Method for monitoring at least one rotating working machine on basis of sensor

Abstract

The invention relates to a method for monitoring at least one rotating working machine, in particular a centrifugal pump, comprising at least one sensor for detecting at least one measured variable which is dependent on the operation of the working machine and a central evaluation unit which is communicatively connected to the at least one sensor, the method comprises the following method steps: transmitting at least one measured value and / or a measured value pre-processed on the sensor side by the sensor to the central evaluation unit; evaluating the at least one measured value and / or the pre-processed measured value by an evaluation unit and generating at least one instruction for dynamically and remotely configuring the sensor behavior of the at least one sensor on the basis of the evaluation result; and transmitting the instructions to the at least one sensor by the evaluation unit.

Description

Technical Field

[0001] The present invention relates to a method for monitoring at least one rotating working machine, in particular a centrifugal pump, the working machine comprising at least one sensor for collecting at least one measurement parameter related to the operation of the working machine and a central evaluation unit communicatively connected to the at least one sensor. Background Art

[0002] Automated condition monitoring of working machines, in particular rotating working machines, is already known. For this purpose, for the purpose of condition monitoring, one or more sensors are assembled near or at the machine to be monitored. The measured values of the sensors are collected periodically and then transmitted directly or after preprocessing in the sensor to the cloud or another device.

[0003] Machine monitoring is usually associated with an automatic alarm, which should alarm the operator of the machine or facility as quickly and reliably as possible in the event of a critical state. In machines with dynamic operating behavior, in order to reliably collect and identify critical states, it is often necessary to perform measurements at small measurement intervals, because only in this way can it be ensured that critical states can also be collected and identified by the implemented measurements. Similar problems also exist when monitoring machines with short-cycle operation. Ideally, the execution of the measurement falls on a time point during which the machine is actually active. If, instead, the measured values are collected at the time point when the machine is switched off, then efficient monitoring cannot be meaningfully carried out.

[0004] However, continuous monitoring measurements carried out at short time intervals are not reasonable in terms of energy and cannot be meaningfully implemented precisely in solutions with battery-operated sensors, because there is a huge energy requirement due to the short measurement intervals and the resulting high-frequency data transmission to the central evaluation unit, which greatly shortens the battery life.

[0005] Another problem is the influence of the machine's surrounding environment on the measured values. If the machine to be monitored is part of a large facility, the vibrations, noises, etc. of adjacent machines may cause the measured values to be distorted and trigger false detections. It is desirable to filter out these influences as well as possible to avoid false alarms. Summary of the Invention

[0006] Therefore, the present invention is described with the following task of optimizing the existing sensor-based monitoring method (in which there is data exchange between the sensor and the central evaluation unit) such that it can be implemented more efficiently, more quickly and also more energy-efficiently.

[0007] This task is solved by a method according to the features of claim 1. Advantageous embodiments of the method are the subject matter of the dependent claims.

[0008] According to the present invention, a dynamic remote configuration of the sensor behavior is proposed by means of a central evaluation unit. The remote configuration of the sensor behavior should be understood as the adaptation of the sensor operation or some operating parameters of the sensor, and the adaptation involves or affects the measured value acquisition and the measured value evaluation. In this regard, the "sensor" should not only be understood as a pure measured value receiver, but also as a sensor controller or other unit for preprocessing and / or evaluating the measured values. In addition, the sensor can be equipped with an integrated communication module for communicating with the central evaluation unit, or be communicatively connected to an external communication device for indirect communication with the evaluation unit. The data communication between the evaluation unit and the sensor can be carried out through at least one intermediate node, in particular a gateway installed within the direct reception range of the sensor. The gateway can preferably have an intermediate buffer for temporarily storing the communication data that should be exchanged between the sensor and the evaluation unit. At least one sensor can call the temporarily stored data in the gateway when needed.

[0009] The transmission of instructions from the evaluation unit can be achieved by the evaluation unit sending the instructions to the sensor, or alternatively can be achieved by the sensor calling the instructions in the evaluation unit.

[0010] The dynamic remote configuration of the sensor is a necessary prerequisite in order to optimize the measurement operation or the evaluation process of the sensor during the monitoring of a running machine, taking into account the previous measured value evaluation, so that the entire measurement process and monitoring can be dynamically adapted to the actual situation. Overall, a faster, more reliable and more energy-efficient monitoring can be achieved.

[0011] For the implementation of the method, first at least one acquired measured value and / or the measured value preprocessed on the sensor side must be transmitted by the sensor to the central evaluation unit. Then, on the evaluation unit side, the received measured value or the preprocessed measured value is evaluated, and instructions for dynamically remotely configuring the sensor behavior of at least one sensor are generated according to the evaluation result, and the instructions are transmitted to the sensor for remote configuration.

[0012] For example, there is a possibility that at least one repeated measurement is triggered on the sensor side by means of an instruction. A repeated measurement is understood as a measurement to be performed in the short term after obtaining and evaluating a previous measurement value in order to be able to verify the previous measurement. For this purpose, the repeated measurement must be carried out as soon as possible after the first measurement, depending on the application, especially at intervals of a few seconds, for example within 60 seconds after the measurement to be verified, ideally after 30 seconds or less. It is also conceivable that, by means of an instruction, not only the repeated measurement is triggered, but instead the number of repeated measurements to be carried out can also be configured, or the time interval between at least two successive repeated measurements can be set remotely. After receiving such an instruction, the received configuration of the process flow for the repeated measurement is converted on the sensor side.

[0013] It is also imaginable that a specific threshold for threshold monitoring inside the sensor is configured by means of an instruction. After receiving the corresponding instruction with at least one threshold, the sensor performs the corresponding process steps for threshold monitoring, for example by comparing the currently respectively acquired measurement values with the remotely configured threshold, or by comparing other operating parameters of the sensor with the remotely configured threshold. After exceeding, at least one subsequent measure can be triggered.

[0014] Threshold monitoring can preferably be used for fault identification. For example, when a configurable threshold is exceeded, the sensor can assume an alarm state and / or generate and output an alarm notification. It is also conceivable that at least one repeated measurement is triggered due to the threshold exceedance in order to verify the previous measurement by means of the repeated measurement.

[0015] It is also imaginable that exceeding the threshold causes the sensor to communicate data with a central evaluation unit or with other communication partners. This ensures that only relevant measurement values are transmitted to the evaluation unit, which reduces the total energy requirement due to less communication.

[0016] It is also imaginable that the sensor includes two or more, preferably different, measurement value receivers. For reasons of diversity, the measurement value receivers can acquire the same measurement values. However, it is also conceivable that the measurement value receivers acquire different measurement parameters. By means of an instruction from the central evaluation unit, it is possible to switch between the measurement value receivers, or to change the redundancy or diversity in the measurement operation. In principle, one or more measurement value receivers can be selectively deactivated or activated.

[0017] Furthermore, it is imaginable that the sensor can be placed in a sleep or low-power mode by means of an instruction. In such a sleep or low-power mode, for example, the measurement value acquisition is completely deactivated or reduced to a minimum in order to temporarily reduce the energy consumption in the sensor to a minimum. Of course, the sensor can also operate in the low-power mode with a reduced measurement frequency and / or sampling rate, etc.

[0018] It is also conceivable that during the sleep mode or the low-power mode, the data communication between the sensor and the central evaluation unit is completely interrupted or at least reduced to a minimum. By means of an instruction, not only can such a sleep or low-power phase be triggered actively, but also the duration of such a sleep or low-power phase can be determined by means of an instruction.

[0019] According to an advantageous embodiment, the manner of measuring value acquisition in the sensor can be configured by means of the generated instruction of the central evaluation unit. This relates, for example, to the sampling rate and / or the measurement duration of a single measurement carried out by the measurement value receiver of the sensor. It is also possible to remotely configure the preprocessing of the measurement values inside the sensor, that is to say, for example, it can be configured that the measurement values of a measurement sequence are summarized and only the average value of the measurement sequence is communicated to the evaluation unit. Preferably, the data size of the measurement data or the raw data to be transmitted can be configured by means of an instruction.

[0020] In principle, the method according to the invention can be applied to any type of measurement value. Exemplarily, the application to mechanical vibration measurement and monitoring in a rotating machine is mentioned here. The application of the method in temperature measurement or temperature monitoring is also particularly preferred. Particularly preferred is the application in vibration monitoring, which is supplemented by temperature measurement or temperature monitoring. If the remotely configurable sensor is suitable for acquiring mechanical vibrations, it is preferably possible to set or deactivate / activate the vibration axis to be measured by remote configuration. In addition, it is also meaningful to remotely configure the limit frequency.

[0021] It is also advantageous that at least one sensor is a battery-operated sensor. The data communication between the sensor and the central evaluation unit is preferably carried out by means of a radio connection, where any data transmission standard can be used here. Especially in the case of a battery-operated sensor, it is advantageous to remotely configure the sensor operation in order to optimize the sensor operation in terms of energy consumption and to ensure as long a battery operation time as possible with sufficient measurement accuracy and measurement reliability.

[0022] According to a preferred embodiment of the method, it can be provided that after the sensor has sent one or more measured values to the central evaluation unit, it first transitions into a first sleep mode. During such a first sleep mode, the energy consumption of the sensor is reduced to a minimum, for example, by stopping the data communication with the central evaluation unit and / or the acquisition of measured values during the sleep mode. After the end of the first sleep mode, the sensor can receive at least one instruction from the evaluation unit and / or perform an internal evaluation of the previously acquired measurement data on the sensor side. The duration of the first sleep mode is selected such that the sensor is woken up in time to receive an instruction from the central evaluation unit and to ensure the activation of the data communication for the instruction reception. The process duration required for the evaluation of the measured values on the evaluation unit side is known or can be well estimated, so that the duration of the first sleep mode can be matched to this process duration.

[0023] If the data communication between the evaluation unit and the sensor is carried out indirectly via an intermediate unit, in particular a gateway, there is the possibility that certain data packets, in particular instructions, are temporarily stored in the gateway and made available for the sensor to call. In this case, the duration of the sleep mode can be designed more flexibly, because the sensor can call the data for the sensor from the gateway as needed, that is, according to a personalized selection of the sleep duration.

[0024] Then, it can be decided based on the evaluation result whether a repeated measurement should be carried out. This decision is preferably made in the evaluation unit. However, if there is an internal evaluation of the measured values on the sensor side, it is also conceivable to make the decision inside the sensor. If the evaluation is carried out on the central evaluation unit side, it is also decided in the evaluation unit whether a repeated measurement should be carried out, and then the repeated measurement is prompted by means of a remote configuration of the sensor. On the contrary, if the evaluation of the measured values is carried out inside the sensor, the corresponding decision regarding the implementation of a repeated measurement can also be made inside the sensor. If the previous measurement and at least one repeated measurement provide consistent or at least comparable results, the measurement is considered to be verified and taken into account for the decision on possible subsequent measures, such as for fault detection and, if necessary, for triggering an alarm.

[0025] It is also conceivable that the number of repeated measurements to be carried out is remotely configured by means of the central evaluation unit. Thus, multiple repeated measurements can be advantageous in order to increase the available data volume and to improve the quality of the data evaluation and the state recognition of the working machine.

[0026] If it is determined after the end of the method that repeated measurements are not required, the sensor is preferably placed in a second sleep mode, in which neither measurement value acquisition nor data transmission to the central evaluation unit takes place. The duration of the second sleep mode can hereby be defined to be significantly longer than the duration of the first sleep mode.

[0027] According to another embodiment of the invention, it can be provided that the sensor acquires measurement values at a first measurement interval in a first operating mode. In the first operating mode, the acquired measurement values are transmitted to the central evaluation unit via a communication channel after each measurement carried out. The evaluation unit receives the measurement values and evaluates the measurement values belonging to the measurement sequence that have been received. Depending on the evaluation result, it can be provided that the central evaluation unit causes the sensor to operate in a second operating mode.

[0028] For example, it can be imagined that at least one recognition threshold is determined by evaluating the measurement sequence and transmitted to the sensor. With the aid of such a recognition threshold, it is possible to distinguish, for example, between different machine states, in particular whether the machine is currently in operation or in a stationary phase.

[0029] Preferably, for example, the sensor acquires measurements at a second measurement interval in the second operating mode, where the second measurement interval is selected to be shorter than the first measurement interval, in particular at least by a factor of 5, preferably 10, shorter than the first measurement interval. Furthermore, it is provided that during the second operating mode, no measurement data transmission to the evaluation unit takes place, or at least less data is transmitted to the central evaluation unit than in the first operating mode. The second operating mode is particularly advantageous for monitoring machines with short cycle operation, because the shortened measurement interval ensures that measurement values can be acquired even during the active operation of the machine. The increased energy requirement due to the shortened measurement interval is in turn compensated by an interrupted or reduced data communication with the central evaluation unit.

[0030] Preferably, during the second operating mode, the acquired measurement values are compared in the sensor with the recognition threshold, in particular with the recognition threshold previously determined by the evaluation unit. When the recognition threshold is exceeded, the sensor switches to the first operating mode in order to continuously transmit the acquired measurement values to the central evaluation unit.

[0031] If the measurement value drops below the recognition threshold again, the sensor preferably returns to the second operating mode.

[0032] It is particularly advantageous that the recognition threshold allows a distinction to be made between active or inactive machine states. By defining such a recognition threshold, it is ensured that the sensor only transmits measurement data to the central evaluation unit when it has been fully determined based on the measurement data that the machine is in operation.

[0033] This way of behaving is particularly preferred in sensors for mechanical vibration measurement, because the amplitude of the vibration collected already provides an indication of whether the machine is in operation or instead at rest.

[0034] In addition to the method according to the invention, the invention also relates to a system which comprises at least one central evaluation unit, in particular a cloud-based evaluation unit, and at least one sensor, wherein the evaluation unit and the sensor are configured to carry out the method according to the invention described above. The system can additionally comprise an intermediate gateway.

[0035] In addition to the overall system, the invention also relates to a sensor for such a system, wherein the sensor is configured to be remotely configured dynamically by the central evaluation unit. Finally, the invention also relates to a central evaluation unit for a system according to the invention, wherein the central evaluation unit is configured to generate and transmit instructions for the dynamic remote configuration of the sensors of the communication connection based on the evaluation result of one or more measured values.

[0036] For the system, the sensor and the central evaluation unit, the same advantages and characteristics as those previously demonstrated by means of the method according to the invention thus result, so that a repeated description is omitted at this point. Description of the Drawings

[0037] The other advantages and characteristics of the method are shown in more detail below by means of different embodiments implementing the invention. Among them:

[0038] Figure 1 A flow chart of a sensor device with repeated measurements is shown,

[0039] Figure 2 A conventional time diagram for a short-running machine with a sensor is shown,

[0040] Figure 3 A corresponding flow chart according to the invention for the intelligent identification of the pump state by means of a sensor device is shown. Detailed Description of the Invention

[0041] The method according to the invention can be applied to all machines that are to be monitored by sensors, in particular battery-operated sensors. Here, for the purpose of condition monitoring, one or more sensors are installed in the vicinity of the machine to be monitored, such as a pump, in particular a centrifugal pump. The measured values of the sensors are automatically collected periodically and then sent directly or after preprocessing in the sensor to a central evaluation unit, in particular to a cloud-based solution of the evaluation unit or to other instruments. For machine monitoring, at least one sensor for collecting mechanical vibrations can be used, for example. The sensors, in particular battery-operated sensors, include an integrated communication module that enables data exchange between the sensor and the central evaluation unit. In particular, in larger facilities with a large number of machines to be monitored, in particular pumps, it is recommended to use a gateway installed near the machines and connected to a large number of sensors. The signals received by the sensors via a wireless standard are then conveyed by the gateway via an Internet connection to a cloud-based evaluation unit.

[0042] Machine monitoring is usually linked to an automatic alarm that should alert the operator of the machine or facility as quickly and reliably as possible in the event of a critical state. The present invention has the following objectives: to significantly improve the reliability of such alarm notifications. Another important objective is to improve the accuracy of the associated operating hour counter. Through the temporal measurement intervals of the sensors used and the dynamicization according to the invention of the measurement behavior, a more accurate diagnosis of the state of the monitored machine can be made in an energy-efficient manner. For this purpose, there are different specific application scenarios, which are described in more detail below.

[0043] An important feature of the present invention is the possibility of dynamically changing the behavior of the sensors by the central evaluation unit (i.e., the cloud) without the associated additional energy requirements. As Figure 1 shown, one possibility for controlling the sensor behavior can work as follows:

[0044] The sensor first collects measurement data during machine operation, such as mechanical vibration data and / or temperature values in the environment of the machine to be monitored, and undergoes preprocessing and pre-evaluation of the data (block 10). Subsequently, the sensor sends the measurement data or the preprocessed data to the cloud (block 20).

[0045] Subsequently, the sensor transitions into a "short-term sleep mode" for a short period (e.g., 30 seconds) (block 30) to save energy. Based on the current sensor data and the past sensor data from the same sensor, the future desired behavior of the sensor is automatically determined in the cloud (reference numeral 40). This calculation or evaluation is completed on the server side in the cloud with an extremely short calculation time and can, for example, include comparing the received sensor values with stored thresholds. Similarly, an alarm notification can already be generated on the cloud side based on the evaluation result at this point (block 45), where the alarm notification is preferably only temporarily stored first, and instead, repeated measurements are required through the sensor to verify the alarm status.

[0046] Now, the cloud sends the desired sensor behavior as a reply back to the sensor, and the sensor receives one or more instructions for internally matching the sensor behavior (block 50). Due to the short processing time in the cloud, it is ensured that there is a result before the end of the "short-term sleep mode" of the sensor. If there is also a gateway on the communication path between the sensor and the cloud (which transmits data in both directions), then the gateway can receive the reply determined for the sensor from the cloud and temporarily store it. Once the sensor resumes from the "short-term sleep mode", the sensor either:

[0047] a) directly from the cloud, or

[0048] b) obtains instructions for its further behavior from the gateway that has temporarily stored the information determined for the sensor from the cloud.

[0049] In case b), the process functions significantly faster and thus more energy-efficiently than in case a). Corresponding to the reply received from the cloud, the sensor matches its future desired mode of operation (block 50).

[0050] For example, with remote configuration, repeated measurements can then be triggered in the sensor and its implementation can be configured (block 60).

[0051] If it is decided in block 60 that the sensor does not perform repeated measurements because the sensor value is, for example, below the limit value:

[0052] The sensor returns (without further action) to the normal (longer-duration) sleep mode (block 70), for example, for more than 60 minutes, especially 120 minutes or longer, because the current data does not require special interaction;

[0053] For the case where at least one repeated measurement should be implemented in block 60 (either through an instruction from the cloud or through an internal evaluation of the sensor), then according to the preset from the cloud, for example, the execution can occur as follows:

[0054] a. The sensor immediately performs a repeated measurement (and also directly transmits the data to the cloud again).

[0055] b. The sensor performs multiple repeated measurements at short intervals.

[0056] c. The sensor changes the way of data acquisition (e.g., in the case of vibration measurement: sampling rate, limit frequency, measurement duration, selection of vibration axis, way of data preprocessing, size of the raw data transmitted), and repeats the measurement when the data acquisition changes.

[0057] d. The sensor activates another measurement value receiver (e.g., switches to acoustic measurement via a microphone), and performs repeated measurements using the alternative measurement value receiver.

[0058] If the alarm state (re - exceeding the limit value) is also confirmed by at least one repeated measurement, then an alarm notification can also be output (block 45).

[0059] Summary:

[0060] Therefore, after each data transmission to the cloud, the sensor subsequently also receives a reply from the cloud. The sensor obtains this reply within a pre - determined short time interval. Thus, the sensor dynamically checks which behavior it should adopt currently. Basically, hereby not only can the measurement frequency of the sensor be dynamically changed, but also the working mode of the sensor can be changed. Thus, the behavior of the sensor is actively controlled by the cloud, and for the sensor, this does not require a significantly higher energy demand.

[0061] Next, the execution of repeated measurements and its significance in machine monitoring are explored in more detail again.

[0062] Performing repeated measurements in machine monitoring is particularly meaningful in situations where monitoring is carried out by comparing one or more sensor values with the associated limit values, and external influences may affect the sensor values due to the machine's surroundings. In many machines, such as pumps, there are important components that should be monitored by corresponding sensing devices in order to obtain an alarm as early as possible before a failure (e.g., the bearings of a pump). The sensor collects physical parameters here, such as vibration, pressure, or temperature. Therefore, meaningful limit values are determined for each of these physical parameters, and exceeding the limit values indicates a critical state (e.g., the maximum bearing temperature). Monitoring the set limit values of the parameters can be carried out selectively "at the edge" (i.e., in the sensor or in the industrial Internet of Things device itself) or alternatively in the cloud. Due to greater flexibility and the almost infinitely available computing power, the cloud is preferably used here.

[0063] In the event that the current limit value is exceeded, a message should generally be output in the form of an alarm notification ( Figure 1 , block 45), in order to prevent subsequent damage at the machine or, respectively, at the process (caused by a machine fault). However, many physical parameters are influenced not only by the operation of the machine, but additionally by its surrounding environment. Therefore, the measured values of the sensors do not indicate the actual machine state at every point in time. For example, a connected pipeline may induce vibrations into the pump from the outside; likewise, during machine startup or shutdown and due to other temporarily occurring events, short-term high, yet non-critical vibration values may occur. These non-critical events should be masked accordingly in order to avoid false alarms. On the other hand, in the case of actually high vibration values, for example due to a sudden damage at the machine itself, an alarm notification should be output as quickly as possible in order to be able to initiate targeted countermeasures in a timely manner. This is important in order to prevent expensive subsequent damage as much as possible.

[0064] By means of repeated measurements, non-critical temporary short-term events should be quickly and reliably distinguished from suddenly occurring critical events.

[0065] In traditional wired sensing devices, the measured values can be simply retrieved again in order to increase their credibility. However, in the case of battery-operated sensing devices that are commonly used in industrial Internet of Things environments, the sensors usually automatically switch back to the "sleep mode" immediately after transmitting the measurement data in order to save energy and thus achieve the longest possible battery service life. The next measurement is then (as planned) often carried out only after several hours. By means of the method of dynamic remote configuration, it is now possible to perform one or more measurement repetitions energy-efficiently even in battery-operated sensors or industrial Internet of Things devices in the event of a limit value being exceeded, while the additional energy requirements required for this are minimized.

[0066] If there are indications for an alarm notification in the evaluated data, then this alarm state is initially only stored temporarily internally (block 45). Subsequently, the cloud or, alternatively, the industrial Internet of Things device itself directly initiates one or more repeated measurements (block 60) in order to improve the credibility of the individual measured values. If the previously determined indications for an alarm notification are confirmed by the new measured values, then this alarm notification is now actually also output (block 45). However, in other cases (i.e., in the case of non-critical repeated measurements), the first measured value is regarded as interference, which is triggered by a single external event. Then, the temporarily stored alarm state is immediately cleared again, and no notification is output.

[0067] Intelligent and energy-efficient identification of the machine state of a short-cycle operation machine

[0068] This application scenario can be advantageously used in particular in machines that are typically switched on only for short periods. For example, there are pumps that only run for a few minutes to empty a container and are then switched off again for a long time. If the container is filled only very slowly, the duration of the pump being switched off is even several times longer than the duration of it being switched on.

[0069] If a battery-operated sensing device is used at such a machine, continuous monitoring can only be achieved to a limited extent. Because if a long battery life is to be achieved, a battery-operated sensor only takes a single measurement value at long time intervals (e.g., every hour). Between two measurements, the sensor is then in a sleep mode to save energy. In the sleep mode, neither measurement nor data transmission takes place.

[0070] Figure 2 The conventional time course for a machine that is switched on (multiple times) for short periods is shown. The solid line 1 symbolizes the machine state, which changes between the off and on states. The acquisition of the measurement values to be monitored is carried out periodically in a timer-triggered manner, which is indicated by the arrow 2. The upper dashed arrow 3 represents the data transmission from the sensor to a cloud-based evaluation unit.

[0071] In Figure 2 it can be seen that sensor data is very frequently acquired and transmitted in the off state of the machine. In contrast, in the on state, measurement values are only occasionally obtained. Depending on the actual operating frequency and duration of the machine, this can lead to insufficient monitoring of the machine because the relevant critical states are always only to be expected in the on state. Many short on-phases are neither visible to the sensor nor in the cloud. If the operating hours of the machine are to be additionally determined from the sensor signals, another disadvantage results from this. This function has a high degree of inaccuracy and is therefore not even usable in individual cases.

[0072] Although the on state of a short-cycle operating machine can also be reliably acquired by a very high sampling rate, from an energy perspective, this is not meaningful, and especially in the case of a battery-operated sensor, due to the shortened battery life caused by more frequent measurements and data transmissions, this is not a targeted method.

[0073] However, with the method according to the invention and the possibility of dynamic remote configuration of the sensor, the problem can be solved in another way, which is shown in Figure 3 First, the sensor operates with normal sensor behavior (A). This behavior corresponds to the one already explained with Figure 2 It has a low measurement rate (measurement interval T A)Periodic measurements are carried out, and the measurement data is always directly transmitted to the cloud for evaluation thereafter.

[0074] However, once the machine is identified as a short-cycle operating machine on the cloud side based on previously collected sensor data, and a clear identification threshold between "on" and "off" is also determined for the machine, then the behavior of the sensor changes. When the mentioned preconditions are met and the machine now switches to the "off" state, then the cloud controls the sensor into mode (B) "measurement comparison sleep" via instruction 5. At the same time, the cloud transmits the on and off limit values to the sensor. The limit values are determined by observing and evaluating the sensor data of the machine in the cloud for a long time. In the case of vibration data, the on-off limit value can be, for example, the root mean square value or the energy value of the vibration.

[0075] Now, the special feature of the sensor in mode (B) is that the sensor values are acquired in an energy-saving manner with a significantly shorter measurement interval T B In the case of a vibration sensor, this is achieved, for example, by a shorter measurement duration compared to mode (A). In addition, no data is sent to the cloud in mode (B), so no energy needs to be consumed for this either. Eigenvalues can be derived from the sensor values, and the eigenvalues can be directly compared with the existing limit values inside the sensor. In the case of vibration data, these eigenvalues can be, for example, the root mean square value or the energy value of the vibration.

[0076] With the on-off limit values transmitted to the sensor, the sensor can evaluate locally whether the machine is still off (all observed eigenvalues remain below the limit values), or since when the machine has been turned on (at least one of the eigenvalues exceeds the relevant limit value). Therefore, once the exceeding of the limit value is recognized (reference numeral 6), the sensor automatically switches back to mode (A) again, that is, the sensor now acquires all sensor data again with the full measurement duration and transmits these sensor data to the cloud. The device now continues to work in mode (A) until it receives an instruction to switch to mode (B) again from the cloud.

[0077] Next, briefly summarize again the advantages of the method according to the present invention.

[0078] When using battery-operated sensors or generally industrial Internet of Things devices to monitor machines, the quality of machine monitoring can be improved by applying the method described herein. Through targeted interaction between the cloud and the sensor / industrial Internet of Things device, for different application scenarios, the credibility of the physical values at the machine collected by the sensor can be significantly increased. The special feature of this method is that only a small additional energy requirement is needed for this purpose, so the service life of the battery of the sensor / industrial Internet of Things device is only insignificantly affected thereby.

[0079] Specifically, advantages are obtained for the following application scenarios:

[0080] a) The machine is monitored by an industrial Internet of Things device comprising one or more sensors. The sensor / industrial Internet of Things device can be connected to the cloud. If the characteristic value determined by the sensor exceeds a determined limit value, false alarms can be avoided by automatically prompting a repeated measurement. Instead of a direct alarm notification, it is output only after direct verification by means of one or multiple repeated measurements.

[0081] b) A typically short-cycle switched-on machine is monitored by an industrial Internet of Things device comprising one or more sensors. The industrial Internet of Things device should be connected to the cloud. In an energy-efficient manner, the monitoring of such a machine can be significantly improved. In addition, the desired operating duration measurement can be performed significantly more precisely based on the collected sensor data.

Claims

1. A method for monitoring at least one rotating working machine, in particular a centrifugal pump, the working machine comprising at least one sensor for collecting at least one measurement parameter related to the operation of the working machine and a central evaluation unit in communication connection with the at least one sensor, the method having the following method steps: a. Transmitting at least one measured value and / or a measured value preprocessed on the sensor side from the sensor to the central evaluation unit; b. Evaluating the at least one measured value and / or the preprocessed measured value by the evaluation unit and generating at least one instruction for dynamically remotely configuring the sensor behavior of the at least one sensor based on the evaluation result; c. Transmitting the instruction by the evaluation unit to the at least one sensor.

2. The method according to claim 1, wherein Triggering repeated measurements on the sensor side by means of the instruction, and / or remotely configuring the number of repeated measurements to be performed and / or the time interval between the repeated measurements to be performed.

3. The method according to any one of the preceding claims, characterized in that, Configuring the threshold for threshold monitoring inside the sensor by means of the instruction.

4. The method according to claim 3, characterized in that When a configurable threshold is exceeded, the sensor assumes an alarm state and / or generates and outputs an alarm notification and / or performs at least one repeated measurement and / or performs data communication to transmit the measurement data to the central evaluation unit.

5. The method according to any one of the preceding claims, characterized in that, The at least one sensor includes two or more, preferably different, measured value receivers, and one or more measured value receivers can be deactivated or activated by means of the instruction.

6. The method according to any one of the preceding claims, characterized in that, Putting the sensor into a sleep or low-power mode by means of the instruction, and / or configuring the duration of the sleep or low-power phase by means of the instruction.

7. The method according to any one of the preceding claims, characterized in that, Configuring the process of measured value acquisition by means of the instruction, such as the sampling rate and / or the measurement duration and / or the preprocessing of the measurement data and / or the data size of the measurement data or raw data to be transmitted.

8. The method according to any one of the preceding claims, characterized in that, The sensor collects mechanical vibrations and / or temperature values within the scope of the working machine.

9. The method according to any one of the preceding claims, characterized in that, In the case of collecting vibration data by the sensor, the selection of the vibration axis and / or the configuration of the limit frequency are configured by means of the instruction.

10. The method according to any one of the preceding claims, characterized in that, The at least one sensor is a battery-operated sensor.

11. The method according to at least one of the preceding claims, characterized in that, After sending one or more measured values to the central evaluation unit, the sensor transitions to a first sleep mode and receives at least one instruction from the evaluation unit and / or performs an internal sensor evaluation of the measurement data after the end of the first sleep mode, and performs at least one repeated measurement to verify the previous measurement data according to the internal sensor evaluation and / or the evaluation performed by the evaluation unit. In the case where repeated measurements should not be performed, the sensor transitions to a second sleep mode.

12. The method according to any one of the preceding claims, characterized in that, The sensor collects measured values at a first measurement interval in a first operating mode and transmits the measured values to the central evaluation unit after each measurement, wherein the evaluation unit prompts the sensor to operate in a second operating mode after evaluating the measurement sequence of the received measured values.

13. The method according to claim 12, wherein At least one recognition threshold is determined by evaluating a measurement sequence and transmitted to the sensor, wherein different machine states can be distinguished by means of the recognition threshold.

14. The method according to claim 12 or 13, characterized in that, The sensor acquires measurement values at a second measurement interval in a second operating mode, wherein the second measurement interval is shorter than the first measurement interval, and during the second operating mode, no data is transmitted to the central evaluation unit, or at least less data is transmitted to the central evaluation unit than in the first operating mode.

15. The method according to any one of claims 13 or 14, characterized in that If the current measurement value or a plurality of successive measurement values exceed the recognition threshold, the sensor switches to the first operating mode.

16. A system, comprising at least one central evaluation unit, in particular a cloud-based evaluation unit, and at least one sensor, wherein, The evaluation unit and the sensor are configured to perform the method according to any one of the preceding claims 1 to 15.

17. A sensor for a system according to claim 16, the sensor being configured to perform dynamic remote configuration according to the method according to any one of claims 1 to 15.

18. A central evaluation unit for a system according to claim 16, the central evaluation unit being configured to perform dynamic remote configuration of the sensor according to claim 17 according to the method according to any one of claims 1 to 15.

Images