Method for monitoring the operation of a fan for a stall, device, fan and computer program product

Abstract

The invention relates to a method for monitoring the operation of a fan for a stall, the method including the following steps: 1a) continuously or repeatedly detecting at least one measurement variable of the fan over a first time period during a first operating state (1) in which there is a stall, and / or 1b) continuously or repeatedly detecting at least one measurement variable of the fan over a second time period during a second operating state (2) in which there is no stall, 2) applying a machine pattern recognition to the data of the at least one measurement variable over the first and second time period detected in steps 1a) and / or 1b) in order to assign at least one pattern to at least one of the first operating state (1) and the second operating state (2) by means of which pattern this operating state can be uniquely distinguished from the respective other operating state, 3a) continuously or repeatedly detecting at least one measurement variable of the fan during operation, 3b) classifying a current operating state of the fan to the first or the second operating state based on the data of the at least one measurement variable of the fan during operation captured in step 3a).

Description

[0001] Method for monitoring the operation of a fan for a stall, device, fan and computer program product

[0002] The invention relates to a method for monitoring the operation of a fan for flow stall, a device for monitoring the operation of a fan for flow stall and a fan with such a device.

[0003] Methods and devices for monitoring the operation of fans of the type in question have been known in practice for years.

[0004] When operating fans, for example an axial fan, the fan can reach or approach an operating point or operating state in which the flow at the rotor or rotor blades breaks off.

[0005] Stall represents an aeronautical, non-stationary operating point or condition in which performance and / or efficiency losses may occur. Therefore, operation in this condition should be avoided.

[0006] With regard to permanently efficient operation, it is advantageous if this transient operating point can be detected and the operator warned so that corrective measures can be initiated.

[0007] Furthermore, there are also regulatory requirements in building services engineering to ensure sustained efficiency in fan operation. Improvements in efficiency can be achieved significantly through demand-based control.

[0008] For fans that have sensors for speed, current, etc. – for example, EC fans with preferably internal sensors – the operating state in which a stall occurs can be determined via fluctuations in measured variables, such as speed and motor current. However, there is no clear mathematical relationship between measured variables and operating state that allows for a simple classification as to whether the fan is in a first operating state in which a stall occurs or in a second operating state in which no stall occurs.

[0009] The present invention is therefore based on the object of designing and developing a method and a device for monitoring the operation of a fan for a flow stall of the type mentioned above in such a way that operation can be further optimized over the service life of the fan and operating states in which a flow stall occurs can be detected and preferably avoided. Furthermore, an improved fan with a correspondingly configured device and a computer program product are to be provided.

[0010] According to the invention, the above object is achieved by the features of claim 1. According to this, the method in question for monitoring the operation of a fan for a stall comprises the following steps: la) Continuous or repeated recording of at least one measured variable of the fan over a first period of time during a first operating state in which a stall is present, and / or lb) Continuous or repeated recording of at least one measured variable of the fan over a second period of time during a second operating state in which no stall is present,

[0011] 2) Applying machine pattern recognition to the data of the at least one measured variable acquired in steps 1a) and / or 1b) over the first and second time periods in order to assign at least one pattern to at least one of the first operating state and the second operating state, by means of which this operating state can be clearly distinguished from the other,

[0012] 3a) Continuously or repeatedly recording at least one measured variable of the fan during operation, 3b) Classifying a current operating state of the fan as the first or the second operating state based on the data of the at least one measured variable of the fan during operation recorded in step 3a).

[0013] According to the invention, it was first recognized that the fluctuations in the measured values ​​are expressed in patterns. These patterns can change over a longer period of time. For the continuous or repeated recording of at least one measured variable, steps 1a) and 1b) can be carried out cumulatively or alternatively, whereby a temporal sequence is not important. By applying machine pattern recognition according to step 2), in particular from the field of artificial intelligence (K1), such patterns can be recognized. The clear classification of the operating state according to step 3b) can then be used to maintain the efficiency of the fan's operation. Steps 3a) and 3b) can be carried out continuously or repeatedly during the operation of the fan or over the service life of the fan.

[0014] This provides a method for further optimizing operation over the fan's lifetime. Operating conditions in which flow stall occurs can be detected and avoided.

[0015] Preferably, the at least one measurement variable may comprise one or more of the following:

[0016] Speed ​​information, such as a rotation rate, a motor current,

[0017] Vibration information, for example acceleration in one, two or three spatial axes,

[0018] Sound information, for example structure-borne sound, a temperature, a sound pressure, a flow velocity, a torque, a pressure, in particular an operating or ambient pressure, a humidity value and / or a measured force.

[0019] Data can be obtained from electrical parameters associated with the respective measured quantity. A time signal with a fixed sampling frequency (fs) can be recorded for each measuring point.

[0020] Applying machine pattern recognition to the acquired data of measured variables is also referred to as training. Training can be supervised or unsupervised. Supervised training requires measurement points from both operating states, inside and outside the stall. For unsupervised training, data from either the first or second operating state is sufficient.

[0021] Preferably, the machine pattern recognition may comprise processing the acquired data using a function library. This function library can detect whether the first or second operating state is present based on the patterns and can feed this result back to the fan's closed-loop or closed-loop control system.

[0022] Further preferably, the function library can be modified by machine learning during the execution of the method. Machine learning (ML) can be used to classify a current operating state of the fan as belonging to the first or second operating state. In particular, an ML model can be used as the decision function. This model can be trained using recorded data. In other words, a neural network can be trained. The training of the ML model can take place "on-edge" on the fan controller, on a third-party device, for example, an external computing unit, and / or in a cloud.

[0023] According to one embodiment, the classification of a current operating state of the fan can be used to enable highly precise volume flow control. In known volume flow-controlled systems, an operating state in which a flow stall occurs can occur without being noticed. Reliable detection of the operating state enables particularly high-precision volume flow control, which offers advantages, for example, in dosing applications where the volume flow control must function as precisely as possible and incorrect dosing should not occur in the event of malfunctions.

[0024] During long-term operation of volume flow control, changes in the fan's system can occur, for example, due to contamination. These changes can cause the fan to drift from its operating point or operating state at the time of commissioning. In the worst case, there is a risk that the fan will enter an operating state in which flow stall occurs.

[0025] For particularly efficient operation, the classification of a current operating state of the fan can be used to prevent the first operating state from occurring during fan operation. In other words, pattern recognition can be used to prevent the fan from entering an operating state in which a stall occurs.

[0026] The method can be executed entirely locally, for example, on a fan controller of the associated fan to be monitored. Preferably, data, in particular data obtained by applying machine pattern recognition in step 2), can be transmitted via a cloud service. Connections to third-party devices and / or a cloud function can be established.

[0027] Applying machine pattern recognition to the acquired data can be performed on a third-party device. There are cases where a fan or fan controller does not have sufficient processing power to perform pattern recognition, but the sensors and an interface are available to transmit sensor values ​​to a third-party device. In these cases, it is advantageous to perform pattern recognition on the third-party device. This third-party device can be an external device, such as an external processing unit. It can be a programmable logic controller (PLC) or a comparable system with the appropriate processing power.

[0028] With regard to the device for monitoring the operation of fans, the aforementioned object is achieved by a device having at least one sensor for detecting at least one measured variable of the fan and a computing unit configured to carry out the method described above. The computing unit can be, for example, the fan controller or an external computing unit.

[0029] With regard to the improved fan, the object mentioned at the outset is achieved by a fan which comprises a described device for monitoring its operation.

[0030] With regard to the computer program product, the object mentioned at the outset is achieved by a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method according to claim 1. Thus, according to the object, an improved fan with a correspondingly configured device for monitoring the operation of fans and a computer program product are specified.

[0031] There are now various possibilities for advantageously embodying and developing the teaching of the present invention. For this purpose, reference is made, on the one hand, to the claims subordinate to claim 1 and, on the other hand, to the following explanation of preferred embodiments of the invention with reference to the drawing. In conjunction with the explanation of the preferred embodiments of the invention with reference to the drawing, generally preferred embodiments and developments of the teaching are also explained. The drawing shows

[0032] Fig. 1 shows a schematic flow diagram of an embodiment of the method according to the invention. Fig. 1 generally shows a flow diagram of a method for monitoring the operation of a fan for stall. In the upper part of the flow diagram, a first operating state 1 is shown on the left, in which a stall is present. A second operating state 2 is shown in the top center, in which no stall is present.

[0033] In the first operating state 1 and / or the second operating state 2, at least one measured variable of the fan is recorded. The flowchart shows how the data associated with the measured variables is recorded – preferably as a synchronous time signal. Recording 3 of data from the first operating state 1 is distinguished from recording 4 of data from the second operating state 4.

[0034] The recorded data from both operating states 1 and 2 are saved 5. With the data thus obtained, an ML model for machine pattern recognition is trained 6. In other words, machine pattern recognition according to step 2) is applied to the data of the at least one measured variable recorded in steps 1a) and / or 1b) over the first and second time periods in order to assign at least one pattern to at least one of the first operating state 1 and the second operating state 2, by means of which pattern this operating state can be clearly distinguished from the other.

[0035] The ML model is trained using the data obtained from recordings 3 and 4, resulting in a trained model. This model can be used to classify newly recorded measurement points. According to the exemplary embodiment, the model is applied directly to the device, in other words, "on edge" to the fan controller. According to the exemplary embodiment, there is no connection to a third-party device and / or a cloud function. However, it is also conceivable that the model could be exchanged / updated via a cloud service. This also makes it possible to train a model customized for the respective device—exclusively using the data from that individual device.

[0036] The upper part of the flowchart on the right also shows the recording 7 of data from an unknown current operating state. This data is obtained by continuously or repeatedly recording at least one measured variable of the fan during operation according to step 3a). With the help of machine pattern recognition, it is then possible to classify the current operating state of the fan as the first or second operating state based on the data of the at least one measured variable of the fan during operation recorded in step 3a). In this case, the ML model 8 is applied in the exemplary embodiment to classify the current operating state of the fan. This classification provides a clear binary result 9 as to whether the current operating state is a first operating state 1 or a second operating state 2, or in other words as to whether a flow stall is present or not.The user is therefore provided with the binary result 9, which classifies the currently processed measuring point as “in the break” or “out of the break”.

[0037] The training of ML model 6 can also be continued during the application of ML model 8, so that machine pattern recognition is further optimized over the lifetime of the fan.

[0038] With regard to further advantageous embodiments of the device according to the invention, reference is made to the general part of the description and to the appended claims in order to avoid repetition.

[0039] Finally, it should be expressly pointed out that the embodiments of the device according to the invention described above serve only to explain the claimed teaching, but do not limit it to the embodiments.

[0040] List of reference symbols

[0041] First operating state

[0042] Second operating state

[0043] Recording of data in the first operating state

[0044] Recording data in the second operating state

[0045] Storing data

[0046] Training an ML model

[0047] Recording data from an unknown current operating state

[0048] Applying the ML model

[0049] Binary result

Claims

Claims 1. A method for monitoring the operation of a fan for a stall, the method comprising the following steps: la) Continuous or repeated recording of at least one measured variable of the fan over a first period of time during a first operating state (1) in which a stall is present, and / or lb) Continuous or repeated recording of at least one measured variable of the fan over a second period of time during a second operating state (2) in which no stall is present, 2) applying machine pattern recognition to the data of the at least one measured variable acquired in steps 1a) and / or 1b) over the first and second time periods in order to assign at least one pattern to at least one of the first operating state (1) and the second operating state (2), by means of which this operating state can be clearly distinguished from the other, 3a) Continuous or repeated recording of at least one measured variable of the fan during operation, 3b) Classifying a current operating state of the fan as the first or second operating state based on the data of the at least one measured variable of the fan during operation recorded in step 3a).

2. The method according to claim 1, wherein the at least one measured variable comprises one or more of the following: Speed ​​information, such as a rotation rate, a motor current, Vibration information, for example acceleration in one, two or three spatial axes, Sound information, such as structure-borne sound, temperature, a sound pressure, a flow velocity, a torque, a pressure, in particular an operating or ambient pressure, a humidity value and / or a measured force.

3. The method of claim 1 or claim 2, wherein machine pattern recognition comprises processing the acquired data using a function library.

4. The method according to claim 3, wherein the function library is modified by machine learning (6, 8) during the execution of the method.

5. Method according to one of claims 1 to 4, wherein the classification of a current operating state of the fan is used to enable highly accurate volume flow control.

6. Method according to one of claims 1 to 5, wherein the classification of a current operating state of the fan is used to prevent the first operating state (1) from occurring during operation of the fan.

7. The method according to any one of claims 1 to 6, wherein Data, in particular data obtained by applying machine pattern recognition in step 2), is transferred via a cloud service.

8. The method according to any one of claims 1 to 7, wherein the application of machine pattern recognition to the data acquired in steps 1a) and / or 1b) is carried out on a third-party device.

9. Device for monitoring the operation of a fan for flow separation, comprising: at least one sensor for detecting at least one measured variable of the fan and a computing unit which is designed to carry out the method according to one of claims 1 to 8.

10. Fan with a device according to claim 9.

11. A computer program product comprising instructions which, when executed by a computer, cause the computer to carry out the method according to claim 1.

Images