Geothermal and aquifer energy storage coupling system operation optimization method

Abstract

The present application relates to a kind of geothermal and aquifer energy storage coupling system operation optimization method, comprising: with system total energy consumption minimization of whole cycle operation as objective function, construct system mathematical model;The constraint space of mathematical model is defined optimization feasible region by system energy conservation and hydraulic balance equation;Dynamic optimization solving is carried out to mathematical model using decentralized difference evolution algorithm, obtains the equipment power setting value and flow distribution scheme of system level optimization;According to the preset control strategy of rolling time domain cycle, carry out low temperature cold source call of system, in combination with the optimization solving result of objective function, control energy storage supply mode and equipment operating parameter, and based on the control result, mathematical model is carried out online dynamic optimization.The present application solves the problems of high regulation delay, heavy calculation burden and poor scalability of traditional centralized control by decoupling complex global optimization problem into local optimization and global collaboration based on subsystem.

Description

Technical Field

[0001] This invention relates to the field of HVAC operation optimization and control technology, and in particular to an operation optimization method for a geothermal and aquifer energy storage coupled system. Background Technology

[0002] Ground source heat pump systems convert low-grade underground heat energy into directly usable high-grade energy by consuming a small amount of high-grade energy, with a coefficient of performance (COP) typically reaching 3.5–5.5. Aquifer energy storage systems (ATES), as efficient auxiliary air conditioning systems, utilize groundwater as a heat transfer medium, enabling seasonal energy storage cycles of "cool in winter and used in summer" and "hot in summer and used in winter." Coupled systems that combine geothermal energy utilization with aquifer energy storage technology not only improve the stability of single geothermal energy utilization but also effectively alleviate the temporal and spatial imbalance of building heating and cooling loads. However, the complex hydraulic-thermal coupling characteristics of Geothermal-Aquifer Thermal Energy Storage Coupling Systems (GAESCS) pose significant challenges to their efficient operation. Because the system involves multiple heterogeneous intelligent devices such as main refrigeration equipment (e.g., ground source heat pump units, chillers), ground source side circulation pumps, aquifer pumping / priming pumps, and terminal air conditioning equipment, the coupled regulation of the entire system exhibits significant hysteresis and strong nonlinearity. Online optimization control is widely recognized as a key means to improve the energy efficiency of such systems; however, existing control schemes generally adopt centralized control methods, which reveal the following shortcomings in practical engineering applications: The system logic is overly complex and computationally inefficient: As the number of connected devices increases, the computational load of the centralized controller model grows exponentially, resulting in long single-time optimization calculation times and making it difficult to meet the real-time optimization and rapid response requirements of systems in complex dynamic environments. High-dimensional parameter optimization is difficult: When performing multi-parameter coupled optimization, the search space is extremely large due to the large number of variables to be optimized, making centralized algorithms prone to getting trapped in local optima, preventing the system from operating in the true global efficiency range. The system lacks scalability and flexibility: When the system needs to add or replace equipment, such as adding a new aquifer pumping well or heat pump unit, the existing centralized control system optimization program often requires large-scale remodeling and program modification, resulting in a huge workload and high subsequent maintenance costs.

[0003] In summary, existing control schemes are no longer sufficient to meet the comprehensive requirements of geothermal and aquifer energy storage coupling systems in terms of complexity, real-time performance, and efficiency. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an operation optimization method for a geothermal and aquifer energy storage coupled system, in order to solve the technical problems of heavy computational burden, response lag, and formation thermal imbalance caused by traditional centralized control when dealing with large-scale heterogeneous equipment groups.

[0005] The technical solution adopted in this invention is as follows: An operation optimization method for a geothermal and aquifer energy storage coupled system includes: A mathematical model of the system is constructed with the goal of minimizing the total energy consumption of the system throughout its entire life cycle. The total energy consumption is coupled from the power consumption of the main cooling equipment, the ground source circulating pump, and the aquifer energy storage pump. The constraint space of the mathematical model is defined by the system energy conservation and hydraulic balance equations to define the feasible region for optimization. A distributed differential evolution algorithm is used to dynamically optimize the mathematical model to obtain system-level optimized equipment power setpoints and flow allocation schemes. In the population initialization stage, the distributed differential evolution algorithm randomly generates an initial population of decision variables within the effective physical domain of each control variable. The decision variables of each individual in the population include the host power setpoint, the power or flow control of the ground source side circulating pump, the power or flow control of the aquifer pump, and the operating temperature difference. The preset control strategy is executed in a rolling time domain cycle to call the low temperature cold source of the system. Combined with the optimization solution of the objective function, the energy storage supply mode and equipment operating parameters are controlled, and the mathematical model is dynamically optimized online based on the control results. The discretization formula for the objective function is as follows:

[0006] In the formula, This represents the total power consumption of the system. , , The first t Step time i Instantaneous power of the main cooling equipment, ground source side circulation pump, and aquifer energy storage pump; N Total number of devices; T For the total system operating cycle, t This is the index for the number of off-steps within the runtime cycle.

[0007] The preferred technical solution is: The first t Step time i Instantaneous power of the main unit's cooling equipment The calculation is based on an energy consumption prediction model that combines physical characteristics with multivariate data regression. Its analytical expression is:

[0008] In the formula, For the first Real-time power consumption of the main unit's cooling equipment; This refers to the inlet temperature of the cooling water. This is the temperature for chilled water supply; The heat load borne by the main refrigeration equipment; to The energy consumption characteristic coefficient is obtained by calibration based on the historical data of the main unit refrigeration equipment under all operating conditions.

[0009] Instantaneous power , The calculation is based on a quadratic polynomial energy consumption model that is deeply coupled with the frequency conversion and pipeline resistance characteristics. Its characteristic formula is:

[0010] in, For the first Real-time power consumption of the pump; For the first Real-time mass flow rate of the pump; , , The pump efficiency coefficient is obtained by fitting different operating conditions.

[0011] The constraint space includes inter-component heat transfer balance constraints, which include:

[0012] In the formula, To meet the total heat exchange requirements on the cooling side; To share the heat exchange burden on the ground source side; To share the heat exchange burden on the aquifer side; The total cooling or heating load on the end building side; The total power consumption of all main unit cooling equipment; This represents the total power consumption of the water pumps in the distribution system.

[0013] The constraint space includes real-time heat transfer constraints on the ground source side, which include:

[0014] In the formula, Real-time heat exchange on the ground source side; This refers to the mass flow rate of water drawn from the ground source side. For fluid density; The specific heat capacity at constant pressure of the fluid; and These are the return water temperature and supply water temperature on the ground source side, respectively.

[0015] The constraint space includes the flow constraints of the ground-source water pump and the flow constraints of the aquifer intake pump, and their calculation formulas based on the law of conservation of energy are as follows: , , in, , These are the target flow rates for the ground-source water pumps and the target flow rates for the aquifer intake pumps, respectively. , These correspond to the return water temperature and supply water temperature on the ground source side, respectively. , These correspond to the return water temperature and supply water temperature on the aquifer side, respectively. The heat load borne by the main refrigeration equipment. The specific heat capacity at constant pressure of the fluid. This refers to the power consumption of the main cooling equipment.

[0016] After population initialization, the distributed differential evolution algorithm introduces a representative individual selection mechanism: for the entire population, the K-means clustering algorithm is used to cluster the population space according to the spatial Euclidean distance of the individuals; in each cluster, the individual closest to the cluster centroid is found as the representative individual of the subsystem, and only the operating parameters of the representative individual are used as a lightweight carrier for interaction between subsystems, so as to reduce the communication load in the distributed control network.

[0017] The control strategy includes: The system's low-temperature cold source is determined based on temperature difference priority: the real-time difference between the pumping temperature of the ground source heat exchange branch and the pumping temperature of the aquifer energy storage branch is used. With threshold In comparison, if Then, combining seasonal or pattern signals, the controller determines the mode switching conditions. If the switching conditions are met, the target branch is different from the current dominant branch, and the criteria are continuously satisfied for a preset duration, the controller is allowed to switch between the ground source dominant mode, the aquifer dominant mode, or the mixed regulation mode; if... Maintain the current mode or enter the hybrid adjustment mode; Cross-seasonal energy storage coupling regulation: When the switching conditions are met on the source side or aquifer side and the aquifer energy storage branch meets the injection switching and reinjection constraints, the aquifer side is switched to the dominant energy supply or energy storage branch, and the reinjection branch valve is opened simultaneously to inject circulating water carrying heat into the reinjection well for cross-seasonal heat storage; as the accumulated operation causes changes in the thermal state of the aquifer, when its pumping temperature is inferior to that of the source side, the pumping and injection status touches the constraint boundary, the reinjection capacity is insufficient or the switching conditions are no longer met, the control system maintains the current mode, switches back to the source dominant mode or enters the hybrid regulation mode, and simultaneously adjusts the valve opening, pump frequency and main unit operation command of the corresponding branch.

[0018] Adaptive frequency conversion adjustment includes: Real-time monitoring of terminal load changes, flow upper and lower limits, and equipment alarm status; real-time correction of the start / stop status of each main unit refrigeration unit and the frequency conversion frequency of each pump by solving the objective function; when abnormal temperature difference, flow exceeding limit, or equipment fault signal occurs, protective limiting or degraded operation control is prioritized.

[0019] The switching conditions include: The candidate branch has a temperature advantage in the current season. Specifically, when cooling in summer, the pumping temperature of the candidate branch is lower than that of another branch. The corresponding switching strategy is to prioritize the side with the lower pumping temperature as the low-temperature cold source. When heating in winter, the pumping temperature of the candidate branch is higher than that of another branch. The corresponding switching strategy is to prioritize the side with the higher pumping temperature. The flow rate, pressure, heat exchange rate, and supply and return water temperature difference of the candidate branch are all within the safety boundary defined by the constrained space. No alarm signals were detected in the relevant pumps, valves, main refrigeration equipment, and measuring points; When the candidate branch is an aquifer energy storage branch, it also meets the constraints of pumping-injection balance, recharge permission, recharge capacity and groundwater level fluctuation.

[0020] The technical solution of the present invention can achieve at least some of the following beneficial effects: This invention adapts to multivariate dynamic working conditions, has strong global optimization capabilities, and can effectively reduce computational and communication complexity, providing technical support for the sustainable synergy between building energy supply and underground energy storage resources.

[0021] This invention achieves approximately 5% energy savings through multi-source collaborative optimization of geothermal and aquifer energy storage, under the same seasonal conditions, similar load range, and the same statistical period. The mathematical model for system construction, through multi-dimensional physical constraints and dynamic relay logic, helps alleviate geothermal imbalance and ensures the long-term safety and sustainable development of underground resources. The distributed architecture decouples the high-dimensional optimization problem, reducing computational load and communication latency, making it more suitable for real-time control under dynamic operating conditions. It also enhances the system's robustness and scalability, reducing the impact of single-point failures on overall operation. Attached Figure Description

[0022] Figure 1 This is a flowchart of the method according to an embodiment of the present invention.

[0023] Figure 2 This is a diagram of the overall physical architecture of the system to be optimized in an embodiment of the present invention.

[0024] Figure 3 for Figure 1 The solid lines in the architecture represent the water supply, return, and energy pathways.

[0025] Figure 4 for Figure 1 Diagram showing the relationship between measurement point layout, communication aggregation, and variable mapping in the architecture.

[0026] Figure 5 This is a block diagram of the two-stage data interaction and collaborative optimization logic based on the distributed differential evolution algorithm in an embodiment of the present invention.

[0027] Figure 6 This is a schematic diagram of the adaptive temperature difference sensing scheduling and switching logic based on the distributed differential evolution algorithm in an embodiment of the present invention.

[0028] Figure 7 This is a block diagram of the mode switching and protection logic of the control strategy in an embodiment of the present invention.

[0029] Figure 8 This is a comparison chart of the running results of Experimental Example 1 and the Comparison Example in the embodiments of the present invention.

[0030] Figure 9 This is a comparison chart of the running results of Experimental Example 2 and the Comparative Example in the embodiments of the present invention. Detailed Implementation

[0031] The specific embodiments of the present invention are described below with reference to the accompanying drawings.

[0032] See Figure 1 This embodiment provides an operation optimization method for a geothermal and aquifer energy storage coupled system, including: S1. A mathematical model of the system is constructed with the goal of minimizing the total energy consumption of the system throughout its entire life cycle. The total energy consumption is coupled from the power consumption of the main cooling equipment, the ground source circulating pump, and the aquifer energy storage pump. The constraint space of the mathematical model is defined by the system energy conservation and hydraulic balance equations to define the feasible region for optimization.

[0033] This embodiment deeply analyzes the energy conversion characteristics of heterogeneous equipment in the system, establishes a digital mapping covering generating units, transmission and distribution networks and underground energy storage media, and thus constructs a decoupled mathematical model that integrates physical mechanisms.

[0034] As a specific example, the overall physical architecture and control network of the system in this embodiment are described in [reference needed]. Figure 2 As shown. For details on the solid line water supply and return routes and energy pathways, please refer to [link / reference needed]. Figure 3 For details regarding the layout of measurement points, communication aggregation, and variable mapping relationships, please refer to [link / reference needed]. Figure 4 .

[0035] like Figure 4As shown, temperature sensors, flow meters, and pressure sensors are installed in the main inlet and outlet water pipes of the main unit, the ground source side circulation branch, the aquifer pumping and injection well group, and key heat exchange nodes, forming representative measuring points such as T1 to T4, F1 to F4, and P1 to P2, where T1 corresponds to the main unit's water supply temperature. F1 corresponds to the main unit's water supply flow rate. P1 corresponds to the secondary pressure difference in the heat exchange. Representative mapping relationships are established. At the control layer, each monitoring point is mapped to a unified data label and device address, and data aggregation is completed through the ground source side data acquisition terminal A, aquifer data acquisition terminal B, edge computing gateway, and distributed control nodes to achieve a one-to-one correspondence between "measuring point - device - model variable".

[0036] Based on the above physical architecture, the construction of the mathematical model specifically includes: S11. Constructing a sub-model of the main refrigeration equipment's energy consumption: By collecting historical operating data of the unit under different operating conditions, the mapping relationship between the unit's energy consumption and the cooling water inlet temperature, chilled water supply temperature, and the real-time heat load borne by the main refrigeration equipment is analyzed using a multivariate nonlinear regression method. The specific analytical expression is as follows:

[0037] In the formula, For the first Real-time power consumption of the main unit's cooling equipment; The inlet temperature of the cooling water reflects the thermal state of the return water from the ambient or source side. The temperature of the chilled water supply is directly related to the quality of the supply to the end load; The real-time heat load borne by the main refrigeration equipment; to The energy consumption characteristic coefficient is obtained by calibration based on the historical data of the main unit refrigeration equipment under all operating conditions.

[0038] This sub-model can characterize the nonlinear characteristics of the main unit refrigeration equipment under different load rates and temperature difference conditions. In particular, it can reflect the energy efficiency change law of the main unit when it is running under partial load, thus providing a model basis for determining the optimal operating point of global energy efficiency at the system level.

[0039] During the total system operation cycle, the distance from the walk t Time i Instantaneous power of the main unit's cooling equipment It can be calculated based on the above energy consumption prediction model.

[0040] S12. Establish the pump and aquifer control sub-model: Based on the fluid machinery similarity law and combined with measured data of pipeline resistance in the field, a quadratic polynomial expression for the relationship between pump power consumption and flow rate is established:

[0041] in, For the first Real-time power consumption of the pump; For the first The real-time mass flow rate of the pump is mapped in real time through the frequency output of the frequency converter and the closed-loop feedback of the flow sensor. , , The pump efficiency coefficient, obtained by fitting different operating conditions, reflects the relationship between pump head and power consumption under a specific pipeline topology.

[0042] This model quantifies the power consumption variation of the ground source circulating pump during variable frequency regulation, i.e., it quantifies the nonlinear coupling relationship between flow regulation and power consumption. This ensures that the control system, while meeting the heat exchange temperature difference and hydraulic balance of each branch, can minimize the excess head loss of the distribution system through coordinated optimization of the variable frequency pump unit speed, achieving deep energy savings in the power auxiliary links. It also characterizes the head-power consumption characteristics of the aquifer pump at different groundwater depths and correlates them with… Figure 1 The physical architecture shown corresponds one-to-one with the ground source-side circulation branch and the aquifer pumping branch. The system's total operating cycle includes a distance-to-ground circulation branch. t Time i Instantaneous power of the ground source side circulation pump and the aquifer energy storage pump , It can be calculated based on the above models respectively.

[0043] S13. Construct the objective function based on the sub-model built in S11 and S12. This objective function emphasizes the cumulative integral of the power of all active energy-consuming components within the system's operating period T. The discretized calculation formula is as follows:

[0044] In the formula, This represents the total power consumption of the system. , , The first t Step time i Instantaneous power of the main cooling equipment, ground source side circulation pump, and aquifer energy storage pump; N Total number of devices; T For the total system operating cycle, tThis is the index for the off-line count within the operating cycle. The objective function is designed to establish an energy balance between the dynamic load on the building side and the energy storage status on the underground side by dynamically decoupling and optimizing the unit power and pumping power. Through intelligent scheduling, it balances the long-term thermal energy benefits of the cross-seasonal energy storage side with the current power consumption, and finds the operating combination with the lowest total energy consumption, thereby maximizing the energy efficiency of the GAESCS system throughout its entire life cycle.

[0045] S14. Define a multi-dimensional constraint space, which simultaneously constructs control boundaries covering heat exchange balance between components, heat exchange limit on the ground source side, dynamic safety range of equipment flow, and geological constraints of groundwater level fluctuation, to ensure that the optimization trajectory is always within the physical safety envelope.

[0046] In a specific manner, the constraint space includes: (1) Heat transfer balance constraints between components, including:

[0047] In the formula, To meet the total heat exchange requirements on the cooling side; To share the heat exchange burden on the ground source side; To share the heat exchange burden on the aquifer side; The total cooling or heating load on the end building side; The total power consumption of all main unit cooling equipment; This represents the total power consumption of the water pumps in the distribution system.

[0048] (2) To address the unsteady heat transfer characteristics of underground media, a real-time heat transfer constraint is established on the ground source side, which includes:

[0049] In the formula, Real-time heat exchange on the ground source side; This refers to the mass flow rate of water drawn from the ground source side. For fluid density; The specific heat capacity at constant pressure of the fluid; and These are the return water temperature and supply water temperature on the ground source side, respectively.

[0050] This constraint is based on the law of conservation of energy, ensuring that the heat exchange capacity can cover the total heat load demand at the end in real time, thereby ensuring that the heat dissipation on the buried pipe side is within the acceptable range of the ground heat capacity and preventing ground temperature drift caused by long-term operation.

[0051] (3) The flow constraints of the ground source pump and the aquifer intake pump are calculated based on the law of conservation of energy, respectively: , , in, , These are the target flow rates for the ground-source water pumps and the target flow rates for the aquifer intake pumps, respectively. , These correspond to the return water temperature and supply water temperature on the ground source side, respectively. , These correspond to the return water temperature and supply water temperature on the aquifer side, respectively. The heat load borne by the main refrigeration equipment. The specific heat capacity at constant pressure of the fluid. This refers to the power consumption of the main cooling equipment.

[0052] This constraint mandates that the flow rates at the source and aquifer must always be maintained between 40% and 100% of the rated design range to prevent hydraulic imbalance and abnormal fluctuations in the groundwater level.

[0053] (4) The dynamic stability limit of the source flow and the aquifer flow is defined as follows: the ratio of the measured value of each flow to its rated design value is maintained within the preset safe operating range, preferably 0.4 to 1.0.

[0054] This physical constraint can not only prevent formation seepage damage and drastic fluctuations in groundwater level caused by excessive flow velocity, but also reduce the risk of underground thermal short circuits by controlling the heat exchange temperature difference of each well group, prevent system hydraulic imbalance or equipment idling damage, and ensure the sustainable use of formation resources.

[0055] After this step is completed, the system obtains the objective function and parameter boundary that satisfy the constraints, thus providing initial conditions and decision inputs for the subsequent optimization solution of S2 and the online execution of the multi-source coupled control strategy of S3.

[0056] S2. The distributed differential evolution algorithm (DDE) is used to dynamically optimize the mathematical model to obtain system-level optimized equipment power setpoints and flow allocation schemes. During the population initialization phase, the DDE randomly generates an initial population of decision variables within the effective physical domain of each control variable. The decision variables for each individual in the population include the main unit power setpoint, the power or flow control of the ground source circulating pump, the power or flow control of the aquifer pump, and the operating temperature difference. Each candidate solution contains a decision vector composed of all decision variables. Specifically, the operating temperature difference refers to the supply and return water temperature difference of the heat exchange branch participating in the optimization, including at least the supply and return water temperature difference on the ground source side and the supply and return water temperature difference on the aquifer side. Each candidate solution contains a decision vector composed of all decision variables.

[0057] The process of finding the optimal solution using the DDE algorithm includes population initialization, the first data interaction (local fitness), the second data interaction (global collaborative verification), and a loop of offspring generation and greedy selection until the convergence condition is met, and the optimal control scheme is output.

[0058] Specifically, the initial value for each individual can be calculated using the following formula:

[0059] in: , These are the upper and lower bounds of the decision variable, respectively. From uniform distribution Random variables generated in the process, For population size, i This is the index of the individual numbers in the population.

[0060] As a preferred approach, the distributed differential evolution algorithm introduces a representative individual selection mechanism after population initialization: K-means clustering is used to cluster the entire population according to the spatial Euclidean distance of individuals; the individual closest to the cluster centroid in each cluster is selected as the representative individual of the subsystem, and only the operating parameters of the representative individual are used as a lightweight carrier for interaction between subsystems. This mechanism allows each distributed controller to exchange only key operating characteristic data and energy efficiency data, significantly reducing redundant communication load in the distributed sensing and control network and ensuring the real-time performance and high response frequency of the control logic in a massive data environment. The subsystems represent functional units with independent local models, local constraints, and communication interfaces, including the host unit, the ground source-side circulation unit, the aquifer well group unit, and the distribution pump and valve unit.

[0061] Specifically, the optimization process of the algorithm in this embodiment achieves global collaboration among the distributed nodes through an innovative two-stage data interaction mechanism.

[0062] See Figure 5 In the first phase of interaction, each subsystem node exchanges energy consumption components, local constraint states, and candidate control variables with its physical topology neighbors. The subsystem then calculates its local fitness, including a neighborhood penalty factor, based on its local physical model. This is used to quantify the initial impact of local decisions on the energy consumption components of neighboring nodes. Specifically, it quantifies the cross-influence of neighboring node control actions on the local source-side water pump power consumption and the total system energy consumption.

[0063] In the second phase of interaction, the subsystems further exchange fitness assessment information, neighborhood feedback quantities, and global contribution assessment results required for the second interaction, and receive auxiliary feedback operators from neighboring nodes. This phase focuses on evaluating the global contribution of local control logic actions to the aquifer energy storage system. Through this secondary flow of information, the distributed control nodes can perceive global energy consumption trends, and the results from each node are aggregated into a total fitness score. For the follow-up This provides a basis for the greedy selection mechanism. This interaction mechanism ensures that the various decentralized control nodes can reach a global consensus on energy efficiency, enabling the algorithm to continuously approach the global solution with the lowest total system energy consumption even in a decentralized network architecture through differential mutation, crossover operators, and greedy selection strategies.

[0064] Based on the feedback information obtained from the two-stage interaction, each node uses the differential mutation operator and the crossover operator to generate mutated offspring. After completing the second interaction, the algorithm first calculates the global fitness. This is used to quantify the overall effectiveness of each candidate solution in reducing the total energy consumption of the system, and then according to... Figure 5 The relationship shown will Input a greedy selection module to compare parent and child generations. The greedy selection strategy optimizes and updates the parent and child generations, driving the system state to approach the Nash equilibrium point with lower global energy consumption until the convergence condition is met, and finally obtains and outputs the global optimal operating condition control of the device.

[0065] S3. Execute the preset control strategy in a rolling time-domain cycle to call the system's low-temperature cold source. Combined with the optimization solution of the objective function, control the energy storage supply mode and equipment operating parameters, and perform online dynamic optimization of the mathematical model based on the control results.

[0066] The control strategy of this embodiment achieves coordinated control of geothermal heat exchangers and aquifer energy storage well groups through dynamic scheduling logic based on the pumping temperature on the heat source side. This logic includes temperature difference judgment and valve adjustment to ensure cross-seasonal energy storage efficiency. By utilizing the dynamic complementary characteristics of geothermal temperature and aquifer temperature difference, a heat scheduling and seasonal relay scheme is formulated to achieve efficient matching of cold and heat loads.

[0067] See Figure 6 As a preferred embodiment, the control strategy includes: Firstly, the system's low-temperature cold source is selected based on temperature difference priority: the real-time difference between the pumping temperature of the ground source heat exchange branch and the pumping temperature of the aquifer energy storage branch is used. With threshold In comparison, if Then, combined with seasonal or mode signal input mode switching conditions, if the switching conditions are met and the target branch is different from the current dominant branch, and the criterion is continuously met for a preset duration, the controller is allowed to switch between the ground source dominant mode, the aquifer dominant mode, or the mixed regulation mode; if The current mode can be maintained or a hybrid adjustment mode can be entered. The corresponding threshold judgment and mode switching relationship can be combined. Figure 7 To further understand.

[0068] Specifically, the ground-source dominant mode means that the heat exchange task is mainly undertaken by the ground source side; the aquifer dominant mode means that the heat exchange task is mainly undertaken by the aquifer side; and the mixed regulation mode means that the heat exchange task is jointly undertaken by both sides according to an optimized allocation ratio, which is used to reduce the disturbance caused by frequent switching.

[0069] Specifically, when the temperature advantage of the target branch after switching disappears, the flow or pressure exceeds the limit, an equipment alarm occurs, or the aquifer side no longer meets the pumping / reinjection constraints, the controller performs a rollback judgment, maintaining the current mode, switching back to the original dominant mode, or entering a mixed regulation state.

[0070] Specifically, the switching conditions include: (1) The candidate branch has a temperature advantage in the current season. Specifically, when cooling in summer, the pumping temperature of the candidate branch is lower than that of another branch. The corresponding switching strategy is to prioritize the side with the lower pumping temperature as the low-temperature cold source. When heating in winter, the pumping temperature of the candidate branch is higher than that of another branch. The corresponding switching strategy is to prioritize the side with the higher pumping temperature. (2) The flow rate, pressure, heat exchange, and supply and return water temperature difference of the candidate branch are all within the safe boundary of the constrained space; (3) No alarm signals were detected in the relevant pumps, valves, main refrigeration equipment, and measuring points; (4) When the candidate branch is an aquifer energy storage branch, it should also meet the constraints of pumping-injection balance, recharge permission, recharge capacity and groundwater level fluctuation.

[0071] Secondly, cross-seasonal energy storage coupling regulation: When using aquifer energy storage resources for energy supply, that is, when the source side or aquifer side meets the mode switching conditions and the aquifer branch meets the injection switching and reinjection constraints, the aquifer side is switched to the dominant energy supply or energy storage branch, and the reinjection branch valve is opened simultaneously to inject circulating water carrying heat into the reinjection well for cross-seasonal heat storage.

[0072] Specifically, when the source side or aquifer side meets the above-mentioned mode switching conditions, and the aquifer branch simultaneously meets the conditions for the linkage opening of pumping wells / recharge wells, recharge flow constraints, recharge temperature constraints, and groundwater level safety constraints, the system will switch the aquifer side to the dominant energy supply / storage branch and simultaneously open the relevant valves and pumping / recharge channels for energy storage. At this time, the controller allocates the heat exchange ratio undertaken by the aquifer side according to the optimization results and coordinates the actions of the main unit, water pumps, and valves to avoid excessive instantaneous impact on the pumping / recharge branch. As the accumulated operation leads to changes in the thermal state of the aquifer, when the pumping temperature is inferior to that of the ground source side, the pumping and irrigation state reaches the constraint boundary, the reinjection capacity is insufficient, or the mode switching conditions are no longer met, the control system maintains the current mode, switches back to the ground source dominant mode, or enters the mixed regulation mode, and simultaneously adjusts the valve opening, pump frequency, and main unit operation commands of the corresponding branch.

[0073] Thirdly, adaptive frequency conversion adjustment: The system monitors terminal load changes, flow upper and lower limits, and equipment alarm status in real time. It corrects the start / stop status of each main chiller and the frequency conversion frequency of each pump in real time by solving an objective function. When abnormal temperature differences, flow exceeding limits, or equipment fault signals occur, protective limiting or degraded operation control is prioritized. Specifically, abnormal temperature differences refer to the deviation of the supply and return water temperature difference on the ground source side or aquifer side from the preset normal operating range, preferably determined by comparing the measured supply and return water temperature difference with a preset threshold range. Protective limiting includes at least limiting the main chiller power limit, limiting the pump frequency limit, or limiting the valve opening change rate. Degraded operation control includes at least switching to the current safe dominant branch, reducing the aquifer side's load ratio, or suspending the operation of the faulty branch while maintaining basic power supply.

[0074] The above control strategy logic can be deployed on the edge controller and executed in a rolling time-domain loop, followed by online optimization and dynamic adjustment.

[0075] The optimal power setpoint, branch flow target, and regulating valve setpoint obtained through optimization are sent to the underlying programmable logic controller (PLC). The actuator adjusts the circulating pump frequency and actuator opening in real time according to the instructions. The system can record the energy-saving effect and verify the operation results through the monitoring interface and historical operation logs.

[0076] The effectiveness of the method in this embodiment will be illustrated and verified by specific experiments below.

[0077] The system under control in this experiment is a geothermal and aquifer energy storage coupled cooling system applied to office and supporting buildings in a certain park. The system includes two parallel-operating heat pump units, two sets of ground source side circulation pumps, two sets of aquifer pumping pumps, corresponding branch regulating valves, and a network of temperature, flow, and pressure sensors.

[0078] Experimental and comparative cases were set up, with the same equipment configuration, sensor layout and external load boundary conditions. The statistical objects were selected from the cooling operation conditions of the same summer working day, with the average terminal load rate maintained at 65% to 70%, and the statistical period was 24 consecutive hours.

[0079] In the experimental case, the model and control logic were established according to the steps outlined in this embodiment. A distributed differential evolutionary optimization strategy was used to collaboratively optimize the main unit power setpoint, the flow targets of each branch, and the valve opening. The total power consumption of the system, the main unit power consumption, the total pump power consumption, and the temperature fluctuation range on the aquifer pumping side were recorded using a unified sampling method. In the comparative case, the equipment configuration and operating conditions remained unchanged; only the control method was replaced with a conventional empirical control strategy, and the data acquisition method was consistent with that of the experimental case for statistical analysis. To ensure the comparability of the results, the monitoring data sampling period was uniformly set to 5 minutes, and energy consumption statistics were generated by continuous 24-hour integration or accumulation.

[0080] Taking Experiment Example 1 at a certain representative sampling time Taking a single operating host and its corresponding branch as an example for specific explanation. At this sampling time, the real-time cooling load undertaken by the cooling equipment of a single host is taken as... kW, cooling water inlet temperature is taken ℃, chilled water supply temperature is taken ℃; corresponding unit energy consumption model calibration coefficient is taken as , , , , , Substituting into the energy consumption sub-model of the main unit's refrigeration equipment, we get:

[0081] The result of the above formula is in kW. Further, the temperature difference between the supply and return water on the ground source side is taken. ℃, specific heat capacity at constant pressure The calculated load corresponding to real-time heat exchange on the ground source side is Based on the flow constraint calculation formula, the source-side flow target can be obtained as follows:

[0082] If the power consumption model coefficients of the ground source side circulating pump are taken , , Then the power consumption of the ground source side circulating pump is:

[0083] Similarly, if the heat exchange is shared by the aquifer side at that sampling time... kW, temperature difference between supply and return water on the aquifer side At ℃, the target flow rate on the aquifer side is:

[0084] If the coefficients of the aquifer control submodel are taken , , The power consumption of the aquifer pump is:

[0085] Therefore, at this representative sampling moment in Experiment Example 1, the total instantaneous power of the system can be obtained as follows:

[0086] Since the sampling period is 5 minutes, that is h, therefore the power consumption at this sampling time is:

[0087] At the same sampling time, the comparative examples were calculated using the same formula and statistical methods. If the operating parameters under conventional empirical control are taken as... ℃ ℃ ℃ kW ℃, then respectively can be obtained kW kW kW, therefore the total instantaneous power of the system at this sampling time is approximately kW, corresponding to a power consumption of approximately 5 minutes 5.31kWh. As can be seen, under the same load boundary, Experimental Example 1 already showed lower host power consumption and pump power consumption at a single sampling time.

[0088] For the statistical results over a continuous 24 hours, the above model was substituted point by point for all 288 sampling times and all operating main units and pump sets, and the results were accumulated using the following formula:

[0089] Accordingly, the power consumption of the main unit and the power consumption of the pump can also be calculated separately as follows: and Statistical analysis was conducted. Following this statistical method, experimental case one and the control case were analyzed separately. After accumulating, the daily cumulative power consumption result can be obtained.

[0090] It should be noted that the above representative sampling times The substitution calculation is used to demonstrate the specific application of the host refrigeration equipment energy consumption sub-model, water pump power consumption model, and flow constraint formula.

[0091] Figure 8 The bar chart shows the cumulative results of all 288 sampling times over 24 hours under the same statistical caliber. The left side of the chart compares power consumption, while the right side compares temperature fluctuations on the aquifer pumping side. Figure 8 As shown, under the same seasonal conditions, similar load range, and the same statistical period, compared with the comparative example, the total power consumption of the system in Experimental Example 1 was reduced from... Down to The decrease was approximately The host's power consumption is from Down to The decrease was approximately The total power consumption of the pump is from Down to The decrease was approximately Meanwhile, the temperature fluctuation range on the aquifer pumping side decreased from... Reduce to Therefore, under the same statistical caliber, the strategy of this invention can not only reduce system energy consumption, but also mitigate operational fluctuations on the energy storage side.

[0092] As an extended example, the heating operation conditions of a winter workday can be selected in the same system as experimental example two. Under the condition that the equipment configuration, statistical period and sampling caliber are consistent with the aforementioned comparative example, the operation results of the strategy of the present invention and the conventional experience control strategy are compared.

[0093] In Experiment Example 2, the average load factor at the terminal can be maintained between 60% and 68%, and the statistical period is also a continuous 24 hours. The following shows a representative sampling time for Experiment Example 2. The calculation process.

[0094] At this sampling time, the equivalent real-time load corresponding to a single main unit cooling device is taken as follows: Model input variables take , The calibration coefficients for the corresponding unit energy consumption model are taken as follows: , , , , , Substituting into the energy consumption sub-model of the main unit's refrigeration equipment, we get:

[0095] If the temperature difference between the supply and return water on the source side is taken Specific heat capacity at constant pressure Then, from the flow constraint formula, the source-side flow target can be obtained as:

[0096] If the power consumption model coefficients of the ground source side circulating pump are taken , , Then we have:

[0097] If the heat exchange is shared from the aquifer side... Temperature difference between supply and return water on the aquifer side The target flow rate on the aquifer side is:

[0098] If the coefficients of the aquifer control submodel are taken , , Then we have:

[0099] Therefore, at this representative sampling moment in Experiment Example 2, the total instantaneous power of the system is:

[0100] correspond The power consumption during the sampling period is:

[0101] At the same sampling time, compare the results of the calculations. , , , , Calculated using the same model and statistical methods, we can obtain , , Therefore, the total instantaneous power of the system at that moment is approximately ,correspond Power consumption is approximately Winter working conditions The cumulative statistical method is the same as in Experiment Example 1, that is, the model is still substituted point by point according to all sampling times and the integration is performed.

[0102] Figure 9 The bar chart also corresponds to the 24-hour cumulative statistical results of all sampling times under winter conditions. The statistical method is consistent with that of Experiment Example 1. The left side of the figure shows the comparison of power consumption, and the right side shows the comparison of temperature fluctuation on the aquifer pumping side.

[0103] like Figure 9 As shown, under winter heating conditions, compared with the comparative example, the total power consumption of the system in experimental example two is reduced from [previous example]. Down to The decrease was approximately The host's power consumption is from Down to The decrease was approximately The total power consumption of the pump is from Down to The decrease was approximately Meanwhile, the temperature fluctuation range on the aquifer pumping side decreased from... Reduce to The results show that, under different seasonal boundary conditions, the strategy of the present invention can also achieve both energy-saving effect and operational stability.

[0104] Further experimental example three was set up, which selected the mixed regulation condition during the transition season as the statistical object to verify... Figure 6 and Figure 7 The example illustrates the mixed regulation state within the "Ground Source Priority / Aquifer Priority / Mixed Regulation" logic. In this case study, the system maintains coordinated operation of one main unit and two pump sets, with the average terminal load rate maintained at 50% to 58%, and the statistical period is also continuous. .

[0105] Taking Experiment Example 3 at a representative sampling time For example, if we take the real-time load of a single host... , , And still using the energy consumption sub-model coefficients of the main cooling equipment, we can obtain:

[0106] If the temperature difference between the supply and return water on the source side is taken Then we have:

[0107] When taking the power consumption model coefficients of the ground source side circulating pump , , At that time, we can obtain:

[0108] If the heat exchange is shared from the aquifer side... Temperature difference between supply and return water on the aquifer side Then we have:

[0109] When the coefficients of the aquifer control submodel are taken , , At that time, we can obtain:

[0110] Therefore, the total instantaneous power of the system at that moment in Embodiment 3 is:

[0111] correspond Power consumption is:

[0112] If the comparison examples are taken at the same sampling time , , , , Then, the same method can be used to calculate... , , Therefore, the total instantaneous power of the system at that moment is approximately ,correspond Power consumption is approximately .

[0113] Following the same method as Experimental Example 1 The cumulative statistical method can further obtain the daily cumulative results of Example 3 under the mixed regulation conditions of the transition season. Compared with the comparative example, the total power consumption of the system in Example 3 can be reduced by approximately dropped to about The host's power consumption can be reduced from approximately dropped to about The total power consumption of the pump can be approximately dropped to about Meanwhile, the temperature fluctuation range on the aquifer pumping side can be reduced from approximately Reduce to approximately This indicates that, under mixed regulation conditions, the strategy of this invention can also achieve good energy consumption control and operational stability.

[0114] As can be seen from the above-mentioned summer experimental example 1, winter extended experimental example 2, and transitional season mixed regulation example 3, the distributed differential evolutionary optimization control strategy proposed in this invention can simultaneously reduce energy consumption and improve operational stability while meeting different seasonal operating conditions, different load levels, and equipment constraints. The above values ​​are exemplary implementation results used to illustrate the control effect; specific results can be adjusted accordingly based on changes in project scale, seasonal boundary conditions, and load levels.

[0115] In summary, this invention achieves global optimization of geothermal and aquifer energy storage coupling systems through the deep integration of distributed differential evolution algorithm and physical mechanism modeling. This not only solves the optimization problem of high-dimensional nonlinear systems, but also improves system energy efficiency and helps ensure the long-term safe utilization of underground resources.

[0116] It will be understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing the operation of a geothermal and aquifer energy storage coupled system, characterized in that, include: A mathematical model of the system is constructed with the goal of minimizing the total energy consumption of the system throughout its entire life cycle. The total energy consumption is coupled from the power consumption of the main cooling equipment, the ground source circulating pump, and the aquifer energy storage pump. The constraint space of the mathematical model is defined by the system energy conservation and hydraulic balance equations to define the feasible region for optimization. A distributed differential evolution algorithm is used to dynamically optimize the mathematical model to obtain system-level optimized equipment power setpoints and flow allocation schemes. In the population initialization stage, the distributed differential evolution algorithm randomly generates an initial population of decision variables within the effective physical domain of each control variable. The decision variables of each individual in the population include the host power setpoint, the power or flow control of the ground source side circulating pump, the power or flow control of the aquifer pump, and the operating temperature difference. The preset control strategy is executed in a rolling time domain cycle to call the low temperature cold source of the system. Combined with the optimization solution of the objective function, the energy storage supply mode and equipment operating parameters are controlled, and the mathematical model is dynamically optimized online based on the control results. The discretization formula for the objective function is as follows: ， In the formula, This represents the total power consumption of the system. , , The first t Step time i Instantaneous power of the main cooling equipment, ground source side circulation pump, and aquifer energy storage pump; N Total number of devices; T For the total system operating cycle, t This is the index for the number of off-steps within the runtime cycle.

2. The method according to claim 1, characterized in that, The first t Step time i Instantaneous power of the main unit's cooling equipment The calculation is based on an energy consumption prediction model that combines physical characteristics with multivariate data regression. Its analytical expression is: ， In the formula, For the first Real-time power consumption of the main unit's cooling equipment; This refers to the inlet temperature of the cooling water. This is the temperature for chilled water supply; The heat load borne by the main refrigeration equipment; to The energy consumption characteristic coefficient is obtained by calibration based on the historical data of the main unit refrigeration equipment under all operating conditions.

3. The method according to claim 1, characterized in that, Instantaneous power , The calculation is based on a quadratic polynomial energy consumption model that is deeply coupled with the frequency conversion and pipeline resistance characteristics. Its characteristic formula is: ， in, For the first Real-time power consumption of the pump; For the first Real-time mass flow rate of the pump; , , The pump efficiency coefficient is obtained by fitting different operating conditions.

4. The method according to claim 1, characterized in that, The constraint space includes inter-component heat transfer balance constraints, which include: ， In the formula, To meet the total heat exchange requirements on the cooling side; To share the heat exchange burden on the ground source side; To share the heat exchange burden on the aquifer side; The total cooling or heating load on the end building side; The total power consumption of all main unit cooling equipment; This represents the total power consumption of the water pumps in the distribution system.

5. The method according to claim 1, characterized in that, The constraint space includes real-time heat transfer constraints on the ground source side, which include: ， In the formula, Real-time heat exchange on the ground source side; This refers to the mass flow rate of water drawn from the ground source side. For fluid density; The specific heat capacity at constant pressure of the fluid; and These are the return water temperature and supply water temperature on the ground source side, respectively.

6. The method according to claim 1, characterized in that, The constraint space includes the flow constraints of the ground-source water pump and the flow constraints of the aquifer intake pump, and their calculation formulas based on the law of conservation of energy are as follows: ， ， in, , These are the target flow rates for the ground-source water pumps and the target flow rates for the aquifer intake pumps, respectively. , These correspond to the return water temperature and supply water temperature on the ground source side, respectively. , These correspond to the return water temperature and supply water temperature on the aquifer side, respectively. The heat load borne by the main refrigeration equipment. The specific heat capacity at constant pressure of the fluid. This refers to the power consumption of the main cooling equipment.

7. The method according to claim 1, characterized in that, After population initialization, the distributed differential evolution algorithm introduces a representative individual selection mechanism: for the entire population, the K-means clustering algorithm is used to cluster the population space according to the spatial Euclidean distance of the individuals; in each cluster, the individual closest to the cluster centroid is found as the representative individual of the subsystem, and only the operating parameters of the representative individual are used as a lightweight carrier for interaction between subsystems, so as to reduce the communication load in the distributed control network.

8. The method according to claim 1, characterized in that, The control strategy includes: The system's low-temperature cold source is determined based on temperature difference priority: the real-time difference between the pumping temperature of the ground source heat exchange branch and the pumping temperature of the aquifer energy storage branch is used. With threshold In comparison, if Then, combining seasonal or pattern signals, the controller determines the mode switching conditions. If the switching conditions are met, the target branch is different from the current dominant branch, and the criteria are continuously satisfied for a preset duration, the controller is allowed to switch between the ground source dominant mode, the aquifer dominant mode, or the mixed regulation mode; if... Maintain the current mode or enter the hybrid adjustment mode; Cross-seasonal energy storage coupling regulation: When the switching conditions are met on the source side or aquifer side and the aquifer energy storage branch meets the injection switching and reinjection constraints, the aquifer side is switched to the dominant energy supply or energy storage branch, and the reinjection branch valve is opened simultaneously to inject circulating water carrying heat into the reinjection well for cross-seasonal heat storage; as the accumulated operation causes changes in the thermal state of the aquifer, when its pumping temperature is inferior to that of the source side, the pumping and injection status touches the constraint boundary, the reinjection capacity is insufficient or the switching conditions are no longer met, the control system maintains the current mode, switches back to the source dominant mode or enters the hybrid regulation mode, and simultaneously adjusts the valve opening, pump frequency and main unit operation command of the corresponding branch.

9. The method according to claim 8, characterized in that, Adaptive frequency conversion adjustment includes: Real-time monitoring of terminal load changes, flow upper and lower limits, and equipment alarm status; real-time correction of the start / stop status of each main unit refrigeration unit and the frequency conversion frequency of each pump by solving the objective function; when abnormal temperature difference, flow exceeding limit, or equipment fault signal occurs, protective limiting or degraded operation control is prioritized.

10. The method according to claim 1, characterized in that, The switching conditions include: The candidate branch has a temperature advantage in the current season. Specifically, when cooling in summer, the pumping temperature of the candidate branch is lower than that of another branch. The corresponding switching strategy is to prioritize the side with the lower pumping temperature as the low-temperature cold source. When heating in winter, the pumping temperature of the candidate branch is higher than that of another branch. The corresponding switching strategy is to prioritize the side with the higher pumping temperature. The flow rate, pressure, heat exchange rate, and supply and return water temperature difference of the candidate branch are all within the safety boundary defined by the constrained space. No alarm signals were detected in the relevant pumps, valves, main refrigeration equipment, and measuring points; When the candidate branch is an aquifer energy storage branch, it also meets the constraints of pumping-injection balance, recharge permission, recharge capacity and groundwater level fluctuation.

Images