A multi-connected air conditioner pipe control method with energy consumption characteristic calibration

Abstract

The present application relates to air conditioner control technology field, especially to a kind of energy consumption characteristic calibration multi-connected air conditioner control method.The present application includes: obtaining the air conditioner operating data of multiple time periods generates energy consumption mapping dataset, constructs energy consumption response surface based on energy consumption mapping dataset;Energy consumption response surface is judged whether current energy consumption response surface needs correction based on energy consumption mapping dataset and current energy consumption response surface, if not need correction, then do not trigger group control strategy;If need correction, then based on current energy consumption response surface triggers corresponding group control strategy, according to the air conditioner operating data of real-time measurement energy consumption response surface is corrected;To the energy consumption response surface of correction good repeatedly whether need correction is judged, until air conditioner is shut down.The present application is based on real-time data and intelligent model multi-connected air conditioner system energy consumption characteristic accurate calibration and dynamic group control optimization technical goal, reach the technical effect of improving system energy efficiency, reducing energy consumption fluctuation, enhancing user comfort and realizing partition fine control.

Description

Technical Field

[0001] This invention relates to the field of air conditioning control technology, and in particular to a method for controlling multi-split air conditioners with calibrated energy consumption characteristics. Background Technology

[0002] With the continuous improvement of the level of intelligence in modern buildings, multi-split air conditioning systems have become the mainstream choice for air conditioning solutions in large commercial buildings, office buildings, and public places due to their flexible combination capabilities and high energy efficiency. However, in actual operation, traditional multi-split air conditioning systems still face many challenges in energy consumption management and system control, which seriously affect their energy-saving effect and user experience.

[0003] Currently, existing technologies mainly rely on preset operating parameters and simple control logic, which makes it difficult to accurately reflect changing load demands and environmental conditions. This results in inaccurate energy consumption data, sluggish system adjustments, and an inability to achieve real-time, dynamic energy consumption optimization. Furthermore, traditional control strategies lack detailed analysis of heat load variations in different rooms, making precise zoning control impossible. This can easily lead to over-operation of air conditioning in some areas while providing insufficient comfort in others, resulting in reduced overall system efficiency.

[0004] In summary, existing technologies suffer from technical problems due to their reliance on preset parameters and simple control logic, which makes it difficult to accurately reflect varying load demands and environmental conditions. This results in inaccurate energy consumption data and delayed system adjustment responses, further affecting the real-time dynamic energy consumption optimization and precise zoning control of multi-split air conditioning systems. Summary of the Invention

[0005] The purpose of this invention is to solve the problem that existing technologies rely on preset parameters and simple control logic, which makes it difficult to accurately reflect variable load demands and environmental conditions, resulting in inaccurate energy consumption data and sluggish system adjustment response, further affecting the real-time dynamic energy consumption optimization and precise zoning control of multi-split air conditioning systems. The invention proposes a multi-split air conditioning management and control method with energy consumption characteristic calibration.

[0006] To achieve the above objectives, the technical solution provided by this invention is as follows:

[0007] A method for controlling multi-split air conditioning units with calibrated energy consumption characteristics includes:

[0008] The energy consumption mapping dataset is generated by acquiring air conditioner operation data from multiple time periods, and an energy consumption response surface is constructed based on the energy consumption mapping dataset.

[0009] Based on the energy consumption mapping dataset and the current energy consumption response surface, determine whether the current energy consumption response surface needs to be corrected. If no correction is needed, do not trigger the group control strategy, and continue to determine whether correction is needed.

[0010] If correction is needed, the corresponding group control strategy will be triggered based on the current energy consumption response surface, and the energy consumption response surface will be corrected according to the real-time measured air conditioning operation data.

[0011] The corrected energy consumption response surface is repeatedly evaluated to determine whether further correction is needed, until the air conditioner is turned off.

[0012] Furthermore, the step of acquiring air conditioner operation data from multiple time periods to generate an energy consumption mapping dataset, and constructing an energy consumption response surface based on the energy consumption mapping dataset, includes:

[0013] Obtain air conditioner operation data for multiple time periods;

[0014] Based on the air conditioning operation data, operating parameters and unit energy consumption are extracted and mapped to generate an energy consumption mapping dataset.

[0015] Construct an energy consumption regression model with energy consumption ratio as the target variable based on the energy consumption mapping dataset;

[0016] The energy consumption response surface is obtained through an energy consumption regression model.

[0017] Furthermore, the step of determining whether the current energy consumption response surface needs correction based on the energy consumption mapping dataset and the current energy consumption response surface includes:

[0018] Based on the difference between the air conditioner operation data in the energy consumption mapping dataset and the predicted air conditioner operation values ​​predicted by the energy consumption response surface, a prediction bias vector is constructed.

[0019] Determine whether the energy consumption response surface has entered the degradation range based on the prediction deviation vector;

[0020] If the current energy consumption response surface is in the degradation range, it needs to be corrected; if it is not in the degradation range, it does not need to be corrected.

[0021] Furthermore, the prediction bias vector is constructed based on the difference between the air conditioner operation data in the energy consumption mapping dataset and the predicted air conditioner operation values ​​predicted by the energy consumption response surface, including:

[0022] Calculate the numerical deviation between the air conditioning operation data of multiple dimensions in the energy consumption mapping dataset and the predicted values ​​of the corresponding dimensions under the corresponding conditions of the energy consumption response surface prediction.

[0023] The numerical deviations of each dimension are used to construct a prediction deviation vector.

[0024] Furthermore, the method for formulating the corresponding group control strategy includes:

[0025] First formulation method:

[0026] Construct heat load variation models for multiple rooms based on the energy consumption response surface;

[0027] A group control strategy is formulated based on the heat load change model and the equipment capacity boundary, wherein the equipment capacity boundary is the upper and lower limits of the operation of each air conditioning unit.

[0028] Second formulation method:

[0029] Load feature vectors of multiple rooms are extracted based on air conditioning operation data;

[0030] Clustering the load feature vectors yields load clusters;

[0031] Generate a corresponding group control strategy template for the load cluster; match the current load cluster during actual operation to obtain the current group control strategy template;

[0032] Determine the group control strategy based on the current group control strategy template;

[0033] The first formulation method and the second formulation method are used in conjunction.

[0034] Furthermore, the step of correcting the energy consumption response surface based on real-time measured air conditioning operation data includes:

[0035] Real-time reading of evaporation temperature, superheat, and return gas pressure;

[0036] The compressor frequency in the current group control strategy is obtained based on the load change rate, and the priority adjustment objects are adjusted accordingly.

[0037] Based on the cooling demand and operating status of each air conditioner in the sub-loop, the valve opening in the current group control strategy is obtained and the priority adjustment object is adjusted.

[0038] Calculate the changes in evaporation temperature, superheat, and return gas pressure readings before and after adjusting the compressor frequency and valve opening.

[0039] The energy consumption optimization target is determined by analyzing the changes in evaporation temperature, superheat, and return gas pressure readings. If the target is achieved, the energy consumption response surface is corrected based on the current evaporation temperature, superheat, and return gas pressure. If not, the compressor frequency and valve opening are adjusted, and the energy consumption optimization target is reassessed.

[0040] Furthermore, the adjustment of the priority adjustment object includes:

[0041] Obtain the thermal imbalance difference between the set temperature and the actual temperature of each room, and calculate the impact of the thermal imbalance difference on the load distribution of the entire air conditioning system, expressed by the formula:

[0042] ΔL=Σ(w_i×ΔT_i)

[0043] Where w_i is the heat capacity weight of the i-th room, ΔT_i is the heat imbalance difference of the i-th room, and ΔL is the load distribution influence.

[0044] The room with the largest product of heat capacity weight and the difference in heat imbalance is selected as the priority for adjustment.

[0045] Use the preset user comfort threshold as the adjustment threshold;

[0046] Adjust the air conditioner of the priority target within the adjustment threshold.

[0047] Furthermore, the step of determining whether the adjustment behavior has achieved the energy consumption optimization target based on changes in evaporation temperature, superheat, and return gas pressure readings includes:

[0048] If the change in evaporation temperature is greater than or equal to 0.5℃, the change in superheat is between -1K and 0K, and the change in return gas pressure is less than or equal to 5%, then the regulation behavior is judged to have achieved the energy consumption optimization target.

[0049] Otherwise, it is determined that the adjustment behavior has not achieved the energy consumption optimization goal.

[0050] Furthermore, it also includes:

[0051] Acquire user's power-on / off habits and temperature settings;

[0052] By combining sunshine and wind speed data from weather forecasts with the aforementioned habitual data and mapping them to the group control strategy, a personalized control strategy is generated.

[0053] Replace the group control strategy with a personalized control strategy.

[0054] Furthermore, it also includes:

[0055] Energy consumption is calculated based on real-time measured air conditioning data, and the control path is recorded when energy consumption exceeds the normal range;

[0056] By comparing the control path with the standard energy consumption curve, anomalies can be identified.

[0057] Extract external interference information that causes anomalies, and generate adjustment and repair measures based on the external interference information;

[0058] The regulatory pathways and adjustment and repair measures will be visualized and retrospectively displayed.

[0059] Compared with existing technologies, the significant advantages of this invention are: by collecting data from multiple time periods to construct an energy consumption response surface and using prediction bias for correction, the energy consumption regression model better reflects actual energy consumption characteristics. This achieves the technical goal of accurately calibrating the energy consumption characteristics of multi-split air conditioning systems and dynamically optimizing group control based on real-time data and intelligent models, thereby improving system energy efficiency, reducing energy consumption fluctuations, enhancing user comfort, and realizing refined zoned management. Attached Figure Description

[0060] Figure 1 This is a flowchart illustrating a multi-split air conditioner control method for energy consumption characteristic calibration according to the present invention. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0062] like Figure 1 As shown, the present invention provides a method for controlling multi-split air conditioners with calibrated energy consumption characteristics, specifically including the following steps:

[0063] S1: Obtain air conditioner operation data from multiple time periods to generate an energy consumption mapping dataset and construct an energy consumption response surface.

[0064] Acquire air conditioning operation data for multiple time periods; extract operating parameters and unit energy consumption based on the air conditioning operation data, and generate an energy consumption mapping dataset; construct an energy consumption regression model with energy consumption ratio as the target variable based on the energy consumption mapping dataset; obtain the energy consumption response surface for scheduling optimization through the energy consumption regression model.

[0065] Specifically, operational information of air conditioning equipment is collected over different time periods, such as power consumption, temperature changes, and equipment status, reflecting the energy consumption of air conditioners in actual use. Then, an energy consumption mapping dataset is generated based on this data. The mapping establishes a link between operating parameters and corresponding energy consumption data, forming a systematic database used to describe energy consumption performance under different operating conditions. Next, an energy consumption response surface is constructed based on the energy consumption mapping dataset. The energy consumption response surface is a curve or surface in a multi-dimensional space used to describe the trend and relationship of air conditioning equipment energy consumption as operating parameters change.

[0066] Acquiring air conditioning operation data across multiple time periods refers to collecting operational records from the air conditioning system during different time periods. This data may include key operational information such as compressor on / off status, supply air temperature, set temperature, indoor and outdoor temperatures, operating frequency, current, and voltage. The time periods should cover high-load, medium-load, and low-load scenarios. For example, collecting data from 8:00 AM to 9:00 PM throughout the day helps to comprehensively reflect the energy consumption behavior of the air conditioning system under different environmental conditions and loads.

[0067] Next, operating parameters and unit energy consumption are extracted based on the air conditioner operation data. This means calculating key variables that directly affect energy consumption performance from the original operating data, such as compressor frequency, evaporation temperature, and fan speed, as operating parameters. At the same time, unit energy consumption (i.e., electrical energy consumed per unit time or unit of cooling capacity) is calculated. For example, the electricity consumed in 30 minutes is divided by the output cooling capacity (the unit is usually kilowatt-hours / kilowatt of cooling capacity). These parameters and unit energy consumption are paired through a mapping method to form samples, and finally, an energy consumption mapping dataset is constructed.

[0068] Constructing an energy consumption regression model with energy consumption ratio as the target variable based on an energy consumption mapping dataset refers to using mathematical modeling or machine learning methods, such as multinomial regression, random forest, or support vector regression algorithms, to use energy consumption ratio (e.g., the electrical energy required per unit of cooling capacity, or the system cooling efficiency per unit of compressor power output) as the target variable to be predicted, and using the aforementioned operating parameters as input features to train a mathematical model that can represent the trend of system energy consumption changes.

[0069] Subsequently, the energy consumption response surface for scheduling optimization is obtained through the energy consumption regression model. This model is then applied to simulate and predict under different parameter combinations, outputting a multivariate energy consumption result. For example, it can describe "the expected energy consumption performance of the system under different group control strategies, load conditions, or indoor temperature settings".

[0070] S2: Determine whether the current energy consumption response surface needs to be corrected based on the energy consumption mapping dataset and the current energy consumption response surface.

[0071] Based on the difference between air conditioner operation data from multiple time periods and the predicted air conditioner operation values ​​obtained through the energy consumption response surface, a prediction deviation vector is constructed. The prediction deviation vector is used to determine whether the energy consumption response surface has entered the degradation range. If it is in the degradation range, the group control strategy is triggered to correct the energy consumption response surface. If it is not in the degradation range, the current energy consumption response surface does not need to be corrected.

[0072] Specifically, by analyzing the energy consumption status corresponding to the current operating parameters of the air conditioner using energy response surface methodology, a group control strategy is triggered when the energy consumption performance of the equipment reaches a certain preset condition or threshold. Energy consumption performance is a combination of multiple indicators, such as power and temperature; the group control strategy is triggered when the preset threshold is reached. The group control strategy refers to adjusting the activation or deactivation schemes of compressors, fans, and other key components in the air conditioning system. For example, when the energy consumption response surface shows an abnormally high energy consumption at a certain operating point, it will automatically select to shut down some compressors or adjust the fan speed, thereby reducing energy consumption and achieving energy-saving effects.

[0073] Based on the differences between air conditioning operation data from multiple time periods and the predicted values ​​of air conditioning operation obtained through the energy consumption response surface, a prediction deviation vector is constructed. This involves comparing the actual energy consumption performance in operation with the predicted values ​​under corresponding conditions in the energy consumption response surface and calculating the numerical deviation between the two. This deviation includes multiple dimensions such as unit energy consumption, electricity consumption, and the response amplitude of compressor frequency to energy consumption. These differences form a vector called the prediction deviation vector, which reflects the degree of deviation between the actual system state and the model prediction. For example, if the actual unit energy consumption measured in a certain period is 1.2 kWh / kW of cooling capacity, while the predicted value is 0.9 kWh / kW of cooling capacity, then the deviation is 0.3 kWh / kW of cooling capacity, indicating that there is a certain error between the model and reality.

[0074] Next, determining whether the energy response surface has entered the degradation zone based on the prediction deviation vector involves analyzing the trend and magnitude of the prediction deviation vector to assess the reliability of the current energy response model. The degradation zone refers to the area where the model's predictive ability significantly declines, typically occurring when the model has not been updated for a long time, or due to factors such as aging system hardware or changes in the operating environment. Judgment criteria can include deviations exceeding a set threshold over multiple consecutive time periods, such as an average deviation exceeding 0.25 kWh / kW of cooling capacity over three consecutive hours, or a systematic shift in the direction of the deviation vector, indicating that the response surface can no longer effectively reflect the system's current true energy consumption characteristics.

[0075] Therefore, if the energy consumption response surface is corrected by triggering a group control strategy during the degradation range, it means that once the energy consumption response surface is detected to deviate from the actual system performance, the start-stop strategy of the compressor and indoor unit is automatically triggered. New data is collected through systematic sampling to re-correct or retrain the energy consumption regression model, thereby updating the energy consumption response surface. The group control strategy refers to the combination of starting and stopping multiple compressors and indoor units within a certain time period. Adjusting the start-stop status allows the system to cover more operating scenarios, thereby obtaining more comprehensive correction data. For example, if the old model is based on the alternating operation of three compressors, but the actual measured energy consumption is higher due to the decreased efficiency of some equipment or a deteriorating external environment, it is necessary to collect new energy consumption data by switching different start-stop strategies for model updates.

[0076] S3: If no correction is needed, the group control strategy will not be triggered. If correction is needed, the corresponding group control strategy will be triggered based on the current energy consumption response surface, and the energy consumption response surface will be corrected according to the real-time measured air conditioning operation data.

[0077] Specifically, the system collects and updates real-time operating data of the air conditioning equipment, such as temperature, pressure, and frequency. Based on this real-time data, the group control strategy is evaluated and adjusted. The group control strategy refers to the specific control scheme adopted for the on / off states of equipment such as compressors, fans, or valves. If real-time data detects a sudden rise in the evaporator temperature in a certain area, it feeds back to the control module to automatically adjust the compressor frequency or valve opening in the group control strategy to maintain optimal energy consumption and stable indoor temperature.

[0078] The group control strategy can be formulated using the following method (the first formulation method), including: constructing a heat load change model for multiple rooms based on the energy consumption response surface; and formulating a group control strategy based on the heat load change model and the equipment capacity boundary.

[0079] Specifically, constructing heat load variation models for multiple rooms based on the energy response surface refers to using the input-output characteristics of the air conditioning system described in the energy response surface to deduce the changing trends of cooling or heating demand in different rooms and at different times. Heat load refers to the energy required to maintain a set temperature in a space, and the variation model reflects the dynamic fluctuations of this load under changes in factors such as time, external environment, and indoor activities. For example, the heat load of a southeast-facing room typically increases from morning to noon due to increased sunlight, rising from 1.5 kW of cooling load per hour to 2.8 kW; while the heat load of a north-facing, windowless room remains relatively stable, hovering around 1.0 kW throughout the day. These models are typically established through historical data fitting, physical heat conduction models, or data-driven methods, serving as the basis for subsequent control decisions.

[0080] Next, based on the heat load variation model and equipment capacity boundaries, a group control strategy is formulated. This involves coordinating and optimizing the overall system by combining the room's heat load demand with the maximum adjustable range and operational limitations of the compressors and indoor units. Equipment capacity boundaries refer to the upper and lower limits of each device's operation; for example, a certain model of compressor operates most efficiently within the 5 kW to 12 kW range, and a certain indoor unit has a maximum cooling capacity of 3 kW. The group control strategy determines how to uniformly allocate different equipment to respond to load changes in different rooms at the system-wide level, thereby avoiding overload, frequent start-stop, or inefficiency of individual devices. The group control strategy specifies which compressors and indoor units should operate simultaneously or be suspended to achieve optimal matching. For example, if the southeast-facing room has a load as high as 2.8 kW, while the north-facing room has a load of only 1.0 kW, the system might arrange for one main compressor to operate at high frequency to meet the high-load area, while simultaneously reducing the airflow of indoor units in the low-load area to avoid energy waste.

[0081] Furthermore, the present invention also includes: real-time reading of evaporation temperature, superheat and return gas pressure to obtain reading data; adjusting the compressor frequency in the group control strategy according to the load change rate; adjusting the valve opening in the group control strategy based on the sub-refrigerant circuit status; obtaining reading changes of the reading data based on adjusting the compressor frequency and adjusting the valve opening; judging the energy consumption impact of the adjustment behavior in the group control strategy through the reading changes and providing feedback correction.

[0082] Specifically, real-time reading of evaporation temperature, superheat, and return gas pressure to obtain data refers to the continuous collection of key operating parameters by sensors during the operation of the air conditioning system to determine the system's thermal state. Evaporation temperature refers to the temperature at which the refrigerant changes from a liquid to a gaseous state in the evaporator, reflecting the evaporator's heat absorption efficiency; superheat indicates the degree to which the refrigerant continues to heat up after complete evaporation, serving as an indicator for assessing the presence of liquid slugging or overheating; return gas pressure refers to the pressure level of the gaseous refrigerant before it returns from the evaporator to the compressor.

[0083] Subsequently, adjusting the compressor frequency in the group control strategy based on the load change rate refers to dynamically adjusting the compressor's operating speed after detecting a load change trend. The load change rate represents the degree of increase or decrease in cooling load per unit time. For example, if the cooling load of a room increases from 2.0 kW to 2.5 kW, the change rate is 0.5 kW per unit time. Compressor frequency adjustment is achieved through inverter technology, for example, increasing from 45 Hz to 55 Hz, to enhance cooling capacity to match the increased load. This avoids frequent compressor starts and stops, improves energy efficiency ratio, and reduces equipment wear.

[0084] Next, based on the valve opening in the sub-refrigerant loop status adjustment group control strategy, in a multi-split system, the opening of the electronic expansion valve or electric regulating valve is controlled according to the cooling demand and operating status of each indoor unit in the sub-loop. A sub-refrigerant loop refers to multiple branches branching off downstream of the main refrigerant pipeline, each branch connecting to one or more indoor units. Valve opening directly affects the refrigerant flow and pressure distribution. For example, when the load on a room increases, the valve in its corresponding loop can be opened from 40% to 70% to increase the refrigerant flow and enhance the cooling effect.

[0085] Then, the changes in data obtained by adjusting the compressor frequency and regulating the valve opening refer to the real-time response changes in evaporation temperature, superheat, and return gas pressure through the aforementioned coordinated control of the compressor and valves. The changes in data describe the immediate feedback of the control behavior on the system's thermodynamic state. For example, after the compressor frequency is increased, the return gas pressure increases by 10 kPa, and the evaporation temperature increases by 2 degrees Celsius.

[0086] Finally, judging the energy consumption impact of the adjustment behavior in the group control strategy by reading changes and providing feedback for correction refers to analyzing the changes in the aforementioned state parameters, evaluating whether the adjustment behavior has achieved the energy consumption optimization target, and making strategy adjustments accordingly. Its judgment logic is based on the comprehensive changing trend of system operating parameters and dynamically comparing it with the preset energy consumption optimization target value. This process combines real-time feedback of the system's operating status and uses a multi-parameter joint evaluation mechanism to determine whether to correct the energy consumption response surface or readjust the control strategy, ensuring that the adjustment behavior always moves towards energy consumption optimization. For example, the change in evaporation temperature reading is greater than or equal to 0.5℃, the change in superheat reading is between -1K and 0K, and the change in return gas pressure reading is less than or equal to 5%. Frequently increasing the compressor frequency under low load conditions may lead to increased energy consumption with little effect. In this case, the system will identify it as a regulation failure and automatically reduce the frequency or valve opening, thereby achieving adaptive energy efficiency optimization.

[0087] Furthermore, the present invention also includes: acquiring the thermal imbalance difference between the set temperature and the actual temperature of multiple rooms, and analyzing the impact on the air conditioning load; determining the priority adjustment target of the air conditioning based on the impact of thermal imbalance; determining the adjustment threshold according to the user comfort threshold; adjusting the priority adjustment target of the air conditioning within the adjustment threshold, and outputting the group control strategy.

[0088] Specifically, obtaining the thermal imbalance difference between the set temperature and the actual temperature in multiple rooms and analyzing its impact on the air conditioning load involves comparing the target temperature set by the user in each room with the currently measured actual room temperature, calculating the difference between the two, and evaluating the effect of these differences on the load distribution of the entire air conditioning system. This evaluation process can be represented as:

[0089] ΔL=Σ(w_i×ΔT_i)

[0090] Where w_i is the heat capacity weight of the i-th room, ΔT_i is the heat imbalance difference of the i-th room, and ΔL is the load distribution influence.

[0091] The thermal imbalance difference is the result of subtracting the actual temperature from the set temperature. A positive value indicates the room is too hot, while a negative value indicates it is too cold. For example, if room A is set to 24 degrees Celsius but actually reaches 27 degrees Celsius, the difference is 3 degrees Celsius. This type of difference often manifests as imbalance in multi-room systems, potentially leading to overheating or cooling in some areas, thus increasing the overall load on the air conditioning system. This model accurately quantifies the impact of thermal imbalance on the overall system load distribution, providing a basis for subsequent adjustments to group control strategies.

[0092] Determining priority air conditioning targets based on the impact of thermal imbalance refers to selecting the room with the greatest or least coordinated impact on the load as the priority adjustment target after identifying the temperature difference distribution in each room. The thermal imbalance difference is the basic data for analyzing the impact of thermal imbalance and is used to determine the priority air conditioning target. Prioritizing the air conditioning target means selecting the room with the greatest impact on the system load for adjustment based on the degree of influence of the thermal imbalance difference. The air conditioning target is the specific execution point in the control strategy; for example, prioritizing the reduction of the temperature in the room with the largest temperature difference to alleviate the peak system load. For example, if the temperature difference in room B is 5 degrees Celsius, much higher than the approximately 2-degree Celsius temperature difference in other rooms, then the operating status of the indoor unit in room B will be adjusted first, including increasing the airflow or refrigerant flow. Other rooms still participate in system adjustment according to their operating status and cooling needs, but the adjustment priority and magnitude will be dynamically adjusted based on the effect of the priority adjustment target, thereby achieving overall system optimization.

[0093] Subsequently, the adjustment threshold is determined based on the user comfort threshold. This means setting the upper and lower limits of the adjustment based on the range of comfortable temperature changes that users can accept during the control process. The comfort threshold represents the range of human sensitivity to temperature fluctuations. For example, most users can accept a change range of ±1 degree Celsius. Therefore, the adjustment threshold should also be limited to this range to ensure that excessive adjustments do not affect the living experience. This not only protects the user's comfort but also avoids increased energy consumption of the equipment due to frequent adjustments.

[0094] Within the adjustment threshold, the system prioritizes adjusting the air conditioning target and outputs a group control strategy. This means that, under the condition of meeting comfort requirements, it specifically executes temperature control adjustments for the corresponding rooms and generates corresponding start-stop control schemes. The group control strategy refers to a set of paired on / off states and operating parameters of the compressor and indoor units. For example, it activates the main compressor and simultaneously turns on the indoor units of rooms B and C, achieving overall energy efficiency optimization by increasing the fan speed in room B and decreasing the cooling capacity in room C. After the group control strategy is output, it will be executed by the system to guide hardware-level actions.

[0095] The output group control strategy can also be implemented using the following method (the second formulation method), which specifically includes: extracting load feature vectors from multiple rooms based on air conditioning operation data; clustering the load feature vectors to obtain load clusters; generating corresponding group control strategy templates for the load clusters; matching the current load clusters in actual operation to obtain the current group control strategy template; and determining the group control strategy based on the current group control strategy template.

[0096] Specifically, extracting load feature vectors from multiple rooms based on air conditioning operation data refers to collecting air conditioning operation data from different rooms at different time periods, such as set temperature, actual temperature, compressor operating frequency, indoor unit fan speed, and indoor-outdoor temperature difference, to construct a multi-dimensional data structure reflecting the heat load characteristics of each room. The feature vector here represents the load status of each room at a certain time period using several key values. For example, the load feature vector of a room at 10:00 AM could include dimensions such as the current load of 2.0 kW, a temperature difference of 5 degrees Celsius, and a fan speed of 0.8 meters per second. This vectorization is the foundation for subsequent intelligent analysis and control.

[0097] Next, the load feature vectors are clustered to obtain load clusters. This refers to using clustering algorithms to group rooms with similar characteristics into the same group for unified control. Clustering is an unsupervised learning technique used to identify naturally occurring grouping relationships in data. A load cluster is a set of multiple rooms with similar load characteristics within the same time period. For example, in the morning, south-facing rooms generally experience high loads due to direct sunlight and can be classified into a high-load cluster, while rooms in corridors or on the north side may be classified into a low-load cluster due to a more stable environment.

[0098] Next, corresponding group control strategy templates are generated for the load clusters. This refers to pre-setting matching air conditioning group control methods based on the characteristics of different load clusters. These strategy templates include equipment start-stop logic, compressor frequency allocation, and fan speed adjustment strategies. For example, for high-load clusters, the group control template may set the compressor to run at high frequency and the indoor unit to operate at maximum fan speed; while for low-load clusters, it recommends intermittent compressor operation and maintaining low fan speed in the indoor unit, thereby achieving energy-saving control.

[0099] Subsequently, in actual operation, the current load cluster is matched to obtain the current group control strategy. This means that in each control cycle, the latest load characteristic vector of each room is analyzed in real time and compared with the existing load clusters to determine which cluster the current state should belong to. Then, the corresponding strategy template is retrieved and applied. For example, at 3 pm, the loads of rooms A and B are 2.5 kW and 2.8 kW respectively, which highly match the high load cluster, so the corresponding high load group control strategy is applied.

[0100] Finally, determining the group control strategy based on the current group control strategy means issuing specific control commands to the compressors and indoor units according to the matched group control strategy, deciding which devices are activated, which are temporarily deactivated, and their operating parameter configurations. The group control strategy represents the optimal set of devices to operate in the current state, such as activating 2 main compressors, shutting down 1 auxiliary compressor, and adjusting 3 out of 4 indoor units to operate at medium fan speed.

[0101] The method of extracting load feature vectors from air conditioning operation data and generating group control strategies, and the method of generating group control strategies based on heat load change models, are independent in theoretical foundation and application logic, but in practical systems they are often coordinated through hierarchical control or hybrid decision-making. Upper-level decision-making: The heat load model is used for global planning (such as chiller plant start-up combinations and ice storage scheduling) to determine the total cooling capacity demand of the system. Lower-level execution: The load cluster strategy is used for local optimization (such as terminal airflow allocation and temperature setting) to match the real-time needs of specific rooms. The two methods can be used in conjunction or independently.

[0102] The two approaches are both independent and complementary. Their independence lies in the fundamental difference between data-driven and model-driven approaches; their complementarity lies in the collaborative optimization of policy generation and the improvement of system robustness. The future development direction may be hybrid intelligent control that integrates physical models and machine learning.

[0103] This embodiment optimizes the group control strategy based on user-specific data to generate a personalized control strategy.

[0104] Specifically, by leveraging user habits during air conditioner use, such as frequently used on / off times, preferred temperature settings, and daily usage patterns, we can reflect users' actual needs and preferences. Optimizing group control strategies means adjusting the air conditioner's operating parameters and control methods by analyzing individual data to better suit users' habits and comfort requirements. The optimized adjustment schemes are then concretized to create customized control plans for different users, thereby improving user experience and energy efficiency.

[0105] Furthermore, the present invention also includes: acquiring user's habitual data on power on / off and temperature settings; combining sunshine and wind speed from weather forecasts with the habitual data and mapping it to a group control strategy to generate a personalized control strategy.

[0106] Specifically, acquiring user habits regarding power on / off and temperature settings means monitoring and recording specific user behaviors when using air conditioning, such as the time of day when the air conditioner is turned on and off, and the temperature settings. This reflects the user's daily routine and preferences. For example, a user might habitually turn on the air conditioner at 7 a.m., set the temperature to 26 degrees Celsius, and turn it off at 10 p.m. This habitual data provides the basis for subsequent intelligent control.

[0107] Next, combining weather forecasts regarding sunshine and wind speed with user habits and mapping them to the group control strategy involves analyzing meteorological information such as solar radiation intensity and airflow speed in conjunction with user habits. Sunlight intensity affects the natural variation of indoor temperature, while wind speed affects the efficiency of indoor and outdoor heat exchange. By combining these environmental factors with user behavior data, air conditioning load demand can be predicted more accurately, leading to corresponding group control strategies. For example, if the weather forecast indicates that the sunshine intensity is as high as 800 watts per square meter and the wind speed is low, the system will consider starting the air conditioning earlier and adjusting the set temperature to maintain comfort.

[0108] Based on a comprehensive analysis of user habits and weather effects, a customized air conditioning control scheme was developed for this user and the current environment. This strategy not only meets the user's comfort needs but also considers energy efficiency. For example, during prolonged periods of high temperatures in summer, the air conditioning on / off times and temperature settings are adjusted according to user habits and weather forecasts, resulting in a reduction of energy consumption by more than 10% while maintaining indoor comfort.

[0109] Furthermore, the present invention also includes: recording the control path under abnormal energy consumption; comparing the control path with the standard energy consumption curve to identify abnormal points; extracting the external interference that caused the abnormal points, suggesting feasible adjustment and repair measures, and visually displaying the backtracking on the user interface.

[0110] Specifically, recording control paths under abnormal energy consumption means automatically saving all relevant control operation steps and parameter change trajectories during the period when the energy consumption of the air conditioning system exceeds the normal range. The control path reflects the specific adjustment process under abnormal conditions, including multiple control behaviors such as group control strategy adjustment, frequency changes, and valve opening, facilitating subsequent analysis of the cause of the anomaly.

[0111] Subsequently, the control path is compared with the standard energy consumption curve to identify anomalies. This involves comparing the energy consumption changes corresponding to the actual control path with the pre-set standard energy consumption curve to find key time points or operational stages where there are significant deviations, abnormal increases, or decreases in energy consumption. The standard energy consumption curve represents the energy consumption trend of the equipment under ideal or normal operating conditions. By comparing these curves, the location and time period of anomalies can be quickly pinpointed.

[0112] Next, the external interferences causing the anomalies are extracted, feasible adjustment and remediation measures are suggested, and a visual retrospective display is provided on the user interface. This involves analyzing the possible causes behind the anomalies, including changes in the external environment such as sudden temperature changes, power grid fluctuations, or equipment failures and operational errors. The analysis method involves comparing the control path with the standard energy consumption curve to identify energy consumption deviations, and then combining this with recorded control operations and environmental data to infer the external interference factors causing the anomalies.

[0113] Based on the analysis results, specific adjustment suggestions are generated, such as adjusting the compressor frequency or resetting the group control strategy. Simultaneously, the abnormal control path and repair suggestions are displayed graphically and on a timeline in the user interface, facilitating user understanding and tracking of the problem-solving process. Table 1 shows a partial control record of the most recent energy consumption anomaly.

[0114] Table 1: Partial Control Records of the Most Recent Energy Consumption Anomaly

[0115] Time point Control parameters Actual energy consumption (kilowatt-hours) Standard energy consumption (kWh) Energy consumption difference (kWh) Anomaly marker External interference factors Recommended adjustment measures Remark 08:00 The compressor frequency was adjusted to 60Hz. 12.5 11.0 +1.5 no none none Normal operation 08:15 Adjust the valve opening to 30%. 13.8 11.5 +2.3 yes The outside temperature rose by 5 degrees Celsius Increase ventilation and lower room temperature Abnormal energy consumption began to appear 08:30 Group control strategy adjustment 14.2 11.2 +3.0 yes The number of people indoors increased by 20%. Adjust the group control strategy to reduce compressor load. Abnormal peak value 08:45 The compressor frequency drops to 50Hz 11.8 11.0 +0.8 no External temperature drops Maintain existing settings Energy consumption has returned to normal. 09:00 Adjust the valve opening to 25%. 11.2 10.8 +0.4 no none none Normal operation

[0116] In summary, the energy consumption characteristic calibration method for multi-split air conditioning systems provided by this invention has the following technical effects: by achieving the technical goal of accurate calibration and dynamic group control optimization of the energy consumption characteristics of multi-split air conditioning systems based on real-time data and intelligent models, it achieves the technical effects of improving system energy efficiency, reducing energy consumption fluctuations, enhancing user comfort, and realizing refined zone control.

[0117] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for controlling a multi-split air conditioner with calibrated energy consumption characteristics, characterized in that, The energy consumption characteristic calibration method for multi-split air conditioning control includes: The energy consumption mapping dataset is generated by acquiring air conditioner operation data from multiple time periods, and an energy consumption response surface is constructed based on the energy consumption mapping dataset. Based on the energy consumption mapping dataset and the current energy consumption response surface, determine whether the current energy consumption response surface needs to be corrected. If no correction is needed, do not trigger the group control strategy, and continue to determine whether correction is needed. If correction is needed, the corresponding group control strategy will be triggered based on the current energy consumption response surface, and the energy consumption response surface will be corrected according to the real-time measured air conditioning operation data. The corrected energy consumption response surface is repeatedly judged to determine whether further correction is needed until the air conditioner is turned off; in, The step of correcting the energy consumption response surface based on real-time measured air conditioning operation data includes: Real-time reading of evaporation temperature, superheat, and return gas pressure; The compressor frequency in the current group control strategy is obtained based on the load change rate, and the priority adjustment objects are adjusted accordingly. Based on the cooling demand and operating status of each air conditioner in the sub-loop, the valve opening in the current group control strategy is obtained and the priority adjustment object is adjusted. Calculate the changes in evaporation temperature, superheat, and return gas pressure readings before and after adjusting the compressor frequency and valve opening. The energy consumption optimization target is determined by the changes in the readings of evaporation temperature, superheat, and return gas pressure. If so, the energy consumption response surface is corrected based on the current evaporation temperature, superheat, and return gas pressure. If not, the compressor frequency and valve opening are adjusted, and the energy consumption optimization target is reassessed. The adjustment of the priority adjustment object includes: Obtain the thermal imbalance difference between the set temperature and the actual temperature of each room, and calculate the impact of the thermal imbalance difference on the load distribution of the entire air conditioning system, expressed by the formula: ΔL=Σ(w_i×ΔT_i) Where w_i is the heat capacity weight of the i-th room, ΔT_i is the heat imbalance difference of the i-th room, and ΔL is the load distribution influence. The room with the largest product of heat capacity weight and the difference in heat imbalance is selected as the priority for adjustment. Use the preset user comfort threshold as the adjustment threshold; Adjust the air conditioner of the priority target within the adjustment threshold.

2. The multi-split air conditioning control method for energy consumption characteristic calibration according to claim 1, characterized in that, The process of acquiring air conditioning operation data over multiple time periods to generate an energy consumption mapping dataset, and constructing an energy consumption response surface based on the energy consumption mapping dataset, includes: Obtain air conditioner operation data for multiple time periods; Based on the air conditioning operation data, operating parameters and unit energy consumption are extracted and mapped to generate an energy consumption mapping dataset. Construct an energy consumption regression model with energy consumption ratio as the target variable based on the energy consumption mapping dataset; The energy consumption response surface is obtained through an energy consumption regression model.

3. The multi-split air conditioning control method for energy consumption characteristic calibration according to claim 1, characterized in that, The step of determining whether the current energy consumption response surface needs correction based on the energy consumption mapping dataset and the current energy consumption response surface includes: Based on the difference between the air conditioner operation data in the energy consumption mapping dataset and the predicted air conditioner operation values ​​predicted by the energy consumption response surface, a prediction bias vector is constructed. Determine whether the energy consumption response surface has entered the degradation range based on the prediction deviation vector; If the current energy consumption response surface is in the degradation range, it needs to be corrected; if it is not in the degradation range, it does not need to be corrected.

4. The multi-split air conditioning control method for energy consumption characteristic calibration according to claim 3, characterized in that, The prediction bias vector is constructed based on the difference between the air conditioner operation data in the energy consumption mapping dataset and the predicted air conditioner operation values ​​predicted by the energy consumption response surface, including: Calculate the numerical deviation between the air conditioning operation data of multiple dimensions in the energy consumption mapping dataset and the predicted values ​​of the corresponding dimensions under the corresponding conditions of the energy consumption response surface prediction. The numerical deviations of each dimension are used to construct a prediction deviation vector.

5. The multi-split air conditioning control method for energy consumption characteristic calibration according to claim 1, characterized in that, The methods for formulating the corresponding group control strategy include: First formulation method: Construct heat load variation models for multiple rooms based on the energy consumption response surface; A group control strategy is formulated based on the heat load change model and the equipment capacity boundary, wherein the equipment capacity boundary is the upper and lower limits of the operation of each air conditioning unit. Second formulation method: Load feature vectors of multiple rooms are extracted based on air conditioning operation data; Clustering the load feature vectors yields load clusters; Generate a corresponding group control strategy template for the load cluster; match the current load cluster during actual operation to obtain the current group control strategy template; Determine the group control strategy based on the current group control strategy template; The first formulation method and the second formulation method are used in conjunction.

6. The multi-split air conditioning control method for energy consumption characteristic calibration according to claim 1, characterized in that, The method of determining whether the adjustment behavior has achieved the energy consumption optimization target based on changes in evaporation temperature, superheat, and return gas pressure readings includes: If the change in evaporation temperature is greater than or equal to 0.5℃, the change in superheat is between -1K and 0K, and the change in return gas pressure is less than or equal to 5%, then the regulation behavior is judged to have achieved the energy consumption optimization target. Otherwise, it is determined that the adjustment behavior has not achieved the energy consumption optimization goal.

7. The multi-split air conditioning control method for energy consumption characteristic calibration according to claim 1, characterized in that, Also includes: Acquire user's power-on / off habits and temperature settings; By combining sunshine and wind speed data from weather forecasts with the aforementioned habitual data and mapping them to the group control strategy, a personalized control strategy is generated. Replace the group control strategy with a personalized control strategy.

8. The multi-split air conditioning control method for energy consumption characteristic calibration according to claim 1, characterized in that, Also includes: Energy consumption is calculated based on real-time measured air conditioning data, and the control path is recorded when energy consumption exceeds the normal range; By comparing the control path with the standard energy consumption curve, anomalies can be identified. Extract external interference information that causes anomalies, and generate adjustment and repair measures based on the external interference information; The regulatory pathways and adjustment and repair measures will be visualized and retrospectively displayed.

Images