Efficient cross-season energy storage energy pile

Abstract

The invention discloses an efficient cross-season energy storage energy pile, and relates to the technical field of new energy and energy conservation. The problems that in an existing energy storage system, the thermal load prediction error is large, underground thermal diffusion attenuation is caused, the heat exchange capacity is lowered, and the geological adaptability is insufficient are solved. According to the scheme, an LA mixed time sequence prediction model is adopted to optimize load prediction, a distributed optical fiber temperature measurement array and a finite element inversion algorithm are combined to accurately reconstruct a stratum temperature field, and the geological type is identified based on a support vector machine classifier to realize self-adaptive regulation and control of heat exchange parameters; meanwhile, the heat pump power, the circulating pump frequency and the heat charging and discharging rate of the phase change material are optimized through a depth deterministic strategy gradient algorithm and a gradient heat release strategy; the heat exchange efficiency, the long-term stability and the complex environment adaptability of the cross-season energy storage system are remarkably improved, and the energy-saving effect of a building energy supply system is improved.

Description

Technical Field

[0001] The present invention relates to the field of new energy and energy-saving technology, and more specifically to an energy pile for efficient cross-season energy storage. Background Art

[0002] Energy piles for cross-seasonal energy storage are a type of energy storage technology that uses the heat capacity of underground soil or water to store and release heat between seasons. Energy piles combine building foundation piles with ground-source heat pump technology, using the thermal buffering capacity of underground soil to store heat in summer and release it in winter, thereby reducing the building's external energy dependence, improving energy efficiency and reducing carbon emissions. In recent years, with the advancement of material technology, heat exchange efficiency and intelligent control algorithms, the performance of energy pile systems has been continuously optimized, and the scope of application has also expanded from single buildings to regional heating systems.

[0003] At present, the energy pile system with efficient cross-seasonal energy storage is mainly based on the combination of ground source heat pump (GSHP) and structural heat exchange pile (Energy Piles) for energy storage and release. The core technology of this system involves the burial method of heat exchange pipes, the selection of heat transfer media, the optimization design of geological thermophysical properties, and the regulation of intelligent control systems. Usually, energy piles are laid with heat exchange pipes simultaneously during the construction of building foundations to form an underground thermal energy storage network. In summer, waste heat from buildings or solar thermal collectors are used to store heat in underground media (such as sand, clay or rock formations); in winter, the stored heat is extracted by ground source heat pumps for building heating.

[0004] In the existing related patent technologies, CN116094154A proposes an energy storage management system, which continuously collects the operating status data of the energy storage device, processes and judges the operating status data, so as to control the operating status of the energy storage device; another patent CN117791687B proposes an energy management method for a photovoltaic energy storage system, which estimates the cumulative power generation within a certain period by building a photovoltaic power generation model to compensate for the energy consumption demand gap of the electricity user; however, these technologies have caused some problems when applied to energy piles for cross-season energy storage:

[0005] First, the existing control system has a large thermal load prediction error. Due to the dynamic changes in building usage patterns and the uncertainty of weather conditions, it is difficult for the control system to accurately match heat input and output, resulting in excessive heat storage in summer or insufficient heating in winter. After long-term operation, the nonlinear effect of underground heat diffusion will cause energy storage to decay, and the heat exchange efficiency will also decrease year by year. In addition, the existing intelligent control algorithm still lacks adaptive optimization capabilities, making it difficult to achieve efficient energy storage management under various geological and climatic conditions. Most systems rely on fixed heat input / output strategies and ignore the dynamic changes in formation temperature, resulting in a decrease in heat exchange capacity and even requiring additional energy compensation, which reduces the energy saving effect of the overall system. At the same time, since the energy pile relies on the thermal conductivity characteristics of the underground soil, different geological conditions have a greater impact on the system performance. For example, in sandy soil or high groundwater flow areas, heat may be quickly dissipated, reducing the overall energy storage capacity; while in clay or rock formations, the heat accumulation effect may cause abnormal ground temperature increases, affecting the system's heat exchange efficiency and even causing foundation deformation problems. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention discloses an energy pile for efficient cross-seasonal energy storage, aiming to solve the problems raised in the background technology.

[0007] In order to achieve the above technical effects, the present invention adopts the following technical solutions:

[0008] An energy pile with efficient inter-seasonal energy storage, comprising: a dynamic coupling prediction module, which is used to perform multi-scale prediction of building heat load through an LA hybrid time series prediction model; when the building energy consumption monitoring system detects that the start and stop frequency of indoor temperature control equipment exceeds a threshold, the LA hybrid time series prediction model reconstructs the weight of historical load samples through a sliding time window mechanism, and simultaneously inputs meteorological forecast data into a three-dimensional heat conduction correction coefficient matrix, and outputs an inter-seasonal heat storage demand prediction value;

[0009] The formation temperature field reconstruction module is used to collect formation temperature field gradient data and axial and radial temperature gradients in the energy pile using a distributed optical fiber temperature measurement array. When it is detected that the formation temperature gradient change rate exceeds a preset critical value, the module reconstructs the three-dimensional non-steady-state temperature field through a finite element inversion algorithm based on the heat flux density sensor and temperature sensor data in the energy pile, calculates the heat diffusion loss rate, and corrects the heat exchange tube flow rate and phase change material filling density of the underground heat storage body in real time;

[0010] The geological parameter identification module includes a transient thermal response test unit and a pore water pressure monitoring array, which is used to identify the sand, clay or bedrock geological type through a support vector machine classifier, map heat exchange parameters, and invert the equivalent thermal conductivity and heat storage density correction coefficient through a thermo-seepage coupling model when the borehole radar detects abnormal fluctuations in the dielectric constant of the formation;

[0011] An adaptive optimization control module is used to adopt a convection-dominated heat exchange strategy when the groundwater velocity output by the geological parameter identification module exceeds a preset threshold value or the thermal conductivity is lower than a preset threshold value, generate an optimal control strategy through a deep deterministic strategy gradient algorithm, adjust the heat pump power and the circulation pump frequency in real time, and adjust the trigger threshold of the phase change material according to the real-time formation temperature distribution;

[0012] The global heat balance module is used to adopt a gradient heat release strategy when the formation temperature field reconstruction module detects that the temperature gradient in the heat accumulation area exceeds a preset threshold value, and adjusts the heat pump working fluid flow rate through a nonlinear PID controller for active thermal compensation, while dynamically adjusting the heat charging and releasing rate of the phase change material according to the real-time heat storage and release power ratio.

[0013] As a further technical solution of the present invention, the LA hybrid time series prediction model includes a hierarchical time series decomposition layer, a multimodal fusion layer, a dynamic weight reconstruction layer, a three-dimensional heat conduction correction layer, an abnormal frequency judgment layer and a cross-season coupling output layer;

[0014] The hierarchical time series decomposition layer is used to adopt an improved adaptive noise set empirical mode decomposition algorithm, generate an eigenmode function set by superimposing adaptive white noise iteration, and screen the components with sample entropy lower than a preset threshold of 0.8 and energy proportion exceeding 75%, and output high-frequency noise, medium-frequency daily cycle and low-frequency seasonal components;

[0015] The multimodal fusion layer is used to calibrate the meteorological and load time series offsets through a dynamic time warping algorithm, stretch the meteorological data time axis when the peak offset of the cross-correlation function of the meteorological and load time series exceeds a preset threshold, and fuse the temperature, humidity, radiation parameters and load components into a three-dimensional space-time tensor through a multi-head attention mechanism;

[0016] The dynamic weight reconstruction layer is used to mark outliers when the Mahalanobis distance exceeds the preset confidence interval threshold based on the KL divergence distribution of samples in the sliding window through Bayesian optimization, and solve the maximum a posteriori probability weight matrix for non-outlier samples through Gaussian process regression; the three-dimensional heat conduction correction layer is used to construct an unstructured formation heat conduction grid model using the finite element method, and when the meteorological temperature change rate ΔT / Δt>2°C / h, the correction coefficient matrix is ​​iteratively solved through the transient heat conduction control formula; the transient heat conduction control formula is:

[0017]

[0018] In formula (1), ρ is the soil density in kg / m 3 , C p is the soil specific heat capacity, in J / (kg·K), T is the formation temperature field, t is the time variable, in seconds, k is the soil thermal conductivity, in W / (m·K), Qsolar is the solar radiation heat source term, in W / m 3 , is the heat flux correction matrix;

[0019] The abnormal frequency decision layer is used to define the equipment start-stop event sequence as an observation value based on the hidden Markov model. The implicit states include normal, transition and abnormal. If the proportion of abnormal state Viterbi paths is greater than the preset threshold, the sample pool is cleared and the 30-day historical data is reloaded;

[0020] The cross-seasonal coupling output layer is used to input the predicted value of heat flux density into the pile strain formula:

[0021] ε=aΔT+β+δ thermal / Δt (2)

[0022] In formula (2), ε is the strain value of the pile, a is the thermal expansion coefficient of the pile material, ΔT is the temperature change, and β is the thermal stress coupling coefficient, the unit is MPa. -1 , δ thermal is the thermal stress value in MPa; if ε>0.2%, the Pareto optimal solution is solved by the sequential quadratic programming algorithm, and the heat storage demand value is output and sent to the heat pump control unit.

[0023] As a further technical solution of the present invention, the working method of the finite element inversion algorithm is as follows: based on the coordinates of the abnormal area marked by clustering, the local grid of the target area is encrypted by an unstructured grid generation algorithm, the heat flux density sensor data and the cluster partition temperature data are loaded, and the heat conduction forward model is used, and the nonlinear least squares algorithm is used to iteratively invert the formation thermal physical property parameters, including the thermal conductivity γ and the heat capacity c, to reconstruct the three-dimensional non-steady-state temperature field, and calculate the heat diffusion loss rate n loss :

[0024]

[0025] In formula (3), γ is the dynamic correction value of formation thermal conductivity; q sensor is the heat flux density sensor data; v is the flow rate of the working medium in the heat exchange tube; is the temperature gradient; if n loss When the preset safety threshold is exceeded, the flow rate of the heat exchange tube is dynamically adjusted through the PID controller, and based on the latent heat temperature curve of the phase change material, the optimal filling density under the current temperature field is matched through a linear interpolation algorithm.

[0026] As a further technical solution of the present invention, the working steps of the deep deterministic policy gradient algorithm include:

[0027] Step 1: Based on the state-action value function, define the optimal time difference of the target strategy and calculate the strategy value. The formula is:

[0028]

[0029] In formula (4), is the optimal time difference target value, R t is the instantaneous benefit of the heat exchange control strategy at time t; δ is the discount factor; Q' is the Q value estimate of the target network; a' is the optimal strategy for the next step; ω 1 is the strategy smoothing factor; is the execution action at the previous moment; θ Q' is the target Q network parameter; S t+1 Indicates that action a is executed at the current time t t Afterwards, the change of the environmental state at time t+1; represents the policy network parameter θ u The calculated value of the gradient;

[0030] Step 2: Use the policy gradient update function to optimize the policy network and calculate the gradient. The calculation formula is:

[0031]

[0032] In formula (5), J(θ u ) is the policy objective function; u(S|θ u ) is the policy network; ω 2 is the gradient smoothing factor; is the rate of change of strategy update; E represents the expected value of all future state S and action a samples;

[0033] Step 3: Store historical state data through the experience replay mechanism, and introduce the entropy regularization term through the adaptive entropy regularization strategy to optimize and improve the exploration ability of the heat exchange strategy. The formula is:

[0034]

[0035] In formula (6), T entropy is the entropy regularization loss; ω 3 is the regularization coefficient; π(a i |S) is the strategy probability distribution; N is the number of samples in the strategy space;

[0036] Step 4: Combine the adaptive adjustment of the circulation pump frequency with the nonlinear mapping of the heat pump power to calculate the optimal heat exchange control parameters As the final implementation strategy.

[0037] Based on the above technical solution, the positive and beneficial effects of the present invention are:

[0038] 1. The present invention dynamically adjusts the load prediction sample weights through the LA hybrid time series prediction model, and optimizes the heat exchange control strategy in real time based on the deep deterministic policy gradient algorithm, so that the heat storage and release of the energy pile are more accurate, ensuring that the energy storage system adapts to the dynamic changes in building usage patterns and the uncertainty of meteorological conditions, thereby effectively reducing the heat load prediction error, improving the supply and demand matching of cross-season energy storage, and avoiding the problem of excessive heat storage in summer or insufficient heating in winter due to prediction deviations.

[0039] 2. The formation temperature field reconstruction module and the geological parameter identification module work together to monitor the formation temperature field and groundwater flow characteristics in real time through the distributed optical fiber temperature measurement array and the transient thermal response test unit, and calculate the underground heat diffusion loss rate in combination with the finite element inversion algorithm, and use the support vector machine classifier to identify the geological type, and then establish a thermal seepage coupling model to optimize the heat exchange parameters, so that the energy pile can maintain the optimal heat exchange capacity in different formation environments. This synergistic mechanism not only enhances the system's adaptability to changes in the geological environment, but also reduces the energy storage attenuation caused by the nonlinear effect of underground heat diffusion through accurate correction of formation thermal conductivity and heat storage density, and improves the energy storage stability in long-term operation.

[0040] 3. The adaptive optimization control module and the global heat balance module jointly construct a dynamic heat exchange control and thermal field equilibrium optimization mechanism. When the groundwater flow rate exceeds the standard or the formation thermal conductivity is lower than the preset threshold, the system uses a convection-dominated heat exchange strategy to adjust the heat pump power and circulation pump frequency, and optimizes the formation heat release path based on the gradient heat release strategy, so that the heat pump working fluid flow and the phase change material charging and releasing heat rate can be dynamically adjusted according to the formation temperature. This control strategy effectively alleviates the abnormal formation temperature caused by insufficient heat exchange capacity or heat accumulation effect, improves the long-term operation safety of the system in complex environments, and reduces the demand for additional energy compensation, improving overall energy efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work, among which:

[0042] Figure 1 This is a structural diagram of an energy pile for efficient cross-season energy storage according to the present invention;

[0043] Figure 2 It is a structural diagram of the LA hybrid time series prediction model of the present invention;

[0044] Figure 3It is a working framework diagram of the finite element inversion algorithm of the present invention;

[0045] Figure 4 It is a working principle diagram of the support vector machine classifier of the present invention;

[0046] Figure 5 It is a working framework diagram of the thermal seepage coupling model of the present invention;

[0047] Figure 6 This is a step diagram of the working method of the deep deterministic policy gradient algorithm of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] An energy pile for efficient cross-season energy storage, comprising: a dynamic coupling prediction module, a formation temperature field reconstruction module, a geological parameter identification module, an adaptive optimization control module and a global thermal balance module;

[0050] In the energy pile of this efficient cross-season energy storage, Figure 1 As shown in the figure, the cross-seasonal heat storage demand prediction value of the dynamic coupling prediction module is output to the input end of the adaptive optimization control module as the decision basis for heat exchange control. At the same time, the building heat load prediction data of this module is transmitted to the formation temperature field reconstruction module to guide the control strategy of the underground temperature field. The formation temperature field reconstruction module calculates the formation temperature gradient change rate and heat diffusion loss rate based on the distributed optical fiber temperature measurement array, and outputs them to the global heat balance module for adjusting the gradient heat release strategy. At the same time, the temperature distribution data of this module is input to the geological parameter identification module to assist in the correction of formation thermal conductivity. The geological parameter identification module outputs the geological type, equivalent thermal conductivity and heat storage density correction coefficient to the adaptive optimization control module through the support vector machine classifier to optimize the heat exchange parameters, and at the same time feeds back the geological characteristic data to the formation temperature field reconstruction module to improve the temperature field inversion accuracy. The adaptive optimization control module uses the deep deterministic strategy gradient algorithm to adjust the heat pump power and circulation pump frequency according to the formation thermal conductivity and groundwater flow rate data of the geological parameter identification module, and outputs the optimized heat exchange flow control parameters to the global heat balance module. The global heat balance module receives the formation temperature gradient change rate, heat exchange flow control parameters and heat storage and release power ratio, uses a nonlinear PID controller to adjust the heat pump working fluid flow, and feeds back the temperature state data after heat release adjustment to the adaptive optimization control module to achieve dynamic optimization of the system.

[0051] Among them, the dynamic coupling prediction module is used to perform multi-scale prediction of building heat load through the LA hybrid time series prediction model; when the building energy consumption monitoring system detects that the start and stop frequency of indoor temperature control equipment exceeds the threshold, the LA hybrid time series prediction model reconstructs the historical load sample weight through the sliding time window mechanism, and inputs the meteorological forecast data into the three-dimensional heat conduction correction coefficient matrix to output the cross-season heat storage demand prediction value; Figure 2 As shown, further, the LA hybrid time series prediction model includes a hierarchical time series decomposition layer, a multimodal fusion layer, a dynamic weight reconstruction layer, a three-dimensional heat conduction correction layer, an abnormal frequency judgment layer, and a cross-season coupling output layer;

[0052] The hierarchical time series decomposition layer is used to adopt an improved adaptive noise set empirical mode decomposition algorithm, generate an eigenmode function set by superimposing adaptive white noise iteration, and screen the components with sample entropy lower than a preset threshold of 0.8 and energy proportion exceeding 75%, and output high-frequency noise, medium-frequency daily cycle and low-frequency seasonal components;

[0053] The multimodal fusion layer is used to calibrate the meteorological and load time series offsets through a dynamic time warping algorithm, stretch the meteorological data time axis when the peak offset of the cross-correlation function of the meteorological and load time series exceeds a preset threshold, and fuse the temperature, humidity, radiation parameters and load components into a three-dimensional space-time tensor through a multi-head attention mechanism;

[0054] The dynamic weight reconstruction layer is used to mark outliers when the Mahalanobis distance exceeds the preset confidence interval threshold based on the KL divergence distribution of samples in the sliding window through Bayesian optimization, and solve the maximum a posteriori probability weight matrix for non-outlier samples through Gaussian process regression; the three-dimensional heat conduction correction layer is used to construct an unstructured formation heat conduction grid model using the finite element method, and when the meteorological temperature change rate ΔT / Δt>2°C / h, the correction coefficient matrix is ​​iteratively solved through the transient heat conduction control formula; the transient heat conduction control formula is:

[0055]

[0056] In formula (1), ρ is the soil density in kg / m 3 , C p is the soil specific heat capacity, in J / (kg·K), T is the formation temperature field, t is the time variable, in seconds, k is the soil thermal conductivity, in W / (m·K), Q solar is the solar radiation heat source term, in W / m 3 , is the heat flux correction matrix;

[0057] The abnormal frequency decision layer is used to define the equipment start-stop event sequence as an observation value based on the hidden Markov model. The implicit states include normal, transition and abnormal. If the proportion of abnormal state Viterbi paths is greater than the preset threshold, the sample pool is cleared and the 30-day historical data is reloaded;

[0058] The cross-seasonal coupling output layer is used to input the predicted value of heat flux density into the pile strain formula:

[0059] ε=aΔT+β+δ thermal / Δt (2)

[0060] In formula (2), ε is the strain value of the pile, a is the thermal expansion coefficient of the pile material, ΔT is the temperature change, and β is the thermal stress coupling coefficient, the unit is MPa. -1 , δ thermal is the thermal stress value in MPa; if ε>0.2%, the Pareto optimal solution is solved by the sequential quadratic programming algorithm, and the heat storage demand value is output and sent to the heat pump control unit.

[0061] Among them, adaptive noise ensemble empirical mode decomposition (EEMD) superimposes adaptive white noise on the original building load time series signal to decompose the data into multiple intrinsic mode functions (IMFs) at different time scales. In each decomposition process, EEMD uses an ensemble averaging mechanism to eliminate the problem of modal aliasing. Finally, the dominant energy components are screened based on sample entropy calculation, and only the components with sample entropy less than 0.8 and energy accounting for more than 75% are retained to suppress short-term random fluctuations and improve the accuracy of long-term trend prediction. Dynamic time warping (DTW) is used to compare the time alignment errors of meteorological data and load data. When the peak offset of the cross-correlation function between the two exceeds the threshold, the time axis stretching or compression transformation is used to make the time scales of different data sources consistent. This method uses the dynamic programming (DP) algorithm to calculate the optimal alignment path, and adjusts the meteorological data based on the cumulative distance matrix to synchronize it with the building load data, thereby improving the load prediction accuracy. The multi-head attention mechanism (MHA) processes the spatiotemporal associations between temperature, humidity, solar radiation and building load in parallel through multiple independent attention heads. Each attention head uses query-key-value (QKV) mapping to calculate the attention score, and obtains the influence coefficient of different meteorological variables on load forecasting through weighted summation. Finally, MHA stabilizes the data distribution through residual connection and layer normalization to improve the adaptability of the load forecasting model to environmental variables. The KL divergence anomaly detection technique is used to calculate the degree of change of data distribution within the sliding window to detect anomalies in building load data. First, the probability density function (PDF) of the data distribution is calculated in the current window, and then the KL divergence value of the distribution relative to the previous window is calculated. If the value exceeds the preset confidence interval threshold, the data is judged to be abnormal, and the abnormal data points are further screened by Mahalanobis distance. Bayesian optimization is used to predict the optimal value of the objective function based on a probability model (usually a Gaussian process) during the correction of abnormal data. During the optimization process, the acquisition function (AF) is used to search for the optimal solution in the data sample space, and the expectation improvement (EI) strategy is used to dynamically adjust the threshold of KL divergence detection to ensure the stability of the anomaly detection algorithm. Gaussian process regression (GPR) uses kernel functions to map the nonlinear relationship of data, and optimizes the weight matrix of building load prediction through maximum a posteriori probability (MAP) estimation. In the weight reconstruction process, GPR calculates the gradient of the likelihood function and uses Laplace approximation to solve the optimal weight distribution to improve the dynamic adaptability of load prediction. Finite element analysis (FEA) is used to simulate the thermal diffusion process of the underground temperature field. Unstructured grids are used to divide the stratum area, and the heat diffusion loss rate is calculated based on the transient heat conduction control equation.FEA calculates the temperature change of each grid unit through time-stepping iteration, and corrects the formation heat conduction parameters in combination with the finite element interpolation function to improve the accuracy of temperature prediction. The hidden Markov model (HMM) is used to analyze the state changes of the start and stop events of building temperature control equipment by constructing the implicit state-observation value probability transfer matrix. The model calculates the optimal implicit state path based on the Viterbi algorithm, and decides whether to clear the sample pool and reload historical data based on the proportion of abnormal states to improve the stability of load forecasting. Sequential quadratic programming (SQP) uses a second-order optimization method to solve the Pareto optimal solution of cross-seasonal heat storage demand. This method calculates the optimization direction through Lagrangian dual optimization and solves the quadratic approximate subproblem in each iteration to ensure the optimal regulation of heat storage demand when the thermal strain of the pile exceeds the limit, thereby improving the long-term stability of the building energy storage system.

[0062] In engineering applications, the implementation of this solution needs to follow the following process: First, the start and stop signals of indoor temperature control equipment, meteorological parameters (temperature, humidity, radiation) and pile strain data are collected in real time through the building energy consumption monitoring system, and the original database is constructed with a granularity of 1 minute, and the training set and the validation set are divided by a sliding time window (window 7 days, step length 1 day). Secondly, the CEEMDAN algorithm is called to perform modal decomposition on the historical load data, and the low-frequency seasonal component is extracted as the baseline of the cross-seasonal heat storage demand. The dynamic time warping algorithm is used to align the meteorological and load time series phases, and the multi-head attention network is implemented through the PyTorch framework to generate a multimodal tensor that integrates three-dimensional spatiotemporal features. Subsequently, the KL divergence and Mahalanobis distance are calculated based on Scikit-learn, and the Bayesian optimization is performed in combination with the GPyOpt toolkit to remove outlier samples, and the weight matrix is ​​dynamically allocated through Gaussian process regression. At the same time, a finite element model of formation heat conduction is constructed in COMSOL Multiphysics, and a Python script is embedded to update the heat flux correction coefficient in real time to ensure the model accuracy under the meteorological mutation scenario. The abnormality handling link uses the hmmlearn library to train the hidden Markov model, sets the abnormal state threshold (30%), and calls the SQL script to clear the sample pool in the Redis cache when the abnormality is triggered, and reloads the historical data of the InfluxDB time series database. Finally, the optimized heat storage demand value is sent to the heat pump PLC controller through the Modbus protocol, and the pile strain sensor data is monitored in real time. If the thermal stress exceeds the limit (ε>0.2%), the IPOPT solver is called to recalculate the Pareto optimal solution, dynamically adjust the working fluid flow and the phase change material charging and discharging rate, and realize the efficient and safe operation of the energy storage system.

[0063] Compared with the existing technology, this solution significantly improves the prediction accuracy of cross-seasonal heat storage demand through hierarchical time series decomposition and multimodal fusion, and enhances the generalization ability of the model by combining dynamic weight optimization and heat conduction physical correction; the abnormal judgment mechanism effectively avoids the risk of system-level disturbances, and the thermal-mechanical coupling control strategy balances structural safety and energy storage efficiency, which improves the long-term operation stability and energy utilization of energy piles as a whole, and provides reliable technical support for low-carbon and intelligent energy supply of buildings.

[0064] In implementation, the hardware environment of the dynamic coupling prediction module includes:

[0065] Temperature control equipment monitoring sensors (installed in the air conditioning / floor heating circuits in the building to collect start and stop frequencies); meteorological station (arranged on the roof of the building, containing temperature, humidity, and radiation sensors); pile heat flux meters (embedded at depths of 5m, 10m, and 15m underground in the energy pile to monitor heat flux density); strain sensors (mounted on the surface of the pile to detect axial / radial strain); data acquisition modules (centrally process sensor signals and transmit them to the server via RS485); heat pump PLC controller (connecting the working fluid circulation pump and the phase change material heat storage unit).

[0066] Based on the above hardware environment, a comparative experiment was designed. Group A enabled the dynamic coupling prediction module, and Group B used the ARIMA+static model; that is, the ARIMA time series model was combined with the static thermal conductivity coefficient (fixed value 0.8W / (m·K)), ignoring the dynamic weight optimization and abnormal judgment mechanism. Groups A / B shared the same sensors and heat pump equipment, and recorded the heat load prediction error, equipment start and stop times, pile strain value, energy consumption and heat storage efficiency every 5 minutes; the sliding window of Group A = 7 days, the ARIMA order of Group B (p, d, q) = (3, 1, 2), and the thermal conductivity coefficient was fixed; during the experiment, the outdoor temperature fluctuated from -5℃ to 5℃, and the building heat load demand was constant (30kW); the experimental data records are shown in Table 1:

[0067] Table 1 Module comparison experiment record table

[0068]

[0069] The experimental results show that the heat load prediction error of group A (RMSE = 1.24kW) is significantly lower than that of group B (3.86kW), because dynamic weight optimization and multimodal fusion improve the generalization ability of the model; the start-stop frequency of group A (2.10 times / hour) is 57.3% lower than that of group B (4.92 times / hour), and the dynamic abnormal judgment mechanism effectively suppresses invalid regulation; the maximum strain value of the pile body of group A (0.164%) is lower than that of group B (0.290%), and the three-dimensional heat conduction correction and Pareto optimization alleviate the accumulation of thermal stress; the average daily energy consumption of group A (58.92kWh) is 21.4% lower than that of group B (74.98kWh), and the heat storage efficiency is increased by 26.3% (82.08% vs 64.92%). Therefore, the dynamic coupling prediction module significantly improves the prediction accuracy, operation stability and energy efficiency of the energy pile through multi-scale optimization and physical correction.

[0070] The formation temperature field reconstruction module is used to collect the formation temperature field gradient data and the axial and radial temperature gradients in the energy pile using a distributed optical fiber temperature measurement array. When it is detected that the formation temperature gradient change rate exceeds the preset critical value, the three-dimensional non-steady-state temperature field is reconstructed through the finite element inversion algorithm based on the heat flux density sensor and temperature sensor data in the energy pile, the heat diffusion loss rate is calculated, and the heat exchange tube flow rate and phase change material filling density of the underground heat storage body are corrected in real time; further, Figure 3 As shown in the figure, the working method of the finite element inversion algorithm is as follows: based on the coordinates of the abnormal area marked by clustering, the local grid of the target area is encrypted through the unstructured grid generation algorithm, the heat flux density sensor data and the cluster partition temperature data are loaded, and the heat conduction forward model is used. The nonlinear least squares algorithm is used to iteratively invert the formation thermal physical parameters, including the thermal conductivity γ and the heat capacity c, to reconstruct the three-dimensional non-steady-state temperature field, and calculate the heat diffusion loss rate n loss :

[0071]

[0072] In formula (3), γ is the dynamic correction value of formation thermal conductivity; q sensor is the heat flux density sensor data; v is the flow rate of the working medium in the heat exchange tube; is the temperature gradient; if n loss When the preset safety threshold is exceeded, the flow rate of the heat exchange tube is dynamically adjusted through the PID controller, and based on the latent heat temperature curve of the phase change material, the optimal filling density under the current temperature field is matched through a linear interpolation algorithm.

[0073] Among them, the unstructured grid generation algorithm is used to discretize the formation temperature field with high precision. First, the abnormal temperature gradient area is identified based on clustering marks, and the density peak clustering method is used to determine the center point of the abnormal area according to the local density and relative distance of the temperature data, and generate cluster partitions. Subsequently, Delaunay triangulation is used to generate local high-density grids in the abnormal area, and uniform grid distribution is maintained in the normal area to reduce the amount of calculation and improve the accuracy of finite element solutions.

[0074] The heat conduction forward model is used to simulate the thermal diffusion behavior of the formation and serves as the forward calculation framework for finite element inversion. It is based on the Fourier heat conduction equation, describes the relationship between heat flux density, thermal conductivity and temperature gradient, and constructs the time-dependent heat diffusion control equation. During the calculation process, the implicit finite difference method IFDM is used to control the time step, and the heat exchange intensity between the heat exchange pile and the soil interface is corrected in combination with adaptive boundary conditions to ensure that the simulated temperature field can accurately reflect the actual formation heat diffusion.

[0075] The nonlinear least squares inversion algorithm is used to solve the thermophysical parameters of the formation, including thermal conductivity and specific heat capacity. First, based on the Levenberg-Marquardt optimization (LM Optimization) method, the objective function is constructed to minimize the mean square error (MSE) between the simulated temperature and the actual temperature measured by the sensor. Secondly, the multi-scale iterative regularization MSIR method is combined to prevent the parameters from overfitting or converging to the local optimal solution during the inversion process. Finally, the partial derivatives are solved by the Newton iteration method, and the parameter update step size is optimized, so that the inversion calculation can achieve the best balance between convergence accuracy and computational efficiency.

[0076] The heat diffusion loss rate is used to quantify the heat transfer performance of underground heat storage. First, the heat dissipation mode in the heat transfer process is analyzed based on Laplace transform, and the heat flux loss of different grid cells is calculated by combining the finite volume method FVM. Subsequently, the entropy increase analysis is used to evaluate the irreversible energy loss in the formation heat diffusion process, and the heat diffusion loss rate is ensured to be maintained within a safe range through adaptive temperature gradient threshold control.

[0077] In the implementation of this module: optical fiber temperature measurement nodes are arranged every 0.5m along the axial direction inside the pile foundation, and a ring temperature measurement array is set radially around the pile body; a heat flux density sensor (range 0~1000W / m 2), and a meteorological station is configured on the surface to monitor temperature, humidity and solar radiation data. The temperature gradient, heat flux density and meteorological parameters are collected in real time, and the temperature change rate is extracted through a sliding time window (window length 24h). The K-means clustering algorithm is used to mark abnormal areas (standard deviation σ ≥ 1.2℃). The Delaunay grid generator is called to locally encrypt the abnormal area, load the forward model, and use the finite element inversion algorithm to iteratively correct γ and c until the temperature residual converges. When n loss ≥10%, the PID controller outputs the heat exchange tube flow rate adjustment instruction (v = 0.51200kg / m 3 ), and is sent to the circulation pump and phase change material injection system in real time. If the inversion iteration number exceeds 50 and still has not converged, the grid reconstruction process is automatically triggered, the local encrypted area is reset and the data is reloaded.

[0078] Compared with the traditional static temperature field monitoring method, this module achieves high-resolution thermal field reconstruction through distributed optical fiber and dynamic inversion algorithm, and combines PID and phase change material optimization strategies to significantly reduce heat diffusion losses and improve heat storage efficiency. Local grid encryption and non-steady-state parameter inversion in abnormal areas enhance the model's adaptability to changes in formation thermal properties, effectively avoiding the risk of overheating or heat leakage, and providing accurate and reliable thermal management capabilities for inter-seasonal energy storage systems.

[0079] In order to verify the effectiveness of the finite element inversion algorithm in the cross-season energy storage pile, a comparative experiment was designed: Group A used the finite element inversion algorithm, and Group B used the traditional static empirical model (fixed thermal conductivity λ = 1.5W / m·K, temperature field estimated based on linear interpolation). The experiment was carried out in the same sand-clay alternating stratum site. Both groups deployed distributed optical fiber temperature measurement arrays, heat flux density sensors and circulating pump systems. The heat source input a constant heat flux density of 500W / m 2 The temperature field data, heat diffusion loss rate and energy consumption were collected every 30 minutes for 24 hours. Group A used Delaunay grid local encryption and Levenberg-Marquardt optimization to invert thermal conductivity and heat capacity, dynamically adjust the heat exchange tube flow rate (0.5-2.0m / s) and phase change material filling density (800-1200kg / m 3 ); Group B operated according to a fixed strategy (flow rate 1.0 m / s, filling density 1000 kg / m 3 ). The experimental record form is shown in Table 2:

[0080] Table 2 BP finite element inversion algorithm experimental record

[0081]

[0082] The experimental results show that group A using the finite element inversion algorithm is significantly better than group B using the traditional static model in core indicators such as temperature field reconstruction accuracy, thermal conductivity inversion error and thermal diffusion loss rate.

[0083] The geological parameter identification module includes a transient thermal response test unit and a pore water pressure monitoring array, which is used to identify the sand, clay or bedrock geological type through a support vector machine classifier, map heat exchange parameters, and invert the equivalent thermal conductivity and heat storage density correction coefficient through a thermal seepage coupling model when the borehole radar detects abnormal fluctuations in the dielectric constant of the formation. The transient thermal response test unit uses pulsed heat flow excitation to apply short-term heat pulses to the formation around the borehole, and collects formation temperature response data through fiber Bragg grating sensing, calculates the transient thermal diffusion coefficient, and identifies the initial thermal physical properties of the formation. The pore water pressure monitoring array extracts the groundwater flow trend through a dynamic water pressure gradient analysis method, and uses adaptive permeability regression APR to calculate soil permeability parameters to obtain a correction factor for the influence of groundwater on thermal diffusion. Further, if Figure 4 As shown in the figure, the working method of the support vector machine classifier is as follows: first, the dielectric property data of the formation is projected into the high-dimensional feature space through kernel function mapping, then, the Lagrangian dual optimization method is used to maximize the data category interval, and the soft interval support vector is used to enhance the adaptability of the model to the geological transition zone; in the classification process, the support vector machine classifier uses KKT condition screening to dynamically update the support vector and eliminate redundant data points; then, the support vector weight distribution is optimized through the adaptive weight adjustment mechanism, so that the decision boundary of the classifier between sand, clay and bedrock categories conforms to the trend of the change of the formation thermophysical properties; finally, the support vector machine classifier uses the heat exchange parameter mapping mechanism to call the corresponding thermal diffusivity, heat capacity ratio and permeability from the predefined thermophysical property parameter library according to the classification results.

[0084] Among them, the transient thermal response test unit uses pulse heat flow excitation technology to apply short-term heat pulses to the formation around the borehole to stimulate the transient thermal response behavior of the formation. The change curve of the formation temperature over time is recorded by fiber Bragg grating sensing, and the transient thermal diffusion coefficient of the formation is calculated using the time-temperature response model. In the data processing process, the heat flux density correction algorithm is used to compensate for the heat loss of the formation surface, and the measurement accuracy is optimized in combination with the partial differential temperature gradient inversion method to accurately identify the initial thermophysical parameters of the formation, including thermal conductivity, specific heat capacity and thermal diffusion coefficient. The pore water pressure monitoring array obtains the groundwater flow trend through the dynamic water pressure gradient analysis method. First, the groundwater infiltration rate is analyzed using the transient water level disturbance response, and the local permeability change is calculated based on the Poisson equation. Subsequently, the adaptive permeability regression model is used to dynamically fit the permeability distribution in multiple time windows to obtain the correction factor of the impact of groundwater flow on formation thermal diffusion. Finally, the permeability parameters are adjusted through the multi-scale permeability correction mechanism so that the calculation results can adapt to the dynamic changes of different geological conditions.

[0085] The support vector machine classifier is used to identify the formation type and map the corresponding heat transfer parameters based on the formation dielectric property data detected by borehole radar. First, the kernel function mapping method is used to project the low-dimensional dielectric property data into the high-dimensional feature space to enhance the linear separability of the data. Subsequently, the objective function of the support vector machine is solved by Lagrangian dual optimization to maximize the hyperplane interval of the data category, and the soft margin support vector enhancement strategy is combined to improve the adaptability of the classifier to the geological transition zone. During the classification process, the support vector is dynamically updated based on the KKT condition screening, and redundant data points are eliminated to reduce the computational complexity. Then, the support vector weight is optimized through adaptive weight adjustment so that the classification boundary conforms to the trend of the formation thermophysical property change. Finally, the classification result calls the predefined thermophysical property parameter library through the heat exchange parameter mapping mechanism, and automatically matches the corresponding thermal diffusivity, heat capacity ratio and permeability to optimize the heat transfer performance of the energy storage system.

[0086] In the specific application of high-efficiency cross-season energy storage piles, technicians in this field can implement the geological parameter identification function according to the following steps:

[0087] Fiber Bragg grating temperature sensors (FBG-T) and pressure sensors (FBG-P) were arranged at 1m intervals in the borehole, and transient thermal response test units (power 2kW, adjustable pulse width) and data acquisition terminals were installed on the surface. The transient thermal response test unit applied heat pulses to the formation, and the FBG-T recorded the temperature response curve at a sampling rate of 1kHz. The temperature attenuation characteristics were extracted by fast Fourier transform (FFT); the pore water pressure monitoring array collected water pressure gradient data in real time, and the sliding average filter was used to eliminate noise and generate a dynamic water pressure gradient time series. Then, the temperature attenuation curve was fitted based on the non-Fourier heat conduction model to invert the initial thermal conductivity and heat capacity ratio; the permeability and non-Darcy flow coefficient were iteratively solved by the Darcy-Forchheimer equation, and the groundwater correction factor was output; the dielectric constant and loss tangent were used as feature vectors to pre-train the RBF kernel SVM model and optimize the soft interval parameters and kernel width. When the borehole radar detects abnormal fluctuations in the dielectric constant, the SVM classifier is triggered, the real-time feature vector is input, and the probability of sand, clay or bedrock is output; the predefined thermophysical property parameters are called according to the classification results, and the equivalent thermal conductivity and heat storage density correction coefficient inverted by the thermal seepage coupling model are input. Finally, the equivalent thermal conductivity and heat storage density correction coefficient inverted by the thermal seepage coupling model are sent to the heat storage control unit, the flow rate of the heat exchange tube is adjusted, and the groundwater correction factor is synchronously updated to the heat diffusion loss rate calculation module.

[0088] like Figure 5 As shown in the figure, the workflow of the thermal seepage coupling model starts with the input of the classified geological type. First, the non-steady-state thermal seepage coupling equation (combined energy equation and Darcy's law) is constructed. If the finite element multiscale inversion method (FEMI) successfully discretizes the temperature field-flow velocity-heat exchange relationship matrix, it enters the Poisson equation constraint (PEC) correction stage; if the discretization fails, the energy equation is re-parameterized and the equation is rebuilt. In the velocity correction link, when the error of the water velocity after Poisson correction exceeds the 2% threshold, the coupling coefficient of Darcy's law is automatically adjusted and cyclically corrected until the accuracy requirements are met. The correction parameters are verified by sensitivity analysis. If the error exceeds 5%, the geological type reclassification process is triggered and the model inversion is restarted; if the verification passes, the parameters are pushed to the thermal storage control system to dynamically adjust the operation strategy. The entire process forms a closed loop through discretization fault tolerance, velocity correction cycle and parameter feedback mechanism to ensure that the model output is dynamically matched with the thermal properties of the formation.

[0089] During implementation, the present invention significantly improves the accuracy and robustness of formation thermal physical parameter identification through the fusion of multi-source data of transient thermal response test and pore water pressure monitoring, combined with the dynamic optimization decision of support vector machine classifier. Compared with the traditional empirical model, it can accurately identify the sand-clay transition zone and bedrock formation, dynamically correct thermal conductivity and permeability parameters, effectively suppress the interference of groundwater seepage on thermal diffusion, and optimize the heat storage density control strategy, thereby enhancing the geological adaptability and long-term operation stability of the cross-season energy storage system, and providing technical guarantee for efficient thermal energy management under complex formation conditions.

[0090] In order to verify the advantages of the geological parameter identification module (Group A) over the traditional geological empirical model (Group B) in terms of stratigraphic classification accuracy and thermophysical parameter inversion, a comparative experiment was conducted; Group B adopted an empirical classification method based on drill core samples (manually distinguishing sand / clay / bedrock), combined with fixed thermal conductivity (sand λ = 1.5 W / m·K, clay λ = 1.2 W / m·K, bedrock λ = 2.8 W / m·K) and permeability (k = 1e-4 m / s) parameters, without a dynamic correction mechanism.

[0091] The experiment was conducted on two groups of pile foundations (Group A and Group B) in the same area. The strata were alternating layers of sand-clay-bedrock (thickness 0.5-2m). The heat source input was a constant heat flux density of 400W / m 2 , lasting 48 hours; the experimental data records are shown in Table 3:

[0092] Table 3 Experimental record of geological parameter identification module

[0093]

[0094] As can be seen from data table 3, the geological parameter identification module (Group A) is significantly superior to the traditional empirical model (Group B) in key indicators such as formation classification accuracy (increased by 32.9%), thermal conductivity inversion error (reduced by 79.0%) and permeability correction error (reduced by 75.6%). It accurately identifies the sand-clay transition zone and bedrock formations through support vector machine dynamic classification and thermal seepage coupling inversion, adaptively corrects thermophysical parameters, reduces thermal diffusion loss rate (reduced by 58.4%) and system energy consumption (reduced by 30.4%), and improves temperature field stability (σ reduced by 64.3%) and phase change material filling accuracy (error reduced by 79.3%). Experiments have shown that the module can effectively cope with the challenges of thermophysical parameter identification under complex formation conditions and provide reliable technical support for the efficient operation of inter-seasonal energy storage systems.

[0095] The adaptive optimization control module is used to adopt a convection-dominated heat exchange strategy when the groundwater flow rate output by the geological parameter identification module exceeds a preset threshold value or the thermal conductivity is lower than a preset threshold value, and generate an optimal control strategy through a deep deterministic policy gradient algorithm, so as to adjust the heat pump power and the circulation pump frequency in real time, and adjust the trigger threshold of the phase change material according to the real-time formation temperature distribution; the specific working method is: based on the abnormal signals of the groundwater flow rate and formation thermal conductivity of the geological parameter identification module, the constraint conditions are established through the boundary constraint mechanism; then, the convection-dominated heat exchange strategy is used to optimize and adjust the heat pump power and the circulation pump frequency; during the control process, a multi-step timing strategy is used to evaluate and calculate the cumulative reward value under different heat exchange strategies, and the control parameters are corrected through the policy gradient adaptive update method; then, the optimal trigger temperature of the phase change material is calculated based on the real-time formation temperature distribution prediction method, and the phase change trigger threshold dynamic adjustment mechanism is used to dynamically adjust the phase change heat storage process; finally, the heat exchange parameters are iteratively adjusted through deep reinforcement learning feedback optimization to optimize the long-term heat storage stability of the energy pile. Further, such as Figure 6 As shown, the working steps of the deep deterministic policy gradient algorithm include:

[0096] Step 1: Based on the state-action value function, define the optimal time difference of the target strategy and calculate the strategy value. The formula is:

[0097]

[0098] In formula (4), is the optimal time difference target value, R t is the instantaneous benefit of the heat exchange control strategy at time t; δ is the discount factor; Q' is the Q value estimate of the target network; a' is the optimal strategy for the next step; ω 1 is the strategy smoothing factor; is the execution action at the previous moment; θ Q' is the target Q network parameter; S t+1 Indicates that action a is executed at the current time t t Afterwards, the change of the environmental state at time t+1; represents the policy network parameter θ u The calculated value of the gradient;

[0099] Step 2: Use the policy gradient update function to optimize the policy network and calculate the gradient. The calculation formula is:

[0100]

[0101] In formula (5), J(θ u ) is the policy objective function; u(S|θ u ) is the policy network; ω 2 is the gradient smoothing factor; is the rate of change of strategy update; E represents the expected value of all future state S and action a samples;

[0102] Step 3: Store historical state data through the experience replay mechanism, and introduce the entropy regularization term through the adaptive entropy regularization strategy to optimize and improve the exploration ability of the heat exchange strategy. The formula is:

[0103]

[0104] In formula (6), τ entropy is the entropy regularization loss; ω 3 is the regularization coefficient; π(a i |S) is the strategy probability distribution; N is the number of samples in the strategy space;

[0105] Step 4: Combine the adaptive adjustment of the circulation pump frequency with the nonlinear mapping of the heat pump power to calculate the optimal heat exchange control parameters As the final implementation strategy.

[0106] Among them, the boundary constraint control first determines whether it exceeds the set heat exchange adaptation threshold based on the groundwater flow rate and thermal conductivity data of the geological parameter identification module. The system uses the constrained Lagrange multiplier method to establish the heat exchange boundary constraint conditions, and constructs the penalty term optimal solution method to limit the heat exchange parameters that exceed the safety range. At the same time, through the boundary search strategy based on gradient descent, the heat exchange parameters are kept in the optimal distribution within the feasible solution range to ensure that the heat exchange process will not cause energy storage imbalance due to fluctuations in geological characteristics. The system monitors the formation temperature field and groundwater flow rate data in real time, and uses the fluid momentum balance equation to calculate the dominant mode of formation heat exchange. In the case of high groundwater flow rate, the system gives priority to the convection heat exchange enhancement strategy, and combines the flow field adaptive heat exchange regulation model to calculate the optimal allocation plan of the heat pump power. At the same time, through the heat exchange working medium temperature gradient adaptive control method, the circulation pump frequency is dynamically adjusted to optimize the fluid heat transfer efficiency. In a low flow rate environment, the system switches to the heat conduction dominant mode, and optimizes the heat storage time based on the formation heat diffusion equation to improve the uniformity of underground heat storage.

[0107] When evaluating the timing strategy, the system uses a multi-step timing return calculation method based on reinforcement learning to evaluate the long-term benefits under different heat exchange control strategies. First, a state-action value function is constructed, and the expected benefit of the current heat exchange strategy is calculated in combination with the discounted return optimization model. During the optimization process, the system uses a time series data smoothing strategy to perform regression analysis on the heat exchange data within different historical time steps to improve the stability of strategy optimization. At the same time, through the dynamic step update mechanism, the weight of the heat exchange strategy is adjusted according to the change of groundwater flow rate to ensure that the system heat exchange control meets the long-term optimal energy distribution.

[0108] During the policy optimization process, the system uses a policy gradient update mechanism to calculate the impact of the heat exchange strategy on the overall benefit. First, based on the Gaussian distribution policy sampling method, different heat exchange parameters are selected in the policy space for exploration, and their heat exchange efficiency is calculated. Subsequently, the gradient approximation method is used to calculate the optimal update direction of the strategy, and combined with the adaptive learning rate adjustment mechanism, the strategy update amplitude is dynamically optimized within each time step. In addition, the system also uses a policy pruning method based on the optimal solution constraint to eliminate non-optimal heat exchange strategies to improve optimization efficiency.

[0109] Adaptive threshold control is based on the real-time formation temperature distribution prediction method to calculate the optimal trigger temperature of the phase change material. When the temperature changes drastically, the system uses the phase change thermodynamic equilibrium model to optimize the filling density of the phase change material, and combines the nonlinear adjustment strategy of the phase change energy storage threshold to dynamically correct the phase change heat storage process under different heat storage requirements. At the same time, the system uses a filling optimization method based on the minimum thermal resistance principle to adjust the spatial distribution structure of the phase change material to improve the overall heat storage efficiency.

[0110] Based on the reinforcement learning feedback optimization mechanism, the system continuously optimizes the heat exchange parameters under different underground heat exchange environments. First, the experience replay mechanism is used to store the heat exchange data under different working conditions, and the adaptive pruning method of the reinforcement learning network structure is combined to optimize the complexity of the learning model and improve the computational efficiency. During the heat exchange control process, the system searches for the optimal strategy in the heat exchange strategy space through a cross-entropy-based strategy exploration method, and combines the dynamic convergence mechanism based on the strategy adversarial network to ensure that the heat exchange control parameters can stably converge to the optimal solution under different environments.

[0111] In the specific implementation, the implementation process of this module is as follows: First, deploy distributed optical fiber temperature measurement arrays and thermal conductivity sensors in the formation to collect groundwater flow velocity and thermal conductivity data in real time. When the flow velocity is greater than 0.1m / s or the thermal conductivity is less than 1.2W / (m·K), the control module is triggered. The DDPG network is built through the TensorFlow framework. The Actor network input includes heat pump power, circulation pump frequency, formation temperature distribution and phase change material state. The output is the heat pump power increment (range ±10%) and the circulation pump frequency adjustment (±5Hz); the Critic network evaluates the state action value and generates the Q value gradient. In the training phase, the experience replay pool is used to store 100,000 sets of historical data, the batch size is 256, the learning rate is set to Actor 1e-4, Critic 3e-4, and the entropy regularization coefficient is initialized to 0.2 and decays with training. In real-time control, the formation temperature field data is collected every 5 minutes, and the trained strategy network is called to generate control instructions, which are sent to the heat pump inverter and circulating pump PLC through the Modbus protocol to synchronously adjust the phase change material trigger threshold (error tolerance ±0.5℃). When the heat pump power exceeds the limit (>120% of the rated value) or the frequency overshoots (>50Hz) in the abnormal handling link, it switches to the PID bottom control mode and triggers the strategy network to fine-tune online. During the system maintenance phase, the experience pool data is updated once a month to adapt to seasonal geological parameter changes through transfer learning.

[0112] Compared with traditional PID and static threshold control methods, this module realizes adaptive optimization of heat exchange strategies through deep reinforcement learning, significantly improving the stability of heat storage in complex geological environments; the dynamic phase change trigger mechanism and multi-objective benefit function design effectively balance energy efficiency and equipment life, avoid thermal stress concentration and over-adjustment, and provide highly robust intelligent control capabilities for cross-seasonal energy storage systems.

[0113] The adaptive optimization control module of the present invention is applied to energy piles for efficient cross-season energy storage, and its hardware environment includes: borehole radar monitors changes in stratum dielectric properties, and combines fiber Bragg grating (FBG) temperature sensors and pore water pressure monitoring arrays to collect groundwater velocity, stratum thermal conductivity and temperature gradient data in real time. All sensors are arranged along the energy pile to achieve three-dimensional stratum thermal field monitoring. In terms of the heat exchange system, an efficient dual-source heat pump unit is configured to form a complete underground heat exchange loop with the heat exchange pipe network and the circulating pump, and a phase change energy storage material module is filled inside the energy pile to optimize the heat storage regulation capability. The data acquisition and control part is composed of an industrial-grade programmable logic controller (PLC), an edge computing unit (ECU) and a deep reinforcement learning control server, wherein the PLC is used for real-time heat exchange control, the ECU is responsible for data parsing and algorithm execution, and the control server runs a deep deterministic policy gradient algorithm (DDPG), and optimizes the calculation in combination with real-time data. The entire system is remotely monitored by a SCADA system for data storage and remote control to ensure long-term stable operation of the energy pile in a complex environment.

[0114] Based on the above hardware operating conditions, experiments were conducted in the underground heat storage test field. Two groups of energy piles with high-efficiency cross-seasonal energy storage of the same specifications were selected, and the adaptive optimization control module (Group A) and the traditional PID control method (Group B) were used to conduct heat exchange performance comparison experiments to verify the advantages of the present invention in practical applications. Group A uses the deep deterministic policy gradient algorithm (DDPG) to optimize the heat exchange strategy, and adaptively adjusts the heat pump power, circulation pump frequency and phase change material heat storage threshold based on real-time geological data, while Group B uses fixed gain PID control to adjust the heat exchange parameters based on the preset target temperature, but cannot adapt to the dynamic changes of geological conditions. During the experiment, the same temperature, flow and pressure sensors were arranged around the two groups of energy piles to ensure the comparability of the data, and the same initial formation temperature was set before the experiment. The experiment selected two typical geological environments (sand areas with high groundwater flow rates and clay areas with low flow rates) to test the adaptability of the two control strategies under different heat exchange conditions. The experimental record table is shown in Table 4:

[0115] Table 4 Adaptive optimization control module application experiment record

[0116]

[0117] The experimental results show that the energy piles of group A using the adaptive optimization control module are significantly superior to those of group B using the traditional PID control in terms of heat exchange response speed, heat storage efficiency and formation temperature stability.

[0118] The global heat balance module is used to adopt a gradient heat release strategy when the formation temperature field reconstruction module detects that the temperature gradient in the heat accumulation area exceeds a preset threshold value, and adjusts the heat pump working fluid flow rate through a nonlinear PID controller for active thermal compensation, while dynamically adjusting the heat charging and releasing rate of the phase change material according to the real-time heat storage and release power ratio. The working principle of the gradient heat release strategy is as follows: based on the temperature gradient data of the formation temperature field reconstruction module, the spatial distribution of the heat accumulation area is calculated by the multi-scale heat partition clustering method, and the heat diffusion path is identified based on the gradient direction field fitting method GDFF to generate a dynamic heat release priority weight matrix; then, based on the dynamic feedback of the formation thermal diffusion coefficient and the heat pump working fluid flow rate, the nonlinear PID heat flow control method is used to adjust the heat exchange flow rate in real time. The nonlinear PID heat flow control method suppresses the temperature fluctuation caused by non-steady-state heat exchange through adaptive disturbance compensation; then, the heat storage and release power ratio adjustment method HSRP-AA is used to calculate the deviation between the current heat storage rate of the phase change material and the formation heat release demand, and the phase change dynamic charging and discharging control method PCD-CDR is called to optimize the heat release rate of the phase change material so that the heat release matches the local temperature gradient change trend; finally, based on the formation heat release rate, working fluid flow fluctuation and heat storage stability, the heat pump flow control strategy is adaptively updated through the multi-objective thermal balance iterative optimization method MOTB-IO.

[0119] Among them, the multi-scale thermal zoning clustering method is used to calculate the spatial distribution of the heat accumulation area. First, based on the adaptive K-means clustering method, the temperature gradient data output by the formation temperature field reconstruction module is clustered and analyzed. The local density peak detection is used to identify the high temperature area, and the multi-scale data fusion strategy is used to optimize the clustering boundary so that the temperature anomaly areas at different scales are consistent. Subsequently, the system constructs the formation heat accumulation feature space based on the heat diffusion time scale mapping to form a three-dimensional distribution model of the heat accumulation area for the dynamic adjustment of the subsequent heat release strategy. The gradient direction field fitting method is used to identify the heat diffusion path to optimize the heat release order. First, based on the finite difference thermal gradient calculation method, the direction field of the formation temperature change is constructed, and the Gaussian curvature constraint fitting algorithm is used to calculate the gradient change trend. Then, the adaptive Laplace smoothing operator is used to optimize the gradient direction to eliminate abnormal gradient points and improve the accuracy of the heat diffusion direction prediction. Finally, through the dynamic heat release priority calculation, the heat diffusion paths of different regions are weighted to form a heat release priority weight matrix. The nonlinear PID heat flow control method is used to optimize the heat exchange flow according to the dynamic feedback of the formation thermal diffusion coefficient and the working fluid flow of the heat pump. First, based on the variable gain PID adjustment strategy, the temperature-flow dual parameter adaptive adjustment model is adopted to dynamically adjust the PID gain parameters with temperature changes. Then, the disturbance suppression compensation control is used to adaptively suppress the temperature fluctuation caused by non-steady-state heat exchange by calculating the flow regulation overshoot in real time. The system further combines the second-order differential advance correction mechanism to smooth the heat flow control instructions and improve the response stability of the heat exchange system. The heat storage and release power ratio adjustment method optimizes the heat storage and release process by calculating the deviation between the heat storage rate of the phase change material and the heat release demand of the formation. The system first calculates the real-time heat exchange power of the phase change material based on the heat storage and release ratio calculation model and matches it with the heat release demand of the formation. Subsequently, the power mismatch least squares optimization method is adopted to introduce the dynamic thermal resistance balance factor in the objective function to optimize the heat charging and releasing rate of the phase change material so that it can dynamically adapt to the changing trend of the formation temperature distribution. The phase change dynamic charging and releasing control method is used to optimize the heat release rate of the phase change material so that it matches the local temperature gradient change. Based on the thermodynamic non-equilibrium phase change model, the system calculates the latent heat release characteristics of the phase change material under different temperature fields, and uses the dynamic optimal heat charging and releasing algorithm to adjust the heat charging and releasing rate of the phase change material. Subsequently, a linear interpolation control strategy is adopted to optimize the heat charging and releasing curves within different temperature ranges to improve the heat transfer efficiency of the phase change material. The multi-objective thermal balance iterative optimization method is used to optimize the heat pump flow control strategy based on the formation heat release rate, working fluid flow fluctuation and heat storage stability. First, a multi-objective optimization function is constructed, and heat transfer stability, energy utilization efficiency and formation heat diffusion uniformity are taken as optimization objectives. A multi-objective optimization algorithm based on Pareto frontier search is used to solve the optimal heat transfer parameter combination.Secondly, the adaptive weight adjustment mechanism is used to adjust the weight ratios of different optimization objectives in real time, so that the heat exchange strategy can dynamically adapt to changes in the external environment and improve the long-term heat storage balance of the system.

[0120] In engineering applications, the implementation of this module needs to follow the following process: First, deploy distributed optical fiber temperature measurement arrays (spacing 0.5m) and heat flux density sensors along the axial and radial directions of the energy pile to collect formation temperature gradient data in real time. When the local temperature gradient is detected to be greater than 3℃ / m, the gradient heat release strategy is triggered. Multi-scale thermal zoning clustering is performed through Python's Scikit-learn library to divide the heat accumulation sub-areas, and the gradient direction field fitting plug-in (GDFFToolbox) of COMSOL Multiphysics is called to generate the heat release priority matrix. A nonlinear PID controller is built in MATLAB / Simulink, and the thermal diffusion coefficient and working fluid flow feedback signals are input to output the heat pump inverter adjustment instructions (frequency range 25-50Hz). The HSRP-AA algorithm is written in LabVIEW for the phase change material charging and releasing heat control link. The temperature-latent heat curve of the phase change material is read through the OPC UA protocol, and the charging and releasing heat rate deviation is calculated in real time and sent to the phase change unit controller. The multi-objective optimization module integrates the NSGA-II algorithm (population size 100, iterations 500), performs global optimization every 6 hours, and generates joint control parameters for the heat pump and phase change material. The abnormal handling mechanism is set to switch to safe mode and trigger historical data rollback when the heat pump power exceeds the limit (>130% of the rated value) or the phase change material temperature changes suddenly (ΔT>5℃ / min). During the system maintenance phase, the temperature drift error of the optical fiber temperature measurement array is calibrated once a month, and the fitness function weight coefficient of NSGA-II is updated to adapt to seasonal changes in heat storage demand.

[0121] The present invention solves the problems of lag in heat accumulation control, low utilization rate of phase change materials and imbalanced energy consumption in traditional systems through the coordination of multi-scale thermal field analysis, dynamic regulation and multi-objective optimization. Compared with the existing technology, it can accurately identify the heat diffusion path and dynamically allocate heat release priority, significantly improve the thermal balance response speed and regulation accuracy, effectively suppress abnormal fluctuations in temperature gradients, and optimize the matching degree of charge and discharge rates of phase change materials, thereby enhancing the stability and energy efficiency of long-term operation of the system, and providing efficient and reliable technical support for cross-season energy storage under complex geological conditions.

[0122] When the global thermal balance module of the present invention is applied to an energy pile for efficient cross-seasonal energy storage, its hardware environment includes: a distributed optical fiber temperature measurement array (DTS), arranged axially and radially along the energy pile to monitor the underground temperature gradient; a heat flux density sensor, installed at the interface between the heat exchange pipe wall and the formation, to monitor the heat exchange flow and heat diffusion in real time; a heat pump system, including a variable frequency heat pump and an adjustable speed circulation pump, to adjust the flow of the heat exchange medium; a phase change energy storage material module, embedded in the energy pile, to optimize the heat storage / release rate through controllable filling; an intelligent control unit, including a nonlinear PID controller and a deep optimization algorithm calculation server, to dynamically adjust the heat exchange strategy and receive feedback data from the remote SCADA monitoring system to ensure the long-term stability of the energy storage system.

[0123] Two groups of energy piles with the same structure for efficient cross-season energy storage were selected, and experiments were carried out using the global heat balance module (Group A) and the traditional fixed threshold heat exchange control (Group B) to verify the heat distribution optimization capability and long-term heat storage stability of the present invention. Group A adopts a gradient heat release strategy, combined with nonlinear PID regulation, heat storage and release power ratio optimization, and phase change charging and releasing heat optimization, to dynamically adjust the heat exchange flow under different formation temperature changes. Group B adopts fixed threshold heat exchange control, that is, the heat exchange flow is adjusted according to the preset temperature threshold, without considering the dynamic changes of formation heat diffusion, and cannot be optimized and adjusted according to the local heat accumulation situation. During the experiment, the same number of temperature sensors and heat flux density sensors were arranged in the axial and radial directions of the two groups of energy piles to ensure the accuracy and comparability of data acquisition. The experimental environment was set to a typical winter heating period, and each group of experiments was repeated five times to reduce the influence of environmental factors. At the beginning of the experiment, the same initial formation temperature was set, and the two groups of systems were operated under the same load conditions. The experimental record table is shown in Table 5:

[0124] Table 5 Global thermal balance module experimental record

[0125]

[0126] The experimental results show that the energy piles of group A using the global thermal balance module are significantly better than those of group B using fixed threshold heat exchange control in terms of heat release uniformity, heat pump regulation response speed, storage and release heat power ratio matching accuracy, and formation temperature stability. Therefore, the global thermal balance module of the present invention can dynamically optimize heat release under different formation conditions, improve the stability and heat exchange efficiency of the heat storage system, and has better thermal management capabilities and stronger long-term operation economy than the traditional fixed threshold control method, providing an intelligent optimization strategy for energy piles with efficient cross-season energy storage.

[0127] Although the specific embodiments of the present invention are described above, it should be understood by those skilled in the art that these specific embodiments are only illustrative, and those skilled in the art may omit, replace, and change the details of the above methods and systems in various ways without departing from the principles and essence of the present invention. For example, merging the above method steps so as to perform substantially the same functions in substantially the same manner to achieve substantially the same results is within the scope of the present invention. Therefore, the scope of the present invention is limited only by the appended claims.

Claims

1. An energy pile with high efficiency and cross-season energy storage; characterized by: include: Dynamic coupling prediction module, used to perform multi-scale prediction of building heat load through LA hybrid time series prediction model; When the building energy consumption monitoring system detects that the start and stop frequency of the indoor temperature control equipment exceeds the threshold, the LA hybrid time series prediction model reconstructs the historical load sample weight through the sliding time window mechanism, and inputs the meteorological forecast data into the three-dimensional heat conduction correction coefficient matrix to output the cross-seasonal heat storage demand prediction value; The formation temperature field reconstruction module is used to collect formation temperature field gradient data and axial and radial temperature gradients in the energy pile using a distributed optical fiber temperature measurement array. When it is detected that the formation temperature gradient change rate exceeds a preset critical value, the module reconstructs the three-dimensional non-steady-state temperature field through a finite element inversion algorithm based on the heat flux density sensor and temperature sensor data in the energy pile, calculates the heat diffusion loss rate, and corrects the heat exchange tube flow rate and phase change material filling density of the underground heat storage body in real time; The geological parameter identification module includes a transient thermal response test unit and a pore water pressure monitoring array, which is used to identify the sand, clay or bedrock geological type through a support vector machine classifier, map heat exchange parameters, and invert the equivalent thermal conductivity and heat storage density correction coefficient through a thermo-seepage coupling model when the borehole radar detects abnormal fluctuations in the dielectric constant of the formation; An adaptive optimization control module is used to adopt a convection-dominated heat exchange strategy when the groundwater velocity output by the geological parameter identification module exceeds a preset threshold value or the thermal conductivity is lower than a preset threshold value, generate an optimal control strategy through a deep deterministic strategy gradient algorithm, adjust the heat pump power and the circulation pump frequency in real time, and adjust the trigger threshold of the phase change material according to the real-time formation temperature distribution; The global heat balance module is used to adopt a gradient heat release strategy when the formation temperature field reconstruction module detects that the temperature gradient in the heat accumulation area exceeds a preset threshold value, and adjusts the heat pump working fluid flow rate through a nonlinear PID controller for active thermal compensation, while dynamically adjusting the heat charging and releasing rate of the phase change material according to the real-time heat storage and release power ratio.

2. The energy pile for efficient cross-season energy storage according to claim 1 is characterized by: The LA hybrid time series prediction model includes a hierarchical time series decomposition layer, a multimodal fusion layer, a dynamic weight reconstruction layer, a three-dimensional heat conduction correction layer, an abnormal frequency judgment layer and a cross-season coupling output layer; The hierarchical time series decomposition layer is used to adopt an improved adaptive noise set empirical mode decomposition algorithm, generate an eigenmode function set by superimposing adaptive white noise iteration, and screen the components with sample entropy lower than a preset threshold of 0.8 and energy proportion exceeding 75%, and output high-frequency noise, medium-frequency daily cycle and low-frequency seasonal components; The multimodal fusion layer is used to calibrate the meteorological and load time series offsets through a dynamic time warping algorithm, stretch the meteorological data time axis when the peak offset of the cross-correlation function of the meteorological and load time series exceeds a preset threshold, and fuse the temperature, humidity, radiation parameters and load components into a three-dimensional space-time tensor through a multi-head attention mechanism; The dynamic weight reconstruction layer is used to mark outliers when the Mahalanobis distance exceeds the preset confidence interval threshold based on the KL divergence distribution of samples in the sliding window through Bayesian optimization, and solve the maximum a posteriori probability weight matrix for non-outlier samples through Gaussian process regression; the three-dimensional heat conduction correction layer is used to construct an unstructured formation heat conduction grid model using the finite element method, and when the meteorological temperature change rate ΔT / Δt>2°C / h, the correction coefficient matrix is ​​iteratively solved through the transient heat conduction control formula; the transient heat conduction control formula is: In formula (1), ρ is the soil density in kg / m 3 , C p is the soil specific heat capacity, in J / (kg·K), T is the formation temperature field, t is the time variable, in seconds, k is the soil thermal conductivity, in W / (m·K), Q solar is the solar radiation heat source term, in W / m 3 , is the heat flux correction matrix; The abnormal frequency decision layer is used to define the equipment start-stop event sequence as an observation value based on the hidden Markov model. The implicit states include normal, transition and abnormal. If the proportion of abnormal state Viterbi paths is greater than the preset threshold, the sample pool is cleared and the 30-day historical data is reloaded; The cross-seasonal coupling output layer is used to input the predicted value of heat flux density into the pile strain formula: ε=aΔT+β+δ thermal / Δt (2) In formula (2), ε is the strain value of the pile, a is the thermal expansion coefficient of the pile material, ΔT is the temperature change, and β is the thermal stress coupling coefficient, the unit is MPa. -1 , δ thermal is the thermal stress value in MPa; if ε>0.2%, the Pareto optimal solution is solved by the sequential quadratic programming algorithm, and the heat storage demand value is output and sent to the heat pump control unit.

3. The energy pile for efficient cross-season energy storage according to claim 1 is characterized by: The working method of the finite element inversion algorithm is as follows: based on the coordinates of the abnormal area marked by clustering, the local grid of the target area is encrypted by an unstructured grid generation algorithm, the heat flux density sensor data and the cluster partition temperature data are loaded, and the heat conduction forward model is used to iteratively invert the formation thermal physical parameters, including the thermal conductivity γ and the heat capacity c, to reconstruct the three-dimensional non-steady-state temperature field, and calculate the heat diffusion loss rate n loss : In formula (3), γ is the dynamic correction value of formation thermal conductivity; q sensor is the heat flux density sensor data; v is the flow rate of the working medium in the heat exchange tube; is the temperature gradient; if n loss When the preset safety threshold is exceeded, the flow rate of the heat exchange tube is dynamically adjusted through the PID controller, and based on the latent heat temperature curve of the phase change material, the optimal filling density under the current temperature field is matched through a linear interpolation algorithm.

4. The energy pile for efficient cross-season energy storage according to claim 1 is characterized by: The transient thermal response test unit uses pulsed heat flow excitation to apply short-term thermal pulses to the formation around the borehole, and collects formation temperature response data through fiber Bragg grating sensing, calculates the transient thermal diffusion coefficient, and identifies the initial thermal physical properties of the formation; the pore water pressure monitoring array extracts the groundwater flow trend through a dynamic water pressure gradient analysis method, and uses adaptive permeability regression APR to calculate soil permeability parameters to obtain a correction factor for the influence of groundwater on thermal diffusion.

5. The energy pile for efficient cross-season energy storage according to claim 1 is characterized by: In the geological parameter identification module, the working method of the support vector machine classifier is as follows: first, the dielectric property data of the formation is projected into a high-dimensional feature space through kernel function mapping, then, the Lagrangian dual optimization method is used to maximize the data category interval, and the soft interval support vector is used to enhance the adaptability of the model to the geological transition zone; during the classification process, the support vector machine classifier uses KKT condition screening to dynamically update the support vector and eliminate redundant data points; then, the support vector weight distribution is optimized through an adaptive weight adjustment mechanism, so that the decision boundary of the classifier between sand, clay and bedrock categories conforms to the trend of changes in the thermal properties of the formation; finally, the support vector machine classifier uses a heat exchange parameter mapping mechanism to call the corresponding thermal diffusivity, heat capacity ratio and permeability from a predefined thermal property parameter library according to the classification results.

6. The energy pile for efficient cross-season energy storage according to claim 1 is characterized by: In the formation thermal seepage coupling inversion process, the thermal seepage coupling model first constructs a non-steady-state thermal seepage coupling equation based on the classified geological types, and adopts the energy equation and Darcy's law to jointly solve the dynamic interaction between underground fluid and thermal field; then, the relationship matrix between underground temperature field, pore water flow rate and heat exchange capacity is nonlinearly discretized through the finite element multi-scale inversion method FEMI; in the inversion process, the thermal seepage coupling model uses Poisson's equation constraint PEC to correct the groundwater flow rate; the corrected output parameters are verified through sensitivity analysis and pushed to the control system to dynamically adjust the heat storage operation strategy.

7. The energy pile for efficient cross-season energy storage according to claim 1 is characterized by: The working method of the adaptive optimization control module is as follows: based on the abnormal signals of groundwater flow velocity and formation thermal conductivity of the geological parameter identification module, constraint conditions are established through the boundary constraint mechanism; then, the convection-dominated heat exchange strategy is used to optimize and adjust the heat pump power and circulation pump frequency; during the control process, a multi-step timing strategy is used to evaluate and calculate the cumulative reward value under different heat exchange strategies, and the control parameters are corrected through the strategy gradient adaptive update method; then, the optimal trigger temperature of the phase change material is calculated based on the real-time formation temperature distribution prediction method, and the phase change trigger threshold dynamic adjustment mechanism is used to dynamically adjust the phase change heat storage process; finally, the heat exchange parameters are iteratively adjusted through deep reinforcement learning feedback optimization to optimize the long-term heat storage stability of the energy pile.

8. The energy pile for efficient cross-season energy storage according to claim 1 is characterized by: The working steps of the deep deterministic policy gradient algorithm include: Step 1: Based on the state-action value function, define the optimal time difference of the target strategy and calculate the strategy value. The formula is: In formula (4), is the optimal time difference target value, R t is the instantaneous benefit of the heat exchange control strategy at time t; δ is the discount factor; Q' is the Q value estimate of the target network; a' is the optimal strategy for the next step; ω1 is the strategy smoothing factor; is the execution action at the previous moment; θ Q’ is the target Q network parameter; S t+1 Indicates that action a is executed at the current time t t Afterwards, the change of the environmental state at time t+1; represents the policy network parameter θ u The calculated value of the gradient; Step 2: Use the policy gradient update function to optimize the policy network and calculate the gradient. The calculation formula is: In formula (5), J(θ u ) is the policy objective function; u(S|θ u ) is the policy network; ω2 is the gradient smoothing factor; is the strategy update change rate; E represents the expected value of all future state S and action a samples; Step 3: Store historical state data through the experience replay mechanism, and introduce the entropy regularization term through the adaptive entropy regularization strategy for optimization to improve the exploration ability of the heat exchange strategy. The formula is: In formula (6), τ entropy is the entropy regularization loss; ω3 is the regularization coefficient; π(a i |S) is the strategy probability distribution; N is the number of samples in the strategy space; Step 4: Combine the adaptive adjustment of the circulation pump frequency with the nonlinear mapping of the heat pump power to calculate the optimal heat exchange control parameters As the final implementation strategy.

9. The energy pile for efficient cross-season energy storage according to claim 1 is characterized by: The working principle of the gradient heat release strategy is as follows: based on the temperature gradient data of the formation temperature field reconstruction module, the spatial distribution of the heat accumulation area is calculated by the multi-scale heat partition clustering method, and the heat diffusion path is identified based on the gradient direction field fitting method GDFF to generate a dynamic heat release priority weight matrix; then, based on the dynamic feedback of the formation thermal diffusion coefficient and the heat pump working fluid flow rate, the nonlinear PID heat flow control method is used to adjust the heat exchange flow rate in real time. The nonlinear PID heat flow control method suppresses the temperature fluctuation caused by non-steady-state heat exchange through adaptive disturbance compensation; Subsequently, the heat storage-release power ratio adjustment method HSRP-AA is used to calculate the deviation between the current heat storage rate of the phase change material and the heat release demand of the formation, and the phase change dynamic charging and discharging control method PCD-CDR is called to optimize the heat release rate of the phase change material so that the heat release matches the changing trend of the local temperature gradient; finally, based on the formation heat release rate, working fluid flow fluctuation and heat storage stability, the heat pump flow control strategy is adaptively updated through the multi-objective thermal balance iterative optimization method MOTB-IO.

Images