Dynamic load regulation-liquid nitrogen cold storage integrated high-temperature superconducting power grid capacity expansion system and method

Abstract

The invention discloses a dynamic load adjustment-liquid nitrogen cold storage integrated high-temperature superconducting power grid capacity expansion system and method. The system comprises a data preprocessing system, a dynamic load adjustment device, a liquid nitrogen cold storage device, a state monitoring device, a high-temperature superconducting cable capacity expansion device and a communication coordination system. The method comprises the steps of generating a standardized feature matrix through normalization processing; multi-time-scale load prediction is realized through an LSTM (Long Short-Term Memory) network, a collaborative instruction set is generated by adopting an optimal scheduling engine of mixed integer linear programming, and adaptive closed-loop control based on a fuzzy PID (Proportion Integration Differentiation) algorithm is realized; a multi-physics field coupling optimization model of a real-time cold energy distribution diagram is adopted to realize power grid capacity improvement; and the time sequence consistency of the whole system is coordinated and circulation is formed. Zero-resistance power transmission is realized in a liquid nitrogen temperature region by adopting the high-temperature superconducting cable, the loss is reduced to 0.5% or below, and the transmission capacity is improved by 3-5 times; the load demand is predicted in combination with the AI algorithm, and the renewable energy consumption rate is increased to 95%.

Description

Technical Field

[0001] The present invention relates to the field of power systems and smart grids, and in particular to a dynamic load regulation-liquid nitrogen cold storage integrated high-temperature superconducting grid expansion system and method. Background Art

[0002] Traditional grid expansion technology mainly relies on copper / aluminum conductors, which have problems such as high transmission loss, high expansion cost, and large footprint. Although high-temperature superconducting cables can reduce losses, they are limited by the stability of liquid nitrogen cooling systems and the difficulty of large-scale application, such as low liquid nitrogen circulation efficiency and temperature fluctuations leading to superconducting performance degradation. Existing dynamic load regulation technology has slow response speed and limited capacity, and it is difficult to match the instantaneous fluctuation demand of new energy grid connection. Current technical solutions mostly focus on a single link and lack systematic integrated design, and cannot simultaneously solve the contradiction between grid expansion and renewable energy consumption. The application of existing high-temperature superconducting power line demonstration projects is still limited by manufacturing costs and short-distance scenarios. Summary of the invention

[0003] Purpose of the invention: The purpose of the present invention is to provide a dynamic load regulation-liquid nitrogen cold storage integrated high-temperature superconducting grid expansion system and method.

[0004] Technical solution: The dynamic load adjustment-liquid nitrogen cold storage integrated high-temperature superconducting grid expansion system described in the present invention includes a data preprocessing system, a dynamic load adjustment device, a liquid nitrogen cold storage device, a state monitoring device, a high-temperature superconducting cable expansion device and a communication coordination system; the data preprocessing module collects grid load and temperature parameters in real time, and transmits them to the dynamic load adjustment module after standardized processing. The dynamic load adjustment module predicts load fluctuations based on the AI algorithm to balance the peaks and valleys of the grid. When an overload signal is detected, the liquid nitrogen cold storage module starts the pumpless self-circulation system to drive liquid nitrogen to flow through the superconducting cable cooling tank to maintain temperature stability. The state monitoring module is linked to the pressure valve through a temperature difference sensor to monitor the liquid nitrogen evaporation rate and the cable heat dissipation status in real time. When an abnormality occurs, the redundant cooling channel is triggered. The high-temperature superconducting cable realizes low-loss power transmission under the support of liquid nitrogen cold energy, and the expansion capacity is improved. The communication coordination module integrates the data of each subsystem, and realizes closed-loop control of instructions and strategies through a two-way interactive protocol to achieve integrated operation of grid expansion, dynamic load adjustment and low-temperature energy storage.

[0005] Furthermore, the overload signal receiving module in the liquid nitrogen cold storage device continuously captures the current density distribution and multi-section temperature field data of the core conductor layer of the superconducting cable through a monitoring network composed of distributed optical fiber sensors and Hall current probes. When it is detected that the current in any section exceeds the critical threshold, a three-level interlocking response mechanism is triggered.

[0006] Further, the multi-source heterogeneous sensor network in the state monitoring device collects real-time parameters of the liquid nitrogen cold storage system and the superconducting cable through temperature, pressure, and current sensors; subsequently, a dynamic threshold calculation engine outputs critical state criteria and generates a regulation instruction for the liquid nitrogen pump group and a power grid load scheduling signal.

[0007] Further, the real-time cold energy distribution map generation module in the high-temperature superconducting cable capacity expansion device generates a dynamic cold energy distribution heat map to guide the cold energy distribution strategy; the current-carrying capacity dynamic calculation unit dynamically optimizes the power grid capacity increase amplitude through the feedback mechanism of the critical current density of the superconducting tape and real-time temperature data, combined with the adaptive adjustment of the transmission line impedance parameters.

[0008] The dynamic load regulation-liquid nitrogen cold storage integrated high-temperature superconducting power grid capacity expansion method of the present invention includes the following steps:

[0009] (1) Generate a standardized feature matrix through multi-modal heterogeneous data cleaning, high-dimensional spatio-temporal feature fusion, and adaptive normalization processing;

[0010] (2) Implement multi-time scale load forecasting through an LSTM long short-term memory network, and use an optimization scheduling engine of mixed integer linear programming to generate a collaborative instruction set, including dynamic scheduling strategies for energy storage charging and discharging, multi-objective priority sorting of interruptible loads, and pre-start timing optimization of the liquid nitrogen cold storage pump group;

[0011] (3) Trigger a feedback mechanism by receiving an overload signal in real time, and dynamically adjust the pump group speed in combination with the liquid nitrogen flow-pressure coupling model to achieve adaptive closed-loop control based on the fuzzy PID algorithm;

[0012] (4) Through multi-modal data fusion and dynamic threshold warning mechanism, realize real-time acquisition of multi-source heterogeneous sensor data and analysis of critical state criteria;

[0013] (5) Adopt a multi-physical field coupling optimization model of a real-time cold energy distribution map, and through the dynamic calculation of the current-carrying capacity and the adaptive adjustment mechanism of transmission line parameters, combined with the low-temperature environment maintenance mediated by liquid nitrogen cold storage, achieve power grid capacity improvement;

[0014] (6) Through real-time heterogeneous data interaction, coordinate the timing consistency of the entire system and form a cycle.

[0015] Further, the multi-modal heterogeneous data cleaning in step (1) includes:

[0016] Adopt the IQR anomaly detection algorithm with a sliding window, the window width is adapted to the power grid load fluctuation frequency, and the anomaly threshold K of the IOR anomaly detection algorithm is expressed as:

[0017]

[0018] where Δf is the deviation between the current fluctuation frequency and the reference frequency f base , and α is the frequency sensitivity coefficient.

[0019] Further, the step (2) includes:

[0020] (2.1) Through the LSTM neural network of the multi-time scale load forecasting module, fuse multi-source data streams such as historical 15-minute load data, real-time meteorological monitoring parameters, dynamic electricity price signals, and ambient temperature. After time series feature extraction and non-linear relationship modeling, generate a refined load forecasting result with a 15-minute resolution for the next 10 days;

[0021] (2.2) Input the forecasting result into the optimization scheduling engine in real time. Under the premise of satisfying the superconducting cable temperature rise constraint, the mixed integer linear programming algorithm synchronously solves the multi-objective optimization problem of minimizing the power grid operation cost and maximizing the renewable energy consumption rate, and generates a collaborative instruction set including the charge and discharge power threshold of the energy storage system, the hierarchical response plan of interruptible load, and the pre-start time window of the liquid nitrogen cold storage equipment;

[0022] (2.3) The collaborative instruction set generation module decomposes the execution strategy through the dynamic programming algorithm: implement rolling optimization control of the charge and discharge power for the energy storage device, combine the electricity price fluctuation curve and the load priority evaluation model to perform dynamic hierarchical scheduling on the industrial flexible load; at the same time, predict the cooling demand based on the superconducting cable thermodynamic model and start the preheating process of the liquid nitrogen pump group 6 - 8 hours in advance;

[0023] (2.4) The liquid nitrogen cold storage control module adjusts the rotational speed and flow distribution of the multistage centrifugal pump in real time through a high-speed PID controller, completes the closed-loop regulation of the superconducting cable temperature field within a 200ms delay, and achieves a temperature control accuracy of 0.1K for the entire link through the distributed optical fiber temperature measurement system.

[0024] Further, for the optimization scheduling engine in the step (2), the objective function is:

[0025] min(α·C cost + β·ΔT cable + γ·(1 - R renewable ))

[0026] where C cost is the power grid operation cost, ΔT cable is the superconducting cable temperature rise, R renewable is the renewable energy consumption rate, and the weight coefficients α, β, γ are dynamically adjusted through the fuzzy analytic hierarchy process.

[0027] Furthermore, the prioritization of interruptible loads in step (2) is based on the TOPSIS multi-objective decision-making model. The evaluation indicators include user compensation cost, load importance level, and interruption duration, generating a Pareto optimal solution set;

[0028] The TOPSIS multi-objective decision-making model is constructed as follows:

[0029] a. Data standardization processing to eliminate the influence of different dimensions on decision-making;

[0030] b. Determination of weights;

[0031] c. Construction of a weighted normalized matrix;

[0032] d. Determination of the ideal solution and the negative ideal solution;

[0033] e. Calculation of distance and closeness;

[0034] Euclidean distance formula:

[0035]

[0036] Relative closeness:

[0037]

[0038] The closer C is to 1, the better the solution.

[0039] Furthermore, the construction of the liquid nitrogen flow-pressure coupling model in step (3) includes:

[0040] (3.1) Based on the mass conservation and Bernoulli equations, deduce the relationship between the liquid nitrogen flow velocity and the pipeline pressure drop;

[0041] (3.3) Calibrate the pump speed-flow curve through experiments and establish a non-linear fitting equation;

[0042] (3.3) Combine the Kalman filter algorithm to update the model parameters in real time to adapt to the working conditions of pipeline aging or blockage.

[0043] Beneficial effects: Compared with the prior art, the present invention has the following remarkable advantages: The present invention uses high-temperature superconducting cables to achieve zero-resistance power transmission in the liquid nitrogen temperature range, reducing the loss to below 0.5% and increasing the transmission capacity by 3-5 times; combining with the AI algorithm to predict the load demand, the renewable energy consumption rate is increased to 95%; it not only adapts to the grid connection of wind and solar power generation, but also provides technical reserves for the access of controllable nuclear fusion energy. Its low-loss and high-capacity characteristics can reduce the civil engineering requirements for grid transformation, and the carbon emission is reduced by up to 60%. Description of the Drawings

[0044] Figure 1It is a framework diagram of a dynamic load regulation - liquid nitrogen cold storage integrated high - temperature superconducting power grid expansion system;

[0045] Figure 2 It is a working flowchart of the dynamic load regulation device. Specific implementation manners

[0046] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0047] As Figure 1 shown, the dynamic load regulation - liquid nitrogen cold storage integrated high - temperature superconducting power grid expansion system of the present invention includes:

[0048] The data pre - processing system generates a standardized feature matrix through multi - modal heterogeneous data cleaning, high - dimensional spatio - temporal feature fusion, and adaptive normalization processing;

[0049] The dynamic load regulation device realizes multi - time - scale load forecasting through the LSTM long - short - term memory network, and uses an optimization scheduling engine of mixed - integer linear programming to generate a collaborative instruction set, including dynamic scheduling strategies for energy storage charging and discharging, multi - objective priority sorting of interruptible loads, and pre - start timing optimization of the liquid nitrogen cold storage pump group;

[0050] The liquid nitrogen cold storage device triggers a feedback mechanism by receiving overload signals in real time, and dynamically adjusts the pump group speed in combination with the liquid nitrogen flow - pressure coupling model to achieve adaptive closed - loop control based on the fuzzy PID algorithm;

[0051] The state monitoring device realizes real - time acquisition of multi - source heterogeneous sensor data and critical state criterion analysis through multi - modal data fusion and dynamic threshold warning mechanism, ensuring heat dissipation and zero - resistance characteristics;

[0052] The high - temperature superconducting cable expansion device adopts a multi - physical - field coupling optimization model of a real - time cold energy distribution map, and through dynamic calculation of the current - carrying capacity and adaptive adjustment mechanism of transmission line parameters, combined with the low - temperature environment maintenance mediated by liquid nitrogen cold storage, realizes power grid capacity improvement;

[0053] The communication and coordination system coordinates the timing consistency of the entire system and forms a cycle through real - time heterogeneous data interaction.

[0054] The data pre - processing system uses a multi - modal heterogeneous data cleaning module to eliminate sensor noise and repair data integrity; uses a high - dimensional spatio - temporal feature fusion module to align multi - source data based on the dynamic time warping algorithm; uses an adaptive normalization processing module to dynamically adjust the standardization parameters according to the real - time data distribution; and uses a standardized feature matrix output interface to encapsulate the pre - processed data into a structured data packet.

[0055] The multi-modal heterogeneous data cleaning module adopts the IQR anomaly detection algorithm with a sliding window. The window width adapts to the grid load fluctuation frequency. The anomaly threshold K of the IOR anomaly detection algorithm is expressed as:

[0056]

[0057] where Δf is the deviation between the current fluctuation frequency and the reference frequency f base , and α is the frequency sensitivity coefficient (usually taken as 0.2 - 0.5).

[0058] As Figure 2 shown, the dynamic load regulation device includes:

[0059] First, through the LSTM neural network of the multi-time scale load forecasting module, this device fuses multi-source data streams such as historical 15-minute load data, real-time meteorological monitoring parameters, dynamic electricity price signals, and ambient temperature. After time series feature extraction and non-linear relationship modeling, it generates a refined load forecasting curve with a 15-minute resolution for the next 10 days.

[0060] This forecasting result is input into the optimization scheduling engine in real time. The mixed integer linear programming (MILP) algorithm synchronously solves the multi-objective optimization problem of minimizing the grid operation cost and maximizing the renewable energy consumption rate under the premise of satisfying the superconducting cable temperature rise constraint, and generates a collaborative instruction set including the charge and discharge power threshold of the energy storage system, the hierarchical response scheme of interruptible loads, and the pre-start time window of the liquid nitrogen cold storage equipment.

[0061] Subsequently, the collaborative instruction set generation module decomposes the execution strategy through the dynamic programming algorithm: implements rolling optimization control of the charge and discharge power for the energy storage device, combines the electricity price fluctuation curve and the load priority evaluation model to perform dynamic hierarchical scheduling on industrial flexible loads; at the same time, based on the thermodynamic model of the superconducting cable, predicts the cooling demand and starts the preheating process of the liquid nitrogen pump group 6 - 8 hours in advance.

[0062] Finally, the liquid nitrogen cold storage control module adjusts the speed and flow distribution of the multistage centrifugal pump in real time through a high-speed PID controller, completes the closed-loop regulation of the superconducting cable temperature field within a 200ms delay, and achieves a temperature control accuracy of 0.1K for the entire link through the distributed optical fiber temperature measurement system.

[0063] The multi-time scale load forecasting module uses the LSTM hidden layer structure, and its internal adopts a combined algorithm model of tanh and ReLU. The formula is as follows:

[0064] Forget gate:

[0065] f t = σ(W f ·[h t-1 , x t +bf )

[0066] Among them, f t is the forgetting gate calculation value, W f is the forgetting gate weight matrix, b f is the bias term, and σ is the Sigmoid function (output [0, 1]).

[0067] Input gate:

[0068] i t = σ(W i · [h t-1 , x t + b i )

[0069] i t is the input gate calculation value, x t is the input vector (including 15-minute data such as meteorology, electricity price, temperature, etc.).

[0070] Candidate memory:

[0071]

[0072] C t is the candidate memory calculation value, and the tanh function is used for non-linear feature extraction, with the output range [-1, 1].

[0073] Memory update:

[0074]

[0075] C t is the latest memory value, C t-1 is the memory unit at the previous moment, and this formula realizes the progressive update of long-term dependence.

[0076] Output gate:

[0077] o t = σ(W o · [h t-1 , x t + b o )

[0078] o t is the output gate calculation value, and this formula controls the contribution of the current memory unit to the hidden state.

[0079] Hidden output:

[0080] h t = ReLU{W h · [o t · tanh(C t )] + b h}

[0081] h t To hide the output value, W h is the weight matrix of the output layer, with dimensions matching the prediction target. The ReLU function accelerates gradient propagation.

[0082] Finally, Z-score normalization is adopted:

[0083]

[0084] where x is the original data value (such as 15-minute load data, temperature, humidity, etc.), μ is the historical data mean of the same feature, and σ is the historical data standard deviation of the same feature.

[0085] Furthermore, for the optimization scheduling engine, the objective function is:

[0086] min(α·C cost +β·ΔT cable +γ·(1 - R renewable ))

[0087] where C cost is the grid operation cost, ΔT cable is the temperature rise of the superconducting cable, R renewable is the renewable energy consumption rate, and the weight coefficients α, β, γ are dynamically adjusted through the fuzzy analytic hierarchy process.

[0088] For the collaborative instruction set generation module, the dynamic scheduling strategy for energy storage charging and discharging adopts rolling horizon optimization. The core idea is to achieve real-time optimal control of complex systems through online rolling calculations within a finite time window. The formula is:

[0089]

[0090] where N is the prediction horizon length; P grid，k is the actual power of the grid at time period k; P tagret，k is the target power (calculated from the predicted output of photovoltaic / wind power and load demand); u k is the energy storage charging and discharging instruction (u k >0 for charging, u k <0 for discharging); λ is the weight of the control variable, used to balance tracking accuracy and equipment loss.

[0091] The prioritization of interruptible loads is based on the TOPSIS multi-objective decision-making model, and the construction process is as follows:

[0092] Step 1: Data standardization processing to eliminate the influence of different dimensions on decision-making;

[0093] Step 2: Weight determination;

[0094] Step 3: Construct a weighted normalization matrix;

[0095] Step 4: Determine the ideal solution and the negative ideal solution;

[0096] Ideal solution (V + ): Each index takes the optimal value (the maximum value for benefit type and the minimum value for cost type); Negative ideal solution (V - ): Each index takes the worst value (the minimum value for benefit type and the maximum value for cost type).

[0097] Step 5: Calculate the distance and the closeness degree.

[0098] Euclidean distance formula:

[0099]

[0100] Relative closeness degree:

[0101]

[0102] C i The closer it is to 1, the better the solution.

[0103] The liquid nitrogen cold storage control module adopts a PID-fuzzy composite controller to adjust the pump group speed according to the superconducting cable temperature feedback, with a steady-state error ≤ ±0.5K; the pre-start signal trigger condition is that the predicted load exceeds the critical threshold (≥ 85% of the rated capacity), and the pump group response time ≤ 5 seconds.

[0104] The liquid nitrogen cold storage device includes:

[0105] Overload signal receiving module: Real-time collect the current and temperature data of the superconducting cable of the power grid, and trigger the feedback mechanism when an overload signal is detected;

[0106] Liquid nitrogen flow-pressure coupling model: Establish a dynamic correlation equation between the flow rate, pressure and pump group speed in the liquid nitrogen circulation system to generate pump speed adjustment instructions;

[0107] The construction process of the liquid nitrogen flow-pressure coupling model is as follows:

[0108] Step 1: Based on the mass conservation and Bernoulli equations, deduce the relationship between the liquid nitrogen flow velocity and the pipeline pressure drop;

[0109] Step 2: Calibrate the pump group speed-flow curve through experiments and establish a nonlinear fitting equation;

[0110] Step 3: Combine the Kalman filter algorithm to update the model parameters in real time to adapt to the pipeline aging or blockage conditions.

[0111] The state monitoring device includes:

[0112] The multi-source heterogeneous sensor network collects the real-time parameters of the liquid nitrogen cold storage system and superconducting cables through temperature, pressure and current sensors. Then the dynamic threshold calculation engine outputs the critical state criteria based on the thermal-electric coupling model and generates liquid nitrogen pump group adjustment instructions and grid load dispatching signals.

[0113] The high-temperature superconducting cable capacity expansion device comprises:

[0114] The real-time cooling energy distribution map generation module generates a dynamic cooling energy distribution thermodynamic map based on a multi-physics field coupling optimization model (electromagnetic-thermal-fluid coupling) to guide the cooling energy allocation strategy; then the current carrying capacity dynamic calculation unit dynamically optimizes the increase in grid capacity through the feedback mechanism of the critical current density of the superconducting tape and the real-time temperature data, combined with the adaptive adjustment of the transmission line impedance parameters.

[0115] The communication coordination system consists of a heterogeneous data interface module, a timing controller and a closed-loop feedback unit. It achieves millisecond-level timing synchronization between multiple modules through real-time interaction of liquid nitrogen cold storage temperature control data, superconducting cable load parameters and dynamic load adjustment instructions, and ensures the atomicity of cross-system operations based on transaction management technology.

[0116] The specific implementation method of the present invention realizes efficient capacity expansion of the power grid through the collaborative process of data collection → dynamic load adjustment → liquid nitrogen cold storage → state monitoring → capacity expansion transmission → closed-loop control:

[0117] The data preprocessing module collects grid load, temperature and other parameters in real time, and transmits them to the dynamic load adjustment module after standardized processing. The dynamic load adjustment module predicts load fluctuations based on AI algorithms to balance grid peaks and valleys. When an overload signal is detected, the liquid nitrogen cold storage module starts the pumpless self-circulation system to drive liquid nitrogen to flow through the superconducting cable cooling tank to maintain temperature stability and ensure the superconducting zero resistance characteristics. The state monitoring module is linked to the pressure valve through the temperature difference sensor to monitor the liquid nitrogen evaporation rate and cable heat dissipation status in real time. When abnormal, it triggers the redundant cooling channel to prevent the risk of quenching. High-temperature superconducting cables achieve low-loss power transmission with the support of liquid nitrogen cold energy, and the expansion capacity is increased to 3 times that of traditional power grids57. The communication coordination module integrates the data of each subsystem, realizes closed-loop control of instructions and strategies through a two-way interactive protocol, and finally achieves the integrated operation of grid expansion, dynamic load adjustment and low-temperature energy storage.

Claims

1. A dynamic load regulation - liquid nitrogen cold storage integrated high-temperature superconducting power grid expansion system, characterized in that, It includes a data preprocessing system, a dynamic load regulation device, a liquid nitrogen cold storage device, a status monitoring device, a high-temperature superconducting cable capacity expansion device, and a communication coordination system. The data preprocessing module collects grid load and temperature parameters in real time, and after standardization processing, transmits them to the dynamic load regulation module. The dynamic load regulation module predicts load fluctuations based on AI algorithms to balance the peak and valley of the grid. When an overload signal is detected, the liquid nitrogen cold storage module starts a pump-free self-circulation system to drive liquid nitrogen to flow through the superconducting cable cooling tank to maintain temperature stability. The status monitoring module is linked with a temperature difference sensor and a pressure valve to monitor the liquid nitrogen evaporation rate and the cable heat dissipation status in real time. When abnormal, it triggers a redundant cooling channel. The high-temperature superconducting cable realizes low-loss power transmission with the support of liquid nitrogen cold energy, and the capacity expansion ability is improved. The communication coordination module integrates the data of each subsystem and realizes the closed-loop control of instructions and strategies through a two-way interaction protocol, achieving the integrated operation of grid capacity expansion, dynamic load regulation, and cryogenic energy storage.

2. The dynamic load regulation - liquid nitrogen cold storage integrated high-temperature superconducting power grid expansion system according to claim 1, wherein The overload signal receiving module in the liquid nitrogen cold storage device continuously captures the current density distribution and multi-section temperature field data of the core conductor layer of the superconducting cable through a monitoring network composed of a distributed optical fiber sensor and a Hall current probe. When the current in any section is detected to exceed the critical threshold, a three-level interlock response mechanism is triggered.

3. The dynamic load regulation - liquid nitrogen cold storage integrated high-temperature superconducting power grid expansion system according to claim 1, wherein The multi-source heterogeneous sensor network in the status monitoring device collects the real-time parameters of the liquid nitrogen cold storage system and the superconducting cable through temperature, pressure, and current sensors. Subsequently, the dynamic threshold calculation engine outputs the critical state criterion and generates the liquid nitrogen pump group adjustment instruction and the grid load scheduling signal.

4. The dynamic load regulation - liquid nitrogen cold storage integrated high-temperature superconducting power grid expansion system according to claim 1, wherein The real-time cold energy distribution map generation module in the high-temperature superconducting cable capacity expansion device generates a dynamic cold energy distribution heat map to guide the cold energy distribution strategy. The dynamic ampacity calculation unit dynamically optimizes the grid capacity improvement amplitude through the feedback mechanism of the critical current density of the superconducting tape and the real-time temperature data, combined with the adaptive adjustment of the transmission line impedance parameters.

5. A dynamic load regulation - liquid nitrogen cold storage integrated method for expanding the capacity of a high - temperature superconducting power grid, characterized in that, It includes the following steps: (1) Generate a standardized feature matrix through multi-modal heterogeneous data cleaning, high-dimensional spatio-temporal feature fusion, and adaptive normalization processing. (2) Implement multi-time scale load prediction through an LSTM long short-term memory network, and use an optimization scheduling engine based on mixed integer linear programming to generate a collaborative instruction set, including the dynamic scheduling strategy of energy storage charge and discharge, the multi-objective priority sorting of interruptible loads, and the pre-start timing optimization of the liquid nitrogen cold storage pump group. (3) Trigger a feedback mechanism by receiving an overload signal in real time, and dynamically adjust the pump group speed in combination with the liquid nitrogen flow-pressure coupling model to achieve adaptive closed-loop control based on the fuzzy PID algorithm. (4) Implement the real-time collection and critical state criterion analysis of multi-source heterogeneous sensor data through multi-modal data fusion and dynamic threshold warning mechanism. (5) Adopt a multi-physical field coupling optimization model of a real-time cold energy distribution map, and through the dynamic calculation of ampacity and the adaptive adjustment mechanism of transmission line parameters, combined with the maintenance of a low-temperature environment mediated by liquid nitrogen cold storage, achieve grid capacity improvement. (6) Coordinate the timing consistency of the entire system and form a cycle through real-time heterogeneous data interaction.

6. The dynamic load regulation-liquid nitrogen cold storage integrated high-temperature superconducting power grid expansion method according to claim 5, wherein, The multi-modal heterogeneous data cleaning in step (1) includes: The IQR anomaly detection algorithm using a sliding window, where the window width adapts to the grid load fluctuation frequency, and the anomaly threshold K of the IOR anomaly detection algorithm is expressed as: where Δf is the deviation between the current fluctuation frequency and the reference frequency f base , and α is the frequency sensitivity coefficient.

7. The dynamic load regulation - liquid nitrogen cold storage integrated high-temperature superconducting power grid expansion method according to claim 5, wherein The step (2) includes: (2.1) Through the LSTM neural network of the multi-time scale load forecasting module, fuse multi-source data streams such as historical 15-minute load data, real-time meteorological monitoring parameters, dynamic electricity price signals, and ambient temperature. After temporal feature extraction and non-linear relationship modeling, generate a refined load forecasting result with a 15-minute resolution for the next 10 days. (2.2) Input the forecasting result into the optimal scheduling engine in real time. The mixed integer linear programming algorithm solves the multi-objective optimization problem of minimizing the grid operation cost and maximizing the renewable energy consumption rate under the premise of meeting the superconducting cable temperature rise constraint, and generates a collaborative instruction set including the charge and discharge power thresholds of the energy storage system, the hierarchical response plan for interruptible loads, and the pre-start time window of the liquid nitrogen cold storage equipment. (2.3) The collaborative instruction set generation module decomposes the execution strategy through the dynamic programming algorithm: implement rolling optimization control of the charge and discharge power for the energy storage equipment, and dynamically schedule the industrial flexible loads at different levels in combination with the electricity price fluctuation curve and the load priority evaluation model; at the same time, predict the cooling demand based on the thermodynamic model of the superconducting cable, and start the preheating process of the liquid nitrogen pump group 6 - 8 hours in advance. (2.4) The liquid nitrogen cold storage control module adjusts the rotational speed and flow distribution of the multi-stage centrifugal pump in real time through a high-speed PID controller, completes the closed-loop regulation of the superconducting cable temperature field within a 200ms delay, and achieves a temperature control accuracy of 0.1K for the entire link through the distributed optical fiber temperature measurement system.

8. The dynamic load regulation-liquid nitrogen cold storage integrated high-temperature superconducting power grid expansion method according to claim 5, wherein For the optimal scheduling engine in the step (2), the objective function is: min(α·C cost +β·ΔT cable +γ·(1 - R renewable )) Among them, C cost is the grid operation cost, ΔT cable is the temperature rise of the superconducting cable, R renewable is the renewable energy accommodation rate, and the weight coefficients α, β, and Y are dynamically adjusted by the fuzzy analytic hierarchy process.

9. The dynamic load regulation - liquid nitrogen cold storage integrated high-temperature superconducting power grid expansion method according to claim 5, wherein The priority ranking of interruptible loads in the step (2) is based on the TOPSIS multi-objective decision-making model. The evaluation indicators include the user compensation cost, the load importance level, and the interruption duration, and generate a Pareto optimal solution set. The construction process of the TOPSIS multi-objective decision-making model is as follows: a. Data standardization processing to eliminate the influence of different dimensions on decision-making. b. Weight determination. c. Construction of a weighted normalized matrix. d. Determination of the ideal solution and the negative ideal solution. e. Calculation of distance and closeness. Euclidean distance formula: Relative closeness: C i The closer to 1, the better the solution.

10. The dynamic load regulation - liquid nitrogen cold storage integrated high-temperature superconducting power grid expansion method according to claim 5, characterized in that The construction of the liquid nitrogen flow-pressure coupling model in the step (3) includes: (3.1) Based on the mass conservation and Bernoulli equations, deduce the relationship between the liquid nitrogen flow velocity and the pipeline pressure drop. (3.2) Calibrate the pump group speed-flow curve through experiments and establish a non-linear fitting equation. (3.3) Combine the Kalman filter algorithm to update the model parameters in real time to adapt to the pipeline aging or blockage conditions.

Images