A row-to-row precision air conditioning system

Abstract

The present invention discloses an inter-row precision air conditioning system, which relates to the field of air conditioning, and includes a prediction modeling module, a cooling capacity allocation module, an equipment data module, a parameter control module, and a single output module. The present invention first constructs a prediction modeling module to predict the cooling demand of each time period of the next day in advance. Secondly, on the basis of cooling load prediction, combined with the building cooling and heating load factors, the air conditioning equipment data acquisition module and the key equipment data acquisition module are used to collect the operation status of the central air conditioning equipment and the key equipment respectively, and the terminal model is used to globally optimize the control parameters of multiple equipment of the large central air conditioning system, and then the parameter control module is used to constrain the related equipment of the air conditioning and its equipment performance, and the cooling capacity is redistributed. The allocated cooling capacity is output to the key equipment from different positions and different numbers of single output modules.

Description

Technical Field

[0001] The present invention relates to the technical field of air conditioning, and in particular to an inter-row precision air conditioning system. Background Art

[0002] With the rapid development and widespread application of information technology, the demand for centralized computing equipment such as data centers and computer rooms is growing. Data centers and computer rooms store a large amount of critical data and critical equipment that runs a large amount of data, and have high requirements for environmental conditions. They require stable temperature, humidity, dust control and other environmental conditions to ensure the normal operation of equipment and data security. Inter-row precision air conditioners play a vital role in these important places.

[0003] However, in the prior art, data centers, computer rooms and other places basically use inter-row precision air conditioners based on central air conditioners, most of which rely on manual control of building air conditioning systems by properties, and a small number use air conditioning automatic control systems with control algorithms as the core. The operation effect of manual control systems is greatly affected by individual experience. The current air conditioning automatic control system is difficult to achieve optimal control for large-scale buildings with large time lags and large thermal inertia, resulting in different temperatures and humidity for multiple key devices that store a large amount of key data and run a large amount of data. Both of the above control methods cannot be adjusted in real time according to the operating conditions of key equipment, and cannot provide a long-term stable temperature and humidity environment. In order to ensure that key equipment is not affected by overheating, the temperature of the location will be relatively low. If this continues for a long time, it will not only waste energy, but also affect the normal operation of key equipment. Summary of the invention

[0004] The purpose of the present invention is to provide an inter-row precision air-conditioning system to solve the problem that the above two control methods proposed in the above background technology cannot be adjusted in real time according to the operating conditions of key equipment, and cannot provide a long-term stable temperature and humidity environment. In order to ensure that key equipment is not affected by overheating, the temperature of the location will be relatively low. If this continues for a long time, it will not only waste energy, but also affect the normal operation of key equipment.

[0005] To achieve the above object, the present invention provides the following technical solutions: an inter-row precision air conditioning system, comprising a prediction modeling module, a cooling capacity distribution module, an equipment data module, a parameter control module and a monomer output module;

[0006] The prediction modeling module completes the task of predicting the cooling and heating loads of buildings and meets the requirements for establishing a data model for the central air-conditioning system;

[0007] The cooling capacity allocation module uses the optimal hyperparameter group to train the model under the training set, evaluates the model performance under the test set, and performs the optimal cooling allocation, taking into account the building cooling and heating load factors;

[0008] The equipment data module is divided into an air conditioning equipment data acquisition module and a key equipment data acquisition module;

[0009] The parameter control module ensures the stable operation of the air conditioning system and the cooling capacity meets the demand by applying corresponding constraints to the data collected by the air conditioning equipment data acquisition module;

[0010] The positions and quantities of the single output modules are reasonably arranged according to the size of the site area, and are used for surround cooling of key equipment.

[0011] Preferably, the predictive modeling module relies on an extreme gradient boosting tree algorithm, which is improved based on a gradient boosting decision tree algorithm, and the specific steps are as follows:

[0012] S1. Combining multiple weak learners, usually decision trees, to build a pre-model, whose objective function consists of two parts: training error and regularization term. It uses first-order and second-order derivative information to build decision trees.

[0013] S2, based on the greedy algorithm, selects the optimal node split to recursively build a tree, and finally forms a final model. The algorithm can process large-scale data in a short time and has good prediction performance, completes the task of building cooling and heating load prediction, and meets the needs of establishing a central air-conditioning system data model;

[0014] S3. Add a heuristic optimization algorithm to the final model, which assumes that these particles have positions

[0015] and speed; each particle can remember its own historical optimal position, and the optimal position of the particle group is shared globally; in each iteration, the particle will move a certain distance, which is affected by the particle's historical optimal position and the optimal position of the group, and is also random; the basic form of iteration is as follows:

[0016] ;

[0017] Where: and Indicates Particles are currently Position and velocity at dimension; The optimal position of the group Dimensions; is the optimal position of the individual Dimensions; , , They are inertia weight, individual coefficient, and social coefficient respectively; is a random number between [0, 1];

[0018] in, and These two coefficients determine whether the particle is more inclined to fly to its own optimal position or to the optimal position of the group. If the value is too large, the particles will linger in a small area and fail to converge. If the value is too large, the particles will fly to the local minimum too quickly and fall into the local optimum. Through such basic iterative logic, under convergence, the particles will eventually gather near the global optimal solution.

[0019] Different particles share information with each other and are less sensitive to the initial point. They have a certain ability to escape from the local optimal solution and find the global optimal solution, which helps to globally optimize the control parameters of multiple devices in large central air-conditioning systems.

[0020] Preferably, in the cooling capacity allocation module, five-dimensional features of day type, number of hours, cooling load at the current moment, which characterize the work and rest of personnel, and outdoor dry-bulb temperature at the current moment and outdoor relative humidity at the current moment, which characterize the outdoor environment, are selected as input parameters of the final model, and the output parameter of the final model is the cooling load at the next moment;

[0021] The grid search method is used to optimize the number of weak learners, the maximum tree depth, the learning rate and the gamma coefficient in the final model, wherein the number of weak learners is searched in [50, 100, 150, 200, 250, 300], the maximum tree depth is searched in [4, 5, 6, 7, 8], the learning rate is searched in [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8], and the gamma coefficient is searched in [0, 20, 40, 60, 80]. The model is trained using the optimal hyperparameter group under the training set, and the model performance is evaluated under the test set;

[0022] The central air conditioning energy efficiency and the prediction performance of the terminal model are evaluated by the mean absolute error, root mean square error and root mean square error coefficient of variation, as follows:

[0023] The mean absolute error is as follows:

[0024] ;

[0025] Where: Indicates The observed value of each particle; Indicates The predicted value of each particle;

[0026] The root mean square error is as follows:

[0027] ;

[0028] The root mean square error coefficient of variation can reflect the relative size of the error, as shown in the following formula:

[0029] .

[0030] Preferably, the air conditioning equipment data acquisition module collects the operating status of the central air conditioning equipment based on sensors; and the key equipment data acquisition module collects the operating status of the key equipment based on sensors.

[0031] Preferably, the parameter control module applies corresponding constraints to the data collected by the air conditioning equipment data acquisition module, as follows:

[0032] Cooling tower heat dissipation constraint: The air conditioning system completes the refrigeration task based on the reverse Carnot cycle, and the refrigerant absorbs heat in the evaporator , the compressor generates heat by doing work on the refrigerant , the refrigerant releases heat in the condenser , assuming that under ideal conditions, the relationship between the three should satisfy , in the central air conditioning system, the optimized cooling tower heat dissipation should be greater than or equal to the optimized chiller power plus cooling capacity to ensure the stable operation of the chiller;

[0033] Refrigeration capacity constraint of the chiller: the moment-by-moment cooling capacity after redistribution is input into the terminal model as a known input parameter to achieve global optimization control of the control parameters of the chiller under a given cooling capacity;

[0034] Chilled water supply and return temperature difference constraint: Considering the buildings with large temperature difference design, the chilled water supply and return temperature difference constraint is a hyperparameter, and the temperature difference constraint value is greater than or equal to 2℃ and less than or equal to 6℃;

[0035] Cooling water supply and return water temperature difference constraint: The cooling water supply and return water temperature difference constraint is also a hyperparameter, and the temperature difference constraint value is greater than or equal to 2°C and less than or equal to 6°C;

[0036] Cooling water return approach temperature constraint: The approach temperature is defined as the difference between the cooling water return temperature and the outdoor wet bulb temperature at that moment. The difference is greater than or equal to 2°C and less than or equal to 6°C.

[0037] Frequent start and stop constraint: The duration of the frequent start and stop constraint is also a hyperparameter. A 2h frequent start and stop constraint is set for the device, which means that it is required to run continuously for at least 2h after being turned on.

[0038] Preferably, the single output module comprises a flipping mechanism and a rotating mechanism, the flipping mechanism is used to adjust the cooling direction, and the rotating mechanism is used to drive the flipping mechanism to rotate.

[0039] Preferably, the flipping mechanism includes a flap, a driven sprocket ring, a telescopic tube, an arc-shaped rack, a gear, a motor, a fixed plate and a roller, one end of the telescopic tube is fixedly connected to the lower side of the middle hole of the fixed plate, the other end of the telescopic tube is fixedly connected to the upper side of the middle hole of the flap, and one end of the flap is rotatably connected to one end of the fixed plate.

[0040] Preferably, the driven sprocket ring is fixedly connected to the upper side of the middle hole of the fixed plate, multiple groups of rollers are annularly distributed and rotatably installed on the upper side of the middle hole of the driven sprocket ring, the motor is fixedly installed inside the other end of the fixed plate, the gear is fixedly installed at the output end of the motor, one end of the arc-shaped rack is fixedly connected to the other end of the upper side of the flap, and the arc-shaped rack passes through the fixed plate and is meshed with the gear.

[0041] Preferably, a mounting plate is rollingly connected to the middle recess of the roller, one end of the upper side of the mounting plate is fixedly connected to a mounting frame, the outer side of the driven sprocket ring is transmission connected to a chain, the other end of the chain is transmission connected to a driving sprocket, and a three-stage gear is coaxially installed above the driving sprocket.

[0042] Preferably, the outer side of the three-stage gear is meshedly connected with a secondary gear set, the outer side of the secondary gear set is meshedly connected with a primary gear, a servo motor is fixedly installed on the upper side of the primary gear, the servo motor is fixedly installed on the upper side of the mounting frame, the three-stage gear, the secondary gear set and the primary gear are rotatably installed between the mounting frame and the mounting plate, and a side tensioning sprocket is provided on the outer side of the chain, and the side tensioning sprocket is fixedly installed on the lower side of the mounting plate.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. In the present invention, a prediction modeling module is first constructed to predict the cooling demand for each period of the next day in advance. Secondly, based on the cooling load prediction, combined with the building's cooling and heating load factors, the air-conditioning equipment data acquisition module and the key equipment data acquisition module are used to collect the operating conditions of the central air-conditioning equipment and the key equipment respectively. The terminal model is used to globally optimize the control parameters of multiple devices in the large central air-conditioning system. The parameter control module is then used to constrain the relevant equipment of the air-conditioning and its equipment performance, and the cooling capacity is redistributed to make full use of the cooling capacity and avoid waste caused by uneven distribution. The distributed cooling capacity is output to the key equipment from different positions and different numbers of single output modules, and targeted arrangements are made according to the specific conditions on the spot to provide a stable temperature and humidity environment for the key equipment to ensure the normal operation of data centers, computer rooms, etc.

[0045] 2. In the present invention, the cooling delivery pipeline is connected to the upper side of the middle hole of the mounting plate, and the arc-shaped rack is subjected to a counterclockwise tangential force, so that the flap moves in a circle with the connection point with the fixed plate as the center, and the flap and the fixed plate form a certain angle, thereby driving the telescopic tube to unfold into an arc shape, pointing to the key equipment, and the cooling passes through the middle hole of the driven sprocket ring and the middle hole of the fixed plate into the telescopic tube, directly acting on a certain key equipment, providing it with a stable temperature and humidity environment.

[0046] 3. In the present invention, when the arc-shaped rack rotates and the angle between the flap and the fixed plate reaches the maximum value, the driving sprocket is forced to rotate, and after the chain transmission, it drives the driven sprocket ring to rotate, and the driven sprocket ring drives the multiple groups of rollers connected to rotate above to rotate in the middle hole of the mounting plate. As a result, the entire flipping mechanism rotates with the center of the middle hole of the driven sprocket ring as the origin, so that the outlet direction of the telescopic tube sweeps across the range of the conical surface, acts on multiple key equipment, reduces the temperature of multiple key equipment and the space near them, and provides them with an environment with stable temperature and humidity. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a flow chart of an inter-row precision air conditioning system of the present invention;

[0048] Figure 2 It is a three-dimensional structural schematic diagram of a single output module in an inter-row precision air conditioning system of the present invention;

[0049] Figure 3 It is a schematic diagram of a partial three-dimensional structure of a single output module in an inter-row precision air conditioning system of the present invention;

[0050] Figure 4 It is a schematic diagram of a partial planar structure of a single output module in an inter-row precision air conditioning system of the present invention;

[0051] Figure 5 It is a flow chart of the change of the flipping mechanism of a single output module in an inter-row precision air conditioning system of the present invention;

[0052] Figure 6 It is a schematic diagram of the three-dimensional structure of the rotating mechanism of a single output module in an inter-row precision air conditioning system of the present invention.

[0053] In the figure: 1. Prediction modeling module; 2. Cooling capacity distribution module; 3. Equipment data module; 31. Air conditioning equipment data acquisition module; 32. Key equipment data acquisition module; 4. Parameter control module; 5. Single unit output module; 51. Flipping mechanism; 511. Flip plate; 512. Driven sprocket ring; 513. Telescopic tube; 514. Arc rack; 515. Gear; 516. Motor; 517. Fixed plate; 518. Roller; 52. Rotating mechanism; 521. Servo motor; 522. Mounting frame; 523. Mounting plate; 524. Chain; 525. Side tightening sprocket; 526. Driving sprocket; 527. Third-stage gear; 528. Second-stage gear set; 529. First-stage gear. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the implementation regulations described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Example 1: Reference Figure 1 As shown: a row-to-row precision air conditioning system, comprising a prediction modeling module 1, a cooling capacity distribution module 2, an equipment data module 3, a parameter control module 4 and a single unit output module 5;

[0056] Prediction modeling module 1 completes the task of predicting the cooling and heating loads of buildings and meets the requirements of establishing a data model for the central air-conditioning system. Prediction modeling module 1 relies on the extreme gradient boosting tree algorithm, which is improved based on the gradient boosting decision tree algorithm. The specific steps are as follows:

[0057] S1. Combining multiple weak learners, usually decision trees, to build a pre-model, whose objective function consists of two parts: training error and regularization term. It uses first-order and second-order derivative information to build decision trees.

[0058] S2, based on the greedy algorithm, selects the optimal node split to recursively build a tree, and finally forms a final model. The algorithm can process large-scale data in a short time and has good prediction performance. The algorithm completes the task of predicting the cooling and heating loads of buildings and meets the needs of establishing a data model for the central air-conditioning system;

[0059] S3. Add a heuristic optimization algorithm to the final model, which assumes that these particles have positions

[0060] and speed; each particle can remember its own historical optimal position, and the optimal position of the particle group is shared globally; in each iteration, the particle will move a certain distance, which is affected by the particle's historical optimal position and the optimal position of the group, and is also random; the basic form of iteration is as follows:

[0061] ;

[0062] Where: and Indicates Particles are currently Position and velocity at dimension; The optimal position of the group Dimensions; is the optimal position of the individual Dimensions; , , They are inertia weight, individual coefficient, and social coefficient respectively; is a random number between [0, 1];

[0063] in, and These two coefficients determine whether the particle is more inclined to fly to its own optimal position or to the optimal position of the group. If the value is too large, the particles will linger in a small area and fail to converge. If the value is too large, the particles will fly to the local minimum too quickly and fall into the local optimum. Through such basic iterative logic, under convergence, the particles will eventually gather near the global optimal solution.

[0064] Different particles share information with each other and are less sensitive to the initial point. They have a certain ability to escape from the local optimal solution and find the global optimal solution, which helps to globally optimize the control parameters of multiple devices in large central air-conditioning systems.

[0065] The cooling capacity allocation module 2 uses the optimal hyperparameter group to train the model under the training set, evaluates the model performance under the test set, and performs the optimal cooling allocation, taking into account the building cooling and heating load factors;

[0066] In the cooling capacity allocation module 2, the building cooling and heating load is related to many factors, such as the lighting system, the work and rest of the staff, the window-to-wall ratio, and the exterior wall materials and finishes, etc. For the same building, its window-to-wall ratio, exterior wall materials and finishes have been determined, and the indoor load is closely related to the work and rest of the staff and outdoor environmental factors; the five-dimensional features of the day type, number of hours, and cooling load at the current moment that characterize the work and rest of the staff, and the outdoor dry-bulb temperature and outdoor relative humidity at the current moment that characterize the outdoor environment are selected as the input parameters of the final model, and the output parameter of the final model is the cooling load at the next moment;

[0067] The grid search method is used to optimize the number of weak learners, maximum tree depth, learning rate and gamma coefficient in the final model. The number of weak learners is searched in [50, 100, 150, 200, 250, 300], the maximum tree depth is searched in [4, 5, 6, 7, 8], the learning rate is searched in [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8], and the gamma coefficient is searched in [0, 20, 40, 60, 80]. The model is trained with the optimal hyperparameter group in the training set, and the model performance is evaluated in the test set.

[0068] The prediction performance of the central air conditioning energy efficiency and terminal model was evaluated by the mean absolute error, root mean square error and root mean square error coefficient of variation, as follows:

[0069] The mean absolute error is as follows:

[0070] ;

[0071] Where: Indicates The observed value of each particle; Indicates The predicted value of each particle;

[0072] The root mean square error is as follows:

[0073] ;

[0074] The root mean square error coefficient of variation can reflect the relative size of the error, as shown in the following formula:

[0075] .

[0076] The equipment data module 3 is divided into an air conditioning equipment data acquisition module 31 and a key equipment data acquisition module 32. The air conditioning equipment data acquisition module 31 collects the operation status of the central air conditioning equipment based on sensors; the key equipment data acquisition module 32 collects the operation status of key equipment based on sensors;

[0077] The parameter control module 4 ensures the stable operation of the air conditioning system and the cooling capacity meets the demand by applying corresponding constraints to the data collected by the air conditioning equipment data acquisition module 31, as follows:

[0078] Cooling tower heat dissipation constraint: The air conditioning system completes the refrigeration task based on the reverse Carnot cycle, and the refrigerant absorbs heat in the evaporator , the compressor generates heat by doing work on the refrigerant , the refrigerant releases heat in the condenser , assuming that under ideal conditions, the relationship between the three should satisfy , in the central air conditioning system, the optimized cooling tower heat dissipation should be greater than or equal to the optimized chiller power plus cooling capacity to ensure the stable operation of the chiller;

[0079] Refrigeration capacity constraint of chiller: The moment-by-moment cooling capacity after redistribution is input into the final model as a known input parameter to achieve global optimization control of the control parameters of the chiller under a given cooling capacity;

[0080] Constraint on the supply and return temperature difference of chilled water: In many buildings, the standard value of the supply temperature of chilled water is 7°C, the standard value of the return temperature of chilled water is 12°C, and the supply and return temperature difference is 5°C. When the terminal cooling load remains unchanged, the smaller the chilled water flow rate, the greater the supply and return temperature difference, and vice versa. Since the water pump pressure difference is proportional to the square of the water flow rate, an increase in the chilled water flow rate will cause the energy consumption of the chilled water pump to rise sharply. At the same time, when the water flow rate is increased, the chilled water pump may deviate from the design operating point, resulting in inefficient operation of the chilled water pump. Therefore, the supply and return temperature difference of chilled water should not be too small. Since the terminal specifications have been determined, although reducing the water flow rate and increasing the temperature difference can save water pump energy consumption, a water flow rate lower than the design value will reduce the maximum cooling capacity, thereby greatly affecting the indoor comfort under high load conditions. Therefore, the supply and return temperature difference of chilled water should not be too large. Considering the existence of buildings with large temperature difference designs, the supply and return temperature difference constraint of chilled water is a hyperparameter, and the temperature difference constraint value is greater than or equal to 2°C and less than or equal to 6°C;

[0081] Cooling water supply and return water temperature difference constraint: The cooling water supply and return water temperature difference constraint is also a hyperparameter, and the temperature difference constraint value is greater than or equal to 2°C and less than or equal to 6°C;

[0082] Cooling water return approach temperature constraint: The approach temperature is defined as the difference between the cooling water return temperature and the outdoor wet bulb temperature at that moment. Under normal circumstances, the lower limit of this value tends to 0°C. The higher the approach temperature, the greater the cooling potential of the cooling water leaving the cooling tower, and vice versa. When the approach temperature is very high, it may be caused by poor heat dissipation performance of the cooling tower or excessive cooling water flow; when the approach temperature is very low, it may be caused by too small cooling water flow. When the cooling water flow is too small, resulting in a small approach temperature, it usually only occurs for a short time on the same system and cannot maintain stability; when the cooling water flow is too large, resulting in a high approach temperature, the cooling water pump is not energy-saving and the high return water temperature indirectly reduces the energy efficiency of the chiller and increases the energy consumption of the chiller. Therefore, the approach temperature is a hyperparameter that should change with factors such as the actual installation of the cooling tower, cooling water pump, and rated operating conditions of the cooling side of the chiller in the project. The difference is greater than or equal to 2°C and less than or equal to 6°C;

[0083] Frequent start-stop constraint: When the device starts, the current is large, and frequent start-stop has a certain adverse effect on the life of the device. The duration of the frequent start-stop constraint is also a hyperparameter. A 2h frequent start-stop constraint is set for the device, which requires it to run continuously for at least 2h after being turned on.

[0084] The position and quantity of the single output module 5 are reasonably arranged according to the size of the site area, and are used for surround cooling of key equipment.

[0085] In this embodiment, first, a prediction modeling module 1 is constructed to predict the cooling demand in each time period of the next day in advance. Secondly, based on the cooling load prediction, combined with the building's cooling and heating load factors, the central air-conditioning equipment is taken as the analysis object, and the air-conditioning equipment data acquisition module 31 and the key equipment data acquisition module 32 are used to collect the operating conditions of the central air-conditioning equipment and the key equipment respectively. The terminal model is used to globally optimize the control parameters of multiple devices in the large central air-conditioning system, and then the parameter control module 4 is used to constrain the relevant equipment of the air-conditioning and its equipment performance, and the cooling capacity is redistributed, so that the cooling capacity is fully utilized and waste caused by uneven distribution is avoided. The allocated cooling capacity is output to the key equipment from different positions and different numbers of single output modules 5, and targeted arrangements are made according to the specific conditions on the spot to provide a stable temperature and humidity environment for the key equipment to ensure the normal operation of data centers, computer rooms, etc.

[0086] Embodiment 2: According to Figure 2-5As shown, the single output module 5 is reasonably arranged in position and quantity according to the size of the site area, and is used for surround cooling of key equipment. The single output module 5 includes a flip mechanism 51 and a rotating mechanism 52. The flip mechanism 51 is used to adjust the direction of cooling. The flip mechanism 51 includes a flip plate 511, a driven sprocket ring 512, a telescopic tube 513, an arc-shaped rack 514, a gear 515, a motor 516, a fixed plate 517 and a roller 518. One end of the telescopic tube 513 is fixedly connected to the lower side of the middle hole of the fixed plate 517, and the other end of the telescopic tube 513 is fixedly connected to the flip plate 511. On the upper side of the middle hole of plate 511, one end of the flap 511 is rotatably connected to one end of the fixed plate 517, the driven sprocket ring 512 is fixedly connected to the upper side of the middle hole of the fixed plate 517, multiple groups of rollers 518 are distributed in an annular manner and rotatably installed on the upper side of the middle hole of the driven sprocket ring 512, the motor 516 is fixedly installed inside the other end of the fixed plate 517, the gear 515 is fixedly installed at the output end of the motor 516, one end of the arc-shaped rack 514 is fixedly connected to the other end of the upper side of the flap 511, and the arc-shaped rack 514 passes through the fixed plate 517 and is meshed with the gear 515.

[0087] In this embodiment, a cooling delivery pipeline is connected to the upper side of the middle hole of the mounting plate 523, and the motor 516 is started. The motor 516 drives the gear 515 to rotate and engage the arc rack 514. The arc rack 514 is subjected to a counterclockwise tangential force, which acts on one end of the flap 511, and the other end of the flap 511 is limited due to its rotational connection with the fixed plate 517, so that the flap 511 moves in a circle with the connection point with the fixed plate 517 as the center. The flap 511 and the fixed plate 517 form a certain angle, thereby driving the telescopic tube 513 to expand into an arc and point to the key equipment. The cooling passes through the middle hole of the driven sprocket ring 512 and the middle hole of the fixed plate 517 into the telescopic tube 513, directly acting on a certain key equipment, reducing the temperature of a single key equipment and its space, and providing it with a stable temperature and humidity environment.

[0088] Embodiment 3: According to Figure 2-6As shown, the position and quantity of the single output module 5 are reasonably arranged according to the size of the site area, and are used for surround cooling of key equipment. The single output module 5 includes a flip mechanism 51 and a rotating mechanism 52. The rotating mechanism 52 is used to drive the flip mechanism 51 to rotate. A mounting plate 523 is rollingly connected to the middle depression of the roller 518. One end of the upper side of the mounting plate 523 is fixedly connected to the mounting frame 522. The outer side of the driven sprocket ring 512 is transmission-connected with a chain 524. The other end of the chain 524 is transmission-connected with a driving sprocket 526. The upper side of the driving sprocket 526 is coaxially mounted There is a three-stage gear 527, the outer side of the three-stage gear 527 is meshedly connected with a two-stage gear set 528, the outer side of the two-stage gear set 528 is meshedly connected with a first-stage gear 529, a servo motor 521 is fixedly installed on the upper side of the first-stage gear 529, the servo motor 521 is fixedly installed on the upper side of the mounting frame 522, the three-stage gear 527, the two-stage gear set 528 and the first-stage gear 529 are rotatably installed between the mounting frame 522 and the mounting plate 523, and a side tightening sprocket 525 is arranged on the outer side of the chain 524, and the side tightening sprocket 525 is fixedly installed on the lower side of the mounting plate 523.

[0089] In this embodiment, when the arc rack 514 rotates and the angle between the flap 511 and the fixed plate 517 reaches the maximum value, the servo motor 521 is started, and the output shaft of the servo motor 521 drives the primary gear 529 to rotate, and the primary gear 529 meshes with the large gear in the secondary gear set 528, so that the coaxially installed small gear in the secondary gear set 528 drives the tertiary gear 527 to rotate, and the tertiary gear 527 drives the coaxially installed driving sprocket 526 below to rotate after being subjected to force, and after being transmitted by the chain 524, drives the driven sprocket ring 512 to rotate, and the driven sprocket ring 512 drives the multiple groups of rollers 518 connected to the upper rotation to rotate in the middle hole of the mounting plate 523, thereby the entire flip mechanism 51 rotates with the center of the middle hole of the driven sprocket ring 512 as the origin, so that the outlet direction of the telescopic tube 513 sweeps the range of the conical surface, acts on multiple key equipment, reduces the temperature of multiple key equipment and the space around them, and provides them with a stable temperature and humidity environment.

[0090] The use method and working principle of this device: When using the inter-row precision air-conditioning system, first, construct a prediction modeling module 1 to predict the cooling demand for each time period of the next day in advance. Secondly, based on the cooling load prediction, combined with the building's cooling and heating load factors, take the central air-conditioning equipment as the analysis object, collect the operation status of the central air-conditioning equipment and key equipment through the air-conditioning equipment data acquisition module 31 and the key equipment data acquisition module 32 respectively, and globally optimize the control parameters of multiple devices of the large central air-conditioning system through the terminal model, and then cooperate with the parameter control module 4 to constrain the relevant equipment of the air-conditioning and its equipment performance, complete the redistribution of the cooling capacity, and the allocated cooling capacity is output to the key equipment from different positions and different numbers of single output modules 5;

[0091] The cooling delivery pipeline is connected to the upper side of the middle hole of the mounting plate 523, and the motor 516 is started. The motor 516 drives the gear 515 to rotate and mesh with the arc-shaped rack 514. The arc-shaped rack 514 is subjected to a counterclockwise tangential force, which acts on one end of the flap 511, and the other end of the flap 511 is limited due to the rotation connection with the fixed plate 517, so that the flap 511 moves in a circle with the connection point with the fixed plate 517 as the center. The flap 511 and the fixed plate 517 form a certain angle, thereby driving the telescopic tube 513 to expand into an arc and point to the key equipment. The cooling passes through the middle hole of the driven sprocket ring 512 and the middle hole of the fixed plate 517 into the telescopic tube 513, and directly acts on a certain key equipment to provide it with stable temperature and humidity, etc.;

[0092] When the arc-shaped rack 514 rotates and makes the angle between the flap 511 and the fixed plate 517 reach the maximum value, the servo motor 521 is started, and the output shaft of the servo motor 521 drives the primary gear 529 to rotate, and the primary gear 529 meshes with the large gear in the secondary gear set 528, so that the coaxially installed small gear in the secondary gear set 528 drives the tertiary gear 527 to rotate, and the tertiary gear 527 drives the coaxially installed driving sprocket 526 below to rotate after being subjected to force, and after being driven by the chain 524, drives the driven sprocket ring 512 to rotate, and the driven sprocket ring 512 drives the multiple groups of rollers 518 connected to the upper rotation to rotate in the middle hole of the mounting plate 523, so that the entire flip mechanism 51 rotates with the center of the middle hole of the driven sprocket ring 512 as the origin, so that the outlet direction of the telescopic tube 513 sweeps across the range of the conical surface, acts on multiple key equipment, and provides them with a stable temperature and humidity environment;

[0093] The single output modules 5 are reasonably arranged according to the location and quantity of key equipment. Multiple single output modules 5 operate independently and are adjusted according to the actual conditions of the key equipment. The inter-row precision air-conditioning system meets the use of data centers, computer rooms and other places to the maximum extent.

[0094] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An inter-row precision air conditioning system, characterized in that: It includes a prediction modeling module (1), a cooling capacity allocation module (2), an equipment data module (3), a parameter control module (4) and a single unit output module (5); The prediction modeling module (1) completes the task of predicting the cooling and heating loads of the building and meets the requirements for establishing a data model for the central air conditioning system; The cooling capacity allocation module (2) uses an optimal hyperparameter group to train a model in a training set, evaluates model performance in a test set, and performs optimal cooling capacity allocation, taking into account factors of building cooling and heating loads; The equipment data module (3) is divided into an air conditioning equipment data collection module (31) and a key equipment data collection module (32); The air conditioning equipment data collection module (31) collects the operation status of the central air conditioning equipment based on sensors; the key equipment data collection module (32) collects the operation status of the key equipment based on sensors; The parameter control module (4) applies corresponding constraints to the data collected by the air conditioning equipment data acquisition module (31); The positions and numbers of the single output modules (5) are reasonably arranged according to the size of the site area, and are used for surround cooling of key equipment; The single output module (5) comprises a flipping mechanism (51) and a rotating mechanism (52), wherein the flipping mechanism (51) is used to adjust the direction of cooling, and the rotating mechanism (52) is used to drive the flipping mechanism (51) to rotate; The prediction modeling module (1) relies on the extreme gradient boosting tree algorithm, which is based on the improvement of the gradient boosting decision tree algorithm. The specific steps are as follows: S1, combining multiple weak learners, usually decision trees, to build a preliminary model; S2, based on the greedy algorithm, select the optimal node split to recursively build the tree, and finally form a final model; S3. Add a heuristic optimization algorithm to the final model, which assumes that these particles have positions and speed; each particle can remember its own historical optimal position, and the optimal position of the particle group is shared globally; the particle moves a certain distance in each iteration; the basic form of iteration is as follows: ； Where: and Indicates Particles are currently Position and velocity at dimension; The optimal position of the group Dimensions; is the optimal position of the individual Dimensions; , , They are inertia weight, individual coefficient, and social coefficient respectively; is a random number between [0, 1]; in, and These two coefficients determine whether the particle is more inclined to fly to its own optimal position or to the optimal position of the group. If the value is too large, the particles will linger in a small area and fail to converge. If the value is too large, the particles will fly to the local minimum too quickly and fall into the local optimum. Through such basic iterative logic, under convergence conditions, the particles will eventually gather near the global optimal solution.

2. The inter-row precision air conditioning system according to claim 1, characterized in that: In the cooling capacity allocation module (2), five dimensional features, namely, day type, number of hours, cooling load at the current moment, which characterize the work and rest situation of personnel, and outdoor dry-bulb temperature at the current moment, and outdoor relative humidity at the current moment, which characterize the outdoor environment, are selected as input parameters of the final model, and the output parameter of the final model is the cooling load at the next moment; The grid search method is used to optimize the number of weak learners, the maximum tree depth, the learning rate and the gamma coefficient in the final model, wherein the number of weak learners is searched in [50, 100, 150, 200, 250, 300], the maximum tree depth is searched in [4, 5, 6, 7, 8], the learning rate is searched in [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8], and the gamma coefficient is searched in [0, 20, 40, 60, 80]. The model is trained using the optimal hyperparameter group in the training set, and the model performance is evaluated in the test set. The central air conditioning energy efficiency and the prediction performance of the terminal model are evaluated by the mean absolute error, root mean square error and root mean square error coefficient of variation, as follows: The mean absolute error is as follows: ； Where: Indicates The observed value of each particle; Indicates The predicted value of each particle; The root mean square error is as follows: ； The root mean square error coefficient of variation can reflect the relative size of the error, as shown in the following formula: 。 3. The inter-row precision air conditioning system according to claim 2, characterized in that: The parameter control module (4) imposes corresponding constraints on the data collected by the air conditioning equipment data acquisition module (31), specifically as follows: Cooling tower heat dissipation constraint: Refrigerant absorbs heat in the evaporator , the compressor generates heat by doing work on the refrigerant , the refrigerant releases heat in the condenser , in the central air conditioning system, the optimized cooling tower heat dissipation should be greater than or equal to the optimized chiller power plus cooling capacity to ensure the stable operation of the chiller; Refrigeration capacity constraint of the chiller: the moment-by-moment cooling capacity after redistribution is input into the terminal model as a known input parameter to achieve global optimization control of the control parameters of the chiller under a given cooling capacity; Chilled water supply and return temperature difference constraint: Considering the buildings with large temperature difference design, the chilled water supply and return temperature difference constraint is a hyperparameter, and the temperature difference constraint value is greater than or equal to 2℃ and less than or equal to 6℃; Cooling water supply and return water temperature difference constraint: The cooling water supply and return water temperature difference constraint is also a hyperparameter, and the temperature difference constraint value is greater than or equal to 2°C and less than or equal to 6°C; Cooling water return approach temperature constraint: The approach temperature is defined as the difference between the cooling water return temperature and the outdoor wet bulb temperature at that moment. The difference is greater than or equal to 2°C and less than or equal to 6°C. Frequent start and stop constraint: The duration of the frequent start and stop constraint is also a hyperparameter. A 2h frequent start and stop constraint is set for the device, which means that it is required to run continuously for at least 2h after being turned on.

4. The inter-row precision air conditioning system according to claim 3, characterized in that: The flipping mechanism (51) comprises a flap (511), a driven sprocket ring (512), a telescopic tube (513), an arc-shaped rack (514), a gear (515), a motor (516), a fixed plate (517) and a roller (518); one end of the telescopic tube (513) is fixedly connected to the lower side of the middle hole of the fixed plate (517); the other end of the telescopic tube (513) is fixedly connected to the upper side of the middle hole of the flap (511); and one end of the flap (511) is rotatably connected to one end of the fixed plate (517).

5. The inter-row precision air conditioning system according to claim 4, characterized in that: The driven sprocket ring (512) is fixedly connected to the upper side of the middle hole of the fixed plate (517); a plurality of groups of rollers (518) are annularly distributed and rotatably mounted on the upper side of the middle hole of the driven sprocket ring (512); the motor (516) is fixedly mounted inside the other end of the fixed plate (517); the gear (515) is fixedly mounted on the output end of the motor (516); one end of the arc-shaped rack (514) is fixedly connected to the other end of the upper side of the flap (511); the arc-shaped rack (514) passes through the fixed plate (517) and is meshedly connected with the gear (515).

6. The inter-row precision air conditioning system according to claim 5, characterized in that: A mounting plate (523) is rollingly engaged at a recessed portion in the middle of the roller (518); one end of the upper side of the mounting plate (523) is fixedly connected to a mounting frame (522); a chain (524) is drivingly connected to the outer side of the driven sprocket ring (512); the other end of the chain (524) is drivingly connected to a driving sprocket (526); and a three-stage gear (527) is coaxially mounted above the driving sprocket (526).

7. The inter-row precision air conditioning system according to claim 6, characterized in that: The outer side of the three-stage gear (527) is meshedly connected with a two-stage gear set (528), and the outer side of the two-stage gear set (528) is meshedly connected with a first-stage gear (529). A servo motor (521) is fixedly mounted on the upper side of the first-stage gear (529), and the servo motor (521) is fixedly mounted on the upper side of the mounting frame (522). The three-stage gear (527), the two-stage gear set (528) and the first-stage gear (529) are rotatably mounted between the mounting frame (522) and the mounting plate (523). A side tension sprocket (525) is provided on the outer side of the chain (524), and the side tension sprocket (525) is fixedly mounted on the lower side of the mounting plate (523).

Images