Online evaluation method and system for adjustability of optical storage direct current flexible power distribution and utilization system

Abstract

The invention is suitable for the technical field of optical storage direct-current flexible power distribution systems, and provides an optical storage direct-current flexible power distribution and utilization system adjustable capability online evaluation method and system, and the method comprises the steps: obtaining the historical operation data and real-time operation data of an optical storage direct-current flexible system; a photovoltaic power generation power output model, an energy storage charging and discharging dynamic model and a load adjustable range model are constructed respectively, and a multi-dimensional physical model of the optical storage direct flexible system is generated through a system-level electrical coupling relation; based on the multi-dimensional physical model, calculating an initial adjustable capability evaluation value of the system in a target time period through an optimization algorithm under a preset constraint condition; carrying out online parameter correction on the multi-dimensional physical model by adopting a time sequence deep learning model according to dynamic characteristic parameters in the real-time operation data; and fusing the corrected model output with the initial evaluation value to generate an adjustable capability prediction value updated in real time. The flexibility and reliability of system operation are effectively improved, the energy utilization efficiency is improved, and the operation cost is reduced.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic-storage-DC-flexible power distribution system technology, and in particular to an online assessment method and system for the adjustability of such a system. Background Technology

[0002] With the increasing proportion of renewable energy in the global energy structure, especially the rapid development of photovoltaic (PV) power generation technology, the operating mode of energy systems is undergoing profound changes. PV-storage systems, combining PV power generation with energy storage technology, have become one of the key technologies driving the green and low-carbon energy transition. At the same time, DC power grids, with their significant advantages in reducing energy conversion losses and improving transmission efficiency, are closely integrated with PV-storage systems to create an innovative system architecture: PV-storage-DC-flexible, a new type of energy system composed of PV, energy storage, DC grid, and flexible loads.

[0003] In existing technologies, photovoltaic-storage-DC-flexible systems exhibit significant volatility and uncertainty. From the power generation perspective, photovoltaic power generation is easily affected by changes in natural conditions such as solar radiation intensity and ambient temperature, resulting in obvious intermittent and fluctuating output power. Energy storage devices are constrained by factors such as their own charge-discharge cycle count and operating time, causing their charge-discharge characteristics and efficiency to change dynamically over time, making it difficult to accurately predict the available capacity and regulation capabilities of the energy storage system. Load-side demand is influenced by multiple factors such as user electricity consumption habits and real-time electricity consumption scenarios, exhibiting randomness and unpredictability. Current offline assessment methods cannot collect and process dynamic data during system operation in real time, making it difficult to accurately capture instantaneous fluctuations in photovoltaic power generation, real-time changes in energy storage status, and the randomness of load demand. This leads to deviations between assessment results and actual operating conditions, failing to provide timely and effective support for dispatch decisions and making it difficult to meet the urgent need for real-time assessment in photovoltaic-storage-DC-flexible systems. Summary of the Invention

[0004] This application provides an online assessment method and system for the adjustability of a photovoltaic-storage-DC-flexible power distribution system, which addresses the problem of the difficulty in meeting the urgent need for real-time assessment of such systems.

[0005] The first aspect of this application provides an online assessment method for the adjustability of a photovoltaic-storage-DC-flexible power distribution system, including:

[0006] Acquire historical and real-time operational data of the photovoltaic-storage-direct-drive-flexible system;

[0007] Based on the environmental factors, charging and discharging behavior and load demand in the historical operating data, a photovoltaic power output model, an energy storage charging and discharging dynamic model and a load adjustable range model are constructed respectively. A multi-dimensional physical model of the photovoltaic-storage-DC-flexible system is generated through the system-level electrical coupling relationship.

[0008] Based on the multi-dimensional physical model, the initial adjustability assessment value of the system within the target time period is calculated using an optimization algorithm under preset constraints.

[0009] Based on the dynamic feature parameters in the real-time running data, a time-series deep learning model is used to perform online parameter correction on the multi-dimensional physical model.

[0010] The corrected model output is fused with the initial adjustable capability assessment value to generate a real-time updated adjustable capability prediction value.

[0011] Furthermore, the expression for the photovoltaic power output model includes:

[0012] P pv (t)=η pv ·A pv ·I(t)·[1―α·(t(t)―T ref )]

[0013] Where: P pv (t) represents the photovoltaic output power at time t, and η pv For photovoltaic conversion efficiency, A pv Let I(t) be the effective area of ​​the photovoltaic array, I(t) be the solar irradiance at time t, and T(t) be the ambient temperature at time t. ref The reference temperature is α, and the temperature correction factor is α.

[0014] Furthermore, the expression for the energy storage charging and discharging dynamic model includes:

[0015]

[0016] Where: SOC(t+1) and SOC(t) are the states of charge at times t+1 and t, respectively, and η b For energy storage charging and discharging efficiency, P b (t) represents the actual charge / discharge power of the stored energy at time t, Δt is the time step, and C b For the rated capacity of energy storage, P rated This is the rated power of the energy storage. and These represent the upper and lower limits of the charging and discharging power at time t, respectively.

[0017] Furthermore, the expression for the load adjustable range model includes:

[0018]

[0019] Where: P load (t) represents the actual load power at time t. and These are the lower and upper limits of load power adjustment at time t, respectively.

[0020] Furthermore, the generation of a multi-dimensional physical model of the photovoltaic-storage-direct-drive-flexible system through system-level electrical coupling relationships includes:

[0021] Establish the DC bus power balance equation:

[0022] P pv (t)+P b (t)=P load (t)+P grid (t)

[0023] Where: P grid (t) represents the grid interaction power at time t;

[0024] By integrating the photovoltaic power output model, the energy storage charging and discharging dynamic model, and the load adjustable range model through dynamic constraints, a multi-dimensional physical model containing power coupling and time-series characteristics is generated.

[0025] Furthermore, the preset constraints include: power balance constraints, energy storage operation constraints, load regulation constraints, and state of charge constraints.

[0026] Furthermore, the step of calculating the initial adjustability assessment value of the system within the target time period based on the multi-dimensional physical model and using an optimization algorithm under preset constraints includes:

[0027] C pre (t)=max(P pv (t)+P b (t)―P load (t),P max —P min )

[0028] Where: C pre (t) represents the initial adjustability assessment value at time t, P max and P min These represent the upper and lower limits of the system's power constraints, respectively.

[0029] Furthermore, the step of using a time-series deep learning model to perform online parameter correction on the multi-dimensional physical model based on the dynamic feature parameters in the real-time running data includes:

[0030] The dynamic feature parameters in the real-time running data are processed based on the improved long short-term memory network model. The long short-term memory network model receives photovoltaic power output time-series signals, energy storage state of charge change signals and load power fluctuation signals through a multi-channel input structure, and uses a time weighting mechanism to assign attenuation weights to the inputs at different time steps.

[0031] The input signal is directly coupled to the hidden layer output through the residual connection structure of the long short-term memory network model to generate a dynamic correction value.

[0032] The parameters of the multi-dimensional physical model are adjusted according to the dynamic correction amount.

[0033] Furthermore, the step of fusing the corrected model output with the initial adjustable capability assessment value to generate a real-time updated adjustable capability prediction value includes:

[0034] The updated output value of the multi-dimensional physical model parameters is used as a dynamic correction evaluation value and is time-series superimposed and fused with the initial adjustable capability evaluation value.

[0035] Based on a preset confidence threshold, the fusion ratio of the multi-dimensional physical model and the dynamically corrected evaluation quantity is controlled to generate an adjustable capability prediction value sequence.

[0036] The second aspect of this application provides an online assessment system for the adjustability of a photovoltaic-storage-DC-flexible power distribution system, comprising:

[0037] The operation data acquisition unit is used to acquire historical and real-time operation data of the optical-storage-direct-drive-flexible system;

[0038] The multi-dimensional physical model generation unit is used to construct a photovoltaic power output model, an energy storage charging and discharging dynamic model, and a load adjustable range model based on environmental factors, charging and discharging behavior, and load demand in the historical operating data, and to generate a multi-dimensional physical model of the photovoltaic-storage-DC-flexible system through system-level electrical coupling relationships.

[0039] The initial adjustability assessment value calculation unit calculates the initial adjustability assessment value of the system within the target time period based on the multi-dimensional physical model and through an optimization algorithm under preset constraints.

[0040] The model parameter correction unit performs online parameter correction on the multi-dimensional physical model based on the dynamic feature parameters in the real-time running data using a time-series deep learning model.

[0041] The prediction generation unit merges the corrected model output with the initial adjustable capability assessment value to generate a real-time updated adjustable capability prediction value.

[0042] As can be seen from the above technical solutions, this application has the following advantages:

[0043] This application constructs a multi-dimensional physical model based on acquired historical and real-time operational data, enabling real-time evaluation of the adjustability of a photovoltaic-storage-DC-flexible system during system operation and achieving rapid response. Dynamic correction of the multi-dimensional physical model based on real-time data significantly improves the accuracy of the evaluation results, especially in the face of fluctuations in photovoltaic power generation and changes in the charging and discharging efficiency of energy storage devices. This invention also effectively addresses system uncertainties, such as fluctuations in photovoltaic power generation and changes in load demand, providing stable evaluation results under different operating scenarios. Accurate evaluation of the system's adjustability provides a reliable basis for the optimized scheduling of the photovoltaic-storage-DC-flexible system, effectively improving the flexibility and reliability of system operation, thereby increasing energy utilization efficiency and reducing operating costs. Attached Figure Description

[0044] Figure 1 This is a schematic flowchart of an embodiment of an online assessment method for the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to the present invention;

[0045] Figure 2 This is a schematic diagram of an embodiment of an online assessment system for the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to the present invention. Detailed Implementation

[0046] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “corresponding to,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0047] Example 1

[0048] The implementation method in this embodiment can be implemented in a system, on a server, or on a terminal; no specific limitation is made. The following section will describe the online assessment method for the adjustability of the photovoltaic-storage-DC-flexible power distribution system in this application from the perspective of system implementation. Please refer to... Figure 1 The method provided in this application includes the following steps:

[0049] S11. Obtain historical and real-time operating data of the photovoltaic-storage-direct-drive-flexible system;

[0050] In this embodiment, historical operating data includes photovoltaic power output timing characteristics, energy storage charging and discharging behavior patterns, and load demand variation patterns; real-time operating data includes environmental parameter monitoring values, real-time energy storage state of charge, and load dynamic power parameters. The photovoltaic-storage-DC-flexible system here consists of photovoltaic power generation, energy storage devices, a DC grid, and flexible loads. In this system, environmental factors, charging and discharging behavior, and load demand from historical operating data are the main factors affecting the system's adjustability.

[0051] S12. Based on environmental factors, charging and discharging behavior and load demand in historical operating data, a photovoltaic power output model, an energy storage charging and discharging dynamic model and a load adjustable range model are constructed respectively. A multi-dimensional physical model of the photovoltaic-storage-DC-flexible system is generated through system-level electrical coupling relationship.

[0052] In this embodiment, physical models are constructed based on influencing factors such as environmental factors, charging and discharging behavior, and load demand. The photovoltaic power output model describes the power output of the photovoltaic power station under different environmental conditions; the energy storage charging and discharging dynamic model describes the charging and discharging behavior and characteristics of the energy storage device; and the load adjustable range model describes the changing patterns of user-side electricity demand. The expressions are as follows:

[0053] P pv (t)=η pv ·A pv ·I(t)·[1―α·(T(t)―T ref )]

[0054] Where: P pv (t) represents the photovoltaic output power at time t, and η pv For photovoltaic conversion efficiency, A pv Let I(t) be the effective area of ​​the photovoltaic array, I(t) be the solar irradiance at time t, and T(t) be the ambient temperature at time t. ref The reference temperature is α, and the temperature correction factor is α.

[0055]

[0056] Where: SOC(t+1) and SOC(t) are the states of charge at times t+1 and t, respectively, and η b For energy storage charging and discharging efficiency, P b (t) represents the actual charge / discharge power of the stored energy at time t, Δt is the time step, and C b For the rated capacity of energy storage, P rated This is the rated power of the energy storage. and These represent the upper and lower limits of the charging and discharging power at time t, respectively.

[0057]

[0058] Where: P load (t) represents the actual load power at time t. and These are the lower and upper limits of load power adjustment at time t, respectively.

[0059] Establish the DC bus power balance equation:

[0060] P pv (t)+P b (t)=P load (t)+P grid (t)

[0061] Where: P grid (t) represents the grid interaction power at time t;

[0062] By integrating the photovoltaic power output model, the energy storage charging and discharging dynamic model, and the load adjustable range model through dynamic constraints, a multi-dimensional physical model containing power coupling and time-series characteristics is generated.

[0063] S13. Based on a multi-dimensional physical model, calculate the initial adjustability assessment value of the system within the target time period using an optimization algorithm under preset constraints;

[0064] It should be noted that energy storage devices, including batteries and other equipment, have charging and discharging behaviors that affect the system's regulation capability. The state of a battery is determined by its state of charge, while the adjustable power range of the energy storage device is determined by its current state of charge. Load fluctuations and flexible regulation capabilities affect the system's adjustability. During system operation, adjustability is the total power range that the system can adjust at a given moment. Adjustability requires comprehensive consideration of the actual states and constraints of various parts of the physical model.

[0065] In this embodiment, the preset constraints include: power balance constraints, energy storage operation constraints, load regulation constraints, and state of charge constraints. Their expressions are as follows:

[0066] Power balance constraint: P pv (t)+P b (t)=P load (t)+P grid (t);

[0067] Energy storage operation constraints:

[0068] Load regulation constraints:

[0069] State of charge constraint: SOC min ≤SOC(t)≤SOC max ;

[0070] By solving for the optimal solution of the objective function within the target time period, the initial adjustability assessment value of the system is obtained as follows:

[0071] C pre (t)=max(P pv (t)+P b (t)―P load (t),P max —P min )

[0072] in: and These are the definitions of the energy storage charge and discharge dynamic model at time t; and The load adjustable range model is defined for time t; SOC min and SOC max These are the minimum and maximum allowable values ​​for the state of charge of the energy storage, respectively; C pre (t) represents the initial adjustability assessment value at time t, P max and P min These represent the upper and lower limits of the system's power constraints, respectively.

[0073] S14. Based on the dynamic feature parameters in the real-time running data, a time-series deep learning model is used to perform online parameter correction on the multi-dimensional physical model;

[0074] In this embodiment, the Long Short-Term Memory (LSTM) network model is a special type of recurrent neural network specifically designed for processing sequential data, particularly suitable for capturing long-term dependencies. While traditional LSTM network models can handle time series forecasting tasks, their accuracy and stability may be insufficient when facing the complex nonlinearities and uncertainties of optical-storage-direct-drive flexible systems. Therefore, an improved LSTM network model is proposed to enhance the prediction and dynamic correction of the system's adjustability. The model parameter correction includes the following steps:

[0075] 1. Based on the improved long short-term memory network model, dynamic feature parameters in real-time operation data are processed. The long short-term memory network model receives photovoltaic power output time-series signals, energy storage state of charge change signals and load power fluctuation signals through a multi-channel input structure, and uses a time weighting mechanism to assign attenuation weights to the inputs at different time steps.

[0076] 2. By directly coupling the input signal with the hidden layer output through the residual connection structure of the Long Short-Term Memory network model, dynamic correction values ​​are generated;

[0077] 3. Adjust the parameters of the multi-dimensional physical model according to the dynamic correction amount.

[0078] The improved LSTM model receives three types of dynamic feature parameters through a multi-channel input structure:

[0079] 1. Photovoltaic output time-series signal: based on real-time ambient light intensity change rate 1. Quantifying the impact of sudden changes in sunlight on photovoltaic power output models; 2. Energy storage state of charge (SOC) change signals: calculating the gradient of SOC change in energy storage. Characterizing the dynamic response deviation of energy storage; 3. Load power fluctuation signal: extracting the load power fluctuation rate. This reflects the degree to which load demand deviates from the historical average. A time-weighted mechanism is introduced into the input signal, assigning attenuation weights to historical time-step data. Using the current time t as a reference, an attenuation coefficient λ is applied to the input signal x(t-k) at time t-k. k Where λ∈(0,1) controls the information decay rate. The weighted input is expressed as: x′(t―k)=x(t―k)·λ k .

[0080] The improved LSTM model employs a residual connection structure, directly coupling the input signal to the hidden layer output, thus enhancing model stability. The output of the residual connection is calculated as: h′ t =LSTM Cell (x′(t))+x′(t), where LSTM Cell The improved LSTM cell incorporates time-weighted input and a forget gate optimization design. Through multi-layer hidden state propagation, a dynamic correction variable ΔW(t) is ultimately generated to adjust the core parameters of the multi-dimensional physical model: the photovoltaic conversion efficiency parameter η. pv Correction factor Δη pv =ΔW pv (t); Energy storage charging and discharging power parameter η b Correction factor Δη b =ΔW b (t); Load regulation boundary parameters Correction factor ΔP load =ΔW L (t).

[0081] Incremental adjustments are made to the multi-dimensional physical model based on dynamic corrections:

[0082]

[0083] Where: α is the correction gain coefficient, with a value range of 0.1≤α≤0.3, used to control the parameter update amplitude.

[0084] S15. The corrected model output is fused with the initial adjustability assessment value to generate a real-time updated adjustability prediction value.

[0085] In this embodiment, the corrected model output is fused with the initial adjustable capability assessment value to generate a predicted value sequence. The process is as follows:

[0086] 1. The updated output values ​​of the multi-dimensional physical model parameters are used as dynamic correction evaluation values ​​and are time-series superimposed and fused with the initial adjustable capability evaluation values;

[0087] The updated output value C of the multi-dimensional physical model parameters LSTM (t) serves as the dynamic correction evaluation quantity, compared with the initial adjustability evaluation value C. pre (t) Overlay and fusion according to time steps:

[0088] C final (t)=C pre (t)+γ(t)·C LSTM (t)

[0089] Where: γ(t) is the dynamic fusion coefficient at time t, and its value is determined by the confidence level of the real-time data.

[0090] 2. Based on a preset confidence threshold, control the fusion ratio of the multi-dimensional physical model and the dynamically corrected evaluation quantity to generate an adjustable capability prediction value sequence.

[0091] Preset physics model confidence threshold β th =0.6, when the real-time data volatility δ(t) is below the threshold, the physical model output is used first; otherwise, the correction weight is increased. The fusion ratio is calculated as follows:

[0092]

[0093] Finally, a sequence of adjustable capacity prediction values ​​updated on a minute-by-minute basis is generated {C}. final (t1),C final (t2)}, ensuring that the contribution weight of the physical model is not less than 60%.

[0094] It should be noted that this application is based on the collaborative modeling of photovoltaic, energy storage and load systems, combined with a data-driven dynamic correction mechanism and a formulaic calculation model for real-time evaluation of the system's adjustability, to calculate the system's adjustability in real time; it can dynamically evaluate the adjustability of the photovoltaic-storage-DC-flexible system based on real-time data and system operating characteristics, and further cope with uncertainties and dynamic changes in system operation to achieve accurate online evaluation.

[0095] Example 2

[0096] Please see Figure 2 An embodiment of an online assessment system for the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to the present invention includes the following steps:

[0097] The operation data acquisition unit 101 is used to acquire historical and real-time operation data of the optical-storage-direct-flex system.

[0098] The multi-dimensional physical model generation unit 102 is used to construct a photovoltaic power output model, an energy storage charging and discharging dynamic model, and a load adjustable range model based on environmental factors, charging and discharging behavior, and load demand in historical operating data, and to generate a multi-dimensional physical model of the photovoltaic-storage-DC-flexible system through system-level electrical coupling relationships.

[0099] The initial adjustability assessment value calculation unit 103 calculates the initial adjustability assessment value of the system within the target time period based on a multi-dimensional physical model and through an optimization algorithm under preset constraints.

[0100] The model parameter correction unit 104 uses a time-series deep learning model to perform online parameter correction on the multi-dimensional physical model based on the dynamic feature parameters in the real-time running data.

[0101] The prediction value generation unit 105 merges the corrected model output with the initial adjustable capability assessment value to generate a real-time updated adjustable capability prediction value.

[0102] For specific limitations regarding the evaluation system, please refer to the limitations of the evaluation method above, which will not be repeated here. Each module in the above evaluation system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0103] It is understood that those skilled in the art can combine various implementation methods in the above embodiments under the guidance of the above examples to obtain technical solutions with multiple implementation methods.

[0104] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for online evaluation of the adjustability of a photovoltaic-storage-DC-flexible power distribution system, characterized in that, include: Acquire historical and real-time operational data of the photovoltaic-storage-direct-drive-flexible system; Based on the environmental factors, charging and discharging behavior and load demand in the historical operating data, a photovoltaic power output model, an energy storage charging and discharging dynamic model and a load adjustable range model are constructed respectively. A multi-dimensional physical model of the photovoltaic-storage-DC-flexible system is generated through the system-level electrical coupling relationship. Based on the multi-dimensional physical model, the initial adjustability assessment value of the system within the target time period is calculated using an optimization algorithm under preset constraints. Based on the dynamic feature parameters in the real-time running data, a time-series deep learning model is used to perform online parameter correction on the multi-dimensional physical model. The corrected model output is fused with the initial adjustable capability assessment value to generate a real-time updated adjustable capability prediction value.

2. The method for online evaluation of the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to claim 1, characterized in that, The expression for the photovoltaic power output model includes: P pv (t)=η pv ·A pv ·I(t)·[1―α·(T(t)―T ref )] Where: P pv (t) represents the photovoltaic output power at time t, and η pv For photovoltaic conversion efficiency, A pv Let I(t) be the effective area of ​​the photovoltaic array, I(t) be the solar irradiance at time t, and T(t) be the ambient temperature at time t. ref The reference temperature is α, and the temperature correction factor is α.

3. The method for online evaluation of the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to claim 1, characterized in that, The expression for the energy storage charging and discharging dynamic model includes: Where: SOC(t+1) and SOC(t) are the states of charge at times t+1 and t, respectively, and η b For energy storage charging and discharging efficiency, P b (t) represents the actual charge / discharge power of the stored energy at time t, Δt is the time step, and C b For the rated capacity of energy storage, P rated This is the rated power of the energy storage. and These represent the upper and lower limits of the charging and discharging power at time t, respectively.

4. The method for online evaluation of the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to claim 1, characterized in that, The expression for the load adjustable range model includes: Where: P load (t) represents the actual load power at time t. and These are the lower and upper limits of load power adjustment at time t, respectively.

5. The method for online evaluation of the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to any one of claims 1-4, characterized in that, The process of generating a multi-dimensional physical model of the photovoltaic-storage-direct-drive-flexible system through system-level electrical coupling relationships includes: Establish the DC bus power balance equation: P pv (t)+P b (t)=P load (t)+P grid (t) Where: P grid (t) represents the grid interaction power at time t; By integrating the photovoltaic power output model, the energy storage charging and discharging dynamic model, and the load adjustable range model through dynamic constraints, a multi-dimensional physical model containing power coupling and time-series characteristics is generated.

6. The method for online evaluation of the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to claim 1, characterized in that, The preset constraints include: power balance constraints, energy storage operation constraints, load regulation constraints, and state of charge constraints.

7. The method for online evaluation of the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to claim 6, characterized in that, The calculation of the initial adjustability assessment value of the system within the target time period based on the multi-dimensional physical model and an optimization algorithm under preset constraints includes: C pre (t)=max(P pv (t)+P b (t)―P load (t),P max ―P min ) Where: C pre (t) represents the initial adjustability assessment value at time t, P max and P min These represent the upper and lower limits of the system's power constraints, respectively.

8. The method for online evaluation of the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to claim 1, characterized in that, The step of using a time-series deep learning model to perform online parameter correction on the multi-dimensional physical model based on the dynamic feature parameters in the real-time running data includes: The dynamic feature parameters in the real-time running data are processed based on the improved long short-term memory network model. The long short-term memory network model receives photovoltaic power output time-series signals, energy storage state of charge change signals and load power fluctuation signals through a multi-channel input structure, and uses a time weighting mechanism to assign attenuation weights to the inputs at different time steps. The input signal is directly coupled to the hidden layer output through the residual connection structure of the long short-term memory network model to generate a dynamic correction value. The parameters of the multi-dimensional physical model are adjusted according to the dynamic correction amount.

9. The method for online evaluation of the adjustability of a photovoltaic-storage-DC-flexible power distribution system according to claim 8, characterized in that, The step of fusing the corrected model output with the initial adjustable capability assessment value to generate a real-time updated adjustable capability prediction value includes: The updated output value of the multi-dimensional physical model parameters is used as a dynamic correction evaluation value and is time-series superimposed and fused with the initial adjustable capability evaluation value. Based on a preset confidence threshold, the fusion ratio of the multi-dimensional physical model and the dynamically corrected evaluation quantity is controlled to generate an adjustable capability prediction value sequence.

10. An online assessment system for the adjustability of a photovoltaic-storage-DC-flexible power distribution system, employing the online assessment method for the adjustability of a photovoltaic-storage-DC-flexible power distribution system as described in any one of claims 1-9, characterized in that, include: The operation data acquisition unit is used to acquire historical and real-time operation data of the optical-storage-direct-drive-flexible system; The multi-dimensional physical model generation unit is used to construct a photovoltaic power output model, an energy storage charging and discharging dynamic model, and a load adjustable range model based on environmental factors, charging and discharging behavior, and load demand in the historical operating data, and to generate a multi-dimensional physical model of the photovoltaic-storage-DC-flexible system through system-level electrical coupling relationships. The initial adjustability assessment value calculation unit calculates the initial adjustability assessment value of the system within the target time period based on the multi-dimensional physical model and through an optimization algorithm under preset constraints. The model parameter correction unit performs online parameter correction on the multi-dimensional physical model based on the dynamic feature parameters in the real-time running data using a time-series deep learning model. The prediction generation unit merges the corrected model output with the initial adjustable capability assessment value to generate a real-time updated adjustable capability prediction value.

Images