high-precision method for port wind and solar power output forecasting. Specifically, it addresses the following technical issues: how to eliminate the differences between macroscopic meteorological data and the local microscopic environment of the port, and how to refine the forecasting of wind and solar power output to reflect the true impact of coastal topography and sea surface reflection.

[0081] The wind and solar power output prediction based on terrain perception includes: (1) establishing local wind speed correction in the port area: based on digital elevation model data, wind speed is corrected based on the local wind field acceleration effect caused by terrain, and the corrected wind speed is obtained. (2) Establish local irradiance correction in the port area: Considering the additional irradiance of the photovoltaic array due to sea surface reflection, the irradiance is corrected based on the light enhancement effect of sea surface reflection to obtain the corrected irradiance. (3) Wind and solar power output prediction: The corrected wind speed and corrected irradiance Input the wind and solar power output model to obtain each prediction time within the prediction period. High-precision wind power generation and high-precision photovoltaic power generation .

[0082] Specifically, correcting wind speed based on the local wind field acceleration effect caused by terrain includes: obtaining wind speed from meteorological forecast data. Extracting coastline slope from digital elevation model data (Unit: radians) and coastline orientation; determine topographic correction factors. This is a constant pre-calibrated through computational fluid dynamics (CFD) simulation, typically ranging from 0.12 to 0.18. The above data is then modeled using the local wind field acceleration effect; when sea breezes blow towards the shore, the terrain is lifted (coastline slope). This will produce a lifting and acceleration effect, causing the wind speed at the shore to be greater than the wind speed at the sea surface. An empirical model is used to quantify this effect to correct the wind speed, resulting in the corrected wind speed. The formula for calculating the corrected wind speed is as follows: ; in, This is the corrected wind speed; Wind speed; This is a terrain correction factor, determined based on the coastline slope and coastline orientation; This refers to the slope of the coastline. for The tangent value is used to quantify the steepness of the terrain.

[0083] Terrain correction factor The calibration method includes: establishing a micro-meteorological computational fluid dynamics model of the port area; setting different wind direction and wind speed boundary conditions for simulation; comparing the simulation results with data from on-site anemometer towers; and regression fitting to obtain topographic correction coefficients related to coastline slope and coastline orientation. Subsequently, the topographic correction coefficient can be determined based on the coastline slope and coastline orientation. .

[0084] As another alternative implementation method, terrain correction coefficient It also depends on the port's macro-topographic complexity, roughness, coastline slope, and coastline orientation to determine, specifically including: (1) the relationship with topographic complexity and roughness: combining the port's macro-topographic complexity and roughness, the topographic correction coefficient The value range is set to 0.05~0.30; (2) Relationship with coastline slope: The value can be dynamically assigned according to the range of coastline slope, such as when the coastline slope is <5°. When the value is between 0.05 and 0.10, and the coastline slope is between 5° and 15°, When the value is between 0.10 and 0.18, and the coastline slope is between 15° and 30°, Take 0.18~0.30; (3) Relationship with coastline direction: Calculate the effective angle between the coastline direction and the wind direction (0°~180°). When the angle is between 75° and 105°, Take the corresponding slope range The upper limit of the value range is when the included angle is between 30° and 75° or between 105° and 150°. Take the corresponding slope range The midpoint of the value range, when the included angle is between 0° and 30° or between 150° and 180°. The value should be ≤0.05. This method ensures that the terrain correction factor accurately matches the actual acceleration effect of local terrain on the wind field, resulting in a corrected wind speed. This refers to the high-precision wind speed prediction value after correction for the port area topography.

[0085] Optionally, the above wind speed correction can be activated only when the wind direction is perpendicular or approximately perpendicular to the slope direction. The slope direction is the extension direction of the coastline topography (along the coastline). After calculating the slope aspect based on digital elevation model data, the slope direction is obtained by adding 90° to the slope aspect. When the angle between the wind direction and the slope direction is within the range of 90° ± a preset threshold, the wind direction is considered to be perpendicular or approximately perpendicular to the slope direction, and wind speed correction is performed at this time. The preset threshold can be set to 15° in combination with actual application scenarios such as port engineering and coastal environmental assessment. For some specific scenarios (such as strong storm areas and gentle coastlines), the preset threshold can be further optimized and adjusted to 10°~20°.

[0086] Specifically, correcting irradiance based on the sea surface reflection enhancement effect includes: obtaining irradiance from weather forecast data. ; Obtain sea surface reflectance, which represents the sea surface's ability to reflect sunlight. It is a constant between 0.25 and 0.3; determine the correlation coefficient of component mounting tilt angle. It is used to convert the effective component of the light reflected from the sea surface onto the photovoltaic array plane, and its value is related to the installation tilt angle of the photovoltaic modules in the photovoltaic array. Related, can usually be simplified to =(1-cos( )) / 2; Irradiance For terrestrial environments, the additional irradiance from reflections from nearby sea surfaces is ignored. This embodiment is the first to quantitatively compensate for this effect. The above data is modeled using the sea surface reflection enhancement effect to correct the irradiance, resulting in the corrected irradiance. The formula for calculating the corrected irradiance is as follows: ; in, This is the corrected irradiance. Light irradiance; Install tilt angle correlation coefficients for the components; This represents the sea surface reflectivity. This refers to the additional irradiance caused by reflection from the sea surface, and the corrected irradiance. This is the high-precision predicted value of light irradiance after compensation for the light enhancement effect of sea surface reflection.

[0087] Optionally, the above-mentioned irradiance correction can be activated only when the photovoltaic array is installed close to the shore and has a wide field of view facing the sea.

[0088] Specifically, the wind and solar power output prediction includes: using corrected wind speeds. Corrected irradiance The power-wind speed characteristic curve model of the wind turbine and the conversion efficiency of the photovoltaic array. Perform the calculation.

[0089] The calculation of wind power generation specifically includes: adjusting the corrected wind speed. Substitute the power-speed characteristic curve model of the wind turbine into the calculation of each prediction moment in the prediction period. wind power generation capacity : ; in, A power-wind speed characteristic curve model to characterize the output characteristics of a wind turbine; For the predicted time The corrected wind speed.

[0090] The power-wind speed characteristic curve model can be a piecewise function model, with the following expression: ; in, The output power of the fan (kW); Actual wind speed (m / s); The cut-in wind speed is usually set to 3~4 m / s; To cut off the wind speed, a speed of 20-25 m / s is usually chosen. Design the maximum output power for the wind turbine; The rated wind speed is usually taken as 10~15m / s.

[0091] This embodiment will As Substituting into the power-wind speed characteristic curve model above, the obtained That is, the power generation capacity of wind power. .

[0092] The calculation of photovoltaic power generation specifically includes: adjusting the corrected solar irradiance. Substitute into the photovoltaic power generation model and calculate each prediction time within the prediction period. Photovoltaic power generation : ; in, For the predicted time Corrected irradiance; The effective light-receiving area of ​​the photovoltaic array; This refers to the conversion efficiency of the photovoltaic array.

[0093] This is a predicted wind power output based on terrain-corrected wind speed. The photovoltaic power generation prediction value based on reflection-corrected irradiance can improve the prediction accuracy of wind and solar power output.

[0094] In this embodiment, the wind speed in the weather forecast data is corrected using the coastline slope and coastline orientation to obtain the corrected wind speed at each forecast time within the forecast period. Similarly, the solar irradiance in the weather forecast data is corrected using sea surface reflectance to obtain the corrected solar irradiance at each forecast time within the forecast period. Based on the corrected wind speed at each forecast time within the forecast period, the wind power generation capacity at each forecast time is calculated. Similarly, based on the corrected solar irradiance at each forecast time within the forecast period, the photovoltaic power generation capacity at each forecast time within the forecast period is calculated, resulting in the wind and solar power output prediction. The load prediction results and the wind and solar power output prediction results can be used for early warning and regulation of the port energy system.

[0095] This embodiment provides a method for predicting port energy load and wind and solar power output, which solves the problem of inaccurate prediction caused by unpredictable load changes and local micro-meteorological disturbances of wind and solar power in port energy systems, and provides ultra-short-term high-precision prediction input of ≤15 minutes for local balance control.

[0096] This embodiment is the first to integrate dynamic AIS data, static work plans, and geographic information data. Specifically, it integrates Automatic Identification System (AIS) data, loading and unloading operation plan data, weather forecast data, and weather correction data for port energy forecasting. When performing load forecasting based on a dynamic Bayesian network model, uncertainty quantification (probability) is introduced into load forecasting, enabling probabilistic and quantitative forecasting of planned emergencies. When performing wind and solar power output forecasting based on terrain perception, a dedicated wind and solar power forecasting correction model for port coastal topography and sea surface environment is established for the first time, eliminating systematic biases and improving the forecasting accuracy of load and wind and solar power output, which is beneficial for better early warning and regulation of the port energy system.

[0097] Example 2.

[0098] This embodiment provides a method for early warning of port energy load and wind and solar power output, such as Figure 5 As shown, the port energy load and wind and solar power output early warning method includes the following steps.

[0099] Step T1: Obtain the load power mutation probability, load power mutation amplitude, wind power generation power, and photovoltaic power generation power at each prediction time point within the prediction period determined by the port energy load and wind and solar power output prediction method described in Example 1.

[0100] Step T2: Based on the wind power generation and photovoltaic power generation at each prediction time in the prediction period, calculate the rate of change of wind power generation and photovoltaic power generation at each prediction time in the prediction period.

[0101] Step T3: Based on the load power change probability, load power change amplitude, wind power power change rate, and photovoltaic power change rate at each prediction time in the prediction period, determine whether to trigger an early warning, and generate an early warning trigger signal when an early warning is triggered.

[0102] like Figure 6 As shown, it is a schematic diagram of the predictive-control linkage timing. The following, combined with... Figure 6 The port energy load and wind and solar power output early warning methods used in this embodiment are described in detail.

[0103] Predictive-control linkage is the core bridge connecting the predictive system and the control system. It transforms the prediction results into executable early warning trigger signals, directly triggering subsequent control strategies and achieving a seamless closed loop from prediction to control. This is key to solving the problem of the disconnect between prediction and control strategies in port energy systems. Relying on predictive-control linkage, this embodiment aims to solve the following technical problem: how to transform prediction results into early warning trigger signals that can directly drive the control system to act in advance, forming a closed-loop predictive-control linkage and solving the problem of the disconnect between prediction and control.

[0104] like Figure 6 As shown, the predictive-control linkage specifically includes: the predictive system continuously calculates the probability of load power surges, the magnitude of load power surges, the rate of change of wind power generation, and the rate of change of photovoltaic power generation. When the early warning triggering conditions are detected, an early warning trigger signal is generated and sent to the control system. The control system then activates a preset control strategy (such as activating hydrogen energy) based on the content of the early warning trigger signal to achieve stable power balance, i.e., to perform regulation. The early warning triggering conditions include: Condition 1): The existence of... Make and 15% of the port's total load; Condition 2): The absolute value of the wind and solar power change rate (i.e., the wind power generation change rate and the photovoltaic power generation change rate) exceeds the set value. The early warning trigger signal shall include at least: the early warning type, the expected occurrence time, and the power deviation amplitude, and shall be used to directly trigger subsequent control strategies.

[0105] Specifically, calculating the rate of change of wind power generation and the rate of change of photovoltaic power generation includes: based on and Calculate the rate of change of wind power generation and photovoltaic power generation within 5 minutes (or other time periods) to obtain the rate of change of wind power generation. and photovoltaic power generation change rate The specific calculation method is as follows: for each prediction time... ,Pick and Calculate the rate of change of power at +5 minutes: = ; in, The rate of change of power; for Power at +5 minutes; for Power at that time; This is the rated power.

[0106] wind power generation capacity As Rated power of wind turbine generator set As Substituting into the above formula, the calculated result is... That is, the rate of change of wind power generation. Photovoltaic power generation As Rated power of photovoltaic array As Substituting into the above formula, the calculated result is... That is, the rate of change of photovoltaic power generation. .

[0107] The wind power generation power change rate and photovoltaic power generation power change rate calculated by the above formula have been normalized to the rated power of the corresponding equipment, and the unit is the percentage change per minute (% / min).

[0108] Specifically, the judgment of the early warning triggering conditions includes setting two independent early warning triggering conditions, and generating an early warning triggering signal immediately when and only when any of the following conditions are met.

[0109] Condition 1): High probability of sudden changes in high-power load warning, existing. Make and , The total port load represents the total capacity at some point in the future. At the same time, it meets the following conditions: ① The probability of a sudden change in load power exceeds 80% ( ), and ② the magnitude of the load power change exceeds the total port load. 15% If the condition is met, an early warning will be triggered. This condition provides a high-confidence advance warning for planned but uncertain major load changes such as ship berthing.

[0110] Condition 2): Early warning of drastic changes in wind and solar power. or This indicates that if the absolute value of the rate of change of wind and solar power exceeds ±10% / min ( or If the weather conditions are abnormal, an early warning will be triggered. This condition addresses the intermittency and volatility of renewable energy, and responds quickly to sudden increases or decreases in wind and solar power output caused by weather changes.

[0111] Specifically, the generation and transmission of the early warning trigger signal includes: once any of the above conditions are triggered, an automatically generated structured early warning trigger signal (an early warning trigger signal conforming to predetermined data specifications) is uploaded to the control system in real time via standard communication protocols (such as OPC UA, MQTT).

[0112] The warning trigger signal shall include at least the following fields: (1) (Early warning type): is an enumerated variable used to indicate the trigger source (e.g., 1 = load change warning, 2 = photovoltaic power drop warning, 3 = wind power surge warning, etc.); (2) (Expected time of occurrence): Timestamp, accurate to the second, indicating the time when the disturbance is expected to occur (i.e., the time when the conditions are met). (3) (Power Deviation Amplitude): A floating-point number in kW or MW, indicating the expected power change (i.e., (or changes in wind and solar power); (4) (optional) (Confidence level): Floating-point number (0~1), for example, carrying The value is used by the control system to evaluate the reliability of the early warning system.

[0113] The linkage with the control system may include: a warning trigger signal used to directly trigger the multi-mode switching mechanism, for example, upon receiving a warning signal... (Sudden load change) , After receiving the warning signal, the control system can start the hydrogen fuel cell in advance or adjust the energy storage discharge strategy to achieve zero-impact local balance when a sudden increase in load occurs.

[0114] Through the above process, this embodiment defines early warning triggering conditions and generates a structured early warning triggering signal containing early warning type, expected occurrence time, and power deviation amplitude during prediction-control linkage. It transforms the prediction results into an executable control strategy, realizing a standardized and automated closed loop from prediction to control. This solves the technical problem of the disconnect between prediction and control, embodies the core idea of ​​prediction-control linkage, and by setting quantified early warning triggering conditions, the generation of early warning triggering signals is no longer a subjective judgment, but an objective and quantifiable automated process, enhancing practicality.

[0115] In this embodiment, based on the load power mutation probability, load power mutation amplitude, wind power generation power change rate, and photovoltaic power generation power change rate at each prediction moment within the prediction time period, it is determined whether an early warning should be triggered. If an early warning is triggered, an early warning trigger signal is generated. Specifically, this includes: for each prediction moment within the prediction time period, determining whether the load power mutation probability at the prediction moment is greater than a probability threshold and whether the load power mutation amplitude at the prediction moment is greater than a load threshold; if so, an early warning is triggered, and an early warning trigger signal is generated. The early warning trigger signal includes the early warning type, planned occurrence time, and power deviation amplitude. For each prediction moment within the prediction time period, determining whether the absolute value of the wind power generation power change rate at the prediction moment is greater than a first change rate threshold; if so, an early warning is triggered, and an early warning trigger signal is generated. For each prediction moment within the prediction time period, determining whether the absolute value of the photovoltaic power generation power change rate at the prediction moment is greater than a second change rate threshold; if so, an early warning is triggered, and an early warning trigger signal is generated.

[0116] The core value of this embodiment lies in providing a method for predicting and warning of port energy load and wind and solar power output, which has the following advantages.

[0117] (1) Significantly improved prediction accuracy: 1) For shore power load, it can accurately predict the occurrence time and power magnitude of sudden load change events, reducing the prediction error of the change point time from more than ±45 minutes in the traditional method to within ±8 minutes, and the power amplitude prediction accuracy exceeds 90%. 2) For wind and solar power output, through terrain and reflection correction, the systematic prediction bias caused by the local environment of the port is eliminated, and the short-term prediction error is generally reduced from more than 20% to less than 10%, effectively coping with sudden weather changes. Compared with load prediction methods based on historical load data, this embodiment can predict sudden events that have never occurred before (such as the first berthing of a new route vessel), solving its inherent blind spots, providing a precise time window and power demand for vessel energy consumption for collaborative strategies, and enabling collaborative strategies to move from concept to precise implementation.

[0118] (2) Provide a forward-looking decision window for the control system: by outputting probabilistic and quantitative prediction results ( , It also provides structured early warning trigger signals to the control system 5-15 minutes in advance, enabling the control system to anticipate the future and activate backup resources (such as hydrogen energy and energy storage) in advance, transforming passive response into active defense, fundamentally avoiding power shortage or over-limit accidents, and achieving true local balance.

[0119] (3) High level of automation and intelligence: 1) The entire process, from data acquisition and model calculation to early warning generation, is fully automated and requires no human intervention, overcoming the drawbacks of traditional scheduling that relies on human experience. 2) Based on probability prediction and clear triggering conditions, the generation of early warning triggering signals is scientific and objective, greatly reducing false alarms and missed alarms, and improving the reliability of linkage control.

[0120] (4) Significant economic and safety benefits: 1) Economic efficiency: High-precision forecasting makes energy dispatching plans more efficient, reduces the need for backup resources to cope with uncertainties, and lowers operating costs. At the same time, it reduces the impact of power demand on the upstream power grid and avoids punitive electricity prices. 2) Safety: Early warning allows the system sufficient time to respond, greatly improving the operational safety and reliability of the port power grid and avoiding risks such as voltage drops, frequency fluctuations, and even equipment disconnection caused by power surges.

[0121] Example 3.

[0122] This embodiment provides a port energy load and wind and solar power output prediction and early warning system, such as Figure 7 As shown, it includes the following modules.

[0123] The data acquisition module is used to acquire real-time data from the Automatic Identification System (AIS) of ships in the port area, loading and unloading operation plans, weather forecasts, and weather corrections.

[0124] The load forecasting engine has a built-in dynamic Bayesian network model for load forecasting.

[0125] The wind and solar power prediction engine has a built-in terrain perception correction algorithm to complete the prediction of wind and solar power output.

[0126] The early warning interface module is used to generate and send early warning trigger signals.

[0127] A database is used to store historical data, model parameters, and real-time data streams.

[0128] Example 4.

[0129] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 8 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When executed by the processor, the computer program implements a method for predicting port energy load and wind / solar power output, or a method for early warning of port energy load and wind / solar power output.

[0130] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0131] In one exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the port energy load and wind and solar power output prediction method of Embodiment 1 or the port energy load and wind and solar power output early warning method of Embodiment 2.

[0132] Example 5.

[0133] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the port energy load and wind and solar power output prediction method of Embodiment 1 or the port energy load and wind and solar power output early warning method of Embodiment 2.

[0134] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of the relevant data are carried out in compliance with the relevant data protection laws and policies of the country where the location is located, and with the authorization granted by the owner of the corresponding device.

[0135] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0136] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for predicting port energy load and wind and solar power output, characterized in that, The methods for predicting port energy load and wind and solar power output include: Acquire data from the Automatic Identification System (AIS) of vessels in the port area, including loading and unloading operation plan data, weather forecast data, and weather correction data. The AIS data includes the planned berthing time and vessel identification code of the vessel. The loading and unloading operation plan data includes the planned start time, planned stop time, and rated power of each loading and unloading equipment. The weather correction data includes the coastline slope, coastline direction, and sea surface reflectivity. Vessels whose planned berthing time falls within the predicted time period are selected as target vessels. Multiple historical data sets are extracted based on the target vessel's vessel identification code. The on-time berthing probability of the target vessel is calculated based on all historical data sets. The historical data sets include historical planned berthing time and historical actual berthing time. Based on the loading and unloading operation plan data, the planned power demand and port operation busy index for each forecast time in the forecast period are calculated. Based on meteorological forecast data, determine the weather conditions at each forecast time within the forecast period; Using the on-time berthing probability of the target vessel, the planned power demand, port operation busy index, and weather conditions at each forecast time in the forecast period as inputs, a dynamic Bayesian network model is used to determine the load power mutation probability and load power mutation amplitude at each forecast time in the forecast period, and the load forecast result is obtained. The wind speed in the weather forecast data is corrected by using the slope and orientation of the coastline to obtain the corrected wind speed at each forecast time in the forecast period. The irradiance in the weather forecast data is corrected by using the sea surface reflectance to obtain the corrected irradiance at each forecast time in the forecast period. Based on the corrected wind speed at each forecast moment within the forecast period, the wind power generation at each forecast moment within the forecast period is calculated. Based on the corrected solar irradiance at each forecast moment within the forecast period, the photovoltaic power generation at each forecast moment within the forecast period is calculated, thus obtaining the wind and solar power output forecast results. The load forecast results and the wind and solar power output forecast results can be used for early warning and regulation of the port energy system.

2. The method for predicting port energy load and wind and solar power output according to claim 1, characterized in that, The on-time berthing probability of the target vessel is calculated based on all historical data sets, specifically including: For each historical data set, calculate the difference between the historical planned berthing time and the historical actual berthing time in the historical data set to obtain the deviation of the historical data set; The deviations of all historical data sets are combined into a deviation dataset, and the mean and standard deviation of the deviation dataset are calculated using the maximum likelihood estimation method; the deviation dataset follows a normal distribution. Based on the mean and standard deviation, the probability of the target vessel berthing on time is calculated. The formula for calculating the probability of on-time berthing is: ； in, To increase the probability of berthing on time; It is the cumulative distribution function; It follows a normal distribution. The mean, Standard deviation; For the planned berthing time; This is an acceptable threshold.

3. The method for predicting port energy load and wind and solar power output according to claim 1, characterized in that, Based on loading and unloading operation plan data, the planned power demand and port operation busy index for each forecast time point within the forecast period are calculated, specifically including: For each prediction time within the prediction period, perform the following steps: For each loading and unloading equipment, determine whether the predicted time is between the planned start time and the planned stop time of the loading and unloading equipment. If so, the loading and unloading equipment is recorded as a working loading and unloading equipment. Calculate the sum of the rated power of all loading and unloading equipment to obtain the planned power demand at the predicted time. Based on the number of loading and unloading equipment, the port operation busy index for the predicted time is determined.

4. The method for predicting port energy load and wind and solar power output according to claim 1, characterized in that, The dynamic Bayesian network model includes a Bayesian network model for each prediction time in the prediction time period. The Bayesian network model includes a first input node, a second input node, a third input node, a fourth input node, an intermediate node, a first output node, and a second output node. The input terminal of the first input node is used to receive the on-time berthing probability of the target vessel, and the output terminal of the first input node is connected to the first input terminal of the intermediate node. The first input terminal of the second input node is used to receive the planned power demand at the predicted time, and the output terminal of the second input node is connected to the first input terminal of the second output node. The first input of the third input node is used to receive the port operation busy index at the prediction time. The second input of the third input node is connected to the output of the third input node of the Bayesian network model at the previous prediction time. The output of the third input node is connected to the second input of the second input node. The first input of the fourth input node is used to receive the weather conditions at the prediction time. The second input of the fourth input node is connected to the output of the fourth input node of the Bayesian network model at the previous prediction time. The output of the fourth input node is connected to the second input of the intermediate node. The third input of the intermediate node is connected to the output of the intermediate node of the Bayesian network model at the previous prediction time. The first output of the intermediate node is connected to the input of the first output node. The second output of the intermediate node is connected to the second input of the second output node. The output of the first output node is used to output the probability of load power change at the prediction time. The output of the second output node is used to output the magnitude of load power change at the prediction time.

5. The method for predicting port energy load and wind and solar power output according to claim 1, characterized in that, The revised formula for calculating wind speed is: ； in, This is the corrected wind speed; Wind speed; This is a terrain correction factor, determined based on the coastline slope and coastline orientation; This refers to the slope of the coastline.

6. The method for predicting port energy load and wind and solar power output according to claim 1, characterized in that, The revised formula for calculating irradiance is: ； in, This is the corrected irradiance. Light irradiance; Install tilt angle correlation coefficients for the components; This represents the sea surface reflectivity.

7. A method for early warning of port energy load and wind and solar power output, characterized in that, The port energy load and wind and solar power output early warning methods include: Obtain the load power mutation probability, load power mutation amplitude, wind power generation power and photovoltaic power generation power at each prediction time in the prediction period determined by the port energy load and wind and solar power output prediction method according to any one of claims 1-6. Based on the wind power generation and photovoltaic power generation at each prediction time within the prediction period, the rate of change of wind power generation and photovoltaic power generation at each prediction time within the prediction period are calculated. Based on the probability of load power change, the magnitude of load power change, the rate of change of wind power generation, and the rate of change of photovoltaic power generation at each prediction time within the prediction period, it is determined whether an early warning should be triggered, and an early warning trigger signal is generated when an early warning is triggered.

8. The port energy load and wind and solar power output early warning method according to claim 7, characterized in that, Based on the probability of load power abrupt change, the magnitude of load power abrupt change, the rate of change of wind power generation, and the rate of change of photovoltaic power generation at each forecast time within the forecast period, it is determined whether an early warning should be triggered. If an early warning is triggered, an early warning trigger signal is generated, specifically including: For each prediction time within the prediction period, determine whether the probability of a sudden change in load power at the prediction time is greater than a probability threshold, and whether the magnitude of the sudden change in load power at the prediction time is greater than a load threshold. If so, trigger an early warning and generate an early warning trigger signal. The early warning trigger signal includes the early warning type, the planned occurrence time, and the magnitude of the power deviation. For each prediction time in the prediction period, determine whether the absolute value of the rate of change of wind power generation at the prediction time is greater than the first rate of change threshold. If so, trigger an early warning and generate an early warning trigger signal. For each prediction time within the prediction period, determine whether the absolute value of the rate of change of photovoltaic power generation at the prediction time is greater than the second rate of change threshold. If so, trigger an early warning and generate an early warning trigger signal.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that the processor executes the computer program to implement the port energy load and wind and solar power output prediction method according to any one of claims 1-6 or the port energy load and wind and solar power output early warning method according to any one of claims 7-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the port energy load and wind and solar power output prediction method according to any one of claims 1-6 or the port energy load and wind and solar power output early warning method according to any one of claims 7-8.

Images