Method and system for controlling air conditioning system based on energy efficiency optimization control model

Abstract

The invention discloses a method and system for controlling an air conditioner system based on an energy efficiency optimization control model, the method is executed by an air conditioner controller, and the method comprises the steps that a real-time operation parameter set of the air conditioner system is obtained, and the operation parameter set comprises a temperature parameter, a pressure parameter and a fan operation parameter; the real-time operation parameter set is input into an energy efficiency optimization control model, the energy efficiency optimization control model is obtained by training historical operation data and used for outputting an optimization control target set, and the optimization control target set comprises the target refrigerating capacity, the target switching water temperature and the target system resistance; based on the target refrigerating capacity, a first control signal used for controlling a fan in the air conditioning system, a second control signal used for controlling an electronic expansion valve in the air conditioning system and a third control signal used for controlling a water valve in the air conditioning system are generated, the second control signal enables a refrigerating system in the air conditioning system to output a refrigerating effect conforming to the target refrigerating capacity, and the third control signal is used for assisting in achieving output of the target refrigerating capacity.

Description

Technical Field

[0001] This invention relates to the field of air conditioning, and in particular to a method and system for controlling an air conditioning system based on an energy efficiency optimization control model. Background Technology

[0002] Currently, data centers have two main demands: air cooling and liquid cooling. Air-cooled terminals use chilled water, requiring a standard supply water temperature below 20°C. Liquid-cooled CDUs are newer products, typically designed with a supply water temperature of 35°C. Therefore, dual-source air-cooled terminals sharing the network with CDUs can achieve initial air-liquid compatibility. However, liquid-cooled CDUs and dual-source terminals suffer from water resistance imbalance. Under high loads, when the air-cooled side resistance is low, insufficient flow on the liquid-cooled side can lead to temperature runaway.

[0003] Currently, dual-source air conditioners using natural cooling, including mechanical cooling, hybrid cooling, and natural cooling, generally switch based on the outlet air temperature control. This has the problem of uncontrolled water temperature during switching. Sometimes, after switching, the outlet air temperature does not meet the requirements, and the compressor restarts. Frequent switching occurs during transitional seasons. Sometimes, after switching, the supply air temperature remains below the set target for an extended period, which is not conducive to energy saving. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for controlling an air conditioning system based on an energy efficiency optimization control model.

[0005] To address the aforementioned technical problems, embodiments of the present invention disclose a method for controlling an air conditioning system based on an energy efficiency optimization control model. The method is executed by an air conditioning controller and includes: Obtain the real-time operating parameter set of the air conditioning system, the operating parameter set including temperature parameters, pressure parameters and fan operating parameters; The real-time operating parameter set is input into the energy efficiency optimization control model, which is trained based on historical operating data, and outputs an optimization control target set, which includes target cooling capacity, target switching water temperature, and target system resistance. Based on the target cooling capacity, a first control signal is generated to control the fan in the air conditioning system, a second control signal is generated to control the electronic expansion valve in the air conditioning system, and a third control signal is generated to control the water valve in the air conditioning system. The second control signal enables the refrigeration system in the air conditioning system to output a cooling effect that meets the target cooling capacity, and the third control signal is used to assist in achieving the output of the target cooling capacity.

[0006] An intelligent air conditioning control system includes a processor and a memory. The memory stores a computer program, which, when executed by the processor, implements the steps of the method for controlling an air conditioning system based on an energy efficiency optimization control model as described in any one of the embodiments of this application.

[0007] The beneficial effects of this application are as follows: I. In this application, a comprehensive real-time data foundation for resistance control is provided by acquiring a real-time set of operating parameters, including temperature parameters, pressure parameters, and fan operating parameters. These parameters are then input into an energy efficiency optimization control model trained on historical data, outputting an optimization control target set containing the target system resistance. This target system resistance is generated by the model based on the resistance balance patterns under different loads and operating conditions in history, and can match the actual needs of the current system. Claim 2 further clarifies that "based on the target system resistance, hydraulic balance adjustment is performed by adjusting the water valve opening or fan speed to stabilize the air conditioning water system resistance within the target range." This method differs from traditional fixed adjustment logic, allowing for dynamic adjustment of resistance based on real-time parameters. This avoids "diversion" of the liquid cooling side flow when the air-cooled side resistance is too low, maintaining the liquid cooling side flow within the range that meets cooling requirements, thereby reducing the possibility of liquid cooling side temperature runaway and improving the operational stability of the dual-cold-source network system.

[0008] Second, this application outputs a "target switching water temperature" through an energy efficiency optimization control model. This target switching water temperature is generated by the model based on water temperature-mode adaptation data under different seasons and cooling demands in history. It reflects the balance between the utilization potential of natural cold sources and the system's cooling demand better than the traditional single outlet air temperature. Furthermore, it clarifies that "mode switching instructions are generated based on the comparison results of the target switching water temperature and the real-time water temperature." For example, when the real-time water temperature is higher than the target switching water temperature, it switches to mechanical cooling mode; when it is lower than the target switching water temperature and the water-side cooling capacity meets the demand, it switches to a completely natural cooling mode. This switching logic, centered on the "target switching water temperature," avoids the switching deviations caused by relying solely on outlet air temperature in traditional methods, and reduces frequent start-stop cycles during transitional seasons. Simultaneously, it ensures that the corresponding mode is only switched to when the natural cold source is sufficient to meet the demand, preventing the supply air temperature from being consistently lower than the set target, thus improving the system's energy-saving effect.

[0009] Third, this application explicitly states that "a first control signal for controlling the fan, a second control signal for controlling the electronic expansion valve, and a third control signal for controlling the water valve are generated based on the target cooling capacity." The second control signal ensures that the cooling system output meets the target cooling capacity, while the third control signal assists in achieving the target cooling capacity. This design enables synergy between cooling capacity regulation and water valve control (related to water resistance balance) and fan control (related to resistance regulation and airflow adaptation): On the one hand, when the third control signal of the water valve assists in the cooling capacity output, it can simultaneously adapt to the target system resistance requirements, avoiding resistance imbalance caused by cooling capacity regulation; on the other hand, the synergy between cooling capacity and mode switching (such as adapting the target cooling capacity to the water-side cooling capacity in natural cooling mode) can reduce the mismatch between cooling capacity and demand after mode switching, further improving the stability and energy efficiency of system operation. This contrasts sharply with the "independent adjustment of each link" in traditional control methods in the background technology and is more suitable for the complex operating conditions of dual-source air conditioning systems in data centers. Attached Figure Description

[0010] Figure 1 This invention illustrates a flowchart of a method for controlling an air conditioning system based on an energy efficiency optimization control model, according to an embodiment of the present invention. Figure 2 The diagram shows the operation logic of the water valve according to an embodiment of the present invention. Detailed Implementation

[0011] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Although the description of the present invention will be presented in conjunction with preferred embodiments, this does not mean that the features of the invention are limited to this embodiment. On the contrary, the purpose of describing the invention in conjunction with embodiments is to cover other options or modifications that may be derived based on the claims of the present invention. To provide a deep understanding of the present invention, many specific details will be included in the following description. The present invention may also be implemented without using these details. Furthermore, to avoid confusion or obscuring the focus of the present invention, some specific details will be omitted in the description. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0012] It should be noted that in this specification, similar reference numerals and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0013] In the description of this embodiment, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set up," "connected," and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this embodiment based on the specific circumstances.

[0014] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0015] Reference Figure 1 This application provides a method for controlling an air conditioning system based on an energy efficiency optimization control model. The method is executed by an air conditioning controller and includes: Step 1: Obtain the real-time operating parameter set of the air conditioning system, which includes temperature parameters, pressure parameters, and fan operating parameters; Step 2: Input the real-time operating parameter set into the energy efficiency optimization control model. The energy efficiency optimization control model is trained based on historical operating data and is used to output an optimization control target set, which includes target cooling capacity, target switching water temperature, and target system resistance. Step 3: Based on the target cooling capacity, generate a first control signal for controlling the fan in the air conditioning system, a second control signal for controlling the electronic expansion valve in the air conditioning system, and a third control signal for controlling the water valve in the air conditioning system. The second control signal enables the refrigeration system in the air conditioning system to output a cooling effect that meets the target cooling capacity, and the third control signal is used to assist in achieving the output of the target cooling capacity. Step 4: Based on the target system resistance, adjust the opening of the water valve or the speed of the fan to perform a hydraulic balance adjustment operation to adjust the opening of the water valve, so that the air conditioning water system resistance in the air conditioning system is stabilized within the target range. Step 5: Based on the comparison results of the target switching water temperature and the real-time water temperature, generate a mode switching command to control the air conditioning system to switch between mechanical cooling mode, hybrid cooling mode and completely natural cooling mode.

[0016] In this embodiment, the mode switching operation directly changes the acquisition benchmarks for temperature and pressure parameters in terms of acquiring real-time operating parameter sets. For example, when switching to the completely natural cooling mode, the refrigeration system shuts down the compressor, and the evaporator no longer participates in heat exchange. At this time, the temperature difference between the inlet and outlet of the evaporator collected by the temperature sensor will be significantly reduced, and the refrigerant pipeline pressure monitored by the pressure sensor will also remain stable due to the electronic expansion valve maintaining a minimum opening. In contrast, in the mechanical refrigeration mode, the compressor operation will cause the refrigerant pressure to rise sharply, and the temperature difference between the inlet and outlet air will increase as the cooling capacity output increases. These parameter changes will be fed back to the energy efficiency optimization control model as new inputs to ensure that the model perceives the system characteristics under the current mode. For the output of the energy efficiency optimization control model, mode switching will trigger the model to adaptively adjust the set of optimization control objectives: In mechanical refrigeration mode, the model will focus on increasing the target cooling capacity to match the high-efficiency operating range of the compressor, while reducing the target system resistance to reduce water circulation energy consumption; in hybrid refrigeration mode, the model will dynamically balance the relationship between the target switching water temperature and the target cooling capacity, and reduce repeated mode switching by fine-tuning the water temperature threshold; in the completely natural cooling mode, the model will prioritize reducing the target system resistance to ensure that the water-side flow is maximized to utilize the natural cold source. At this time, the output of the target cooling capacity will rely more on the water valve regulation than the electronic expansion valve. During the control signal generation stage, mode switching determines the collaborative logic of each actuator: In mechanical refrigeration mode, the second control signal (electronic expansion valve control) becomes the core, ensuring efficient compressor refrigeration by adjusting superheat through PID control, while the water valve only serves as an auxiliary regulator; In hybrid refrigeration mode, the first control signal (fan) needs to adapt to the airflow requirements of both mechanical refrigeration and natural cooling, balancing the heat exchange efficiency of both through stepped speed regulation; In the completely natural cooling mode, the role of the electronic expansion valve is weakened, and the third control signal (water valve) dominates the refrigeration output. At this time, the water valve opening adjustment and the hydraulic balance operation of the target system resistance are deeply coupled, directly responding to the refrigeration demand through changes in water flow. The hydraulic balance adjustment operation also changes its adjustment strategy due to mode switching: In mechanical refrigeration, the system resistance is greatly affected by the compressor operating pressure, and the adjustment is "primarily based on the water valve opening and secondarily on the fan speed" to avoid pressure fluctuations impacting refrigeration efficiency; in hybrid mode, the resistance on the water side and the air side interfere with each other, and balance needs to be maintained through coordinated adjustment of the water valve and the fan (such as first adjusting the water valve to stabilize the resistance, and then adjusting the fan to match the air volume); in the completely natural cooling mode, since there is no compressor pressure interference, it can be simplified to a single water valve opening adjustment, and the utilization of natural cooling capacity can be maximized by precisely matching the target system resistance. This linkage mechanism makes mode switching a key node in closed-loop control - from parameter acquisition to model decision-making, and then to execution adjustment, each link is dynamically adapted according to the current mode, which not only ensures the system stability under different cooling modes, but also continuously optimizes the model output through parameter feedback, ultimately forming a control closed loop of "mode switching - parameter change - model adjustment - control adaptation".

[0017] Specifically, step 1: Obtain the real-time operating parameter set of the air conditioning system. Temperature sensors are installed at various key locations in the air conditioning system, such as the indoor return air vent, evaporator inlet and outlet, and condenser inlet and outlet. These sensors collect temperature data in real time and transmit it to the air conditioning controller according to a preset communication protocol. For example, the temperature sensor at the indoor return air vent collects temperature data every 10 seconds and sends it to the air conditioning controller via wired or wireless communication (such as Zigbee or other suitable wireless communication methods) to ensure timely and accurate acquisition of temperature parameters reflecting the indoor environmental conditions.

[0018] Pressure sensors are installed on critical pipes in air conditioning systems, such as refrigerant pipes and cooling water pipes. These pressure sensors need to have appropriate range and accuracy to meet the pressure measurement requirements under different operating conditions of the air conditioning system. These pressure sensors also detect the pressure at their location at certain time intervals (such as every 5 seconds) and feed the pressure data back to the air conditioning controller in real time. This allows the controller to have a comprehensive understanding of the pressure status of each part of the system and thus analyze the system's operating status.

[0019] For the fans in an air conditioning system, operating parameters are obtained by installing devices such as speed sensors and current sensors at the fan motor. The speed sensor measures the real-time speed of the fan, and the current sensor monitors the operating current of the fan motor. The data obtained by the speed and current sensors are transmitted to the air conditioning controller in real time, allowing the controller to obtain the current operating status of the fan, such as whether it is within the normal speed range or whether the motor is overloaded.

[0020] Step 2: Input the real-time operating parameter set into the energy efficiency optimization control model, and output the optimization control target set. Historical operating data of the air conditioning system over a long period (e.g., the past year or several years) is collected. This data covers temperature parameters, pressure parameters, fan operating parameters, and corresponding control target data such as cooling capacity, switching water temperature, and system resistance under different seasons, time periods, and indoor and outdoor environmental conditions. A neural network algorithm is then used as training samples to train an energy efficiency optimization control model. During training, parameters such as model weights are continuously adjusted to ensure the model accurately outputs the corresponding set of optimization control targets based on the input set of operating parameters.

[0021] Once the aforementioned real-time operating parameter set is obtained, the parameters are organized into a specific vector form and input into the pre-trained energy efficiency optimization control model. Based on its internal algorithmic logic and the patterns formed during training, the energy efficiency optimization control model performs calculations and analyses, outputting an optimized control target set that includes the target cooling capacity, target switching water temperature, and target system resistance.

[0022] For example, if the real-time temperature parameters show that the current indoor temperature is high and the outdoor temperature is also at a high level, the model will output a relatively large target cooling capacity after calculation, combined with other information such as pressure parameters, to meet the need for rapid cooling; at the same time, based on the correlation analysis of historical water temperature data, the model will output a suitable target switching water temperature and a target system resistance that matches the current system operating conditions.

[0023] Preferably, in a specific application scenario, the energy efficiency optimization control model comprises four core structural layers: a real-time parameter adaptation layer, a multi-dimensional feature association layer, a dynamic decision-making inference layer, and a control target output layer. The real-time parameter adaptation layer standardizes the input real-time operating parameters to output an enhanced real-time parameter vector, ensuring model compatibility. The multi-dimensional feature association layer calls pre-stored feature mapping relationships to extract and weight the enhanced real-time parameter vector, extracting key correlations between parameters and outputting a real-time feature-target correlation vector. The dynamic decision-making inference layer performs operating condition matching and rule adjustment on the real-time feature-target correlation vector and combines it with real-time operating conditions to match the optimal control strategy, obtaining a set of candidate optimized control targets that meet the constraints. The control target output layer generates and outputs the target cooling capacity, target switching water temperature, and target system resistance based on the set of candidate optimized control targets that meet the constraints.

[0024] Preferably, the specific implementation process of the real-time parameter adaptation layer is as follows: This layer is essentially a data preprocessing module, consisting of a parameter verification unit, a standardization mapping unit, and a dimension expansion unit. The processing object is the set of operating parameters collected in real time by the air conditioning system, including temperature parameters (indoor return air temperature, supply air temperature, return water temperature, outdoor ambient temperature), pressure parameters (refrigerant pressure, current resistance of the air conditioning water system), and fan operating parameters (real-time speed, operating current). First, the parameter validation unit performs format validation on the real-time operating parameters, checking whether the parameters are complete (e.g., whether temperature parameters include data from all preset monitoring points) and whether the data type meets the model requirements (e.g., whether numerical parameters are in floating-point format). Invalid data (e.g., negative values ​​caused by sensor malfunctions) is removed, and validated real-time parameters are generated. The validated real-time parameters are then processed and input into the standardization mapping unit. This unit calls the historical parameter extreme values ​​(maximum and minimum values) pre-stored in the model and converts each parameter to the 0-1 range according to the rule "(real-time parameter value - historical minimum value) / (historical maximum value - historical minimum value)" to eliminate the influence of differences in the magnitude of different parameters and generate a standardized real-time parameter vector. The standardized real-time parameter vector is then processed and input into the dimension expansion unit. This unit adds time dimension identifiers (e.g., current season code, time period code) to form an enhanced real-time parameter vector that includes the operating condition background. This vector will be used as the input to the multi-dimensional feature association layer.

[0025] Preferably, in the specific technical implementation of the multi-dimensional feature association layer: this layer is essentially a Convolutional Neural Network (CNN) feature extraction architecture, composed of a convolutional feature extraction sublayer, an attention mechanism sublayer, and a fully connected mapping sublayer. The processing object is the enhanced real-time parameter vector output by the real-time parameter adaptation layer. First, the convolutional feature extraction sublayer calls the model's pre-stored feature mapping relationship library (this library is formed based on historical data training and contains the association rules between different parameter combinations and control targets), inputting the enhanced real-time parameter vector into three cascaded convolutional layers—the first layer uses a 3×3 convolutional kernel to extract local features of a single parameter (such as the temperature fluctuation amplitude in the past 5 minutes); the second layer uses a 5×5 convolutional kernel to extract the association features of two parameters (such as the ratio of the rate of change of fan speed to return air temperature); the third layer uses a 7×7 convolutional kernel to extract global features of three parameters (such as the coordinated trend of outdoor temperature and system resistance), generating a real-time multi-dimensional feature map through three convolutional operations; the real-time multi-dimensional feature map is used as the processing object and input into the attention mechanism sublayer. The sub-layer contains a preset weight matrix (weight values ​​are determined based on historical training). According to the rule that "fan speed has a higher weight than pressure parameter in cooling capacity correlation, and outdoor temperature has a higher weight than indoor temperature in water temperature switching correlation", the corresponding weights are assigned to the features of each channel in the feature map to strengthen the influence of key features and generate a weighted real-time feature map. The weighted real-time feature map is used as the processing object and input to the fully connected mapping sub-layer. This sub-layer contains two cascaded fully connected layers (128 neurons in the hidden layer and 3 neurons in the output layer). The high-dimensional features are compressed and converted into a 3-dimensional feature vector that matches the dimension of the control target (corresponding to the correlation features of cooling capacity, water temperature switching, and system resistance, respectively). A real-time feature-target correlation vector is generated, which will be used as the input of the dynamic decision inference layer.

[0026] Preferably, in a specific implementation of the dynamic decision-making reasoning layer in a scenario: this layer is essentially a neural network decision architecture that integrates rule-based reasoning, consisting of a working condition matching sublayer, a rule adaptation sublayer, a target adjustment sublayer, and a constraint verification sublayer. The processing object is the real-time feature-target association vector output by the multi-dimensional feature association layer. First, the working condition matching sublayer performs working condition matching: it inputs the real-time feature-target association vector and the model's pre-stored "typical working condition feature library" (containing feature vectors from typical scenarios such as high summer load, medium transition season load, and low winter load) into the cosine similarity calculation unit. By calculating the cosine distance between vectors, it filters out the typical working condition with the highest similarity, using it as a reference benchmark for the current working condition, and generates a working condition matching result. The working condition matching result is then used as the processing object and input into the rule adaptation sublayer. This sublayer calls the pre-stored working condition-decision rule library (stored in the form of if-then rules) to extract the control target adjustment rule for that working condition—for example, when matching "high summer load working condition," the rule is "the target cooling capacity must prioritize meeting the real-time load demand, and the system resistance must be controlled within the low resistance range." Real-time decision rules are generated. These rules are then input into the target adjustment sublayer. This sublayer calculates the initial control target (initial cooling capacity Q0, initial switching water temperature T0, initial system resistance P0) based on the real-time feature-target association vector. The initial control target is then adjusted using a rule mapping matrix (e.g., Q1 = Q0 × 1.05 under summer conditions to enhance cooling capacity), generating a preliminary optimized control target. This preliminary optimized control target is then input into the constraint verification sublayer. This sublayer has built-in system hardware parameter thresholds (e.g., maximum cooling capacity of the compressor, minimum safe resistance of the pipeline) to check whether the target meets the constraints. If it exceeds the limits, it is corrected according to the boundary values, generating a candidate set of optimized control targets that meet the constraints. This candidate set will then be used as the input to the control target output layer.

[0027] Preferably, the specific implementation process of the control target output layer is as follows: This layer is essentially a target transformation and output module, consisting of a priority ranking sublayer, a collaborative correction sublayer, an anti-standardization sublayer, and an instruction encapsulation sublayer. The processing object is the set of candidate optimized control targets (including candidate target cooling capacity Q_c, candidate target switching water temperature T_c, and candidate target system resistance P_c) that conform to the constraints output by the dynamic decision reasoning layer. First, the priority ranking sublayer performs target priority ranking: based on the real-time parameter deviation calculation unit (such as the magnitude of indoor temperature deviation from the set value and the fluctuation magnitude of system resistance), the target priority is determined—for example, when the real-time indoor temperature deviates from the set value by more than 3℃, the target cooling capacity has the highest priority, generating a target priority sequence; the target priority sequence is input into the collaborative correction sublayer, which contains a target conflict detection matrix (preset correlation coefficients between different targets). If a conflict is detected between candidate targets (such as increasing the cooling capacity requires increasing the fan speed, which may lead to an increase in system resistance), then the high-priority target is retained according to priority, and the low-priority target is adaptively adjusted (such as appropriately widening the system resistance control range), generating a collaboratively optimized control target; collaborative optimization… The control target is processed as an input to the destandardization sublayer. This sublayer calls the historical extreme values ​​used by the real-time parameter adaptation layer and converts the standardized target into physical quantities (e.g., cooling capacity in kW, water temperature in °C) according to the rule "actual value = standardized value × (historical maximum value - historical minimum value) + historical minimum value", generating the actual optimized control target set. Finally, the actual optimized control target set is input to the instruction encapsulation sublayer and converted into JSON format instructions that the air conditioning controller can directly parse (e.g., structured data containing "target cooling capacity: 15kW, target switching water temperature: 13°C, target system resistance: 120kPa"), and output to the subsequent control links to generate control signals for the fan, electronic expansion valve, and water valve, completing the model's real-time regulation guidance of the air conditioning system.

[0028] Step 3: Generate corresponding control signals based on the target cooling capacity. The air conditioner controller has a built-in control logic program corresponding to the fan's operating characteristics. Based on the target cooling capacity and the fan's performance curve (pre-stored in the controller's memory), a PID control algorithm calculates the required fan speed and other control parameters. Then, according to the communication protocol and control command format supported by the fan driver, a first control signal for controlling the fan speed is generated and sent to the fan driver. This achieves precise control of the fan speed, enabling the fan to provide the appropriate airflow to achieve the target cooling capacity.

[0029] Based on the target cooling capacity and current refrigerant parameters such as pressure and temperature, these parameters are acquired by sensors installed on the refrigerant pipeline and fed back to the air conditioning controller. Using a pre-stored thermodynamic calculation model of the refrigeration system in the controller, the required opening degree of the electronic expansion valve is calculated. The controller then generates a corresponding pulse width modulation (PWM) signal or other suitable control signal as a secondary control signal, according to the control interface requirements of the electronic expansion valve, and sends it to the electronic expansion valve driver. This precisely adjusts the opening degree of the electronic expansion valve, controlling the refrigerant flow rate to ensure the refrigeration system outputs a cooling effect that meets the target cooling capacity.

[0030] Based on the impact of cooling water or chilled water circulation in the air conditioning system on the cooling capacity, and combined with the target cooling capacity and relevant parameters such as current water temperature and flow rate (obtained through sensors installed on the water pipes), the controller calculates the required opening angle of the water valve using preset water system control logic stored in the controller. The controller generates a third control signal according to the opening control command format of the electric regulating valve and sends it to the water valve driver to adjust the water valve opening, ensuring the water flow is at a suitable level to assist in achieving the target cooling capacity output.

[0031] Step 4: Perform hydraulic balance adjustment based on the target system resistance. Once the target system resistance value is obtained, it is compared with the actual system resistance value monitored by the sensors. If the actual system resistance is greater than the target system resistance, indicating poor water flow in the water system, the air conditioning controller reduces the opening of relevant water valves (such as cooling water valves and chilled water valves) according to a preset adjustment strategy, while simultaneously monitoring changes in system resistance in real time, until the actual system resistance approaches the target system resistance, achieving hydraulic balance and ensuring stable operation of the water system and efficient operation of the refrigeration system. Conversely, if the actual system resistance is less than the target system resistance, the water valve opening is appropriately increased for adjustment.

[0032] In some specialized air conditioning system structures, the fan speed can also affect system resistance. When it is found that adjusting the water valve opening cannot effectively achieve the target system resistance, the fan speed is appropriately adjusted based on the fan performance and the overall system operation, according to a pre-set control strategy linking fan speed and system resistance. This indirectly affects the system resistance by changing factors such as airflow and pressure. Combined with the adjustment of the water valve opening, the hydraulic balance of the system is ultimately achieved, ensuring the stable and efficient operation of the entire air conditioning system.

[0033] Step 5: Based on the comparison results of the target switching water temperature and the real-time water temperature, generate a mode switching command. Water temperature sensors are installed at key circulation points of the cooling water or chilled water in the air conditioning system. The water temperature sensors collect water temperature data in real time and feed it back to the air conditioning controller.

[0034] After receiving real-time water temperature data, the air conditioning controller compares it with the target switching water temperature. If the real-time water temperature is higher than the target switching water temperature and remains higher for a period of time (e.g., more than 5 minutes, to avoid accidental switching due to sudden temperature fluctuations), it indicates that the current natural cooling source is insufficient to meet the cooling demand. In this case, the controller generates a mode switching command according to the preset control program, controlling the air conditioner to switch to mechanical cooling mode or hybrid cooling mode (determined by the specific system design and actual operating conditions), and sends this command to the corresponding cooling module control unit to achieve the mode switch. Conversely, if the real-time water temperature is lower than the target switching water temperature and meets the corresponding stability conditions, the controller generates a command to switch to a completely natural cooling mode, making full use of the natural cooling source for cooling and achieving energy saving.

[0035] By employing the above technical solution, an energy efficiency optimization control model is trained based on historical operating data and outputs a target cooling capacity. This model then generates corresponding control signals for the fan, electronic expansion valve, and water valve. This approach allows for precise adjustment of each component based on the actual operating conditions of the air conditioning system, ensuring that the cooling capacity output matches actual demand. This avoids insufficient or excessive cooling, thereby guaranteeing that the indoor temperature remains stably within a comfortable range while simultaneously improving cooling efficiency and reducing unnecessary energy consumption.

[0036] By performing hydraulic balance adjustments based on the target system resistance, and by monitoring and adjusting the water valve opening or fan speed in real time, the water flow in the system can be ensured to be in a reasonable and stable state. This helps coordinate the operation of all parts of the air conditioning system, avoids problems such as local overheating and poor cooling effect caused by hydraulic imbalance, further improves the overall operating efficiency of the system, and ensures that the air conditioning system provides stable and efficient cooling services for a long time.

[0037] The system generates mode switching commands based on a comparison of the target switching water temperature and the real-time water temperature, enabling the air conditioner to flexibly switch between mechanical cooling mode, hybrid cooling mode, and purely natural cooling mode. When the natural cooling source can meet the cooling demand, it promptly switches to natural cooling mode or hybrid cooling mode to fully utilize the natural cooling source, reduce the operating time of mechanical cooling components such as the compressor, and significantly reduce power consumption, achieving energy saving. The energy-saving effect is even more significant during transitional seasons and other periods when the natural cooling source is more abundant.

[0038] From acquiring real-time operating parameter sets to outputting optimized control target sets from the energy efficiency optimization control model, and then to precise regulation of each stage, the entire process forms a closed-loop optimization control system. The cooperation and coordinated operation of each component ensures that the air conditioning system remains in a relatively ideal operating state, meeting cooling requirements while minimizing energy waste, improving the overall energy efficiency ratio of the system, and reducing long-term operating costs.

[0039] The energy efficiency optimization control model is trained based on a large amount of historical operating data. This means that it can learn the system's operating patterns under various conditions, such as different seasons, different ambient temperatures, and different indoor-outdoor temperature differences. In actual operation, regardless of the complex and ever-changing operating conditions, it can output a suitable set of control targets based on real-time operating parameters, enabling the air conditioning system to adaptively adjust its operating state and always maintain good cooling performance and stability.

[0040] By precisely controlling key components such as fans, electronic expansion valves, and water valves, and by rationally executing operations such as hydraulic balancing and mode switching, system failures or frequent start-ups and shutdowns caused by factors such as incoordination between components or unreasonable parameters are avoided. This ensures the stable and reliable operation of the air conditioning system, reduces maintenance costs, and extends the service life of the equipment.

[0041] In one feasible embodiment, the energy efficiency optimization control model is trained based on historical operating data to output a set of optimization control targets, including: Based on the energy efficiency optimization parameters and real-time operating parameters output by the energy efficiency optimization control model, the real-time system energy efficiency ratio (COP) is calculated. The energy efficiency optimization parameters include the real-time cooling capacity, and the real-time operating parameters include the real-time power consumption of the air conditioning system. The real-time energy efficiency ratio (COP) is calculated based on the real-time cooling capacity and the real-time power consumption of the air conditioning system.

[0042] Specifically, step 2.1: Energy efficiency optimization and COP calculation of the energy efficiency optimization control model. The energy efficiency optimization control model outputs energy efficiency optimization parameters: During training, the energy efficiency optimization control model uses historical operating data such as cooling capacity, power consumption, and environmental parameters as training features. It optimizes model parameters using a gradient descent algorithm, enabling the model to output energy efficiency optimization parameters. In real-time operation, after receiving the current real-time operating parameter set, the model outputs the target cooling capacity, target switching water temperature, and target system resistance, along with the real-time cooling capacity as the core energy efficiency optimization parameter. This real-time cooling capacity is the optimal cooling output value predicted by the model based on the current operating conditions, and its deviation from the target cooling capacity is controlled within ±5%.

[0043] Power consumption data acquisition in real-time operating parameters: The air conditioning controller collects the total power consumption (unit: kW) of the air conditioning system in real time by installing an active power sensor (measurement accuracy class 0.5) in the main power supply circuit. This includes the sum of the real-time power of all electrical equipment such as the compressor, fan, water pump, and electronic expansion valve actuator. The power sensor outputs sampling data once every second, and the controller takes the average of three consecutive sampling values ​​as the current real-time power consumption value to filter out instantaneous fluctuations.

[0044] In one feasible embodiment, the real-time cooling capacity is calculated based on real-time air volume, air density, and the enthalpy difference between the inlet and outlet air. The calculation method for the real-time cooling capacity is as follows: Obtain the real-time speed signal of the current fan; The real-time fan speed is used to query the preset speed-air volume correspondence table, and the real-time air volume V is calculated by interpolation. Calculate the air density ρ based on the real-time supply air temperature; Calculate the enthalpy difference Δh based on the real-time enthalpy values ​​of the incoming and outgoing air; The real-time cooling capacity Q is calculated using the formula Q=ρ×V×Δh.

[0045] In one feasible embodiment, step 2.1.1: Specific calculation method for real-time cooling capacity Obtain the real-time speed signal of the fan: The fan driver's built-in speed sensor converts the fan rotor speed into an electrical signal. The air conditioning controller then converts this analog signal into a digital value via an A / D conversion module, updating the real-time speed value (unit: r / min) every second. For example, when a voltage signal of 5V is detected, the corresponding speed is 1000 r / min.

[0046] Specifically, calculate the real-time air volume V: The controller calls a pre-stored speed-airflow correspondence table, which contains multiple sets of data measured during the system commissioning phase. For example:

[0047] Fan speed and air volume relationship table When the fan speed is at a precise percentage setting (e.g., 30%), directly call up the corresponding air volume value (1.22m³ / s) in the table.

[0048] When the fan speed is between two gears (e.g., 15% is between 10% and 20%), linear interpolation is used for calculation: Let n1 = 10% speed corresponds to airflow V1 = 0.75 m³ / s, and n2 = 20% speed corresponds to V2 = 0.91 m³ / s. If the current speed n = 15%, Real-time air volume: V=V1+(V2-V1)×(n-n1) / (n2-n1)=0.75+(0.91-0.75)×(15-10) / (20-10)=0.83m 3 / s.

[0049] Calculate air density ρ: Based on the real-time supply air temperature T (unit: °C, acquired by a Pt100 temperature sensor installed at the supply air outlet, with a measurement accuracy of ±0.3 °C), the following formula is used for direct calculation: ρ = 1.293 × 273.15 / (273.15 + T) (Unit: kg / m³) For example, when the real-time supply air temperature T=20℃, ρ=1.293×273.15 / (273.15+20)=1.293×273.15 / 293.15≈1.205kg / m³.

[0050] Calculate the enthalpy difference Δh: Inlet air enthalpy h1: The return air temperature t1 (e.g., 26℃) and relative humidity φ1 (e.g., 50%) are obtained by the temperature and humidity sensor in the return air duct. The enthalpy of the humid air in the controller is then consulted to obtain h1 = 58 kJ / kg (dry air).

[0051] The enthalpy of the air supply h2 is obtained by using the temperature and humidity sensor in the air supply duct to obtain the air supply temperature t2 (e.g., 16℃) and relative humidity φ2 (e.g., 90%). The enthalpy table shows that h2 = 40kJ / kg (dry air).

[0052] Enthalpy difference Δh = h1 - h2 = 58 - 40 = 18 kJ / kg = 18000 J / kg.

[0053] Calculate the real-time cooling capacity Q: Substitute into the formula Q=ρ×V×Δh to calculate (units are unified in SI): Q=1.205kg / m³×0.1458m³ / s×18000J / kg≈1.205×0.1458×18000≈3146W≈3.1kW.

[0054] The calculation result is rounded to one decimal place and fed back to the energy efficiency optimization control model as real-time cooling capacity for dynamic calculation of COP and system optimization.

[0055] Step 2.1.2: Specific calculation methods for system energy efficiency ratio (COP) and power usage effect (PUE) Calculation of system energy efficiency ratio (COP): Real-time cooling capacity Q: The real-time cooling capacity (unit: kW) calculated in step 2.1.1 is used, for example, 3.2kW.

[0056] Inter-row active power acquisition: Smart meters with Modbus communication function are installed in the main power supply circuit of the air conditioning system to collect the active power (unit: kW) of the air conditioning system in real time. This power includes the total power consumption of air conditioning equipment such as compressors, fans, water pumps, and electronic expansion valve actuators. The meter sampling frequency is 1 time / second, and the controller takes the average value of the sampling values ​​for 10 consecutive seconds as the current active power value, for example, 2.0kW.

[0057] COP calculation formula: COP = Cooling capacity Q ÷ Active power of inter-row meter For example, if Q = 3.2 kW and the active power of the inter-row meter = 2.0 kW, then COP = 3.2 ÷ 2.0 = 1.6.

[0058] Calculation of Power Usage Effectiveness (PUE): Inter-row active power: same as the total power consumption of the air conditioning system mentioned above (e.g., 2.0kW).

[0059] PDU Active Power Acquisition: Smart meters are installed on the power supply circuits of the 14 PDUs (Power Distribution Units) in the computer room to collect the active power (unit: kW) of each PDU in real time, and are denoted as P1 to P2. 14 The sampling frequency of each meter is consistent with that of the inter-row meters. The controller synchronously acquires and accumulates the total power consumption of the PDU: P_PDU = P1 + P2 + ... + P 14 For example, 10.0kW.

[0060] PUE calculation formula: PUE = (Inter-row active power + P_PDU) ÷ P_PDU For example, if the active power of the inter-row meter is 2.0kW and P_PDU is 10.0kW, then PUE = (2.0 + 10.0) ÷ 10.0 = 1.2.

[0061] The calculated COP and PUE values ​​are stored in the controller's historical database in real time, with a storage interval of 5 minutes per instance. This data is used for continuous training and optimization of the energy efficiency optimization control model (when COP is 5% lower than the historical average or PUE is higher than 1.3, model parameter adjustments are triggered) and generation of system energy efficiency assessment reports.

[0062] In one feasible embodiment, the method further includes a step of coordinating with a cold source-side system: The target cooling capacity, target system resistance, and calculated real-time cooling capacity output by the energy efficiency optimization control model are used as forward-looking demand signals. The demand signal is encapsulated into a data frame with a predefined communication protocol format; The data frame is sent to the control system on the cold source side via the communication interface, so that the cold source side can pre-adjust the output power and water pump frequency in advance.

[0063] Specifically, step 6: Specific implementation steps for coordination with the cold source system. Generate forward-looking demand signals: The air conditioning controller extracts the target cooling capacity (e.g., 8kW), target system resistance (e.g., 130kPa), and real-time cooling capacity (e.g., 7.8kW) from the output results of the energy efficiency optimization control model and real-time calculation data. These three parameters are combined into a forward-looking demand signal. This signal reflects the immediate and short-term demand of the terminal air conditioner for the cooling source. The target cooling capacity represents the expected cooling output, the target system resistance reflects the load characteristics of the terminal water system, and the real-time cooling capacity is used to verify the adjustment effect on the cooling source side.

[0064] Encapsulated into data frames of a predefined communication protocol format: The controller encapsulates the demand signals according to a predefined communication protocol (such as Modbus TCP / IP). The data frame format is defined as follows: frame header (2 bytes, fixed at 0xAA55) + target cooling capacity (2 bytes, unit 0.1kW, 8kW corresponds to 0x0320) + target system resistance (2 bytes, unit 0.1kPa, 130kPa corresponds to 0x0514) + real-time cooling capacity (2 bytes, unit 0.1kW, 7.8kW corresponds to 0x030C) + check bit (1 byte, using the lower 8 bits of the CRC16 algorithm) + frame trailer (1 byte, fixed at 0xEE). Data type conversion (such as floating-point to integer) and check bit calculation are automatically performed during the encapsulation process.

[0065] Send to the cold source-side control system via the communication interface: The controller sends encapsulated data frames to the chilled water source control system (such as a chiller group control system) via an Ethernet communication interface (e.g., an RJ45 interface supporting 100Mbps). The transmission cycle is once per minute to ensure sufficient adjustment response time for the chilled water source. A timeout retransmission mechanism is enabled during communication (5-second timeout, maximum 3 retransmissions). If consecutive failures occur, a communication alarm is triggered. Upon receiving the data, the chilled water source system adjusts the chiller output power (e.g., loading or unloading the compressor) and the operating frequency of the cooling water pump and chilled water pump 5-10 minutes in advance based on the demand signal, dynamically matching the chilled water source output with the terminal demand and reducing system lag.

[0066] In one feasible embodiment, the step of coordinating with the cold source-side system further includes: Receive confirmation signals and actual operating parameters from the cold source side control system; The actual operating parameters are compared with the expected effects of the forward-looking demand signals; When the deviation persists, model correction coefficients are generated and the parameters related to system coordination in the energy efficiency optimization control model are fine-tuned.

[0067] Step 6.1: Cold Source Side Feedback Processing and Model Correction Steps Receive feedback signals and parameters from the cold source side: The air conditioning controller receives feedback data from the cold source-side control system through the aforementioned communication interface, including: Confirmation signal: 1-byte status code (e.g., 0x01 indicates successful reception, 0x02 indicates parameter exceeding limits), used to verify the transmission and parsing status of the required signal.

[0068] Actual operating parameters on the cold source side: including the actual output cooling capacity of the chiller unit (2 bytes, unit 0.1kW), water supply temperature (2 bytes, unit 0.1℃), actual pump frequency (1 byte, unit Hz), and system resistance on the cold source side (2 bytes, unit 0.1kPa). The data format follows the same communication protocol.

[0069] The receiving period is consistent with the sending period (1 minute / time), and the data frame is checked by a check bit. Invalid frames are discarded and an error log is recorded.

[0070] Comparison of actual operating parameters with expected results: The controller calculates the difference between the actual operating parameters fed back from the cold source side and the expected effect of the forward-looking demand signal: Cooling capacity deviation ΔQ = Actual cooling capacity output from the cooling source - Target cooling capacity (e.g., the deviation between the actual 8.5kW and the target 8kW is 0.5kW) Resistance deviation ΔP = Cold source side system resistance - Target system resistance (e.g., the deviation between actual 135 kPa and target 130 kPa is 5 kPa). Set deviation thresholds (e.g., cooling capacity deviation ±1kW, resistance deviation ±10kPa). When all deviations are within the threshold range, the synergistic effect is considered to meet expectations.

[0071] Model correction coefficient generation and parameter fine-tuning: When the deviation persists (e.g., cooling capacity deviation > 1kW for 3 consecutive cycles), the controller activates the model self-correction mechanism and calculates the correction factor: Cooling capacity correction factor Kq = target cooling capacity / actual output cooling capacity of the cold source (e.g., 8kW / 9kW≈0.89) The resistance correction factor Kp = target system resistance / cold source side system resistance (e.g., 130kPa / 140kPa ≈ 0.93) Call the parameter adjustment interface of the energy efficiency optimization control model, apply the correction coefficient to the weight parameters in the model that are related to the cold source (such as the connection weight of the cold demand prediction layer), and limit the fine adjustment range to within ±10% to avoid drastic fluctuations in the model output.

[0072] After correction, the deviation is continuously monitored. If the deviation returns to within the threshold for two consecutive periods, the correction is stopped and the current model parameters are saved. If the deviation still does not improve, a manual intervention alarm is triggered.

[0073] In one implementable embodiment, generating the first control signal for controlling the wind turbine includes: Generate a start signal that allows the fan to start at minimum speed; When the cooling conditions are met, the target speed is calculated according to the target cooling capacity and the preset fan speed-air volume mapping relationship, and a speed regulation signal is generated to gradually adjust to the target speed with a preset step size.

[0074] Specifically, step 3.1: Generate the first control signal for controlling the fan. Generate fan start signal: When the air conditioning controller receives a start command or detects a condition requiring the fan to start, it first generates a start signal. This start signal contains instructions to operate the fan at a preset minimum safe speed, which is the minimum stable operating speed specified by the fan manufacturer (e.g., 800 r / min), to prevent the fan from surging or failing to start at low speeds. The start signal is sent to the fan driver via a fan control interface (such as an RS485 communication interface or a PWM drive circuit). Upon receiving the signal, the driver controls the fan motor to start at the minimum speed. During the start-up process, the fan operating current and vibration parameters are monitored in real time to ensure a smooth start-up.

[0075] Generate fan speed control signal: When the controller detects that the actual indoor temperature is higher than the set temperature (i.e., the cooling conditions are met) through the temperature sensor, and the energy efficiency optimization control model has output the target cooling capacity, it calls the fan speed-airflow mapping table pre-stored in the controller's memory. This mapping table is established using measured data from the fan factory calibration and system commissioning phases, recording the actual airflow values ​​(e.g., 500 m³ / h to 1500 m³ / h) corresponding to different speeds (e.g., 800 r / min to 2000 r / min), and associating the target cooling capacity with the required airflow (e.g., for every 1 kW increase in target cooling capacity, the airflow increases by 100 m³ / h).

[0076] The required air volume is calculated based on the target cooling capacity, and then the corresponding target speed is derived from the mapping table. For example, if the target cooling capacity is 5kW, the required air volume is 1000m³ / h, and the target speed is 1400r / min according to the table.

[0077] The controller generates a speed control signal containing instructions to gradually adjust the fan speed in preset steps (e.g., 50 r / min). Starting from the current operating speed (e.g., 800 r / min at startup), a speed control command is sent every preset time interval (e.g., 2 seconds), increasing the speed by 50 r / min each time, until the target speed of 1400 r / min is reached. During the adjustment process, the fan current, the actual air volume fed back by the air volume sensor, and the system pressure parameters are collected in real time. If abnormal fluctuations occur (e.g., current exceeding limits), speed control is paused and the protection mechanism is triggered.

[0078] In one implementable embodiment, generating the second control signal for controlling the electronic expansion valve includes: Generate an initialization signal to control the electronic expansion valve to open at a preset opening degree; The real-time superheat is calculated based on the real-time monitored evaporator outlet temperature and pressure. A PID control signal is generated to drive the electronic expansion valve so that the real-time superheat approaches the preset superheat setpoint.

[0079] Specifically, step 3.2: Generate a second control signal for controlling the electronic expansion valve. Generate the electronic expansion valve initialization signal: When the air conditioning system starts or the electronic expansion valve needs to be reset, the controller generates an initialization signal. This signal contains an instruction to open the electronic expansion valve to a preset opening degree (e.g., 50 steps, corresponding to 20% of the total valve opening). This preset opening degree is the minimum stable operating opening degree determined during the system commissioning phase, ensuring that the initial refrigerant flow meets the minimum heat exchange requirements of the evaporator and avoiding the risk of a sudden increase in system pressure due to a zero opening degree or the risk of liquid return due to an excessive opening degree. The initialization signal is sent through a dedicated drive circuit for the electronic expansion valve. The drive circuit converts the instruction into a pulse signal (e.g., each step corresponds to a 0.05mm valve core displacement), controlling the valve to complete the preset opening degree action within 3 seconds.

[0080] Calculate real-time superheat: The refrigerant outlet temperature T1 is collected by a temperature sensor (accuracy ±0.5℃) installed on the evaporator outlet pipe, and the corresponding pressure P1 is collected by an evaporator outlet pressure sensor.

[0081] The controller calls up the pre-stored refrigerant saturation temperature-pressure lookup table (e.g., R32 refrigerant at 0.8MPa corresponds to a saturation temperature of 10℃), and obtains the saturation temperature Ts based on the pressure P1.

[0082] The formula for calculating the real-time superheat SH is: SH = T1 - Ts. For example, if T1 is 12℃ and Ts is 10℃, then the real-time superheat is 2℃.

[0083] Generate PID control signals: The controller compares the real-time superheat SH with the preset superheat setpoint (e.g., 5℃, determined according to the refrigerant type and system design) to obtain the deviation value e = setpoint - SH.

[0084] The incremental PID control algorithm is used to calculate the adjustment amount, and the formula is: Δu(k)=Kp[e(k)-e(k-1)]+Ki e(k)+Kd[e(k)-2e(k-1)+e(k-2)], where Kp (proportional coefficient), Ki (integral coefficient), and Kd (differential coefficient) are pre-tuning parameters (e.g., Kp=5, Ki=0.1, Kd=0.5).

[0085] Based on the calculated Δu(k), a pulse control signal is generated. Each pulse corresponds to one step of opening adjustment of the electronic expansion valve (positive pulse for opening, negative pulse for closing). For example, when SH=2℃ (deviation e=3℃), Δu(k)=15 steps are calculated, so 15 valve opening pulses are generated, controlling the valve opening to increase by 15 steps until the real-time superheat approaches 5℃. During the adjustment process, the superheat is sampled every 0.5 seconds to ensure a balance between the adjustment response speed and system stability.

[0086] Reference Figure 2 In one feasible embodiment, the hydraulic balance adjustment operation includes: Real-time monitoring of the current resistance value of the air conditioning water system; Calculate the deviation between the current resistance value and the target system resistance value; If the deviation exceeds the dead zone range, an adjustment command is generated to adjust the water valve opening so that the system resistance is stabilized within the target range.

[0087] Specifically, step 4.1: Step 4: Specific implementation steps for performing hydraulic balance adjustment operation. Real-time monitoring of the current resistance value of the air conditioning water system A differential pressure transmitter (range 0-300 kPa, accuracy class 0.5) is installed between the chilled water supply and return pipes of the air conditioning water system, directly measuring pressure through a metal pipe section. The transmitter outputs a 4-20mA standard signal, corresponding to a resistance value of 0-300 kPa. The data is converted every 500 ms by the analog signal acquisition module of the air conditioning controller to obtain the current resistance value in real time. For example, 4mA corresponds to 0 kPa, 20mA corresponds to 300 kPa, and when a 12mA signal is detected, the current resistance value is 150 kPa.

[0088] Calculate the deviation between the current resistance value and the target system resistance value. The controller calculates the difference between the real-time resistance value (P1) and the target system resistance value (P0) output by the energy efficiency optimization control model: ΔP = P1 - P0. The dead zone is set to ±3 kPa. When |ΔP| > 3 kPa, the water valve adjustment process is triggered; otherwise, the current state is maintained. For example, when P0 = 120 kPa and P1 = 125 kPa, ΔP = 5 kPa, exceeding the dead zone.

[0089] Generation and execution of water valve regulation commands Startup condition verification: The controller first checks the water valve startup conditions (such as no electrical faults and water system pressure ≥ 0.2MPa), and activates the water valve control module when the cooling demand is ≥ [water valve startup demand] (preset to 3kW).

[0090] Initial opening control: Activate the water valve's DO (digital output) signal, which outputs a 0-10V proportional voltage via the D / A conversion module. The water valve first operates at its minimum opening (preset to 10%, corresponding to 1V) for 10 seconds to ensure smooth valve core movement.

[0091] PI Adjustment Phase: Switches to PI control after 10 seconds, with a proportional gain Kp = 0.6 and an integral time Ti = 20 seconds. The adjustment amount is calculated based on ΔP, dynamically adjusting the water valve opening between the maximum opening (preset to 90%, corresponding to 9V) and the minimum opening. For example, when ΔP = 5kPa, the output voltage gradually increases from 1V to 3V, and the opening correspondingly increases to 30%.

[0092] Temperature linkage logic: When the controlled temperature is within the range of [Temperature Setting] ± [Lower Cooling Limit] (e.g., 25℃ ± 1℃), the current opening degree is locked; if the cooling demand is ≤ 0 and the controlled temperature is ≤ [Temperature Setting] - [Upper Cooling Limit] (e.g., 25℃ - 4℃ = 21℃), the 0-10V output is turned off, and the water valve is de-energized and closed.

[0093] The adjustment cycle is 10 seconds per cycle, and the adjustment is stopped when |ΔP|≤3kPa, to ensure that the system resistance is stable within the target range.

[0094] In one implementable embodiment, the rules for the mode switching include: When the real-time return water temperature is higher than the target switching water temperature, the air conditioner is controlled to enter the mechanical cooling mode; When the real-time return water temperature is lower than the target switching water temperature, and the water-side cooling capacity is greater than zero but less than the full load requirement, the air conditioner is controlled to enter the mixed cooling mode. When the real-time return water temperature is lower than the target switching water temperature, and the water-side cooling capacity meets the full load requirement, the air conditioner is controlled to enter the completely natural cooling mode.

[0095] Step 5.1: Specific Execution of Mode Switching Rules Conditions and execution for switching to mechanical refrigeration mode: The controller collects the return water temperature of the air conditioning water system in real time (via a temperature sensor installed on the return water pipe, with a sampling period of 10 seconds) and compares it with the target switching water temperature. If the real-time return water temperature is higher than the target switching water temperature for 30 consecutive seconds (e.g., the target switching water temperature is 12℃, and the real-time return water temperature remains at 13℃), it is determined that the natural cooling source cannot meet the cooling demand. The controller generates a mechanical cooling mode command, which includes the following operations: closing the natural cooling source inlet valve (such as the cooling tower connecting valve), starting the compressor and the matching mechanical cooling module, the electronic expansion valve operating according to the PID parameters of the mechanical cooling condition, and the fan operating at the speed corresponding to the target cooling capacity, ensuring that the system relies entirely on mechanical cooling to output cooling capacity.

[0096] Conditions and execution for switching between hybrid cooling modes: When the real-time return water temperature is lower than the target switching water temperature for 30 consecutive seconds (e.g., real-time return water temperature is 10℃, target switching water temperature is 12℃), the controller simultaneously detects the water-side cooling capacity (calculated via a water flow sensor and the supply / return water temperature difference: cooling capacity = flow rate × specific heat × density × temperature difference). If the calculated water-side cooling capacity is greater than 0 (indicating that the natural cooling source can provide some cooling capacity) but less than the current full-load demand (e.g., full-load demand is 10kW, water-side cooling capacity is 6kW), then the hybrid cooling mode is triggered. The controller generates a hybrid mode command: partially opens the natural cooling source valve (opening degree adjusted according to the proportion of water-side cooling capacity), and simultaneously starts the compressor to supplement the insufficient cooling capacity. By coordinating the control of the electronic expansion valve and the water valve opening, the mechanical cooling capacity and the natural cooling capacity are superimposed to meet the target cooling capacity, and the fan maintains the corresponding speed operation.

[0097] Conditions and execution for switching to fully natural cooling mode: When the real-time return water temperature is lower than the target switching water temperature for 30 consecutive seconds, and the calculated water-side cooling capacity reaches or exceeds the current full-load requirement (e.g., full-load requirement is 10kW, water-side cooling capacity is 12kW), it is determined that the natural cooling source can independently meet the cooling demand. The controller generates a fully natural cooling mode command: stop the compressor, close the valves related to mechanical refrigeration, fully open the valve connecting the natural cooling source, control the natural cooling input only by adjusting the water valve opening, keep the electronic expansion valve at its minimum opening (only used for system pressure maintenance), and adjust the fan speed according to the target cooling capacity to achieve 100% utilization of the natural cooling source to reduce energy consumption.

[0098] While the present invention has been illustrated and described with reference to certain preferred embodiments, those skilled in the art should understand that the above description is a further detailed explanation of the invention in conjunction with specific embodiments, and should not be construed as limiting the specific implementation of the invention to these descriptions. Various changes in form and detail can be made by those skilled in the art, including several simple deductions or substitutions, without departing from the spirit and scope of the invention.

Claims

1. A method for controlling an air conditioning system based on an energy efficiency optimization control model, characterized in that, The method is executed by the air conditioner controller and includes: Obtain the real-time operating parameter set of the air conditioning system, the operating parameter set including temperature parameters, pressure parameters and fan operating parameters; The real-time operating parameter set is input into the energy efficiency optimization control model, which is trained based on historical operating data, and outputs an optimization control target set, which includes target cooling capacity, target switching water temperature, and target system resistance. Based on the target cooling capacity, a first control signal is generated to control the fan in the air conditioning system, a second control signal is generated to control the electronic expansion valve in the air conditioning system, and a third control signal is generated to control the water valve in the air conditioning system. The second control signal enables the refrigeration system in the air conditioning system to output a cooling effect that meets the target cooling capacity, and the third control signal is used to assist in achieving the output of the target cooling capacity.

2. The method for controlling an air conditioning system based on an energy efficiency optimization control model as described in claim 1, characterized in that, The method further includes: Based on the target system resistance, the water valve opening is adjusted or the fan speed is adjusted to perform a hydraulic balance adjustment operation to adjust the water valve opening, so that the air conditioning water system resistance in the air conditioning system is stabilized within the target range. Based on the comparison results of the target switching water temperature and the real-time water temperature, a mode switching command is generated to control the air conditioning system to switch between mechanical cooling mode, hybrid cooling mode and completely natural cooling mode.

3. The method for controlling an air conditioning system based on an energy efficiency optimization control model as described in claim 1, characterized in that, The energy efficiency optimization control model is trained based on historical operating data and is used to output a set of optimization control targets, including: Based on the energy efficiency optimization parameters and real-time operating parameters output by the energy efficiency optimization control model, the real-time system energy efficiency ratio (COP) is calculated. The energy efficiency optimization parameters include the real-time cooling capacity, and the real-time operating parameters include the real-time power consumption of the air conditioning system. The real-time energy efficiency ratio (COP) is calculated based on the real-time cooling capacity and the real-time power consumption of the air conditioning system.

4. The method for controlling an air conditioning system based on an energy efficiency optimization control model as described in claim 3, characterized in that, The real-time cooling capacity is calculated based on real-time air volume, air density, and the enthalpy difference between the inlet and outlet air. The calculation method for the real-time cooling capacity is as follows: Obtain the real-time speed signal of the current fan; The real-time fan speed is used to query the preset speed-air volume correspondence table, and the real-time air volume V is calculated by interpolation. Calculate the air density ρ based on the real-time supply air temperature; Calculate the enthalpy difference Δh based on the real-time enthalpy values ​​of the incoming and outgoing air; The real-time cooling capacity Q is calculated using the formula Q=ρ×V×Δh.

5. The method for controlling an air conditioning system based on an energy efficiency optimization control model as described in claim 4, characterized in that, The method also includes steps that coordinate with the cold source-side system: The target cooling capacity, target system resistance, and calculated real-time cooling capacity output by the energy efficiency optimization control model are used as forward-looking demand signals. The demand signal is encapsulated into a data frame with a predefined communication protocol format; The data frame is sent to the control system on the cold source side via the communication interface, so that the cold source side can pre-adjust the output power and water pump frequency in advance.

6. The method according to claim 5, characterized in that, The steps of coordinating with the cold source system also include: Receive confirmation signals and actual operating parameters from the cold source side control system; The actual operating parameters are compared with the expected effects of the forward-looking demand signals; When the deviation persists, model correction coefficients are generated and the parameters related to system coordination in the energy efficiency optimization control model are fine-tuned.

7. The method for controlling an air conditioning system based on an energy efficiency optimization control model as described in claim 1, characterized in that, The generation of the first control signal for controlling the wind turbine includes: Generate a start signal that allows the fan to start at minimum speed; When the cooling conditions are met, the target speed is calculated according to the target cooling capacity and the preset fan speed-air volume mapping relationship, and a speed regulation signal is generated to gradually adjust to the target speed with a preset step size.

8. The method for controlling an air conditioning system based on an energy efficiency optimization control model as described in claim 1, characterized in that, The generation of the second control signal for controlling the electronic expansion valve includes: Generate an initialization signal to control the electronic expansion valve to open at a preset opening degree; The real-time superheat is calculated based on the real-time monitored evaporator outlet temperature and pressure. A PID control signal is generated to drive the electronic expansion valve so that the real-time superheat approaches the preset superheat setpoint.

9. The method for controlling an air conditioning system based on an energy efficiency optimization control model as described in claim 2, characterized in that, The hydraulic balance adjustment operation includes: Real-time monitoring of the current resistance value of the air conditioning water system; Calculate the deviation between the current resistance value and the target system resistance value; If the deviation exceeds the dead zone range, an adjustment command is generated to adjust the water valve opening so that the system resistance is stabilized within the target range.

10. The method for controlling an air conditioning system based on an energy efficiency optimization control model as described in claim 2, characterized in that, The rules for mode switching include: When the real-time return water temperature is higher than the target switching water temperature, the air conditioner is controlled to enter the mechanical cooling mode; When the real-time return water temperature is lower than the target switching water temperature, and the water-side cooling capacity is greater than zero but less than the full load requirement, the air conditioner is controlled to enter the mixed cooling mode. When the real-time return water temperature is lower than the target switching water temperature, and the water-side cooling capacity meets the full load requirement, the air conditioner is controlled to enter the completely natural cooling mode.

11. An intelligent control system for air conditioning, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, it implements the steps of the method for controlling an air conditioning system based on an energy efficiency optimization control model as described in any one of claims 1 to 10.

Images