Method for a Sensor-Based Monitoring of at Least One Rotating Work Machine

Description

CROSS REFERENCE TO RELATED APPLICATION

This application is a 371 National Stage Application of PCT/2023/081285, filed Nov. 9, 2023, which claims priority from German Patent Application No. 102022130126.5, filed Nov. 15, 2022, the entire disclosure of which is herein expressly incorporated by reference.

BACKGROUND

The disclosure relates to a method for monitoring at least one rotating work machine, in particular a centrifugal pump, comprising at least one sensor for detecting at least one measurement parameter which relates to the operation of the work machine, and a central analysis unit which is communicatively connected to the at least one sensor.

The automated status monitoring of work machines, in particular rotating work machines, is already known. To this end, one or more sensors for the purpose of status monitoring are fitted in the vicinity of or on the machines to be monitored. The measurement values of the sensors are detected cyclically and then transferred directly, or after pre-processing in the sensor, to a cloud or to another device.

Machine monitoring is usually coupled with an automatic alarm system, which is intended to warn the operator of the machine or installation as quickly and reliably as possible when critical statuses occur. For reliable detection and identification of critical statuses in machines with dynamic operating behavior, it is often necessary to carry out the measurements with a small measurement interval, since this is the only way to ensure that a critical status can also be detected and identified by a performed measurement. A similar problem occurs when monitoring machines which run in short cycles. Measurements must ideally be carried out at those times during which the machines are actually active. If, instead, measurement values are detected at times when the machine is switched off, efficient monitoring is not possible in a meaningful way.

Continuous monitoring measurement performed with a short time interval, however, is unreasonable from an energy standpoint and, specifically in the case of solutions with battery-operated sensors, cannot be performed in a meaningful way because the short measurement intervals and the resulting high-frequency data transfer to a central analysis unit give rise to an enormous energy requirement, which greatly shortens the runtime of the battery.

A further problem involves influences from the environment of the machines on the measurement values. If the machines to be monitored are part of larger installations, vibrations, noise from adjacent machines etc. can lead to distortion of the measurement values and trigger an erroneous detection. It is desirable to filter out such influences as effectively as possible in order to avoid false alarms.

SUMMARY

An object of the present disclosure is therefore to optimize existing sensor-based monitoring methods in which there is a data exchange between sensor and central analysis unit such that they can be performed more efficiently, more quickly and also in a more energy-saving manner.

This and other objects are achieved by a method according to the features of the disclosure. Advantageous embodiments of the method are also the subject matter of the disclosure.

According to the disclosure, it is proposed that a dynamic remote configuration of the sensor behavior takes place by means of the central analysis unit. Remote configuration of the sensor behavior is understood to be the adjustment of the sensor operation or any operating parameters of the sensor which relate to or influence measurement value detection and measurement value analysis. In this context, sensor is understood to be not only the measurement value recorder alone, but also a sensor controller or some other unit for pre-processing and/or analyzing the measurement values. In addition, the sensor can be equipped with an integral communication module for communicating with the central analysis unit or be communicatively connected to an external communication device for indirect communication with the analysis unit. Data communication between analysis unit and sensor can take place via at least one intermediate node, in particular a gateway installed in the immediate receiving area of the sensor. The gateway preferably can have an intermediate buffer for temporary caching of communication data which are intended to be exchanged between sensor and analysis unit. The at least one sensor can retrieve the cached data in the gateway when required.

The command can be transferred from the analysis unit by forwarding the command from the analysis unit to the sensor or alternatively by retrieving the command by means of the sensor in the analysis unit.

The dynamic remote configuration of the sensor represents the necessary prerequisite for optimizing the measurement operation or analysis process of the sensor during the ongoing machine monitoring taking previous measurement value analyses into account, with the result that the entire measurement process and monitoring can be adapted dynamically to the actual circumstances. In summary, quicker, more reliable and more energy-saving monitoring can be realized.

For performance of the method, initially at least one detected measurement value and/or a measurement value which is pre-processed by the sensor must be transferred to the central analysis unit by means of the sensor. On the part of the analysis unit, the received measurement value or the pre-processed measurement value is then analyzed and, depending on an analysis result, a command for dynamic remote configuration of the sensor behavior of the at least one sensor is generated and transferred to the sensor for remote configuration.

For example, there is the possibility for at least one repeat measurement on the part of the sensor to be initiated by means of the command. A repeat measurement is understood to be a measurement to be carried out shortly after receipt and analysis of a previous measurement value, in order to be able to verify the previous measurement. For this purpose, the repeat measurement must take place as soon as possible after the first measurement, depending on the application, in particular a few seconds, for example less than 60 seconds after the measurement to be verified, optimally after 30 seconds or less. It is also conceivable for the command to be used not only to trigger a repeat measurement but instead also to be able to configure the number of repeat measurements to be performed or to be able to remotely set the time interval between at least two successive repeat measurements. After receipt of such a command, the received configuration is implemented by the sensor for the process sequence of a repeat measurement.

It is also conceivable that a specific threshold value for sensor-internal monitoring of the threshold value is configured by means of the command. After receipt of the corresponding command with at least one threshold value, the sensor performs corresponding process steps for the threshold value monitoring, by comparing the presently detected measurement value in each case with the remotely configured threshold value or some other operating parameter of the sensor with the remotely configured threshold value, for example. After exceeding, at least one follow-up measure can be initiated.

The threshold value monitoring can be used in a preferred manner for error identification. For example, when the configurable threshold value is exceeded, the sensor can enter an alarm status and/or generate and output an alarm notification. It is also conceivable that, when the threshold value is exceeded, this initiates at least one repeat measurement in order to verify the previous measurement by repeat measurement.

It is also conceivable that exceeding the threshold value prompts the sensor to enter into data communication with the central analysis unit or with some other communication partner. As a result, it can be ensured that only relevant measurement values are transferred to the analysis unit, which reduces the overall energy requirement due to the reduced communication.

It is also conceivable that the sensor comprises two or more, in particular different, measurement value recorders. The measurement value recorders can detect the same measurement value for diversity reasons. However, it is also conceivable that the measurement value recorders detect different measurement parameters. As a result of the command of the central analysis unit, measurement value recorders can be varied or the redundancy or diversity during measurement operation can be altered. In principle, it is possible to selectively deactivate or activate the one or more measurement value recorders.

Furthermore, it is conceivable that the sensor can be placed in a sleep or low-power mode by means of the command. In such a sleep or low-power mode, the measurement value recording is completely deactivated or reduced to a minimum degree, for example, in order to temporarily reduce the energy consumption in the sensor to a minimum. Of course, the sensor could also work with a reduced measurement frequency and/or sampling rate etc. during the low-power mode.

It is also conceivable that the data communication between sensor and central analysis unit is completely interrupted or at least reduced to a minimum during the sleep mode or the low-power mode. The command can be used not only to actively initiate such a sleep or low-power mode, but rather the command can also be used to specify the duration of such a sleep or low-power phase.

According to an advantageous embodiment, the generated command of the central analysis unit can be used to configure the type of measurement value detection in the sensor. This relates, for example, to a sampling rate and/or the measurement duration of a single measurement by the measurement value recorder of the sensor. The remote configuration of the sensor-internal pre-processing of the measurement values may also be possible, i.e. it can be configured, for example, to combine the measurement values of a measurement series and to communicate only the mean value of the measurement series to the analysis unit. The command can preferably be used to configure the data size of the measurement data or raw data to be transmitted.

In principle, the method according to the disclosure can be used for any type of measurement values. At this point, use in the measurement and monitoring of mechanical oscillations in rotating machines may be mentioned by way of example. The use of the method in temperature measurement or temperature monitoring is also particularly preferred. Use in oscillation monitoring supplemented by temperature measurement or temperature monitoring is particularly preferred. If the remotely configurable sensor is suitable for detecting mechanical oscillations, then setting or deactivating/activating the oscillation axes to be measured, preferably by remote configuration, may be possible. Remote configuration of the cut-off frequency is furthermore useful.

It is also advantageous if the at least one sensor is a battery-operated sensor. Data is communicated between sensor and central analysis unit preferably by means of a radio link, wherein any desired data transfer standard can be employed here. Remote configuration of the sensor operation is advantageous particularly in the case of battery-operated sensors, so as to optimize the sensor operation with regard to energy consumption and to guarantee the longest possible battery runtime with sufficient measurement accuracy and measurement reliability.

According to a preferred embodiment of the method, it may be envisioned that the sensor initially transitions into a first sleep mode after the forwarding of the one or more measurement values to the central analysis unit. During such a first sleep mode, the energy consumption of the sensor reduces to a minimum, by the data communication with the central analysis unit and/or measurement value recording during the sleep mode being stopped, for example. After termination of the first sleep mode, the sensor can receive at least one command from the analysis unit and/or undertake a sensor-internal analysis of the previously detected measurement data. The duration of the first sleep mode is selected such that the sensor is wakened in time to receive the command from the central analysis unit and activation of the data communication to receive the command is ensured. The necessary process duration for the measurement value analysis on the part of the analysis unit is known or can be readily estimated, and therefore the duration of the first sleep mode can be adapted to this process duration.

If data is communicated between analysis unit and sensor indirectly via an intermediate unit, in particular a gateway, there is the possibility to cache any data packets, in particular commands, in the gateway and to have them ready for retrieval by the sensor. In this case, the duration of the sleep mode can be made to be more flexible, since the sensor can retrieve the data directed to the sensor from the gateway as required, i.e. individual selection of the sleep duration.

Depending on the analysis result, a decision can then be made as to whether a repeat measurement is to be performed. This decision is preferably made in the analysis unit, but sensor-internal decision-making is also conceivable if a sensor-internal analysis of the measurement values is envisioned. If the analysis is undertaken on the part of the central analysis unit, the decision in the analysis unit is also made as to whether a repeat measurement should be performed, which is then prompted by means of remote configuration of the sensor. In contrast, if the measurement values are analyzed sensor-internally, a corresponding decision to perform repeat measurements can also be made sensor-internally. If the previous measurement and the at least one repeat measurement supply matching or at least comparable results, the measurement is considered verified and is taken into account for the decision on any follow-up measures, for example for error detection and a possible alarm.

It is also conceivable that the central analysis unit is also used to remotely configure the number of repeat measurements to be performed. Thus, a plurality of repeat measurements may be advantageous for increasing the available quantity of data and for improving the quality of the data analysis and status identification of the work machine.

If, after completion of the method, it is established that a repeat measurement is not required, the sensor is preferably placed in a second sleep mode in which neither a measurement detection nor a data transfer to the central analysis unit takes place. The duration of the second sleep mode can be defined to be significantly longer than the duration of the first sleep mode.

According to a further embodiment of the disclosure, it may be envisioned that the sensor detects measurement values with a first measurement interval in a first way of working. In the first way of working, the detected measurement values are transferred to the central analysis unit via the communication channel after each measurement. The analysis unit receives the measurement values and analyzes the received measurement values associated with a measurement series. Depending on the analysis result, it may be envisioned that the central analysis unit prompts the sensor to operate in a second way of working.

It is conceivable, for example, that the analysis unit determines at least one identification threshold value by analyzing the measurement series and transfers it to the sensor. With the aid of such an identification threshold value, a differentiation between different machine statuses may be possible, for example, in particular whether the machine is presently in operation or in a rest phase.

It is preferable, for example, if the sensor detects measurements with a second measurement interval in the second way of working, wherein the second measurement interval is selected to be shorter than the first measurement interval, in particular by at least a factor of 5, preferably 10. It is furthermore envisioned that, during the second way of working, no measurement data are transferred to the analysis unit or at least fewer data are transmitted to the central analysis unit than in the first way of working. The second way of working is advantageous in particular for monitoring machines which run in short cycles, because the shortened measurement interval ensures that measurement values can also be detected during active operation of the machine. In return, the increased energy requirement as a result of the shortened measurement interval is compensated by the interrupted or reduced data communication with the central analysis unit.

During the second way of working, the detected measurement values are preferably compared sensor-internally with an identification threshold value, in particular with the identification threshold value determined previously by the analysis unit. When the identification threshold value is exceeded, the sensor switches to the first way of working so as to transfer the detected measurement values continuously to the central analysis unit. The sensor preferably returns to the second way of working if the measurement values drop below the identification threshold again.

It is particularly advantageous if the identification threshold value allows a distinction to be made between an active and inactive machine status. By defining such an identification threshold value, it is ensured that the sensor transfers measurement data to the central analysis unit only when it has previously been established on the basis of the measurement data that the machine is indeed in operation.

This approach is particularly preferred in sensors for measuring mechanical oscillations, because the amplitude of the detected oscillation already gives an indication as to whether the machine is in operation or instead is in a rest status.

In addition to the method according to the disclosure, the present disclosure also relates to a system comprising at least one central analysis unit, in particular a cloud-based analysis unit, and at least one sensor, wherein analysis unit and sensor are configured to carry out the method of the above-described disclosure. The system can additionally comprise an interconnected gateway.

In addition to the overall system, the disclosure also relates to a sensor for such a system, wherein the sensor is configured to be remotely configured dynamically by a central analysis unit. Finally, the disclosure also relates to the central analysis unit for a system according to the disclosure, wherein the central analysis unit is configured to generate and to transfer a command for dynamic remote configuration of a communicatively connected sensor on the basis of the analysis result from one or more measurement values.

The same advantages and properties as have already been indicated above with reference to the method according to the disclosure therefore apply to the system, the sensor and the central analysis unit, and therefore a repetitive description is dispensed with at this point.

Further advantages and properties of the method are indicated in more detail below with reference to different exemplary embodiments which put the disclosure into effect. In the drawings:

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1: shows a flowchart of a sensor device with repeat measurements,

FIG. 2: shows a conventional time chart for short-running machines with sensor, and

FIG. 3: shows the corresponding flowchart according to the disclosure for the intelligent identification of the pump status by the sensor device.

DETAILED DESCRIPTION

The method according to the disclosure can be used in all machines which are to be monitored by sensors, in particular battery-operated sensors. One or more sensors for the purpose of status monitoring are fitted in the vicinity of the machine to be monitored, for example a pump, in particular a centrifugal pump. The measurement values of the sensors are automatically detected cyclically and then sent directly, or after pre-processing in the sensor, to a central analysis unit, in particular to a cloud-based solution of the analysis unit or to another device.

For the machine monitoring, at least one sensor for detecting mechanical oscillations can be used, by way of example. The sensor, in particular a battery-operated sensor, comprises an integral communication module which enables the exchange of data between the sensor and the central analysis unit. In particular in the case of larger installations with a multiplicity of machines to be monitored, pumps in particular, it is recommended to use a gateway which is installed in the vicinity of the machines, and is connected to the multiplicity of sensors. The signals received by the sensors by radio standard are then communicated from the gateway to the cloud-based analysis unit via an Internet connection.

The machine monitoring is usually coupled with an automatic alarm system, which is intended to warn the operator of the machine or installation as quickly and reliably as possible when critical statuses arise. The aim of the disclosure is to significantly increase the reliability of such alarm notifications. A further important aim consists in improving the accuracy of an operating-hours counter that is connected thereto. As a result of the dynamization, according to the disclosure, of the temporal measurement intervals and also the measurement behavior of the sensors used, more accurate energy-efficiency-related statements regarding the status of the monitored machines can be made. There are various specific applications for this, which are described in more detail in the following text.

A significant special feature of this disclosure consists in the possibility of dynamically altering the behavior of the sensor from the central analysis unit, i.e. the cloud, without a relevant additional energy requirement being necessary. One possibility for controlling the sensor behavior can function as depicted in FIG. 1:

In machine operation, a sensor initially detects measurement data such as mechanical oscillations and/or temperature values in the environment of the machine to be monitored and puts the data through pre-processing and initial analysis (block 10). Then the sensor sends the measurement data or the pre-processed data to the cloud (block 20).

The sensor subsequently transitions for a short time (e.g. 30 seconds) into a “short-term sleep mode” (block 30) in order to save energy. On the basis of the present sensor data and sensor data from the past of the same sensor, the future desired behavior of the sensor is automatically specified in the cloud (reference sign 40). This calculation or analysis takes place on servers in the cloud in a very short computing time and can comprise, inter alia, the comparison of the received sensor value with a stored threshold value, for example. An alarm notification can also be generated here already in the cloud, depending on the analysis result (block 45), wherein this is preferably initially saved only temporarily and instead a repeat measurement is requested by the sensor to verify the alarm status.

The cloud now sends the desired sensor behavior as a response back to the sensor and the sensor receives the command or commands to internally adjust the sensor behavior (block 50). Due to the short processing time in the cloud, it is ensured that the result is present before the end of the “short-term sleep mode” of the sensor. If there is still a gateway (which transfers the data in both directions) on the communication path between sensor and cloud, this gateway can receive the response intended for the sensor from the cloud and cache it. As soon as the sensor comes back out of the “short-term sleep mode”, the sensor fetches the instructions for its further behavior either

• • directly from the cloud or
  • from the gateway which had cached the information intended for the sensor from the cloud.

In case b), this procedure functions even more quickly and thus in a more energy-efficient manner than in case a). In accordance with the response received from the cloud, the sensor adjusts its future desired way of working (block 50).

The remote configuration can then be used, for example, to initiate a repeat measurement in the sensor and to configure the performance thereof (block 60).

If it is decided in block 60 that the sensor does not carry out a repeat measurement, because the sensor value is below the limit value, for example: The sensor returns-with no further action-to a normal (longer lasting) sleep mode (block 70), for example more than 60 minutes, in particular 120 minutes or longer, since the present data do not require any particular interaction.

In the event that, in block 60, at least one repeat measurement is to be carried out, either by command from the cloud or sensor-internal analysis, this can be carried out according to specifications from the cloud, for example, as follows:

• • a. The sensor immediately carries out a repeat measurement (and also transmits the data directly to the cloud again);
  • b. The sensor carries out a plurality of repeat measurements with shorter cycle times;
  • c. The sensor alters the type of data detection (e.g. in the case of oscillation measurements: sampling rate, cut-off frequency, measurement duration, selection of oscillation axes, type of data pre-processing, size of the transmitted raw data) and repeats the measurement with altered data detection;
  • d. The sensor activates another measurement value recorder (e.g. switchover to acoustic measurement via microphone) and performs the repeat measurement with an alternative measurement value recorder.

If the alarm status is also confirmed by the at least one repeat measurement (renewed exceeding of the limit value), the alarm notification can then also be output (block 45).

SUMMARY

The sensor will therefore subsequently also receive a response from the cloud after each transfer of data to the cloud. It fetches this response in a previously specified short time slot. Thus, it checks dynamically which behavior it should presently adopt. As a result, essentially both the measurement frequency and the way of working of the sensor can be altered dynamically. The behavior of the sensor is therefore actively controlled from the cloud, with no appreciably higher energy requirement being necessary for the sensor for this purpose.

Hereinafter, the carrying-out of repeat measurement and the significance thereof in machine monitoring will again be discussed in detail.

Carrying out repeat measurements in machine monitoring is useful in particular where monitoring takes place by comparing one or more sensor values with associated limit values, and external influences can influence the sensor values due to the environment of the machines. In many machines such as pumps, there are important elements which should be monitored by appropriate sensor systems, so as to obtain an early warning before a breakdown (e.g.: the bearings of a pump). Sensors detect the physical parameters here, such as vibration, pressure or temperature, for example. Consequently, for each of these physical parameters useful limit values are specified, and when these are exceeded, a critical status is indicated (e.g. maximum bearing temperature). Monitoring of the set limit value of a parameter can optionally take place “on the edge” (that is to say, in the sensor or IIoT device itself) or alternatively in the cloud. Due to the greater flexibility and virtually unlimited computing power available, the cloud is preferred in this case.

If a present limit value is exceeded, a notice in the form of an alarm notification should normally be output (FIG. 1, block 45), in order to prevent consequential damage to the machine or to the process (by breakdown of the machine). However, many physical parameters are influenced not only by the operation of the machine, but additionally also by their surroundings. Therefore, a measurement value of a sensor does not indicate the actual status of the machine at all times. For example, connected pipelines can induce vibrations from outside into a pump; likewise, high yet non-critical vibration values can arise briefly when the machine is started or stopped and as a result of further temporarily occurring events. These non-critical events should be masked appropriately, so as to avoid false alarms. On the other hand, in the case of genuinely high oscillation values, e.g. as a result of sudden damage to the machine itself, an alarm notification should be output as quickly as possible so as to be able to introduce targeted counter-measures in a timely manner. This is important for preventing expensive consequential damage as far as possible.

With the aid of the repeat measurement, a non-critical temporary short-term event should be differentiated quickly and safely from a suddenly occurring critical event.

In the case of a conventionally wired sensor system, the measurement value could simply be called up again, so as to increase the evidence thereof. In the case of a battery-operated sensor system, however, which is frequently used in the IIoT field, the sensor, after transmission of the measurement data, usually immediately switches automatically to a “sleep mode” again in order to save energy and thus to achieve the longest possible battery service life. The next measurement then often takes place (as planned) not until hours later. By means of the method of dynamic remote configuration, one or more measurement repeats can now be carried out in an energy-efficient manner if a limit value is exceeded, even in battery-operated sensors or IIoT devices, while the additional energy requirement necessary for this is nevertheless reduced to a minimum.

If there is an indication of an alarm notification in the case of the analyzed data, this alarm status is initially cached only internally (block 45). The cloud or alternatively the IIoT device itself then immediately initiates one or more repeat measurements (block 60), in order to improve the evidence of the individual measurement value. If the previously established indication of an alarm notification is confirmed by the new measurement values, then this alarm notification is now also actually output (block 45). But otherwise-that is to say in the case of non-critical repeat measurements-the first measurement value is considered to be a glitch which has been triggered by a unique, external event. The cached alarm status is then immediately deleted again and no notification is output.

Intelligent energy-efficient identification of the machine status of machines which run in short cycles This application is usable advantageously in particular in machines which are mostly switched on only for short cycles. Thus there are pumps, for example, which run for only a few minutes in order to empty a container but afterwards are switched off again for a long time. If the container is filled only very slowly, then the duration for which the pump is in the switched-off status will exceed, even many times over, the duration for which it is in the switched-on status.

If a battery-operated sensor system is used in such a machine, then permanent monitoring is possible only to a limited extent. This is because if a long battery service life is to be achieved, then only individual measurement values are detected in longer time intervals (e.g. hourly) using the battery-operated sensor. The sensor is then in sleep mode between two measurements, so as to save energy. Neither measurement nor data transfer takes place in sleep mode.

FIG. 2 shows a conventional time sequence for a machine that is switched on for a short time (on multiple occasions). The solid line I symbolizes the machine status, which changes between a switched-off and a switched-on status. The measurement values to be monitored are detected in a timer-triggered manner with cyclical repetition, which is characterized by the arrows 2. The upper, dashed arrow 3 stands for the data transfer of the sensor data to the cloud-based analysis unit.

In FIG. 2 it can be seen that the sensor data are detected and transferred very frequently when the machine is in the switched-off status. In contrast, measurement values are acquired only occasionally in the switched-on status. Depending on how often and for how long the machine actually runs, under certain circumstances this leads to insufficient monitoring of the machine, since relevant critical statuses are always to be expected only in the switched-on status. Many short switch-on phases are not visible here either for the sensor or in the cloud. Additionally, if the operating hours of the machine are also to be ascertained from the sensor signal, this produces a further disadvantage. This function is affected by severe inaccuracies and therefore cannot be used at all in some cases.

Although a very high sampling rate would make it possible also to reliably detect the switched-on status of machines which run in short cycles, this would not be appropriate from an energy standpoint and not a target-oriented approach in particular in the case of battery-operated sensors due to the shortened service life of the battery as a result of the more frequent measurements and data transfers.

By means of the method according to the disclosure and the possibility of dynamically remotely configuring the sensor, this problem can however be solved in another way, which is depicted in FIG. 3. Initially the sensor works with normal sensor behavior (A). This behavior corresponds to that which has already been explained above with reference to FIG. 2. Measurement takes place cyclically with a low measurement rate (measurement interval TA) and thereafter the measurement data are always transferred directly into the cloud for analysis.

However, as soon as this machine is identified in the cloud as being a machine running in short cycles, on the basis of the previously collected sensor data, and clear identification thresholds between “switched on” and “switched off” have also been determined for this machine, the behavior of the sensor is altered. If the mentioned preconditions are met and the machine now changes into the “off” status, the cloud then controls the sensor through a command 5 into mode (B) “MeasureCompareSleep”. At the same time, the on and off limit values are transferred from the cloud to the sensor. These limit values were ascertained by longer observation and analysis of the sensor data of this machine in the cloud. In the case of oscillation data, these on-off limit values can be rms values or energy values of the oscillations, for example.

The special feature of mode (B) of the sensor is now that sensor values are detected in an energy-saving manner with a significantly shorter measurement interval TB. In the case of oscillation sensors, this is achieved e.g. by a shortened measurement duration compared to mode (A). Moreover, no data are sent to the cloud in mode (B), and therefore no energy needs to be expended for this. Characteristic values can be derived from the sensor values, which characteristic values are directly comparable sensor-internally with the present limit values. In the case of oscillation data, these characteristic values can be rms values or energy values of the oscillations, for example.

On the basis of the on-off limit values transmitted to the sensor, the sensor can locally analyze whether the machine is still switched off (all observed characteristic values remain below the limit values) or the point from which the machine is switched on (at least one of the characteristic values exceeds the associated limit value). Therefore, as soon as the exceeding of a limit value has been identified (reference sign 6), the sensor changes automatically back to mode (A) again, i.e. it now detects all the sensor data again with a full measurement duration and transmits them to the cloud. The device will now continue to work in mode (A) again until it receives the command from the cloud again to change to mode (B).

Hereinafter, the advantages of the method according to the disclosure will be summarized again briefly.

When using battery-operated sensors or IIoT devices in general to monitor machines, an improvement in the quality of the machine monitoring can be achieved by using the method described here. Through targeted interaction between cloud and sensor/IIoT device, the evidence of the physical values in the machine which are detected by sensors for different applications can be significantly increased. The special feature of the method consists in the fact that only a minor additional requirement of energy is necessary for this, consequently the service life of the battery of the sensor/IIoT device is therefore only slightly affected.

There are specific advantages for the following applications:

• • a) A machine is monitored by an IIoT device which contains one or more sensors. The sensor/IIoT device can be connected to a cloud. If a characteristic value ascertained by a sensor exceeds a specified limit value, false alarms can then be avoided by means of the automatic prompting of repeat measurements. Alarm notifications are output only after an immediate verification by means of one or more repeat measurements instead of directly.
  • b) A machine which is typically switched on in short cycles is monitored by an IIoT device which contains one or more sensors. The IIoT device should be connected to a cloud. The monitoring of such a machine can be significantly improved in an energy-efficient manner. Moreover, this also allows a desired operating duration measurement to be carried out significantly more accurately on the basis of the detected sensor data.

The foregoing disclosure has been set forth merely to illustrate the disclosure and is not intended to be limiting. Since modifications of the disclosed embodiments incorporating the spirit and substance of the disclosure may occur to persons skilled in the art, the disclosure should be construed to include everything within the scope of the appended claims and equivalents thereof.