A central air conditioning system group control energy-saving system based on load prediction and multi-objective optimization

Abstract

The present application relates to air conditioner control technical field, disclose a kind of central air conditioning system group control energy-saving system based on load prediction and multi-objective optimization, comprising: multidimensional data perception module, heat load prediction module, multi-objective collaborative optimization module and group control execution module.Multidimensional data perception module real-time acquisition multidimensional data data;Heat load prediction module is fed forward by calculating temperature rise rate Prediction;Multi-objective collaborative optimization module constructs nonlinear outdoor unit efficiency coupling model, and utilizes rolling horizon strategy to solve multi-objective optimization function, exports global optimal compressor operating frequency and each indoor unit electronic expansion valve's opening degree;Execution module is according to security boundary check strategy to frequency change rate smooth processing and issues instruction to execution mechanism.The present application overcomes control lag problem caused by thermal inertia by feedforward prediction mechanism, realizes on-demand cooling by global collaborative optimization mechanism, reduces energy consumption and prolongs equipment life while ensuring comfort.

Description

Technical Field

[0001] This invention relates to the field of air conditioning control technology, specifically to a group control energy-saving system for central air conditioning systems based on load forecasting and multi-objective optimization. Background Technology

[0002] With the rapid development of modern building technology, VRF / VRV central air conditioning systems have become widely used in commercial office buildings, hotels, and large public buildings due to their advantages such as flexible layout, space saving, and convenient construction. In actual operation, the energy consumption of air conditioning systems accounts for a huge proportion of the total energy consumption of a building. Therefore, improving the operating efficiency and control accuracy of the system has always been a research hotspot in this field.

[0003] In existing technical solutions, traditional central air conditioning multi-split system control systems mainly rely on indoor unit return air temperature sensors for PID (proportional-integral-derivative) feedback regulation. The control logic is typically as follows: when the indoor sensor detects a deviation between the return air temperature and the user-set target temperature, the controller calculates the required adjustment amount using a PID algorithm, and then adjusts the opening of the indoor unit's electronic expansion valve or sends a refrigerant flow request to the outdoor unit in an attempt to bring the room temperature back to the set value.

[0004] However, the aforementioned traditional control methods have some limitations in practical applications, mainly in the following aspects: First, due to the significant thermal inertia of the building envelope (such as walls, floors, and ceilings), and the high randomness and time-varying nature of load changes caused by indoor personnel movement and equipment operation, the indoor thermal environment often deteriorates by the time the sensor detects that the indoor temperature deviates from the set value. When the air conditioning system intervenes to adjust at this point, it not only needs to overcome the sensible heat load of the air but also eliminate the heat already accumulated in the building envelope, resulting in severe adjustment lag and a tendency for temperature overshoot or oscillation, affecting user thermal comfort. Second, the existing control architecture lacks a system-level global coordination mechanism. In conventional systems, each indoor unit typically operates independently, sending refrigerant requests to the outdoor unit independently based on the temperature control needs of its respective area. However, due to the lack of global coordination, the outdoor unit, as a passive actuator, often needs to respond to multiple scattered and disordered load requests in a short period, causing the outdoor unit compressor to frequently operate in an inefficient frequency range or to start and stop frequently under suboptimal conditions. This operating mode not only fails to leverage the high energy efficiency of variable frequency compressors at specific frequencies, resulting in significant energy waste, but also increases mechanical wear and shortens equipment lifespan. Therefore, a central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization is urgently needed to solve these problems. Summary of the Invention

[0005] To address the problems in related technologies, this invention provides a central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization, thereby overcoming the aforementioned technical problems in existing related technologies.

[0006] To solve the aforementioned technical problem, the present invention is achieved through the following technical solution: In a first aspect, embodiments of the present invention provide a group control energy-saving system for a central air conditioning system based on load forecasting and multi-objective optimization, specifically including: a multi-dimensional data sensing module, used to collect real-time data on indoor temperature, outdoor weather, and indoor occupancy status in various indoor areas, and to filter the collected data; a heat load forecasting module, connected to the multi-dimensional data sensing module, used to calculate the total heat load disturbance by combining the real-time collected multi-dimensional data, and to calculate the temperature rise rate based on the regional dynamic heat load balance equation; a multi-objective collaborative optimization module, connected to the heat load forecasting module, used to construct a nonlinear outdoor unit efficiency coupling model to quantify energy consumption characteristics, and to use the temperature rise rate to solve a multi-objective optimization function containing energy consumption indicators, global comfort, and equipment operation losses based on a rolling time domain strategy, outputting the optimal compressor operating frequency and the optimal opening degree of the electronic expansion valve of each indoor unit; and a group control execution module, connected to the multi-objective collaborative optimization module, used to receive the optimal compressor operating frequency and the optimal opening degree of the electronic expansion valve of each indoor unit, and after smoothing the frequency change rate according to a preset equipment safety boundary verification strategy, to issue instructions to the execution mechanism of the central air conditioning equipment.

[0007] As a preferred embodiment of the central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization described in this invention, the temperature rise rate calculated based on the regional dynamic heat load balance equation is as follows: ; In the formula, The specific heat capacity of air, For the first Air quality in each region For a moment The actual cooling capacity provided by the air conditioner This represents the total heat load disturbance. For a moment No. Indoor temperature in the area This represents the rate of temperature rise.

[0008] As a preferred embodiment of the central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization described in this invention, the total heat load disturbance amount Heat transfer from the building envelope Window heat gain from sunlight Human body heat dissipation and equipment heating The composition, and its calculation formula are as follows: ; Among them, the heat transfer of the enclosure structure It is the sum of the heat transfer of all the building envelope surfaces in the region. For any building envelope surface, its heat transfer is the product of the heat transfer coefficient of the building envelope surface, the area of ​​the building envelope surface, and the temperature difference. The window's solar heat gain It is the product of the window's overall solar heat gain coefficient, the solar irradiance acting on the window's outer surface at the current moment, the window's effective light-receiving area, the geometric direction correction coefficient, and the dynamic shading coefficient. Human body heat dissipation It is the product of the real-time estimated number of people in the area at the current moment and the average sensible heat dissipation per person; The device generates heat. It is the sum of the heat dissipation of all electrical equipment in the area. For any electrical equipment, its heat dissipation is the product of the real-time electrical power of the equipment and the heat conversion coefficient of the electrical power into indoor heat.

[0009] As a preferred embodiment of the central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization described in this invention, the nonlinear outdoor unit efficiency coupling model is as follows: ; In the formula, For compressor power, For the compressor operating frequency, Outdoor ambient temperature The evaporation temperature. , , , These are the fitting coefficients based on device characteristics.

[0010] As a preferred embodiment of the central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization described in this invention, the multi-objective optimization function is as follows: ; In the formula, For a multi-objective optimization function, To predict the field of view, For a moment compressor power, The rated power of the system, , , These are the weighting coefficients for energy consumption indicators, overall comfort penalties, and equipment motion loss penalties, respectively. For the total number of regions, For a moment No. Priority coefficients for each region For the first The set temperature for each area This is the temperature tolerance constant. For a moment The penalty value for equipment operation wear.

[0011] As a preferred embodiment of the central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization described in this invention, the weighting coefficients... , , A state-aware normalized evaluation method is used for real-time calculation, and the calculation formula is as follows: ; In the formula, The energy efficiency deviation index corresponds to ; The environmental severity index corresponds to ; For load fluctuation index, corresponding .

[0012] As a preferred embodiment of the central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization described in this invention, wherein the time... No. Priority coefficient of each region As shown below: ; In the formula, For functional normalization factor, For a moment No. Number of people in each region Design the area to accommodate the maximum number of people. This is the lower limit value determined based on the minimum maintenance requirements of the equipment.

[0013] As a preferred embodiment of the central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization described in this invention, the multi-objective collaborative optimization module performs rolling optimization based on a discretized state prediction equation when solving the multi-objective optimization function based on a rolling time-domain strategy. The discretized state prediction equation is as follows: ; In the formula, To predict the first in the time domain The region is Predicted indoor temperature at any given time For the sampling time step, This is the pipeline transmission efficiency coefficient. For the first The region is Predicted total heat load disturbance at any given time. For the outdoor unit in The total cooling capacity provided is based on the operating frequency at all times. For the first The weighting coefficient of the actual cooling capacity allocated to the indoor unit of the table; The solution process must satisfy the following physical constraints: ; in, To predict within the time domain The target operating frequency of the compressor at any given time. To predict within the time domain Time of the first The target opening degree of the electronic expansion valve of the indoor unit of the Taiwan Strait. , For compressor frequency limits, , This refers to the expansion valve opening limit. This is the current air dew point temperature.

[0014] As a preferred embodiment of the central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization described in this invention, the equipment safety boundary verification strategy must satisfy the following: ; In the formula, The optimal compressor operating frequency is obtained by solving the multi-objective optimization function. This refers to the maximum permissible frequency variation rate specified by the compressor manufacturer.

[0015] Secondly, embodiments of the present invention provide a central air conditioning multi-split unit, specifically including: an outdoor unit, an indoor unit, a refrigerant transmission pipeline, and a central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization. The outdoor unit includes a variable frequency compressor and an outdoor heat exchanger, the indoor unit includes an indoor heat exchanger and an electronic expansion valve, the refrigerant transmission pipeline is used to connect the outdoor unit and the indoor unit, and the central air conditioning system group control energy-saving system is used to control the operating frequency of the variable frequency compressor and the opening degree of the electronic expansion valve.

[0016] The present invention has the following beneficial effects: 1. This invention calculates the total heat load disturbance and the temperature rise rate based on the regional dynamic heat load balance equation, thereby determining when to intervene before a significant temperature rise occurs. Specifically, by combining multi-dimensional sensing data such as heat transfer from the building envelope, occupant density, solar irradiance, and equipment power, the system can predict temperature change trends over a future period before actual indoor temperature fluctuations occur. Based on this feedforward prediction mechanism, the system can pre-schedule the outdoor unit frequency and indoor unit valves, thereby eliminating control lag caused by the thermal inertia of the building envelope, such as walls, avoiding drastic fluctuations and overshoot in indoor temperature, and ensuring a stable and comfortable indoor thermal environment.

[0017] 2. This invention introduces a nonlinear outdoor unit efficiency coupling model and a multi-objective collaborative optimization mechanism, so that the system no longer simply responds to the needs of a single indoor unit, but takes a global perspective, comprehensively calculates the needs of all areas, and solves the optimal combination of compressor operating frequency and electronic expansion valve opening. This allows the compressor to always be guided to operate within the high-efficiency frequency range, avoiding frequent start-stops or lingering in the low energy efficiency ratio range due to load fluctuations, thereby achieving secondary energy-saving effects.

[0018] 3. This invention, by introducing a regional priority coefficient and a temperature tolerance constant, changes the traditional one-size-fits-all control approach. By identifying the functional attributes and real-time occupancy of different areas, when cooling resources are limited or extreme energy conservation is desired, it automatically allocates cooling resources to densely populated and temperature-sensitive core areas, while appropriately relaxing control requirements for transitional areas. This differentiated control strategy ensures a high-quality environment in critical areas while maximizing the energy-saving potential of non-critical areas.

[0019] 4. This invention introduces a device motion loss penalty mechanism into the multi-objective optimization function, so that when the system generates control commands, it not only considers temperature control and energy saving, but also endows the control commands with characteristics similar to inertial damping through the algorithm, smoothing the rate of change of compressor frequency, which helps to prevent drastic equipment adjustment caused by small load fluctuations, reduces mechanical wear of variable frequency compressors, and thus extends the service life and maintenance cycle of the entire air conditioning system.

[0020] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This invention provides a schematic diagram of a group control energy-saving system for central air conditioning systems based on load forecasting and multi-objective optimization.

[0023] Figure 2 The flowchart illustrates a group control energy-saving method for central air conditioning systems based on load forecasting and multi-objective optimization, as provided by this invention.

[0024] Figure 3 This is a schematic diagram of a central air conditioning multi-split unit provided by the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Example 1 Traditional VRF / VRV central air conditioning control systems primarily rely on indoor unit return air temperature sensors for PID feedback regulation. However, due to the significant thermal inertia of building envelopes and the randomness of load changes caused by occupant movement, by the time the sensor detects a deviation from the setpoint, the indoor thermal environment has often already deteriorated. Furthermore, each indoor unit independently requests refrigerant from the outdoor unit, lacking global coordination, causing the outdoor unit compressor to frequently operate within an inefficient frequency range, resulting in energy waste.

[0027] To solve the above technical problems, such as Figure 1 As shown, Embodiment 1 of the present invention provides a central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization. It performs spatiotemporal forecasting from the load demand side, pre-schedules the outdoor unit frequency before temperature fluctuations occur, and coordinates the valve opening of each indoor unit through multi-objective optimization to achieve secondary energy savings. Specifically, it includes: a multi-dimensional data sensing module, a heat load forecasting module, a multi-objective collaborative optimization module, and a group control execution module.

[0028] Specific Implementation Example 1: A commercial office building's ground floor office area is equipped with a variable frequency multi-split air conditioning system, comprising one outdoor unit (rated cooling capacity 45kW) and five indoor units (serving one large conference room, three individual offices, and one public corridor, respectively). The system collects indoor temperatures in each area via a ZigBee IoT sensor network. The target summer temperature for the office area is set at 26℃, with a temperature tolerance deviation of ±1.0℃.

[0029] In the specific implementation of Embodiment 1 above, firstly, the multi-dimensional data sensing module collects indoor temperature, outdoor meteorological data, and personnel occupancy status in each area in real time, and filters and extracts features from the original signals through data fusion technology. This method constructs a multi-dimensional environmental sensing network, providing accurate real-time boundary conditions for subsequent physical modeling and avoiding decision-making errors caused by single sensors or noise interference. Secondly, the heat load prediction module calculates the total heat load interference and calculates the temperature rise rate based on the regional dynamic heat load balance equation, thereby determining to intervene in advance before the temperature rise becomes significant. This method uses physical mechanisms to quantify the thermal inertia of the building envelope and the lag effect of personnel flow on room temperature, realizing a shift from reactive post-event adjustment to proactive pre-event defense control logic, and can effectively eliminate temperature overshoot. Then, the multi-objective collaborative optimization module introduces a nonlinear outdoor unit efficiency coupling model. Under the premise of satisfying comfort constraints, a rolling time-domain strategy is used to solve the multi-objective optimization function that includes energy consumption, comfort, and equipment losses. This method breaks through the limitations of traditional single-unit independent control. By globally coordinating the calculation of the compressor's highest energy efficiency point and the optimal opening degree of the electronic expansion valve, it avoids the compressor from fluctuating in the inefficient zone and achieves a system-level secondary energy saving and dynamic balance of comfort. Finally, the group control execution module receives the optimal control command and, based on the preset equipment safety boundary verification strategy, smooths the frequency change rate before issuing the command for execution. This method, while implementing the energy-saving optimization strategy, also constructs a mechanical protection barrier, effectively preventing the risk of liquid slugging or power grid impact caused by drastic load fluctuations, and extending the overall service life of the air conditioning system.

[0030] Furthermore, to better illustrate the technical solution of Embodiment 1 of the present invention, such as... Figure 2 As shown, this paper describes a group control energy-saving method for central air conditioning systems based on load forecasting and multi-objective optimization, and provides a detailed explanation of the air conditioning control system, including the following: S1. Calculate the total heat load disturbance and, based on the regional dynamic heat load balance equation, calculate the temperature rise rate to determine when to intervene before the temperature rise becomes significant. The system no longer relies solely on the return air temperature but establishes a system for each indoor zone. The dynamic thermal equilibrium model, used to quantify the driving force of environmental factors on temperature changes, specifically includes the following sub-steps: S11, Real-time acquisition of the first Indoor temperature in the area Outdoor temperature and solar radiation The sampling period is Furthermore, to eliminate sensor noise, the collected data needs to be filtered. For example, a physical defense boundary can be constructed by using amplitude limiting filtering to eliminate non-physical anomalies caused by signal transmission packet loss or instantaneous voltage fluctuations (such as drastic temperature jumps within 1 second), preventing such outliers from compromising the stability of subsequent prediction models. The low-pass characteristics of a first-order inertial filter can be used to simulate thermal inertial physical processes, attenuating high-frequency jitter components and preserving the true low-frequency trends of environmental parameter changes.

[0031] S12. Calculate the temperature rise rate based on the dynamic heat load balance equation. Its expression is: ; In the formula, The specific heat capacity of air (take) ), For the first Air quality in each region The actual cooling capacity provided by the air conditioner This represents the total heat load disturbance. The rate of temperature rise; Among them, the Air quality in each region Represented as: , air density, For the first Air volume in each region For a moment Effective volume correction factor; For example, in this embodiment 1, time Effective volume correction factor A dynamic fusion evaluation method based on multi-dimensional sensor data is employed to determine the impact factors, specifically through real-time linkage between indoor carbon dioxide concentration sensors and micro-differential pressure sensors. The system first calculates the pressure difference influence factor and the concentration dilution influence factor separately: for the pressure difference dimension, it monitors the indoor and outdoor air pressure difference in real time. When a significant decrease in pressure difference is detected and approaches zero, it determines that the area's airtightness is compromised (e.g., doors and windows are open), thus generating a corresponding pressure difference correction value. For the concentration dimension, it analyzes the rate of change of carbon dioxide concentration. When a sharp decrease in concentration due to non-mechanical ventilation is identified and the rate exceeds the natural settling threshold, it determines that a large influx of external fresh air exists, generating a corresponding concentration dilution correction value. Finally, the system performs a weighted fusion of the two influence factors according to a preset weight ratio, dynamically outputting a correction coefficient reflecting the actual airtightness and connectivity of the current space. This corrects in real-time the deviation from the building's designed volume to the actual effective heat capacity volume caused by actions such as opening and closing doors and windows, ensuring the physical accuracy of the temperature rise rate calculation.

[0032] Specifically, total heat load disturbance Heat transfer from the building envelope Window heat gain from sunlight Human body heat dissipation and equipment heating Composition, which is represented as: .

[0033] For example: Heat transfer in building envelope Used to describe the heat transferred into a room through opaque enclosures such as walls and roofs, it represents the sum of the heat transfer from all enclosure surfaces within a region. For any given enclosure surface, its heat transfer is the product of the heat transfer coefficient of that surface, the area of ​​that surface, and the temperature difference, which can be expressed by the following formula: ; In the formula, For the first The total number of exterior walls and roof surfaces involved in each area; For the first The heat transfer coefficient of each building envelope surface; For the first The area of ​​each enclosure structural surface; For a moment No. Measured indoor temperatures in each area; For the first The thermal delay time of an enclosure structure, i.e. the time required for outdoor heat to penetrate the wall; For a moment Outdoor composite temperature, composed of outdoor temperature With solar radiation The heating effect on the building's exterior surface is a combination of these factors.

[0034] Window heat gain from sunlight The product of the window's overall solar heat gain coefficient, the solar irradiance acting on the window's outer surface at the current moment, the window's effective light-receiving area, the geometric direction correction factor, and the dynamic shading factor can be expressed by the following formula: ; In the formula, The overall solar heat gain coefficient of the form; For a moment The intensity of solar radiation acting on the outer surface of the window was obtained from meteorological station data; For the first Each area corresponds to the effective light-receiving area of ​​the window; For a moment No. The geometric direction correction factor for each window is used to convert horizontal radiation into actual radiation received by the facade. For a moment No. The dynamic occlusion coefficient of each window, with a value range of... For example, the geometric direction correction factor The calculation is based on the principle of spherical geometric projection. The system first calculates the real-time solar altitude angle and solar azimuth angle based on time and geographical coordinates, and then combines the results with the calculation of the second solar altitude angle and solar azimuth angle. The fixed installation orientation of each window is determined by calculating the cosine of the angle between the incident solar ray and the normal to the window surface using trigonometric functions. This converts the total horizontal radiation intensity projection provided by the weather station into the direct radiation component that actually acts vertically on the window facade. A diffuse radiation correction constant is then added to account for the effects of diffuse reflection. The dynamic shading coefficient... Based on pre-set environmental occlusion models in various directions, such as the elevation angle map of surrounding buildings generated by 3D scanning, the system compares the current solar altitude angle with the obstacle elevation angle threshold in real time. If the solar altitude is lower than the obstacle elevation angle, it is determined to be an occlusion state. The system assigns corresponding radiation attenuation weights (such as 0 or semi-transmittance coefficient) according to the type of obstacle (such as solid buildings or sparse vegetation), thereby achieving accurate quantification of local micro-environment lighting conditions.

[0035] Human body heat dissipation The product of the estimated number of people in the area at the current moment and the average sensible heat dissipation per person can be expressed by the following formula: ; In the formula, The average sensible heat loss of an adult male; To estimate the number of people, an infrared counter can be used. However, to eliminate false detections caused by a single infrared counter due to non-human heat sources such as printer heat, sunlight on the ground, or movement of non-target objects such as swaying curtains or running pets, the system uses a low-resolution infrared thermopile array sensor instead of a traditional pyroelectric probe. First, by analyzing the temperature distribution matrix of the heat source, it sets the response range to only correspond to human body temperature ranges (e.g., ...). Furthermore, it identifies thermal images of targets with typical human body contours, thus physically filtering out non-human interference sources that are too high or too low in temperature. Simultaneously, at the algorithm level, a time-sliding window and micro-motion feature extraction logic are introduced. This requires the target to maintain spatial continuity over several consecutive sampling periods and detect characteristic frequencies generated by breathing or limb micro-movements, thereby eliminating fleeting noise or static heat sources and ensuring accurate headcount. Robustness and accuracy.

[0036] Equipment heating This represents the total heat dissipation of all electrical equipment within the area. For any given electrical equipment, its heat dissipation is the product of its real-time electrical power and the heat conversion coefficient from electrical power to indoor heat, which can be expressed by the following formula: ; In the formula, For device index number; For the first Individual electrical equipment at all times Real-time electrical power; For the first The heat conversion coefficient of electrical power from individual electrical equipment to indoor heat.

[0037] Specifically, for example: suppose a large conference room ( air quality The Internet of Things senses the current moment The conference room's occupancy density has surged (20 people entering). Based on outdoor weather data, the total heat load disturbance is predicted to be approximately [amount missing] within the next 15 minutes. If the air conditioner is not activated at this moment (i.e.) The rate of temperature rise is obtained from the dynamic heat load balance equation: ; Converted to minutes Therefore, it rises. It takes about 3.9 minutes, and the system determines that it needs to intervene before the temperature rises significantly.

[0038] In this embodiment 1, the total heat load disturbance of each region is calculated using the monitored multi-dimensional sensing data. The system periodically calculates the rate of temperature rise using the regional dynamic heat load balance equation. The operational approach of this step is to transform traditional hysteresis-based feedback based on temperature deviation into feedforward prediction based on load disturbances. By quantifying the driving forces of environmental factors such as sudden increases in occupancy and changes in solar radiation on room temperature through a physical model, it overcomes the control lag problem caused by the thermal inertia of the building envelope from a physical mechanism perspective. This provides accurate feedforward input variables for the subsequent multi-objective collaborative optimization controller, ensuring that the system can respond in advance before a substantial deterioration in indoor temperature occurs.

[0039] S2. Construct a nonlinear outdoor unit performance coupling model.

[0040] To achieve secondary energy savings, it is necessary to quantify the nonlinear relationship between compressor frequency and power consumption, as well as the impact of ambient temperature on energy efficiency, to avoid the compressor operating in an inefficient range. Specifically: during compressor operation, the compressor operating frequency is collected in real time. Outdoor ambient temperature and evaporation temperature Construct a coupled performance model of the outdoor unit within the condensing temperature range: ; in, For compressor power, , , , These are the fitting coefficients based on device characteristics.

[0041] For example, in this embodiment 1, , , , The least squares regression method based on experimental data is used to determine this, as follows: First, before the air conditioning system is put into operation, or in an enthalpy difference laboratory, a full performance test is conducted on the outdoor unit. During the test, the compressor operating frequency is systematically varied. and outdoor ambient temperature Record the actual power consumption of the corresponding compressor. Control the compressor frequency From the lowest frequency (like (to the highest frequency) (like ), with step size (like (Increase) Control outdoor temperature. Covering the expected operating range (e.g.) to ), with step size (like (Incrementing). Data collected. A set of experimental data samples is denoted as set. .in, For the first The compressor power was measured by a power meter during this test.

[0042] Secondly, according to the formula This is transformed into a linear regression form. The input feature vector is defined. .Will The data sets are constructed in matrix form: ;in for Observed power vector: ; for Design matrix: ; For those in demand Coefficient vector: ; This is the random measurement error vector.

[0043] Finally, to minimize the sum of squared residuals between the model predictions and the measured values ​​(i.e. Solving for the coefficient vector using the normal equation. ; directly obtain through matrix operations The unique optimal solution. Considering that equipment aging can lead to characteristic drift, this system also employs recursive least squares (RLS) to fine-tune the coefficients online during actual operation, defining the time... Gain matrix Covariance Matrix The coefficient update formula is: This step ensures that the coefficients always reflect the current physical state of the equipment, rather than remaining fixed.

[0044] Specifically, for example: the parameters of a certain model of outdoor unit were measured through offline calibration. If the current outdoor temperature is constant, when the frequency At that time, power When the frequency increases to At that time, power This formula shows that the unit energy consumption cost of high-frequency operation is higher (efficiency degradation), so subsequent optimization will favor low-frequency long-term operation rather than high-frequency start-stop.

[0045] In this embodiment 1, by constructing a nonlinear outdoor unit efficiency coupling model, the system can map the theoretical power consumption of the compressor at different operating frequencies based on the real-time monitored outdoor temperature and evaporation temperature. Utilizing data fitting and online correction techniques, the complex variable operating condition energy efficiency characteristics of the outdoor unit are transformed into a calculable mathematical function, providing a quantitative energy consumption assessment basis for subsequent multi-objective optimization solutions. This clarifies the compressor's high-efficiency operating frequency range, achieving not only accurate modeling of the equipment's energy consumption characteristics but also overcoming model inaccuracies caused by equipment aging through an online parameter fine-tuning mechanism. This guides the system to consistently select the frequency point with the optimal energy efficiency ratio in global coordination.

[0046] S3. Solving multi-objective collaborative optimization based on the prediction horizon, including the following sub-steps: S31. To minimize energy consumption, maximize comfort, and reduce equipment wear within the prediction field of view, the following multi-objective optimization function is defined. : In the formula, This is a system energy consumption indicator, representing the system's energy consumption at time [time value missing]. Energy costs; For a moment The compressor power is calculated using the outdoor unit efficiency coupling model; This refers to the system's rated power. The weighting coefficient for energy consumption indicators; This is a global comfort penalty term based on user weights, representing the system's tolerance for user comfort deviations; for Time of the first Priority coefficients for each region; For the first Set temperature for each zone; This is the temperature tolerance constant; This represents the weighting coefficient for the global comfort penalty term; This is a penalty item for equipment operation wear and tear; For a moment The equipment operation wear penalty value is set to prevent drastic frequency fluctuations; The weighting coefficient for the equipment motion loss penalty term; in, Time of the first Priority coefficient of each region It can be calculated using the following formula: ; In the formula, The maximum capacity for a given area should be specified. For example, if the design drawings for a meeting room indicate a maximum capacity of 20 people, then... , express The maximum value is 1.2, which is used to prevent the system from completely cutting off power to other areas due to overloading of personnel; for Time of the first Number of people in each region; The lower limit value was determined based on the minimum maintenance requirements of the equipment and was obtained by consulting the air conditioning unit technical manual. As a functional normalization factor, the core functional areas (meeting rooms, manager's offices) are defined by consulting national standards for building functional zoning, such as the "Code for Design of Office Buildings". Auxiliary functional areas (tea room, printing room) Traffic function areas (corridors, lobby) This formula characterizes the automatic allocation of cooling capacity to densely populated, high-utilization core areas when cooling resources are limited (high... ), while automatically reducing the weight of vacant or low-density areas (low) This allows the air conditioning system to dynamically allocate cooling capacity based on actual usage intensity.

[0047] Among them, time Equipment operation loss penalty value It can be represented as: ; In the formula, for The target frequency of the compressor at any given time; For the previous moment The actual frequency of the compressor; This is a normalized reference constant for frequency variation, representing the typical frequency fluctuation amplitude; for Time of the first The target opening degree of the electronic expansion valve of the indoor unit of the table; For the previous moment No. The actual opening degree of the electronic expansion valve of the indoor unit of the table; Let be the normalized reference constant for the valve opening change, and represent the normal step size of the valve action. This formula constructs a penalty term for equipment action using a differential square form, giving the control system damping characteristics similar to an inertial element. Under non-emergency conditions, it can automatically smooth control commands, significantly reducing mechanical wear of the actuator and extending the service life of the equipment.

[0048] For example, in this embodiment 1, the temperature tolerance constant The function is determined based on the preset functional attributes of the temperature control zones. Specifically, the controller internally stores a functional zone mapping table, when the... When the scene attribute of a partition is configured as "precision control area" (such as a computer room), the system automatically extracts... As When the scene attribute is "transition area" (such as a corridor), extract... As This parameter, used as the normalization denominator, is employed to adjust the sensitivity weights of different regions to temperature deviations in the multi-objective optimization function, ensuring that temperature stability in the precision control zone and the high comfort zone is prioritized when cooling capacity is limited. The functional partition mapping table is shown below: For example, in order to respond to seasonal variations in extreme temperature differences between winter and summer and real-time load fluctuations, the weighting coefficients... , , A state-aware, normalized evaluation method is used for real-time calculation. Specifically, the system defines three orthogonal physical state indices: energy efficiency deviation index... Environmental Severity Index and load fluctuation index And based on this, real-time weights are generated.

[0049] First, calculate the energy efficiency deviation index. This index reflects the energy-saving potential under current operating conditions and determines the energy consumption weight. Using the Carnot cycle principle, the greater the deviation between the theoretical optimal COP and the actual COP at the current outdoor temperature, the higher the urgency of energy-saving optimization. This can be reflected by the following formula: ; Always This represents the system's actual cooling / heating efficiency ratio at the current moment. The theoretical limiting energy efficiency ratio, calculated based on the Carnot cycle at the current outdoor temperature, is given by the following formula for the cooling condition: .

[0050] Secondly, the severity of the computing environment This index automatically senses seasonal differences and determines the comfort weight. In extreme winter and summer weather, the temperature difference between indoors and outdoors approaches the design limit, resulting in strong thermal interference and making comfort control difficult, thus requiring a higher weighting for comfort. In transitional seasons (spring and autumn), the temperature difference is smaller, allowing for a reduction in the comfort weighting to achieve energy savings, which can be reflected in the following formula: ; In the formula, Real-time outdoor temperature; Set the average temperature for all current regions; The maximum indoor-outdoor temperature difference is specified in the design standards for HVAC systems. For example, according to the "Code for Design of Heating, Ventilation and Air Conditioning of Civil Buildings", if the design outdoor temperature in a certain area is 35℃ and the indoor temperature is 26℃ in summer, then the reference temperature difference is 9℃.

[0051] Then, calculate the load fluctuation index. This index reflects the drastic nature of load changes and determines the weighting of equipment losses. When there is drastic movement of people or changes in heat sources, the system tends to increase damping, i.e., increase the loss weight, to prevent frequent load increases and decreases in equipment. This can be reflected by the following formula: ; In the formula, Total heat load disturbance within a past time window Standard deviation; This refers to the rated cooling capacity of the unit.

[0052] Finally, based on the energy efficiency deviation index Environmental Severity Index and load fluctuation index The weights at the current time step are calculated using a normalization method to ensure that the sum of the three is always 1. ; in, , , .

[0053] S32. Perform constraint optimization solution. Specifically, a rolling time-domain control strategy is adopted to transform the continuous control problem into a discrete nonlinear programming problem for solution, including the following sub-steps: S321. Construct discretized state prediction equations. To solve these equations in a digital controller, the dynamic heat load balance equations described in S1 need to be applied in the prediction time domain. The system undergoes discretization. Specifically, it extrapolates future states based on the time-cumulative effect of the temperature rise rate, i.e., prediction in the time domain. The temperature at the next moment is equal to the current temperature plus the product of the temperature rise rate at that moment and the sampling time step. The sampling time step is set to... Then the first The region in the future time( The temperature prediction state equation is defined as follows: ; ; In the formula, To predict the first in the time domain In each region Predicted indoor temperature at any given time; For the first In each region Predicted total heat load disturbance at any given time; For the outdoor unit in The total cooling capacity provided at all times based on the operating frequency is related to the compressor power. There is a mapping relationship ; It is the pipeline transmission efficiency coefficient, used to correct for the loss of cooling capacity during the refrigerant transmission process; To prevent tiny constants with a denominator of zero; For the first The weighting coefficient of the actual cooling capacity allocated to the indoor unit of the table; For the first The rated cooling capacity of the indoor unit; This is the flow characteristic function of the electronic expansion valve; For a moment No. The target opening degree command for the electronic expansion valve of the indoor unit of the table; For discrete-time variables; This represents the total number of indoor units that are currently active in the system. This is the sampling time step; S322. Define the set of physical constraints. To ensure that the solved control commands are within the safe operating range of the equipment, a set of constraints needs to be constructed. The optimization solution must satisfy the following inequality constraints: ; In the formula, To predict within the time domain The target operating frequency of the compressor at any given time; To predict within the time domain Time of the first The target opening degree of the electronic expansion valve of the indoor unit of the table; These are the compressor's minimum and maximum operating frequencies, respectively. For mechanical limiting of the expansion valve; This is the current air dew point temperature.

[0054] S323. Solve for the optimal control sequence and implement the first term. Based on the defined objective function. Given the temperature prediction state equations and constraint sets, sequential quadratic programming (SQP) is used in each control cycle. Solve the following optimization problem: ; Among them, the decision variable sequence After the solution is completed, the system extracts only the first solution from the sequence. The current actual control command is issued to the group control execution module. In the next moment... The system uses the latest sensor measurement data to correct the state and repeats steps S31 to S323 above to achieve closed-loop rolling optimization and output the optimal frequency. And the optimal opening degree of the electronic expansion valve of each indoor unit .

[0055] For example, currently there are people in the main conference room (area 1), with priority... The corridor (area 5) is unoccupied; priority given to this area. The predictive model shows that the main conference room will experience high load and the corridor will experience low load over the next 10 minutes. Using traditional control methods, the compressor might need to be increased to 80Hz for powerful cooling of the conference room. However, under the optimized solution of this invention, the system calculations show that temporarily increasing the target temperature in the corridor would be more effective. (Sacrificing slight comfort in low-priority areas) and utilizing the wall's cold radiation (thermal inertia) to maintain the conference room temperature, the compressor only needs to operate at... The high-efficiency zone can meet the requirements. The minimum value. At this point, the system outputs the optimal control command: the compressor frequency is set to... The electronic expansion valve opening degree in the conference room is set to... The opening degree of the corridor electronic expansion valve is set to .

[0056] In this embodiment 1, based on the multi-objective optimization function and physical constraints within the prediction field, a rolling time-domain control strategy is adopted to periodically solve for the optimal control sequence within a future period. This transforms the complex air conditioning group control problem into a nonlinear programming mathematical solution process within a finite time domain. Through mathematical algorithms, the optimal balance point is found among the three competing objectives of minimizing system energy consumption, maximizing global comfort, and reducing equipment operating losses. This overcomes the limitation of traditional PID single-point independent control lacking a global perspective. The globally optimal solution is calculated that can coordinate the differentiated needs of each region and minimize the overall operating cost of the system. This achieves precise on-demand allocation of cooling resources. Under the premise of prioritizing the thermal comfort of the core area, the compressor is guided to actively operate in the high-efficiency range, and smooth control commands are used to reduce the mechanical wear of the actuator.

[0057] S4. Execution control and safety boundary verification. Specifically, the optimal frequency... And the optimal opening degree of the electronic expansion valve of each indoor unit Issued to the executing agency. Prior to this, a security boundary check is performed to determine the rate execution strategy: ; In the formula, This is the maximum permissible frequency variation rate specified by the compressor manufacturer. For example, let... If the frequency at the previous moment was The optimized calculation yields the current target frequency as follows: The system will not jump instantaneously, but rather at a rate of 1 second per second. The rate of increase is smooth, requiring 15 seconds to reach the target value, to prevent compressor liquid slugging or electrical grid impact. In this embodiment 1, before issuing control commands, the group control execution module determines the optimal frequency based on the rate execution strategy. The final safety verification and smoothing process is then performed. The operational idea behind this step is to construct a safety buffer between mathematical optimization decisions and physical hardware execution, forcibly transforming theoretically abrupt changes into a gradual process consistent with mechanical characteristics. This prevents sudden compressor frequency changes caused by pursuing rapid response, avoids exceeding the equipment's physical limits, and effectively prevents risks such as compressor liquid slugging, poor lubrication, and power grid impacts. This ensures that the air conditioning system achieves high-efficiency energy-saving control while maintaining high operational reliability and safety.

[0058] Example 2 As a second embodiment of the present invention, such as Figure 3As shown, based on Embodiment 1, a central air conditioning multi-split unit is also disclosed, specifically including: an outdoor unit, an indoor unit, refrigerant transmission pipes, and the central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization of Embodiment 1. The outdoor unit includes a variable frequency compressor and an outdoor heat exchanger; the indoor unit includes an indoor heat exchanger and an electronic expansion valve; the outdoor unit is connected in parallel with multiple indoor units through refrigerant transmission pipes to form a closed refrigerant circulation loop. The central air conditioning system group control energy-saving system controls the outdoor unit's variable frequency compressor to adjust its operating frequency and controls the indoor unit's electronic expansion valve to drive the valve needle to make linear up-and-down movements to regulate the refrigerant flow. When the compressor frequency increases and the opening of the electronic expansion valve increases, the refrigerant flow through the indoor heat exchanger increases, and the cooling capacity increases; conversely, the refrigerant flow decreases, and the cooling capacity decreases.

[0059] Specific implementation 2 is as follows: In a daytime operation scenario of a Grade A office building, an outdoor unit connects two indoor units located in the core meeting area (indoor unit A) and the public corridor area (indoor unit B). Assume that at 10:00 AM in summer, a sudden impromptu meeting occurs in the core meeting area (area A), resulting in a large influx of people; while the public corridor area (area B) is sparsely populated at this time. In this embodiment, during operation, the system executes the following collaborative control steps: Step 1: The multi-dimensional data sensing module of the central air conditioning system group control energy-saving system detects a surge in the number of people in area A through infrared sensors. (Upward movement), while simultaneously detecting that no one was present in area B. The heat load prediction module immediately calculated and detected the total heat load disturbance in area A. A significant increase is predicted, based on the regional dynamic heat load balance equation, that region A will experience a rapid temperature rise within the next 10 minutes. (Increase), while the rate of temperature rise in region B is gradual.

[0060] Step 2: The multi-objective collaborative optimization module calculates the priority coefficients of the two regions based on the functional area mapping table and real-time personnel data: Region A is the core functional area and has a high population density, so its priority is calculated as follows: Approaching the upper limit (e.g., 1.0); Area B is a traffic functional zone and uninhabited, so the calculated priority is... Reduced to the lower limit (e.g.) ).

[0061] The system solves a multi-objective optimization function. At this time, a nonlinear outdoor unit performance coupling model is used for global optimization. If conventional control is used, the system might directly control the inverter compressor at its highest frequency (e.g., ...). The compressor operates to suppress the temperature rise in region A, but this causes the compressor to operate in a non-linear range with low energy efficiency. In this invention, calculations show that if the compressor frequency is maintained in a high-efficiency range (such as...),... Although the total cooling capacity is slightly less than the maximum demand, this is achieved by sacrificing a small amount of comfort in low-priority zone B (i.e., reducing the opening of the electronic expansion valve in zone B). Allow its temperature to be at (A brief increase within the range) By centrally transferring the refrigerant resources originally allocated to area B to area A, the high comfort requirements of area A can be met while reducing the overall energy consumption of the system. lowest.

[0062] Step 3: After receiving the optimal control command, the group control execution module performs a safety boundary check, and then issues an action command: On the outdoor unit side, control the inverter compressor to... The frequency operates smoothly, avoiding current surges and high power consumption caused by rapid frequency increases. On the indoor unit side, the opening of the electronic expansion valve controlling indoor unit A is... Increase to Increase the flow rate of liquid refrigerant entering indoor heat exchanger A to quickly absorb heat using latent heat, thus offsetting the heat dissipation load of the people in the conference room; at the same time, control the opening of the electronic expansion valve of indoor unit B. Reduce to The flow rate of refrigerant entering indoor heat exchanger B is limited.

[0063] After the above adjustments, the temperature in the core meeting area (Area A) remained at the set temperature after personnel entered. Within the specified range, no significant fluctuations were observed; although the temperature in the public corridor area (Area B) briefly rose to [a certain level], [the temperature remained relatively stable]. However, it is still within the set temperature tolerance constant ( Within this range, the overall environmental experience was not affected. Compared to traditional control methods, the compressor avoided high-frequency, inefficient operation during this period, improving overall energy efficiency and effectively reducing compressor frequency fluctuations.

[0064] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0065] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A group control energy-saving system for central air conditioning systems based on load forecasting and multi-objective optimization, characterized in that, include: The multi-dimensional data sensing module is used to collect indoor temperature, outdoor weather, and indoor occupancy status data in various indoor areas in real time, and to filter the collected data. The heat load prediction module is connected to the multidimensional data sensing module and is used to calculate the total heat load disturbance by combining the real-time collected multidimensional data, and to calculate the temperature rise rate based on the regional dynamic heat load balance equation. The multi-objective collaborative optimization module, connected to the heat load prediction module, is used to construct a nonlinear outdoor unit efficiency coupling model to quantify energy consumption characteristics, and to use the temperature rise rate to solve a multi-objective optimization function that includes energy consumption index, global comfort and equipment operation loss based on a rolling time domain strategy, and output the optimal compressor operating frequency and the optimal opening degree of each indoor unit's electronic expansion valve. The group control execution module, connected to the multi-objective collaborative optimization module, is used to receive the optimal compressor operating frequency and the optimal opening degree of the electronic expansion valve of each indoor unit, and after smoothing the frequency change rate according to the preset equipment safety boundary verification strategy, it sends the command to the actuator of the central air conditioning equipment.

2. The central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization according to claim 1, characterized in that, The temperature rise rate calculated based on the regional dynamic heat load balance equation is shown below: ； In the formula, The specific heat capacity of air, For the first Air quality in each region For a moment The actual cooling capacity provided by the air conditioner This represents the total heat load disturbance. For a moment No. Indoor temperature in the area This represents the rate of temperature rise.

3. The central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization according to claim 2, characterized in that, The total heat load disturbance Heat transfer from the building envelope Window heat gain from sunlight Human body heat dissipation and equipment heating The composition, and its calculation formula are as follows: ； Among them, the heat transfer of the enclosure structure It is the sum of the heat transfer of all the building envelope surfaces in the region. For any building envelope surface, its heat transfer is the product of the heat transfer coefficient of the building envelope surface, the area of ​​the building envelope surface, and the temperature difference. The window's solar heat gain It is the product of the window's overall solar heat gain coefficient, the solar irradiance acting on the window's outer surface at the current moment, the window's effective light-receiving area, the geometric direction correction coefficient, and the dynamic shading coefficient; Human body heat dissipation It is the product of the real-time estimated number of people in the area at the current moment and the average sensible heat dissipation per person; The device generates heat. It is the sum of the heat dissipation of all electrical equipment in the area. For any electrical equipment, its heat dissipation is the product of the real-time electrical power of the equipment and the heat conversion coefficient of the electrical power into indoor heat.

4. The central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization according to claim 1, characterized in that, The nonlinear outdoor unit performance coupling model is shown below: ； In the formula, For compressor power, For compressor operating frequency, Outdoor ambient temperature The evaporation temperature. , , , These are the fitting coefficients based on device characteristics.

5. The central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization according to claim 4, characterized in that, The multi-objective optimization function is as follows: ； In the formula, For multi-objective optimization functions, To predict the field of view, For a moment compressor power, The rated power of the system, , , These are the weighting coefficients for energy consumption indicators, overall comfort penalties, and equipment motion loss penalties, respectively. For the total number of regions, For a moment No. Priority coefficients for each region For the first The set temperature for each area This is the temperature tolerance constant. For a moment The penalty value for equipment operation wear.

6. The central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization according to claim 5, characterized in that, The weighting coefficient , , A state-aware normalized evaluation method is used for real-time calculation, and the calculation formula is as follows: ； In the formula, The energy efficiency deviation index corresponds to ; The environmental severity index corresponds to ; For load fluctuation index, corresponding .

7. The central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization according to claim 5, characterized in that, The time No. Priority coefficient of each region As shown below: ； In the formula, For functional normalization factor, For a moment No. Number of people in each region Design the area to accommodate the maximum number of people. This is the lower limit value determined based on the minimum maintenance requirements of the equipment.

8. The group control energy-saving system for central air conditioning systems based on load forecasting and multi-objective optimization according to claim 2 or 5, characterized in that, When solving the multi-objective collaborative optimization function based on the rolling time-domain strategy, the multi-objective collaborative optimization module performs rolling optimization based on the discretized state prediction equation, which is shown below: ； In the formula, To predict the first in the time domain In each region Predicted indoor temperature at any given time For the sampling time step, This is the pipeline transmission efficiency coefficient. For the first In each region Predicted total heat load disturbance at any given time. For the outdoor unit in The total cooling capacity provided is based on the operating frequency at all times. For the first The weighting coefficient of the actual cooling capacity allocated to the indoor unit of the table; The solution process must satisfy the following physical constraints: ； in, To predict within the time domain The target operating frequency of the compressor at any given time. To predict within the time domain Time of the first The target opening degree of the electronic expansion valve of the indoor unit of the Taiwan Strait. , For compressor frequency limits, , This refers to the expansion valve opening limit. This is the current air dew point temperature.

9. The central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization according to claim 1, characterized in that, The device security boundary verification strategy must meet the following requirements: ； In the formula, The optimal compressor operating frequency is obtained by solving the multi-objective optimization function. This refers to the maximum permissible frequency variation rate specified by the compressor manufacturer.

10. A central air conditioning multi-split unit, characterized in that, Specifically, it includes: The system comprises an outdoor unit, an indoor unit, a refrigerant transmission pipeline, and a central air conditioning system group control energy-saving system based on load forecasting and multi-objective optimization as described in any one of claims 1 to 9. The outdoor unit includes a variable frequency compressor and an outdoor heat exchanger, the indoor unit includes an indoor heat exchanger and an electronic expansion valve, the refrigerant transmission pipeline is used to connect the outdoor unit and the indoor unit, and the central air conditioning system group control energy-saving system is used to control the operating frequency of the variable frequency compressor and the opening degree of the electronic expansion valve.

Images