A road heterogeneous energy collaborative energy supply control method and system

Abstract

The present application relates to the field of road engineering and intelligent control technology, and specifically provides a road heterogeneous energy collaborative energy supply control method and system, the method comprising: collecting traffic flow, weather, load and energy storage data; predicting piezoelectric output, photovoltaic output and energy storage dispatchable amount; calculating real-time matching degree, flexible adjustment potential and matching degree deviation, taking the difference between predicted matching degree and real-time matching degree as matching degree prediction trend; taking matching degree deviation, matching degree prediction trend, flexible adjustment potential and energy storage state of charge as input, solving output coefficient through fuzzy rule base; switching normal, warning, emergency three modes according to energy storage state of charge, prediction error and external warning signal, reducing prediction dependence or degrading to no prediction mode when prediction error exceeds threshold value; adjusting power electronic converter and executing hierarchical load shedding according to priority. The present application improves the collaborative energy supply efficiency and power supply reliability of heterogeneous energy in road scene.

Description

Technical Field

[0001] This invention relates to the fields of road engineering and intelligent control technology, specifically to a method and system for controlling the coordinated supply of heterogeneous energy sources in roads. Background Technology

[0002] In the field of energy self-sufficiency for road infrastructure, piezoelectric power generation and roadside photovoltaic power generation form a heterogeneous energy complementary system. Piezoelectric output is driven by traffic flow and exhibits high-frequency pulse characteristics; photovoltaic output is affected by weather and is fluctuating. Dynamic matching between these two systems and the road load is crucial to ensuring power supply reliability. Existing multi-energy coordinated control methods for roads are mostly based on passive, real-time power allocation, lacking the ability to predict short-term fluctuations and failing to fully consider the differentiated reliability requirements of road loads and the coupling characteristics of traffic flow and photovoltaics. Furthermore, existing methods have imperfect fault-tolerance mechanisms for prediction errors, making it difficult to guarantee system reliability when predictions fail. Therefore, there is an urgent need to construct a coordinated energy supply control method that integrates short-term prediction, fuzzy decision-making, and fault-tolerant control to achieve an improvement from passive response to proactive pre-scheduling. Summary of the Invention

[0003] To address the aforementioned issues, this invention provides a method and system for controlling the coordinated power supply of heterogeneous energy sources in road environments. It achieves proactive pre-scheduling by directly inputting the difference between the predicted matching degree and the real-time matching degree as the predicted matching degree trend into a fuzzy controller. A hierarchical response mechanism for prediction errors is established to ensure power supply reliability in the event of prediction failure. Furthermore, based on the energy storage charge state, prediction errors, and external early warning signals, it drives adaptive switching between normal, early warning, and emergency operating modes. This achieves an improvement from passive response to proactive pre-scheduling, enhancing the efficiency and reliability of coordinated power supply of heterogeneous energy sources in road scenarios.

[0004] In a first aspect, the technical solution of the present invention provides a method for controlling the coordinated supply of heterogeneous energy sources in roads, comprising the following steps: Collect traffic flow data, meteorological data, load data, and the state of charge of energy storage units for the target road section; Based on the collected traffic flow data and meteorological data, the forecast results for the future period are predicted, including the predicted output of piezoelectric power, the predicted output of photovoltaic power, and the predicted dispatchable amount of energy storage. Based on the collected load data and the actual output of each energy unit, the real-time matching degree index and flexible adjustment potential at the current moment are calculated, and the matching degree deviation is calculated based on the real-time matching degree index. The predicted matching degree is calculated based on the prediction results, and the matching degree prediction trend is calculated based on the difference between the real-time matching degree and the predicted matching degree. The matching degree deviation, the matching degree prediction trend, the flexible adjustment potential, and the state of charge of the energy storage unit are used as inputs, and the piezoelectric output coefficient, photovoltaic output coefficient, and energy storage charge and discharge coefficient are solved by a fuzzy rule base. Based on the state of charge of the energy storage unit, the piezoelectric output prediction error, and the photovoltaic output prediction error, the system will switch to normal mode, early warning mode, or emergency mode; when any prediction error exceeds the corresponding set threshold, the dependence of fuzzy decision on the prediction result will be reduced, or the system will be downgraded to a no-prediction mode. The power electronic converter is adjusted according to the output coefficient and charge / discharge coefficient obtained by the solution, and the graded load reduction is performed according to the preset flexible load reduction priority to realize the coordinated power supply of multiple energy sources to the load.

[0005] Secondly, the technical solution of the present invention provides a road heterogeneous energy collaborative power supply control system, implemented based on the above-mentioned method, including: The data acquisition module is used to collect traffic flow data, meteorological data, load data, and the state of charge of the energy storage unit for the target road section. The supply and demand status calculation module is used to predict the forecast results for future periods based on the collected traffic flow data and meteorological data. The forecast results include piezoelectric predicted output, photovoltaic predicted output and energy storage predicted dispatchable quantity. Based on the collected load data and the actual output of each energy unit, the module calculates the real-time matching degree index and flexible adjustment potential at the current moment, and calculates the matching degree deviation based on the real-time matching degree index. The fuzzy decision module is used to calculate the predicted matching degree based on the prediction results, and then calculate the matching degree prediction trend based on the difference between the real-time matching degree and the predicted matching degree; taking the matching degree deviation, the matching degree prediction trend, the flexible adjustment potential and the state of charge of the energy storage unit as inputs, the piezoelectric output coefficient, the photovoltaic output coefficient and the energy storage charge and discharge coefficient are solved through the fuzzy rule base. The multi-modal operation management module is used to switch the system to normal mode, early warning mode or emergency mode based on the energy storage unit's state of charge, piezoelectric output prediction error and photovoltaic output prediction error. The fault-tolerant control module is used to reduce the dependence of fuzzy decision-making on the prediction results or to downgrade to a no-prediction mode when any prediction error exceeds the corresponding set threshold. The execution module is used to adjust the power electronic converter according to the output coefficient and charge / discharge coefficient obtained by the solution, and to perform graded load shedding according to the preset flexible load shedding priority, so as to realize the coordinated power supply of multiple energy sources to the load.

[0006] As can be seen from the above technical solutions, this application has the following advantages: By predicting the piezoelectric output, photovoltaic output, and energy storage dispatchable quantity in the future period through a predictive model, and directly inputting the difference between the predicted matching degree and the real-time matching degree as the matching degree prediction trend into the fuzzy controller, the system can take countermeasures in advance when the current supply and demand are balanced but the prediction shows that there will be a shortage in the future. This realizes the improvement from "post-event response" to "pre-event scheduling" and improves the collaborative energy supply efficiency of heterogeneous energy in road scenarios. Through a hierarchical response mechanism for prediction errors, the system responds progressively based on the severity of piezoelectric and photovoltaic prediction errors. Specifically, when the prediction error exceeds the first tolerance threshold, the corresponding prediction weight is reduced; when it exceeds the second tolerance threshold, a simplified model is activated; and when it exceeds the third tolerance threshold, the system degrades to a no-prediction mode. This ensures that the system can still operate reliably based on the measured values ​​or the simplified model when the prediction results deviate from the measured values, significantly improving the system's power supply reliability. The system is divided into three operating modes: normal mode, early warning mode, and emergency mode. The control strategy automatically switches between these modes, driven by the state of charge of the energy storage units, prediction errors, and external early warning signals. In normal mode, renewable energy is prioritized for consumption. In early warning mode, flexible loads are proportionally derated, and energy storage is prioritized to maintain a balanced state of charge. In emergency mode, all flexible loads are disconnected, ensuring power supply only to rigid loads. This avoids frequent mode oscillations and achieves adaptive adjustment of the system under different power supply conditions. Attached Figure Description

[0007] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying 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.

[0008] Figure 1 This is a schematic diagram of a road heterogeneous energy collaborative power supply control method provided in an embodiment of the present invention.

[0009] Figure 2 This is a schematic block diagram of a road heterogeneous energy collaborative power supply control system provided in an embodiment of the present invention. Detailed Implementation

[0010] To make the purpose, features, and advantages of this application more apparent and understandable, specific embodiments and accompanying drawings will be used to clearly and completely describe the technical solution protected by this application. Obviously, the embodiments described below are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0011] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this application and in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0012] Figure 1 This is a schematic flowchart of a road heterogeneous energy collaborative power supply control method provided by an embodiment of the present invention. Figure 1 The executing entity can be a road heterogeneous energy collaborative power supply control system. The road heterogeneous energy collaborative power supply control method provided in this embodiment of the invention is executed by computer equipment; correspondingly, the road heterogeneous energy collaborative power supply control system runs on the computer equipment. Depending on different requirements, the order of the steps in this flowchart can be changed, and some steps can be omitted.

[0013] like Figure 1 As shown, the method includes the following steps.

[0014] S1 collects traffic flow data, meteorological data, load data, and the state of charge of the energy storage unit for the target road segment.

[0015] S2, based on the collected traffic flow data and meteorological data, predict the forecast results for the future period, including the predicted output of piezoelectric power, the predicted output of photovoltaic power, and the predicted dispatchable amount of energy storage.

[0016] S3 calculates the real-time matching degree index and flexible adjustment potential at the current moment based on the collected load data and the actual output of each energy unit, and calculates the matching degree deviation based on the real-time matching degree index.

[0017] S4. Calculate the predicted matching degree based on the prediction results, and then calculate the matching degree prediction trend based on the difference between the real-time matching degree and the predicted matching degree. Using the matching degree deviation, the matching degree prediction trend, the flexible adjustment potential, and the state of charge of the energy storage unit as inputs, solve the piezoelectric output coefficient, photovoltaic output coefficient, and energy storage charge and discharge coefficient through a fuzzy rule base.

[0018] S5 switches the system to normal mode, early warning mode, or emergency mode based on the state of charge of the energy storage unit, the piezoelectric output prediction error, and the photovoltaic output prediction error; when any prediction error exceeds the corresponding set threshold, the dependence of fuzzy decision on the prediction result is reduced, or the system is downgraded to no prediction mode.

[0019] S6 adjusts the power electronic converter according to the output coefficient and charge / discharge coefficient obtained by the solution, and performs graded load shedding according to the preset flexible load shedding priority to realize the coordinated power supply of multiple energy sources to the load.

[0020] Furthermore, as a refinement and extension of the specific implementation of the above embodiments, in order to fully illustrate the specific implementation process in this embodiment, a specific embodiment of the road heterogeneous energy collaborative power supply control method is provided, which includes the following steps.

[0021] S101, Data Acquisition.

[0022] This step collects traffic flow data, meteorological data, load data, and the state of charge of energy storage units for the target road segment, serving as the basis for subsequent supply and demand forecasting, real-time matching degree calculation, and multimodal fault-tolerant control.

[0023] Traffic flow data is used to drive the piezoelectric output prediction model. Geomagnetic detectors or piezoelectric sensors are deployed in each lane of the target road segment to collect real-time data on vehicle axle load, speed, and arrival time.

[0024] Meteorological data is used to drive the photovoltaic power output prediction model. Small weather stations are installed on the roadside of the target road section to collect real-time data on light intensity, ambient temperature, and dynamic shading coefficient.

[0025] The dynamic shading coefficient is obtained through a fisheye camera or array-type light sensor installed above the photovoltaic module array. It is used to identify the dynamic shading of the photovoltaic modules by clouds, buildings, trees, etc. The dynamic shading coefficient ranges from [0,1], where 1 represents no shading and 0 represents complete shading. When using a camera solution, the dynamic shading coefficient is obtained by binarizing the captured image and identifying the shading area, and then calculating the ratio of the shading area to the total area.

[0026] Load data is used to calculate real-time matching indexes and flexible adjustment potential. Smart meters are installed in the power supply circuits of each electrical device in the target road section to collect real-time power data for each load. Total load requirements Total load requirements It is obtained by accumulating the real-time power of each load, i.e. .

[0027] Based on power supply reliability requirements, all electrical loads in the target road section are divided into two categories: rigid loads and flexible loads.

[0028] Rigid loads are those whose power outage would cause traffic safety hazards, equipment damage, or communication disruptions, and therefore require high power supply reliability. According to the "Detailed Design Specifications for Highway Tunnel Lighting" and communication industry standards, the power outage time for rigid loads should not exceed 2 seconds, and the voltage fluctuation rate should not exceed 5% / min. Rigid loads include traffic monitoring equipment, emergency lighting, drainage pumps, and communication base stations. Traffic monitoring equipment includes cameras, radar detectors, and traffic lights; emergency lighting includes tunnel emergency lighting and evacuation signs; drainage pumps include tunnel drainage pumps and service area drainage pumps; and communication base stations include 4G / 5G base stations and roadside communication units.

[0029] Flexible loads refer to loads that are allowed to be proportionally derating or briefly interrupted when power is insufficient, with a power interruption tolerance time not exceeding 30 seconds. Flexible loads include conventional LED lighting, information display screens, electric vehicle charging stations, and service area air conditioning. Conventional LED lighting includes road lighting and plaza lighting, with continuous dimming capability from 0% to 100%; information display screens include variable message signs and guidance screens, with brightness that can be reduced to 20%; electric vehicle charging stations, in accordance with the "Technical Specifications for Electric Vehicle Charging Infrastructure," allow charging power adjustment within the range of 30% to 100%; service area air conditioning includes central air conditioning and split air conditioning units, with power adjustment achieved through temperature setting or compressor start / stop.

[0030] The classification results of rigid loads and flexible loads are used as system preset parameters for the mode switching judgment in step S5, that is, in emergency mode, all flexible loads are cut off and only rigid loads are guaranteed to be powered, and for the graded load reduction strategy in step S6, that is, the load reduction target is only flexible loads, and rigid loads do not participate in the load reduction.

[0031] The state of charge (SOC) of the energy storage unit is used for fuzzy decision-making in step S3 and mode switching determination in step S5. The SOC and measured output are acquired in real time through the battery management system (BMS). Specifically, the BMS collects the current and voltage on the DC bus of the energy storage unit, calculates the actual charging and discharging power, and during discharge... It is positive when charging. It is a negative value.

[0032] S102, Supply and Demand Forecast.

[0033] In this embodiment, a two-level prediction architecture combining macro-trend prediction and micro-detail prediction is adopted. The first level is macro-trend prediction, which predicts the average vehicle arrival rate and average irradiance within a future set time period based on historical data. The macro-prediction has a low temporal resolution and is mainly used to capture the overall changing trends of vehicle flow and illumination, providing a benchmark for the second-level prediction. The second level is micro-detail prediction, such as predicting the arrival time of vehicle clusters and short-term irradiance fluctuations based on upstream and downstream detector data. The micro-prediction has a high temporal resolution and is mainly used to capture the random pulse characteristics of vehicle flow and rapid fluctuations in irradiance caused by cloud cover. The prediction results from the two levels are integrated through weighted fusion.

[0034] (1) Piezoelectric output prediction The piezoelectric output is determined by the axle load, vehicle speed, and arrival time of passing vehicles, exhibiting high-frequency pulse characteristics related to vehicle arrival. For piezoelectric output prediction, the first stage uses time series analysis to predict the macroscopic vehicle arrival rate, and the second stage uses a pulse queue model based on upstream detector data to predict the vehicle arrival time sequence. The two stages are fused to generate the piezoelectric output curve.

[0035] Level 1 Macro Traffic Flow Trend Prediction: Based on historical traffic flow data collected in step S1, time series analysis methods are used, such as the Seasonal Differential Autoregressive Moving Average (SARIMA) model, to predict the average traffic arrival rate within the next 15 minutes. (Unit: vehicles / minute). The parameters of the SARIMA model are optimized using the Akaike Information Criterion. For example, the optimized parameters are: autoregressive order p=2, differencing order d=1, moving average order q=2, and seasonality period s=60 (minutes). The model outputs prediction results with a time step of minutes to describe the overall trend of traffic arrival intensity.

[0036] The second level of microscopic vehicle arrival prediction: Based on upstream and downstream detector data, a pulse queue model is used to predict the arrival time sequence of vehicle clusters in future time periods. Specifically, the arrival time of each vehicle is recorded at the upstream detector, and the travel time of vehicles to the downstream target section is estimated based on the road segment length and average vehicle speed, forming an initial prediction sequence of vehicle arrival times. Considering the randomness of vehicle speed, a Gaussian process is used to model the uncertainty of travel time, expanding the initial arrival time sequence into a probability distribution of arrival times. On this basis, a Poisson process is used to model the gaps in traffic flow, generating a vehicle arrival pulse sequence with a time resolution of 10 seconds. For scenarios with upstream and downstream linkages (such as toll station release, traffic light cycles), a periodic traffic flow cluster structure is introduced into the pulse queue model, and the arrival time window of the vehicle cluster is predicted based on the phase of the upstream release cycle.

[0037] The average vehicle arrival rate predicted in the first stage is used as the intensity parameter of the second-stage pulse queue model, ensuring that the overall density of the micro-pulse sequence conforms to the macro-trend prediction. Specifically, the statistical mean of the vehicle arrival pulse sequence generated in the second stage is constrained to be equal to the average vehicle arrival rate output in the first stage. The fused vehicle arrival prediction sequence is input into the piezoelectric power conversion model, which calculates the single-pulse energy based on the axle load and vehicle speed of each vehicle (using historical statistical averages or real-time detection values), and the accumulated energy yields the piezoelectric predicted power curve. The piezoelectric output conversion model can be expressed as:

[0038] in, Let i be the axle load of the i-th vehicle. Let be the speed of the i-th vehicle. The single-pulse energy function is determined based on the piezoelectric unit calibration curve. Let be the Dirac function, representing the time of arrival. Generates pulses.

[0039] (2) Photovoltaic power output forecast Photovoltaic power output is determined by irradiance, ambient temperature, and dynamic shading coefficient, exhibiting fluctuations related to meteorological conditions. For photovoltaic power output prediction, the first stage uses a temporal convolutional network to predict macroscopic irradiance trends, while the second stage uses optical flow to predict short-term irradiance fluctuations. The two stages are then fused and used to generate a photovoltaic power output curve via a photovoltaic physical model.

[0040] First-level macroscopic irradiance trend prediction: Based on historical meteorological data collected in step S1 and numerical weather forecast data released by meteorological departments, a temporal convolutional network is used to predict the average irradiance within the next 15 minutes. and average ambient temperature The input features of the temporal convolutional network include historical light intensity sequences, temperature sequences, humidity sequences, historical dynamic shading coefficient sequences, and extrapolated irradiance data from numerical weather prediction. The network structure consists of four residual blocks, each composed of two dilated convolutional layers with dilation coefficients of 1, 2, 4, and 8, a kernel size of 3, and 64 filters. It is trained using the Adam optimizer, with mean squared error as the loss function. The first-level output provides irradiance and temperature prediction sequences with a temporal resolution of 1 minute, describing the overall trend of light intensity variation.

[0041] The second level of short-term irradiance fluctuation prediction: Based on the high-frequency light intensity data and cloud movement monitoring data collected in step S1, short-term irradiance fluctuations are predicted using the optical flow method. Specifically, sky images are acquired in real time using a fisheye camera, and the cloud movement speed and direction are calculated using the optical flow method. Combined with the current light intensity distribution, the changes in light intensity caused by cloud obstruction within the next 30 seconds to 5 minutes are extrapolated. The core equation of the optical flow method is:

[0042] Where I represents image brightness, x and y are pixel coordinates, and u and v are the components of optical flow in the x and y directions, respectively. By solving the above equations, the cloud motion field is obtained, and thus the sequence of light intensity changes for the target photovoltaic array in the near future can be predicted. The second-level output is an irradiance fluctuation sequence with a time resolution of 10 seconds.

[0043] The short-term irradiance fluctuations predicted by the second level are superimposed on the macroscopic irradiance trend predicted by the first level to obtain the fused irradiance intensity prediction sequence:

[0044] Substituting the fused light intensity sequence into the photovoltaic physical model, the predicted photovoltaic output is calculated:

[0045] in, The photoelectric conversion efficiency of a photovoltaic module. The effective area of ​​the photovoltaic module. For temperature coefficient, For reference temperature, The predicted component operating temperature (estimated based on predicted ambient temperature and light intensity). This is the dynamic shading correction factor (calculated based on the dynamic shading coefficient β(t)). The final output is the photovoltaic predicted power output curve. .

[0046] (3) Prediction of dispatchable energy storage Predictable dispatchable energy storage capacity refers to the maximum discharge or maximum charge that an energy storage unit can provide within a future time period. It is used to determine whether energy storage can compensate for the power output shortfall of piezoelectric and photovoltaic systems. The dispatchable energy storage capacity is calculated forward using the ampere-hour integral method based on the current state of charge of the energy storage and the predicted net power for the future time period.

[0047] The forecast of dispatchable energy storage capacity is based on the current state of charge (SOC(t)) of the energy storage and the predicted net power for future periods. Predicted net power = predicted piezoelectric output + predicted photovoltaic output - predicted load demand. The predicted load demand is obtained from the historical load data collected in step S1 and the traffic volume prediction model.

[0048] The change in the state of charge of energy storage is calculated by using the ampere-hour integral method:

[0049] in, For the rated capacity of energy storage, To predict net power, positive values ​​represent discharging, and negative values ​​represent charging. When When the charge level exceeds the upper limit of the energy storage's state of charge, the energy storage will be unable to continue absorbing power, and the charging capacity of the energy storage's dispatchable capacity will gradually decrease to zero; when When the energy storage charge level is below the minimum, the energy storage will be unable to continue discharging power, and the discharge capacity of the energy storage will gradually decrease to zero.

[0050] Dispatchable energy storage capacity Defined as the maximum discharge power or maximum charging power that energy storage can continuously provide, provided that the energy storage state of charge constraints are met. Its upper limit is also subject to the maximum allowable charge and discharge power reported by the battery management system. and Constraints: (Discharge direction) (Charging direction) Final output energy storage dispatchable quantity curve The unit is kW.

[0051] S103, Supply and demand status calculation.

[0052] This step calculates the real-time matching degree index and flexible adjustment potential at the current moment based on the load data collected in step S1 and the actual output of each energy unit, and calculates the matching degree deviation based on the real-time matching degree index.

[0053] Real-time matching metrics It is used to quantify the degree of matching between the system's current power supply and load demand, and is defined as the ratio of the sum of the actual output of each energy unit to the total load demand.

[0054] Let the current time be t, and the actual output of the piezoelectric element be... The actual output of the photovoltaic modules is The actual output of the energy storage unit is Total load requirement is .in, and Data is collected in real time through their respective electricity metering devices; The data is collected through the battery management system, with positive values ​​taken during discharge and negative values ​​taken during charging. It is obtained by accumulating the real-time power of each load.

[0055] Specifically, real-time matching degree metric The calculation formula is:

[0056] when A value greater than 1 indicates that the current energy supply is sufficient, and excess energy can be stored in the energy storage unit; when When the value is less than 1, it indicates that there is a power deficit that needs to be compensated by discharging energy storage or reducing flexible loads; when... When the value equals 1, it indicates that supply and demand are in balance.

[0057] The flexibility adjustment potential F(t) is used to quantify the total amount of power that can be released or reduced by adjusting the flexible load at the current moment. It is defined as the sum of the products of the current power demand of various flexible loads and their adjustability coefficients.

[0058] Let the current power demand of the j-th type of flexible load be... Its adjustable coefficient is The flexible adjustment potential F(t) is calculated using the following formula:

[0059] Among them, the adjustable coefficient Determined based on the physical characteristics and industry standards of various flexible loads, this reflects the upper limit of the proportion that such loads can be reduced or adjusted without affecting their basic functions.

[0060] The specific values ​​for the adjustability coefficients of various flexible loads are as follows: For conventional LED lighting, it has a continuous dimming capability from 0% to 100%. Considering the basic safety requirements of road lighting, the adjustable coefficient is taken as 0.5, that is, it can be adjusted within a range of no less than 50% brightness.

[0061] For information display screens, the brightness can be reduced to 20% of the rated brightness, with an adjustable coefficient of 0.8, meaning that power can be reduced by up to 80%.

[0062] According to the "Technical Specifications for Electric Vehicle Charging Infrastructure", the charging power of electric vehicle charging piles can be continuously adjusted within the range of 30% to 100% of the rated power, with an adjustment coefficient of 0.3, meaning that the charging power can be reduced by up to 70%.

[0063] For air conditioners in service areas, their power can be adjusted by setting the temperature or starting / stopping the compressor. The adjustable coefficient is 0.5, meaning it can be adjusted within a range of no less than 50% of the cooling / heating capacity.

[0064] The unit of flexible adjustment potential F(t) is kW. At the current moment, if the system experiences a power deficit, the maximum total power that can be reduced by adjusting all flexible loads.

[0065] The matching deviation e(t) is used to quantify the degree of deviation between the current power supply and the load demand. It is defined as the difference between the real-time matching index and the ideal matching value (i.e., 1).

[0066] The formula for calculating the matching deviation e(t) is:

[0067] The matching deviation e(t) ranges from -1 to 1. When e(t) is positive, it indicates a power shortage (supply falls short of demand), and the larger the positive value, the more severe the shortage. When e(t) is negative, it indicates a power surplus (supply exceeds demand), and the larger the absolute value of the negative value, the greater the surplus. When e(t) equals 0, it indicates a balance between supply and demand.

[0068] S104, Collaborative Control Decision.

[0069] This step is used to calculate the predicted matching degree based on the prediction results of step S2, and then calculate the matching degree prediction trend based on the difference between the real-time matching degree and the predicted matching degree. Taking the matching degree deviation, the matching degree prediction trend, the flexible adjustment potential, and the state of charge of the energy storage unit as inputs, the piezoelectric output coefficient, photovoltaic output coefficient, and energy storage charge and discharge coefficient are solved through the fuzzy rule base to provide decision instructions for execution control.

[0070] Predicted matching degree It is used to quantify the degree of predictive matching between the system's power supply and load demand in a future period, and is defined as the ratio of the sum of the predicted piezoelectric output, photovoltaic output and energy storage dispatchable capacity in the future period to the predicted load demand.

[0071] Let the future time be The piezoelectric predicted output is The photovoltaic power output is predicted to be The predicted dispatchable energy storage capacity is The predicted load demand is .in, , and The output of the prediction model in step S2, This was obtained based on historical load data and traffic volume prediction models.

[0072] In some specific embodiments, based on the relatively gradual load changes in road scenarios, the nearest neighbor extrapolation method is used to obtain the predicted load demand, that is, the measured total load demand collected in step S1 at the current time is directly used as the predicted value in the near future: .

[0073] Predicted matching degree The calculation method is to multiply the piezoelectric predicted output by the piezoelectric predicted weight, multiply the photovoltaic predicted output by the photovoltaic predicted weight, add this to the energy storage predicted dispatchable amount, and finally divide by the predicted load demand, expressed as:

[0074] in, For piezoelectric prediction weighting coefficients, These are the photovoltaic prediction weighting coefficients. Under normal operating conditions, and All values ​​are set to 1. When the fault tolerance mechanism in step S5 is triggered, or By reducing the prediction error level to the corresponding preset value, the impact of unreliable prediction results on the matching degree prediction trend can be weakened.

[0075] when A value greater than 1 indicates sufficient energy supply in the future, and excess energy can be stored in energy storage units; when... A value less than 1 indicates a power deficit in the future, requiring advance scheduling intervention; when... When the value equals 1, it indicates that supply and demand will be balanced in the future.

[0076] Matching prediction trend It is used to quantify the direction and rate of change of the supply and demand matching status in the future relative to the current time, and is defined as the difference between the predicted matching degree and the real-time matching degree divided by the prediction time interval.

[0077] Matching prediction trend The calculation formula is:

[0078] in, The real-time matching degree index calculated in step S3 is the current time. For the predicted matching degree at future moments, For the predicted time interval.

[0079] Matching prediction trend The positive or negative sign reflects the direction of future supply and demand trends: when When the value is positive, it indicates that the predicted matching degree has improved compared to the current matching degree, and the future energy supply will be more sufficient; when A negative value indicates that the predicted matching degree has deteriorated relative to the current matching degree, and a power deficit will occur or worsen in the future; when When the value is zero, it indicates that the predicted matching degree is basically the same as the current matching degree.

[0080] In this embodiment, the piezoelectric output coefficient, photovoltaic output coefficient and energy storage charge and discharge coefficient are solved by a fuzzy rule base, specifically including the following steps S104.1 to S104.5. S104.1 Map the matching degree deviation, matching degree prediction trend, flexible adjustment potential, and energy storage unit state of charge to their respective membership functions, and calculate the membership degree of each input variable on each linguistic value. Among them, the matching degree deviation is divided into five linguistic values: negative large, negative small, zero, positive small, and positive large; the matching degree prediction trend is divided into three linguistic values: negative, stable, and positive; the flexible adjustment potential is divided into three linguistic values: small, medium, and large; and the energy storage unit state of charge is divided into three linguistic values: low, medium, and high.

[0081] Specifically, the fuzzy controller has four input variables, all of which are dimensionless. Their value range is mapped to a unified universe of discourse [-1,1] or [0,1] through normalization.

[0082] The first input variable is the matching degree deviation e(t), which reflects the current degree of supply-demand imbalance. This variable is calculated in step S3, with negative values ​​indicating power surplus and positive values ​​indicating power deficit. The fuzzy controller divides e(t) into five linguistic values: negative large, negative small, zero, positive small, and positive large, each corresponding to a different membership function.

[0083] The second input variable is the matching degree prediction trend. This variable reflects the direction and rate of change in future supply and demand. It is obtained through calculation in step S104, and its value range is dynamically adjusted according to actual operating conditions. The fuzzy controller will... It is divided into three language values: negative trend, stable trend, and positive trend.

[0084] The third input variable is the flexibility adjustment potential F(t), which reflects the total power that can be reduced from the current flexible load. This variable is calculated in step S3 and, after normalization, takes a value in the range [0,1]. The fuzzy controller divides F(t) into three linguistic values: small, medium, and large.

[0085] The fourth input variable is the State of Charge (SOC) of the energy storage unit, reflecting the current energy level of the storage. This variable is acquired in real time through step S1, and its value ranges from 0% to 100%. The fuzzy controller divides the SOC into three linguistic values: low, medium, and high.

[0086] For the matching deviation e(t), its universe of discourse is set to [-0.5, 0.5]. For example, the membership function for a large negative value uses a Z-shaped function (left shoulder of a trapezoid), with a membership of 1 when e(t) is less than -0.3, and linearly decreasing to 0 between -0.3 and -0.1; the membership function for a small negative value uses a triangular function, with the center point at -0.15 and a bottom width of 0.2; the membership function for zero uses a triangular function, with the center point at 0 and a bottom width of 0.2; the membership function for a small positive value uses a triangular function, with the center point at 0.15 and a bottom width of 0.2; and the membership function for a large positive value uses an S-shaped function (right shoulder of a trapezoid), with a membership of 1 when e(t) is greater than 0.3, and linearly increasing between 0.1 and 0.3.

[0087] For matching degree prediction trend Its domain is set to [-0.1, 0.1]. For example, the negative function uses a Z-shaped function. When the membership value is less than -0.05, the membership degree is 1, and it linearly decreases to 0 between -0.05 and 0; a triangular function is used for stationary conditions, with the center point at 0 and a base width of 0.1; an S-shaped function is used for positive conditions. When the membership value is greater than 0.05, the membership degree is 1, and it increases linearly between 0 and 0.05.

[0088] For the flexible adjustment potential F(t), its universe of discourse is set to [0, 1] (after normalization). For example, a Z-shaped function is used for small F(t), with a membership degree of 1 when F(t) is less than 0.3, and linearly decreasing to 0 between 0.3 and 0.5; a triangular function is used for medium F(t), with a center point at 0.5 and a bottom width of 0.4; and an S-shaped function is used for large F(t), with a membership degree of 1 when F(t) is greater than 0.7, and linearly increasing between 0.5 and 0.7.

[0089] For the State of Charge (SOC) of the energy storage unit, its universe of discourse is set to [0, 1]. For example, the low-level function uses a Z-shaped function, with a membership degree of 1 when the SOC is less than 0.25, and linearly decreasing to 0 between 0.25 and 0.4; the middle-level function uses a triangular function, with the center point at 0.5 and a bottom width of 0.4; and the high-level function uses an S-shaped function, with a membership degree of 1 when the SOC is greater than 0.75, and linearly increasing between 0.6 and 0.75.

[0090] S104.2, the calculated membership degrees of each input variable are substituted into the pre-established fuzzy rule base. The premise of each rule is connected by the linguistic values ​​of each input variable through an "AND" operation. The activation strength of each rule is calculated by taking the minimum value operation.

[0091] The fuzzy rule base consists of several fuzzy conditional statements in the form of "if...then...", and each rule corresponds to a specific combination of system states and its control strategy.

[0092] The fuzzy rule base covers various typical combinations of supply and demand states. Below are some example rules: Rule 1: If the matching degree deviation is large (severe deficit) and the matching degree prediction trend is negative (continuous deterioration), then the piezoelectric output coefficient, the photovoltaic output coefficient, and the energy storage discharge coefficient are all taken as large (i.e., the energy storage is fully discharged), and the flexible load is triggered to reduce the load.

[0093] Rule 2: If the matching degree deviation is small (slight deficit) and the predicted matching degree trend is negative (tending to deteriorate) and the flexible adjustment potential is large, then the deficit should be compensated by reducing the flexible load first, and the energy storage should be discharged appropriately.

[0094] Rule 3: If the matching degree deviation is negative (large surplus) and the matching degree prediction trend is positive (tending to improve) and the energy storage state of charge is low, then the energy storage charging coefficient should be large, and the excess energy should be stored in the energy storage first.

[0095] Rule 4: If the matching degree deviation is negative (slight surplus) and the matching degree prediction trend is negative (tending to deteriorate) and the energy storage state of charge is high, then the energy storage discharge coefficient should be small to release the stored energy in advance to cope with future shortages.

[0096] Rule 5: If the matching degree deviation is zero (supply and demand are balanced) and the predicted matching degree trend is zero (basically flat), then maintain the current output status and keep the energy storage on standby.

[0097] Rule 6: If the energy storage state of charge is low and the matching degree deviation is positive, then the rigid load should be prioritized for power supply, the flexible load should be reduced, and the energy storage should be discharged appropriately according to the shortage.

[0098] Rule 7: If the energy storage is in a high state of charge and the matching deviation is negative, then the energy storage should be charged first to absorb the excess renewable energy.

[0099] For each input variable, the corresponding membership function is activated based on its current value to obtain the membership degree of each language value. For each rule, the premise satisfaction degree of the rule, i.e., the activation strength, is calculated by using an AND operation (taking the minimum value).

[0100] S104.3, based on the activation intensity of each rule, the minimum value operation is used to derive the output fuzzy set of the rule's conclusion part. The output linguistic values ​​of the piezoelectric output coefficient, photovoltaic output coefficient and energy storage charge and discharge coefficient are each divided into five linguistic values: negative large, negative small, zero, positive small and positive large.

[0101] S104.4, the output fuzzy sets obtained from the derivation of each rule are aggregated by taking the maximum value operation to obtain the overall output fuzzy sets of the piezoelectric output coefficient, photovoltaic output coefficient and energy storage charge and discharge coefficient respectively.

[0102] The output fuzzy set of each rule is obtained by performing an "implication" operation (taking the minimum value) based on the activation strength of each rule. The output fuzzy sets of all rules are then merged by performing an "aggregation" operation (taking the maximum value) to obtain the total output fuzzy set.

[0103] S104.5, the centroid method is used to defuzzify the overall output fuzzy set, and the centroid of the area enclosed by the membership function curve of each output fuzzy set and the horizontal axis is calculated to obtain the accurate values ​​of the piezoelectric output coefficient, photovoltaic output coefficient and energy storage charge and discharge coefficient.

[0104] Defuzzification employs the centroid method, which uses the centroid of the area enclosed by the membership function curve of the output fuzzy set and the horizontal axis as the precise output value. The formula for calculating the centroid is:

[0105] in, The membership function of the output fuzzy set is defined by the integration interval, which is the universe of discourse of the output variable.

[0106] By defuzzifying, three precise output quantities are obtained: piezoelectric output coefficient. Photovoltaic power output coefficient Energy storage charge / discharge coefficient .

[0107] piezoelectric output coefficient The value range is [0,1], and it is used to control the DC / DC converter of the piezoelectric unit. When When =0, the piezoelectric unit does not output power or is in standby mode; when When =1, the piezoelectric unit outputs at maximum available power; When the intermediate value is taken, the piezoelectric unit outputs power according to the corresponding ratio.

[0108] Photovoltaic power output coefficient The value range is [0,1], and it is used to control the operating mode of the photovoltaic inverter. When... When = 1, the photovoltaic inverter operates in maximum power point tracking mode, outputting the maximum available power; when When the value is less than 1, the photovoltaic inverter operates in power-limiting mode, and the output power is limited to 1 / 3 of the rated power. Within times.

[0109] Energy storage charge / discharge coefficient The value range is [-1, 1], and it is used to control the bidirectional energy storage converter. When When the value is greater than 0, the energy storage is in a discharge state, and the discharge power is the maximum allowable discharge power of the energy storage unit. times; when When the value is less than 0, the energy storage is in a charging state, and the charging power is the maximum allowable charging power of the energy storage unit. times; when When the value is 0, the energy storage is in standby mode, neither charging nor discharging.

[0110] The three output coefficients satisfy the normalization constraint: This constraint ensures the physical interpretability of the output coefficients, meaning that under any operating condition, the sum of the absolute values ​​of the three output coefficients does not exceed 1, thus preventing power electronic equipment from malfunctioning due to command conflicts.

[0111] S105, multimodal fault-tolerant control.

[0112] This step switches the system to normal mode, early warning mode, or emergency mode based on the energy storage unit's state of charge, piezoelectric output prediction error, and photovoltaic output prediction error. When any prediction error exceeds the corresponding set threshold, the fuzzy decision-making's dependence on the prediction results is reduced, or the system is downgraded to a no-prediction mode.

[0113] The prediction error is used to quantify the degree of deviation between the prediction result of step S2 and the measured value of step S1, in order to determine the prediction reliability and trigger mode switching.

[0114] Piezoelectric prediction error The calculation method is as follows: divide the absolute value of the difference between the predicted piezoelectric output and the measured piezoelectric output by the measured piezoelectric output, and then multiply by 100%. To avoid division by zero when the measured piezoelectric output is zero, a preset small positive number is added to the denominator. Piezoelectric prediction error Expressed as a percentage, the value ranges from 0% to positive infinity.

[0115] Specifically, piezoelectric prediction error The calculation formula is:

[0116] Photovoltaic prediction error The calculation method is as follows: divide the absolute value of the difference between the predicted photovoltaic power output and the measured photovoltaic power output by the measured photovoltaic power output, and then multiply by 100%. Similarly, add a small positive number to the denominator. To avoid division by zero errors. Photovoltaic prediction errors. Expressed as a percentage, the value ranges from 0% to positive infinity.

[0117] Specifically, photovoltaic prediction error The calculation formula is:

[0118] Based on the energy storage unit's state of charge, prediction errors, and external early warning signals, the operating mode is switched, including normal mode, early warning mode, or emergency mode.

[0119] The normal mode corresponds to a scenario where the system is in a normal state and the prediction results are reliable. When the state of charge of the energy storage unit is greater than or equal to the first state of charge threshold, and the piezoelectric prediction error is less than or equal to the first prediction error threshold, and the photovoltaic prediction error is less than or equal to the first prediction error threshold, the system enters or maintains the normal mode. In the normal mode, the system prioritizes the consumption of renewable energy, flexible loads operate at rated power, the energy storage unit charges and discharges according to the supply and demand matching situation, and the fuzzy decision controller normally uses the prediction results of step S2.

[0120] The early warning mode corresponds to scenarios where the system state enters the alert zone or the prediction reliability decreases. The system enters early warning mode when the energy storage unit's state of charge (SOC) is between the first SOC threshold and the second SOC threshold, or when the piezoelectric prediction error or photovoltaic prediction error is between the first prediction error threshold and the second prediction error threshold. In early warning mode, flexible loads are derated according to a derating factor calculated based on the current SOC of the energy storage unit. Specifically, this factor is the current SOC minus the second SOC threshold, divided by the difference between the first and second SOC thresholds. The energy storage unit prioritizes maintaining a balanced SOC and stops supplying power to the flexible load.

[0121] Emergency mode corresponds to scenarios where the system is in a critical state or where predictions have completely failed. When the state of charge of the energy storage unit falls below the second state of charge threshold, or the piezoelectric prediction error or photovoltaic prediction error exceeds the second prediction error threshold and continues for multiple control cycles, or the communication interruption time exceeds a preset interruption threshold, the system immediately switches to emergency mode. In emergency mode, the system disconnects all flexible loads, ensuring power supply only to rigid loads. The energy storage unit outputs at its maximum allowable discharge power, while the piezoelectric unit and photovoltaic modules output at full power. The fuzzy decision controller stops using prediction results and instead uses measured values ​​as input.

[0122] Preferably, to prevent frequent oscillations of the mode at threshold boundaries, the system includes mode recovery conditions and a hysteresis mechanism. The condition for switching from emergency mode back to early warning mode is that the energy storage unit's state of charge (SOC) recovers to above the third SOC threshold and remains above it for a preset time. The condition for switching from early warning mode back to normal mode is that the energy storage unit's SOC recovers to a level greater than or equal to the first SOC threshold, and both the piezoelectric prediction error and the photovoltaic prediction error recover to a level less than or equal to the first prediction error threshold. The hysteresis setting avoids frequent switching between modes.

[0123] Preferably, the system also considers switching operating modes based on external early warning signals. These external early warning signals originate from extreme weather warnings issued by meteorological departments, power grid dispatch instructions, or mandatory instructions from system maintenance personnel. Upon receiving an external early warning signal, the system takes preventative measures in advance based on the warning level. For meteorological warnings, the system enters early warning mode in advance, increases the state-of-charge reserves of energy storage units, and appropriately reduces the power of flexible loads. For power grid dispatch instructions, the system adjusts the photovoltaic output coefficient or energy storage charge / discharge coefficient accordingly. For mandatory instructions from maintenance personnel, the system unconditionally executes the specified mode switch.

[0124] When the prediction error exceeds a set threshold, the system automatically reduces the fuzzy decision-making's reliance on the prediction result, or downgrades to a no-prediction mode. This embodiment employs a three-level fault-tolerance mechanism, responding progressively according to the severity of the prediction error.

[0125] (1) When the piezoelectric prediction error exceeds the first fault tolerance threshold and continues for multiple control cycles, reduce the piezoelectric prediction weight and use real-time traffic flow as the control input.

[0126] The first level of fault tolerance addresses situations where the piezoelectric prediction error exceeds the first fault tolerance threshold and persists for multiple consecutive control cycles. In this case, the system determines that the reliability of the piezoelectric prediction has significantly decreased. For example, the piezoelectric prediction weights are adjusted accordingly. The input for piezoelectric predictive output in the fuzzy decision controller is reduced by 50% by replacing it with real-time traffic flow as the control input.

[0127] (2) When the photovoltaic prediction error exceeds the second fault tolerance threshold and continues for multiple control cycles, reduce the photovoltaic prediction weight and enable the model based on real-time irradiance.

[0128] The second level of fault tolerance addresses situations where the photovoltaic prediction error exceeds the second fault tolerance threshold and persists for multiple consecutive control cycles. In this case, the system determines that the reliability of the photovoltaic prediction has significantly decreased. For example, the photovoltaic prediction weights are adjusted accordingly. The input to the fuzzy decision controller involving photovoltaic power output prediction has been reduced by 70% by using a simplified model based on real-time irradiance. This simplified model does not rely on future predictions; it estimates the current power output based solely on the current measured light intensity and temperature, combined with a photovoltaic physical model, and uses the current power output as an approximation of the power output in the near future.

[0129] (3) When the piezoelectric prediction error or photovoltaic prediction error exceeds the third fault tolerance threshold and continues for multiple control cycles, the prediction is determined to be in failure, and the mode is downgraded to no prediction mode. The fuzzy controller uses the measured value as input, calls the simplified emergency rule set, and shortens the control cycle.

[0130] The third level of fault tolerance addresses situations where any prediction error exceeds the third fault tolerance threshold and persists for multiple consecutive control cycles. In this case, the system determines that the prediction has completely failed and automatically degrades to a no-prediction mode. In no-prediction mode, the fuzzy decision controller stops using all prediction results from step S2 and instead uses only the measured values ​​from step S1 as input. Simultaneously, the fuzzy controller invokes a simplified emergency rule set, which only takes real-time photovoltaic output, real-time traffic flow, and real-time energy storage state of charge as input, without considering any prediction information. Furthermore, the control cycle is automatically shortened to a preset emergency control cycle to improve the system's response speed to state changes.

[0131] Emergency rules, for example: (1) If real-time photovoltaic output = low AND real-time traffic flow = high AND real-time SOC = low, then prioritize critical loads and reduce the load of non-essential loads by 50%. (2) IF real-time photovoltaic output = high AND real-time traffic flow = low AND real-time SOC = high, THEN photovoltaic power supply is prioritized, and excess energy is stored in energy storage.

[0132] It should be noted that the mode switching focuses on selecting a macro-control strategy based on the overall system state, while the fault tolerance mechanism focuses on gradually adjusting the usage of prediction results as prediction errors increase. Both can ensure reliable system operation under the following three abnormal scenarios: Scenario 1: Moderate prediction error (e.g., 15% to 30%). The system enters early warning mode and selectively reduces the corresponding prediction weight based on the source of the error (piezoelectric or photovoltaic). Flexible loads are derated proportionally, and energy storage prioritizes maintaining a balanced state of charge. Scenario 2: Severe prediction error (e.g., 30% to 35%). The system maintains or enters emergency mode, while degrading to a simplified prediction model, disconnecting some flexible loads, and discharging the stored energy at a higher power. Scenario 3: Fatal prediction error (e.g., exceeding 35%) or communication interruption. The system is forced into emergency mode, while simultaneously degrading to no-prediction mode, cutting off all flexible loads and ensuring power supply only to rigid loads.

[0133] S106, Execution and Feedback.

[0134] This step adjusts the power electronic converter based on the piezoelectric output coefficient, photovoltaic output coefficient, and energy storage charge / discharge coefficient obtained in step S4, and performs graded load shedding according to the preset flexible load shedding priority, so as to realize the coordinated power supply of the load by the piezoelectric unit, photovoltaic module, and energy storage unit.

[0135] The three output coefficients in step S4 are dimensionless continuous values, which need to be converted into specific control signals to drive the power electronic converters of the piezoelectric unit, photovoltaic module and energy storage unit respectively.

[0136] For piezoelectric elements, the piezoelectric output coefficient The value range is 0 to 1. The local controller will Mapped to the output current reference value or output voltage reference value of the piezoelectric DC / DC converter. When When the value equals 0, the controller sends a shutdown command to the DC / DC converter, and the piezoelectric unit stops power output; when... When the value equals 1, the controller sends a maximum power output command, and the piezoelectric unit outputs the maximum available power; when... When taking the intermediate value, the controller calculates the target output power proportionally and adjusts the converter's duty cycle via a PI regulator to ensure the actual output power tracks the target value. Control commands are output as PWM signals, with the PWM frequency set according to the converter's switching frequency.

[0137] For photovoltaic modules, the photovoltaic output factor The value range is from 0 to 1. When When the value equals 1, the controller sends a maximum power point tracking (MPPT) command to the photovoltaic (PV) inverter. The inverter automatically searches for and tracks the maximum power point of the PV array, outputting the maximum available power. When... When the value is less than 1, the controller sends a power-limiting mode command to the photovoltaic inverter, and the target output power is set to 1 / 3 of the rated power. The inverter limits its output power to a target value by adjusting the operating point voltage; when When the value is 0, the controller sends a standby or shutdown command, and the photovoltaic module stops outputting power.

[0138] For energy storage units, the energy storage charge / discharge coefficient The value range is from -1 to 1. When When the value is greater than 0, the controller sends a discharge command to the energy storage bidirectional converter, and the target discharge power is set to the current maximum allowable discharge power of the energy storage unit. times; when When the value is less than 0, the controller sends a charging command, and the target charging power is set to the current maximum allowable charging power of the energy storage unit. times; when When the value equals 0, the controller sends a standby command, and the energy storage unit neither charges nor discharges. The control of charging and discharging power is achieved by adjusting the current loop reference value of the bidirectional converter, and a PI regulator is used to ensure that the actual power quickly tracks the target value.

[0139] The tiered load shedding strategy is used to gradually reduce the power of flexible loads according to a preset priority order when the system power is insufficient, so as to ensure the power supply reliability of rigid loads. The triggering of load shedding is determined by the operating mode of step S5, that is, partial load shedding is performed in early warning mode, and full load shedding is performed in emergency mode.

[0140] The load reduction priority of flexible loads is based on the system's preset parameters. The first-level load reduction target is electric vehicle charging piles, the second-level load reduction target is information display screens, the third-level load reduction target is target LED lighting, and the fourth-level load reduction target is target socket loads.

[0141] The target LED lighting can be conventional LED lighting. If the power is still insufficient after the second-level load reduction, the brightness of conventional lighting will be further reduced. The target socket load is non-essential socket load. If the power is still insufficient after the third-level load reduction and the system enters emergency mode, the power supply to non-essential socket loads will be completely cut off, including conventional sockets in non-critical areas of the service area and non-emergency electrical equipment.

[0142] Throughout the entire control process, the rigid load does not participate in any load reduction operation.

[0143] The foregoing has described in detail an embodiment of a road heterogeneous energy collaborative power supply control method. Based on the road heterogeneous energy collaborative power supply control method described in the above embodiment, this invention also provides a road heterogeneous energy collaborative power supply control system corresponding to the method.

[0144] Figure 2 This is a schematic block diagram of a road heterogeneous energy collaborative power supply control system provided in an embodiment of the present invention. In this embodiment, the road heterogeneous energy collaborative power supply control system can be divided into multiple functional modules according to the functions it performs. A module, as referred to in this invention, is a series of computer program segments that can be executed by at least one processor and perform a fixed function, and is stored in memory.

[0145] The data acquisition module is used to collect traffic flow data, meteorological data, load data, and the state of charge of the energy storage unit for the target road section.

[0146] The supply and demand status calculation module is used to predict the forecast results for future periods based on the collected traffic flow data and meteorological data. The forecast results include piezoelectric predicted output, photovoltaic predicted output and energy storage predicted dispatchable quantity. Based on the collected load data and the actual output of each energy unit, the module calculates the real-time matching degree index and flexible adjustment potential at the current moment, and calculates the matching degree deviation based on the real-time matching degree index.

[0147] The fuzzy decision module is used to calculate the predicted matching degree based on the prediction results, and then calculate the matching degree prediction trend based on the difference between the real-time matching degree and the predicted matching degree; taking the matching degree deviation, the matching degree prediction trend, the flexible adjustment potential and the state of charge of the energy storage unit as inputs, the piezoelectric output coefficient, the photovoltaic output coefficient and the energy storage charge and discharge coefficient are solved by the fuzzy rule base.

[0148] The multi-modal operation management module is used to switch the system to normal mode, early warning mode or emergency mode based on the energy storage unit's state of charge, piezoelectric output prediction error and photovoltaic output prediction error.

[0149] The fault-tolerant control module is used to reduce the dependence of fuzzy decision-making on the prediction results or to downgrade to a no-prediction mode when any prediction error exceeds the corresponding set threshold.

[0150] The execution module is used to adjust the power electronic converter according to the output coefficient and charge / discharge coefficient obtained by the solution, and to perform graded load shedding according to the preset flexible load shedding priority, so as to realize the coordinated power supply of multiple energy sources to the load.

[0151] The road heterogeneous energy collaborative power supply control system of this embodiment is used to implement the aforementioned road heterogeneous energy collaborative power supply control method. Therefore, the specific implementation of this system can be found in the embodiment section of the road heterogeneous energy collaborative power supply control method above. Thus, the specific implementation can be referred to the description of the corresponding embodiment, and will not be elaborated here.

[0152] Furthermore, since the road heterogeneous energy collaborative power supply control system of this embodiment is used to implement the aforementioned road heterogeneous energy collaborative power supply control method, its function corresponds to the function of the above method, and will not be repeated here.

[0153] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for controlling the coordinated supply of heterogeneous energy sources in roads, characterized in that, Includes the following steps: Collect traffic flow data, meteorological data, load data, and the state of charge of energy storage units for the target road section; Based on the collected traffic flow data and meteorological data, the forecast results for the future period are predicted, including the predicted output of piezoelectric power, the predicted output of photovoltaic power, and the predicted dispatchable amount of energy storage. Based on the collected load data and the actual output of each energy unit, the real-time matching degree index and flexible adjustment potential at the current moment are calculated, and the matching degree deviation is calculated based on the real-time matching degree index. The predicted matching degree is calculated based on the prediction results, and then the matching degree prediction trend is calculated based on the difference between the real-time matching degree and the predicted matching degree. Using the matching degree deviation, matching degree prediction trend, flexible adjustment potential, and energy storage unit state of charge as inputs, the piezoelectric output coefficient, photovoltaic output coefficient, and energy storage charge and discharge coefficient are solved by a fuzzy rule base. Based on the state of charge of the energy storage unit, the piezoelectric output prediction error, and the photovoltaic output prediction error, the system will switch to normal mode, early warning mode, or emergency mode; when any prediction error exceeds the corresponding set threshold, the dependence of fuzzy decision on the prediction result will be reduced, or the system will be downgraded to a no-prediction mode. The power electronic converter is adjusted according to the output coefficient and charge / discharge coefficient obtained by the solution, and the graded load reduction is performed according to the preset flexible load reduction priority to realize the coordinated power supply of multiple energy sources to the load. The piezoelectric output coefficient, photovoltaic output coefficient, and energy storage charge / discharge coefficient are solved using a fuzzy rule base, specifically including: The matching degree deviation, matching degree prediction trend, flexible adjustment potential, and energy storage unit state of charge are mapped to their respective membership functions, and the membership degree of each input variable on each linguistic value is calculated. Among them, the matching degree deviation is divided into five linguistic values: negative large, negative small, zero, positive small, and positive large; the matching degree prediction trend is divided into three linguistic values: negative, stable, and positive; the flexible adjustment potential is divided into three linguistic values: small, medium, and large; and the energy storage unit state of charge is divided into three linguistic values: low, medium, and high. The calculated membership degrees of each input variable are substituted into the pre-established fuzzy rule base. The premise of each rule is connected by the linguistic values ​​of each input variable through an "AND" operation. The activation strength of each rule is calculated by taking the minimum value. Based on the activation intensity of each rule, the output fuzzy set of the conclusion part of the rule is derived by taking the minimum value operation. The output linguistic values ​​of the piezoelectric output coefficient, photovoltaic output coefficient and energy storage charge and discharge coefficient are divided into five linguistic values: negative large, negative small, zero, positive small and positive large. The output fuzzy sets obtained from the derivation of each rule are aggregated by taking the maximum value operation to obtain the overall output fuzzy sets of the piezoelectric output coefficient, photovoltaic output coefficient and energy storage charge and discharge coefficient respectively. The centroid method is used to defuzzify the overall output fuzzy set, and the centroid of the area enclosed by the membership function curve of each output fuzzy set and the horizontal axis is calculated to obtain the accurate values ​​of the piezoelectric output coefficient, photovoltaic output coefficient and energy storage charge and discharge coefficient.

2. The road heterogeneous energy collaborative power supply control method according to claim 1, characterized in that, The forecast results are obtained using a two-level forecasting framework that combines macro-trend forecasting with micro-detail forecasting. For piezoelectric output prediction, the first stage uses time series analysis to predict the macroscopic traffic flow arrival rate, and the second stage uses pulse queue model based on upstream detector data to predict the vehicle arrival time sequence. The two stages are fused to generate the piezoelectric output curve. For photovoltaic power output prediction, the first stage uses a time convolutional network to predict the macroscopic irradiance trend, and the second stage uses the optical flow method to predict short-term irradiance fluctuations. The two stages are fused and then used to generate a photovoltaic power output curve through a photovoltaic physical model. The dispatchable energy storage capacity is calculated forward using the ampere-hour integral method based on the current energy storage state of charge and the predicted net power for future periods.

3. The road heterogeneous energy collaborative power supply control method according to claim 1, characterized in that, The real-time matching degree index is calculated by dividing the sum of the current actual output of the piezoelectric unit, the current actual output of the photovoltaic module, and the current actual output of the energy storage unit by the current total load demand; the matching degree deviation is calculated by subtracting the real-time matching degree index from 1.

4. The road heterogeneous energy collaborative power supply control method according to claim 1, characterized in that, The flexibility potential is obtained by multiplying the current power demand of each type of flexible load by its respective adjustability factor and then summing the results.

5. The road heterogeneous energy collaborative power supply control method according to claim 1, characterized in that, In step S4, the predicted matching degree is obtained by multiplying the piezoelectric predicted output by the piezoelectric predicted weight and the photovoltaic predicted output by the photovoltaic predicted weight, then adding them to the energy storage predicted dispatchable amount, and finally dividing by the predicted load demand; wherein, the piezoelectric predicted weight and the photovoltaic predicted weight are both 1 under normal operating conditions. The matching degree prediction trend is calculated by dividing the difference between the predicted matching degree and the real-time matching degree by the prediction time interval.

6. The road heterogeneous energy collaborative power supply control method according to claim 1, characterized in that, The system switches to normal mode, early warning mode, or emergency mode, specifically including: When the state of charge of the energy storage unit is not lower than the first state of charge threshold and the piezoelectric prediction error and photovoltaic prediction error are not higher than the first prediction error threshold, the normal mode is triggered, with the goal of prioritizing the consumption of renewable energy and the flexible load operating at the rated power. When the state of charge of the energy storage unit is between the first state of charge threshold and the second state of charge threshold, or when the piezoelectric prediction error or photovoltaic prediction error is between the first prediction error threshold and the second prediction error threshold, the early warning mode is triggered. The flexible load is derated according to the derating factor ratio, and the energy storage unit prioritizes maintaining the state of charge balance. When the state of charge of the energy storage unit is lower than the second state of charge threshold, or the piezoelectric prediction error or photovoltaic prediction error exceeds the second prediction error threshold and continues for multiple control cycles, or the communication interruption time exceeds the preset interruption threshold, the emergency mode is triggered to cut off all flexible loads and only ensure the power supply to rigid loads.

7. The road heterogeneous energy collaborative power supply control method according to claim 1, characterized in that, Reduce the dependence of fuzzy decision-making on prediction results, or downgrade to a no-prediction mode, specifically including: When the piezoelectric prediction error exceeds the first fault tolerance threshold and continues for multiple control cycles, the piezoelectric prediction weight is reduced and real-time traffic flow is used as the control input. When the photovoltaic prediction error exceeds the second fault tolerance threshold and continues for multiple control cycles, the photovoltaic prediction weight is reduced and the model based on real-time irradiance is enabled. When the piezoelectric prediction error or photovoltaic prediction error exceeds the third fault tolerance threshold and continues for multiple control cycles, the prediction is deemed to have failed, and the system is downgraded to a no-prediction mode. The fuzzy controller then uses the measured value as input, calls the simplified emergency rule set, and shortens the control cycle.

8. The road heterogeneous energy collaborative power supply control method according to claim 1, characterized in that, The preset flexible load reduction priorities include: the first level of load reduction targets electric vehicle charging piles, the second level of load reduction targets information display screens, the third level of load reduction targets LED lighting, and the fourth level of load reduction targets socket loads.

9. A road heterogeneous energy collaborative power supply control system, characterized in that, The method based on any one of claims 1 to 8 includes: The data acquisition module is used to collect traffic flow data, meteorological data, load data, and the state of charge of the energy storage unit for the target road section. The supply and demand status calculation module is used to predict the forecast results for future periods based on the collected traffic flow data and meteorological data. The forecast results include piezoelectric predicted output, photovoltaic predicted output and energy storage predicted dispatchable quantity. Based on the collected load data and the actual output of each energy unit, the module calculates the real-time matching degree index and flexible adjustment potential at the current moment, and calculates the matching degree deviation based on the real-time matching degree index. The fuzzy decision module is used to calculate the predicted matching degree based on the prediction results, and then calculate the matching degree prediction trend based on the difference between the real-time matching degree and the predicted matching degree; taking the matching degree deviation, the matching degree prediction trend, the flexible adjustment potential and the state of charge of the energy storage unit as inputs, the piezoelectric output coefficient, the photovoltaic output coefficient and the energy storage charge and discharge coefficient are solved by the fuzzy rule base. The multi-modal operation management module is used to switch the system to normal mode, early warning mode or emergency mode based on the energy storage unit's state of charge, piezoelectric output prediction error and photovoltaic output prediction error. The fault-tolerant control module is used to reduce the dependence of fuzzy decision-making on the prediction results or to downgrade to a no-prediction mode when any prediction error exceeds the corresponding set threshold. The execution module is used to adjust the power electronic converter according to the output coefficient and charge / discharge coefficient obtained by the solution, and to perform graded load shedding according to the preset flexible load shedding priority, so as to realize the coordinated power supply of multiple energy sources to the load.

Images