Method for predicting air volume and air pressure for each vent of duct ventilation system
A machine learning-based method predicts airflow and air pressure for duct ventilation systems, eliminating the need for sensor installations and maintenance, thereby optimizing damper control during emergencies.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- KOREA ELECTRONICS TECH INST
- Filing Date
- 2025-10-15
- Publication Date
- 2026-05-21
AI Technical Summary
Existing duct ventilation systems require significant effort and cost for installing and maintaining sensors at each vent to monitor airflow and air pressure, which are necessary for controlling dampers during emergencies like fires.
A method using machine learning to predict airflow and air pressure for each vent without installing sensors, by calculating loss coefficients and utilizing an airflow prediction model trained on supply airflow and loss coefficients.
Enables accurate prediction of airflow and air pressure for damper control during emergencies, reducing the need for vent-specific sensor installations and maintenance, thus lowering operational effort and costs.
Smart Images
Figure KR2025016240_21052026_PF_FP_ABST
Abstract
Description
Method for Predicting Airflow and Pressure by Vent in a Duct Ventilation System
[0001] The present invention relates to a duct ventilation system, and more specifically, to a method for predicting airflow and air pressure for each vent in a duct ventilation system installed within a station.
[0002] The duct ventilation system installed within the subway station facilities not only maintains the air quality within the station in a pleasant state, but also functions as a smoke control and exhaust system that prevents the spread of fire by regulating the amount and pressure of air distributed to each section in the event of a fire and minimizes casualties caused by smoke.
[0003] Accordingly, when a fire occurs in a specific area, the manager controls the amount of air supplied to and exhausted into that area. Generally, this can be done by simply controlling the speed of the air handling unit (AHU) to control the total ventilation volume, or by controlling the opening and closing of each damper installed at the branching point of the ventilation duct.
[0004] To control the opening and closing of dampers, it is necessary to know the airflow and air pressure information for each vent of the duct ventilation system. To achieve this, sensors must be installed at each vent, and the sensor data must be monitored. However, this requires significant effort and cost for both equipment installation and maintenance.
[0005] The present invention has been devised to solve the above-mentioned problems, and the objective of the present invention is to provide a method for predicting the airflow and air pressure for each vent based on machine learning without installing or maintaining sensors for each vent, in order to predict the airflow and air pressure for each vent required for controlling the opening and closing of dampers for each vent of a duct ventilation system in the event of an emergency situation such as a fire.
[0006] A method for predicting airflow conditions per vent of a duct ventilation system according to an embodiment of the present invention for achieving the above objective comprises: a ventilation control system acquiring air pressure and airflow volume for each vent of the duct ventilation system; a ventilation control system calculating a loss coefficient for each vent from the acquired air pressure and airflow volume for each vent; a ventilation control system storing the calculated loss coefficient for each vent; and a ventilation control system predicting the airflow volume for each vent from the supply air volume of the duct ventilation system and the stored loss coefficient for each vent.
[0007] The acquisition step may involve operating the duct ventilation system at a predetermined standard supply air volume and acquiring the air pressure and air volume for each vent.
[0008] The acquisition step may involve temporarily installing a pressure gauge for each vent to acquire the wind pressure for each vent, and temporarily installing a flow meter for each vent to acquire the airflow for each vent.
[0009] The acquisition step may involve temporarily installing a pressure gauge for each vent to acquire the air pressure for each vent, and acquiring the air volume for each vent through calculation or simulation from the standard supply air volume by referring to the design structure and specifications of the duct.
[0010] The step for calculating the loss coefficient for each vent is,
[0011] Calculate using the following formula,
[0012]
[0013] Here, P i is the pressure for each vent, k1 is the loss coefficient for each vent, Q i This may be the airflow for each vent.
[0014] The prediction step may predict the airflow for each vent using an airflow prediction model, which is a machine learning model pre-trained to predict the airflow for each vent by receiving the supply airflow and the loss coefficient for each vent as input.
[0015] The prediction step may involve changing the loss coefficient for each vent and re-predicting the airflow for each vent using an airflow prediction model if the prediction result is inaccurate.
[0016] The method for predicting airflow conditions per vent of a duct ventilation system according to the present invention may further include the step of the ventilation control system calculating the pressure for each vent from the predicted airflow volume for each vent and the stored loss coefficient for each vent.
[0017] The method for predicting airflow conditions per vent of a duct ventilation system according to the present invention may further include the step of the ventilation control system outputting the predicted airflow volume per vent and the calculated pressure per vent.
[0018] According to another aspect of the present invention, a wind condition prediction system for a duct ventilation system is provided, comprising: a storage unit for storing a loss coefficient for each vent of a duct ventilation system; a processor for calculating a loss coefficient for each vent from a wind pressure and airflow for each vent and storing it in the storage unit, and predicting an airflow for each vent from the supply airflow of the duct ventilation system and the loss coefficient for each vent stored in the storage unit; and an output unit for outputting the predicted airflow for each vent.
[0019] According to another aspect of the present invention, a method for controlling a duct ventilation system is provided, comprising: a step in which a ventilation control system stores a loss coefficient for each vent calculated from the air pressure and air volume for each vent; a step in which a ventilation control system predicts the air volume for each vent from the supply air volume of the duct ventilation system and the stored loss coefficient for each vent; a step in which a ventilation control system calculates the pressure for each vent from the predicted air volume for each vent and the stored loss coefficient for each vent; a step in which a ventilation control system outputs the predicted air volume for each vent and the calculated pressure for each vent; and a step in which a ventilation control system controls the duct ventilation system according to a control variable input by an administrator with reference to the output air volume and pressure for each vent.
[0020] According to another aspect of the present invention, a duct ventilation control system is provided, comprising: a storage unit for storing a loss coefficient for each vent calculated from the air pressure and air volume for each vent; a processor for predicting the air volume for each vent from the supply air volume of the duct ventilation system and the stored loss coefficient for each vent, and calculating the pressure for each vent from the predicted air volume for each vent and the stored loss coefficient for each vent; an output unit for outputting the predicted air volume for each vent and the calculated pressure for each vent; and a ventilation control unit for controlling the duct ventilation system according to a control variable input by an administrator with reference to the output air volume and pressure for each vent.
[0021] As explained above, according to the embodiments of the present invention, the airflow and air pressure of each vent of a duct ventilation system can be predicted based on machine learning, and thus, in predicting the airflow and air pressure of each vent required for controlling the opening and closing of dampers for each vent in the event of an emergency such as a fire, the installation and maintenance of vent-specific sensors are unnecessary, thereby reducing effort and costs.
[0022] FIG. 1 is an example of a duct ventilation system to which an embodiment of the present invention is applicable,
[0023] FIG. 2 is a ventilation control system according to one embodiment of the present invention,
[0024] FIGS. 3 and 4 are a method for predicting wind conditions by vent of a duct ventilation system according to another embodiment of the present invention,
[0025] Figure 5 is an example of a duct ventilation system in a subway station.
[0026] The present invention will be described in more detail below with reference to the drawings.
[0027] An embodiment of the present invention presents a method for predicting airflow and air pressure for each vent of a duct ventilation system. This is a technology that predicts the airflow and air pressure for each vent required for damper control for each vent of a duct ventilation system in the event of an emergency situation, such as a fire, based on machine learning, without the installation and maintenance of sensors for each vent.
[0028] FIG. 1 is a drawing illustrating a duct ventilation system to which an embodiment of the present invention is applicable. As shown, the duct ventilation system is a facility for performing purification, air conditioning, etc. by circulating indoor air and outdoor air through ducts.
[0029] In a duct ventilation system, air is supplied into the duct by an AHU (Air Handling Unit, not shown), and the supplied air volume is distributed to multiple vents and discharged. In addition, the discharge volume of each vent in the duct ventilation system is controlled by each damper (not shown).
[0030] A system for controlling the duct ventilation system presented in FIG. 1 is illustrated in FIG. 2. FIG. 2 is a diagram illustrating the configuration of a ventilation control system according to an embodiment of the present invention. The ventilation control system according to an embodiment of the present invention is configured to include a storage unit (110), an output unit (120), a processor (130), a ventilation control unit (140), and an input unit (150).
[0031] The storage unit (110) is a storage medium in which loss coefficients and airflow prediction models for each vent of the duct ventilation system are stored. The loss coefficients and airflow prediction models will be described in detail later.
[0032] The processor (130) predicts the airflow for each vent of the duct ventilation system using an airflow prediction model. Additionally, the processor (130) calculates the air pressure for each vent based on the predicted airflow for each vent.
[0033] The output unit (120) is a display that visualizes the airflow for each vent predicted by the processor (130) and the air pressure for each vent calculated by the processor (130). Through the output unit (120), the manager can identify the air conditions (airflow and air pressure) for each vent.
[0034] The control unit (140) is configured to control the AHU and dampers of the duct ventilation system, controls whether the AHU is operating and the supply air volume, and controls the air flow volume in each damper.
[0035] The input unit (140) receives from the manager whether the AHU is operating, the air supply volume, and the air flow volume in each damper, which are control variables by the control unit (150), and transmits them to the control unit (150).
[0036] The process of predicting the airflow and air pressure for each vent by the ventilation control system illustrated in FIG. 2 will be explained in detail below with reference to FIG. 3 and FIG. 4. FIG. 3 and FIG. 4 are diagrams illustrating the flow of a method for predicting air conditions for each vent of a duct ventilation system according to another embodiment of the present invention.
[0037] As described above, first, if the air flow rate of the dampers of the duct ventilation system is controlled by the manager through the ventilation control system (S210-Y), the ventilation control unit (140) of the ventilation control system operates the AHU of the duct ventilation system at a predetermined standard supply air volume (S220). Then, for each vent, the air pressure (P i ) and airflow (Qi Acquires ) (S230).
[0038] In the case where the duct ventilation system is as shown in FIG. 1, in step S230, the wind pressure (P1) and airflow (Q1) in the first duct, the wind pressure (P2) and airflow (Q2) in the second duct, the wind pressure (P3) and airflow (Q3) in the third duct, and the wind pressure (P4) and airflow (Q4) in the fourth duct must be obtained.
[0039] Wind pressure for each vent at stage S230 (P i ) is measured by temporarily installing a pressure gauge for each vent. The airflow (Q) for each vent i While it is possible to temporarily install and measure airflow meters for each vent, it is also possible to obtain the airflow through calculation or simulation from the standard supply airflow based on the duct design structure / specifications.
[0040] In the above case, the wind pressure (P) obtained for each vent i ) and airflow (Q i ) is input by the manager through the input unit (150) and transmitted to the processor (130).
[0041] Then the processor (130) obtains the wind pressure (P) for each vent acquired in step S230. i ) and airflow (Q i Using ), the loss coefficient (Loss coefficient, k) for each vent i Calculate ) (S240).
[0042] According to the Darcy-Weisbach equation, the wind pressure (P) for the i-th vent i ) and airflow (Q i The relationship between ) can be derived as shown in the following Equation 1.
[0043] [Formula 1]
[0044]
[0045] In the above equation, f is the coefficient of friction, L is the duct length, g is the acceleration due to gravity, p is the density of the fluid, i is the average flow velocity, and D is the duct diameter.
[0046] In an embodiment of the present invention, the above Equation 1 is simplified as the following Equation 2, and k i is named the loss coefficient at the i-th event.
[0047] [Equation 2]
[0048]
[0049] According to Equation 2 above, without measuring / calculating the complex parameters (f, L, g, p, i, D) in Equation 1, the wind pressure for each vent (P) obtained in step S230 i ) and airflow (Q i Using ), the loss coefficient (k) for each vent i You can easily calculate ).
[0050] In the case where the duct ventilation system is as shown in Fig. 1, in step S240, the loss coefficient (k1) in the first duct, the loss coefficient (k2) in the second duct, the loss coefficient (k3) in the third duct, and the loss coefficient (k4) in the fourth duct must be calculated.
[0051] The next processor (130) calculates the loss coefficient (k) for each vent calculated in step S240 for each vent. i ) is stored in the storage unit (110) (S250). By doing so, the preparation for the official operation of the duct ventilation system according to damper control by the manager is completed.
[0052] Afterwards, the duct ventilation system is put into full operation, but during full operation, damper control is not performed, and only the supply air volume by the AHU is controlled by the manager through the ventilation control system.
[0053] When the AHU of the duct ventilation system is officially put into operation (S260), the processor (130) uses an airflow prediction model to determine the supply air volume (Q) by the AHU. total) and the loss coefficient for each vent stored in step S250 (k i Air volume for each vent (Q) from ) i Predicts ) (S270).
[0054] The airflow prediction model is the supply airflow (Q total ) and the loss coefficient (k1) for each vent are input, and the airflow (Q) for each vent is input. i It is a machine learning model pre-trained to predict ). The wind volume prediction model can be implemented as a deep learning-based prediction model and can be trained through supervised learning.
[0055] In step S270, the processor (130) receives the supply air volume (Q) from the ventilation control unit (140). total Information regarding ) is received and utilized. In the case where the duct ventilation system is as shown in FIG. 1, the input to the airflow prediction model in step S270 is the supply airflow (Q total ) and the loss coefficients (k1, k2, k3, k4) for each of the four vents, and the output of the airflow prediction model is the airflows (Q1, Q2, Q3, Q4) for each of the vents.
[0056] Meanwhile, the airflow for each vent (Q) according to the airflow prediction model i If the prediction is inaccurate, a re-prediction can be performed by changing the input variables. This is divided into a process of determining whether the prediction by the wind volume prediction model is accurate and a process of performing a re-prediction.
[0057] The accuracy of the prediction depends on the airflow for each vent (Q) predicted by the airflow prediction model. i The sum of ) is the supply air volume (Q total It is determined by whether it matches ). In the case of the above example, if the following condition is satisfied, the prediction is determined to be inaccurate.
[0058] Q1 + Q2 + Q3 + Q4 ≠ Q total
[0059] Re-prediction involves adjusting the loss coefficients (k1, k2, k3, k4) and then using the airflow prediction model to determine the airflow for each vent (Q i It is to predict the airflow (Q) for each vent again. At this time, the airflow for each vent (Q i The sum of ) is the supply air volume (Q total If it is greater than ), the loss coefficients (k1, k2, k3, k4) are increased by a certain proportion. On the other hand, the airflow per vent (Q i The sum of ) is the supply air volume (Q total If it is smaller than ), the loss coefficients (k1, k2, k3, k4) are reduced by a certain proportion.
[0060] Subsequently, the processor (130) determines the airflow (Q) for each vent predicted in step S270. i ) and the calculated loss coefficient for each vent stored in step S250 (k i Pressure for each vent (P) from ) i Calculate ) (S280). Pressure for each vent at step S280 (P i The calculation of ) can be performed using the aforementioned Formula 2.
[0061] The next processor (130) predicts the airflow (Q) for each vent in step S270. i ) and the pressure for each vent calculated in step S280 (P i Outputs ) (S290). Through this, the manager can check the wind conditions for each vent (Q i , P i It becomes possible to grasp ).
[0062] Meanwhile, the above embodiment describes predicting both airflow and air pressure for each vent, but this is merely illustrative. It is also possible to implement it to predict only one of the airflow or air pressure.
[0063] So far, a method for predicting airflow and air pressure for each vent of a duct ventilation system has been described in detail with reference to preferred embodiments.
[0064] In the above embodiment, it is possible to predict the airflow and air pressure of each vent of a duct ventilation system based on machine learning, thereby eliminating the need for installation and maintenance of vent-specific sensors to predict the airflow and air pressure required for controlling the opening and closing of dampers in the event of an emergency such as a fire, thus reducing effort and cost.
[0065] Meanwhile, the structure of the duct ventilation system presented in the above embodiment is exemplary, and other structures of duct ventilation systems can be conceived, and the technical concept of the present invention can also be applied in such cases. In particular, the duct ventilation system of a subway station utilizes a ventilation system having a complex duct structure that branches into sections as shown in FIG. 5.
[0066] In the embodiments of the present invention, the airflow and air pressure information for each vent predicted is referenced for controlling indoor ventilation volume in normal situations, but is also referenced for controlling smoke exhaust / fire control loads in the event of a fire. In particular, air pressure is referenced to prevent fire spread by controlling the differential pressure between compartments in the event of a fire.
[0067] Meanwhile, it goes without saying that the technical concept of the present invention may also be applied to a computer-readable recording medium containing a computer program that enables the device and method according to the present embodiment to perform their functions. Furthermore, the technical concept according to various embodiments of the present invention may be implemented in the form of computer-readable code recorded on a computer-readable recording medium. A computer-readable recording medium may be any data storage device that can be read by a computer and store data. For example, a computer-readable recording medium may be a ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical disk, hard disk drive, etc. Additionally, computer-readable code or a program stored on a computer-readable recording medium may be transmitted through a network connected between computers.
[0068] Furthermore, although preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above. Various modifications are possible by those skilled in the art without departing from the essence of the invention as claimed in the claims, and such modifications should not be understood individually from the technical spirit or perspective of the present invention.
Claims
1. A ventilation control system acquiring air pressure and airflow for each vent of the duct ventilation system; A ventilation control system calculates a loss coefficient for each vent from the acquired air pressure and airflow for each vent; A ventilation control system, a step of storing the calculated loss coefficient for each vent; A method for predicting airflow conditions per vent of a duct ventilation system, characterized by including the step of a ventilation control system predicting airflow per vent from the supply airflow of the duct ventilation system and a stored loss coefficient per vent.
2. In Claim 1, The acquisition stage is, A method for predicting air conditions for each vent of a duct ventilation system, characterized by operating the duct ventilation system at a predetermined standard supply air volume and obtaining air pressure and air volume for each vent.
3. In Claim 2, The acquisition stage is, Temporarily install a wind pressure gauge for each vent to acquire the wind pressure for each vent, and A method for predicting airflow conditions per vent of a duct ventilation system, characterized by temporarily installing an airflow meter for each vent to obtain the airflow for each vent.
4. In Claim 2, The acquisition stage is, Temporarily install a wind pressure gauge for each vent to acquire the wind pressure for each vent, and A method for predicting airflow conditions for each vent of a duct ventilation system, characterized by obtaining the airflow volume for each vent through calculation or simulation from a standard supply airflow volume, by referring to the design structure and specifications of the duct.
5. In Claim 1, The step for calculating the loss coefficient for each vent is, Calculate using the following formula, Here, P i is the pressure for each vent, k1 is the loss coefficient for each vent, Q i A method for predicting airflow conditions per vent of a duct ventilation system, characterized by airflow volume per vent.
6. In Claim 1, The prediction stage is, A method for predicting airflow conditions for each vent of a duct ventilation system, characterized by predicting airflow for each vent using an airflow prediction model, which is a machine learning model pre-trained to receive the supply airflow and the loss coefficient for each vent as inputs and predict the airflow for each vent.
7. In Claim 6, The prediction stage is, A method for predicting airflow conditions by vent of a duct ventilation system, characterized by changing the loss coefficient for each vent and re-predicting the airflow for each vent using an airflow prediction model if the prediction result is inaccurate.
8. In Claim 1, A method for predicting airflow conditions per vent of a duct ventilation system, characterized by further including the step of calculating the pressure per vent from the predicted airflow per vent and the stored loss coefficient per vent of the ventilation control system.
9. In Claim 8, A method for predicting airflow conditions per vent of a duct ventilation system, characterized by further including the step of the ventilation control system outputting a predicted airflow volume per vent and a calculated pressure per vent.
10. A storage unit for storing the loss coefficient for each vent of the duct ventilation system; A processor that calculates a loss coefficient for each vent from the air pressure and airflow for each vent and stores it in a storage unit, and predicts the airflow for each vent from the supply airflow of the duct ventilation system and the loss coefficient for each vent stored in the storage unit; and A wind condition prediction system for a duct ventilation system characterized by including an output unit that outputs the predicted airflow volume for each vent.
11. A ventilation control system storing a loss coefficient for each vent calculated from the air pressure and airflow for each vent; A ventilation control system predicts the airflow volume for each vent from the supply airflow volume of the duct ventilation system and the stored loss coefficient for each vent; A ventilation control system calculates the pressure for each vent from the predicted airflow for each vent and the stored loss coefficient for each vent; A ventilation control system outputs a predicted airflow rate for each vent and a calculated pressure for each vent; and A method for controlling a duct ventilation system characterized by including the step of controlling the duct ventilation system according to control variables entered by an administrator, based on reference to the airflow and pressure for each vent output by the ventilation control system.
12. A storage unit that stores loss coefficients for each vent calculated from wind pressure and airflow for each vent; A processor that predicts the airflow for each vent from the supply airflow of a duct ventilation system and a stored loss coefficient for each vent, and calculates the pressure for each vent from the predicted airflow for each vent and the stored loss coefficient for each vent; An output unit that outputs the predicted airflow for each vent and the calculated pressure for each vent; and A duct ventilation control system characterized by including a ventilation control unit that controls the duct ventilation system according to control variables entered by an administrator, based on reference to the airflow and pressure for each output vent.