Method for checking the properties of a conveyed medium of a centrifugal pump, in particular for detecting a dry run
Patent Information
- Application Number
- EP2023832693
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-01-09
- Filing Date
- 2023-12-12
- Publication Date
- 2025-11-19
AI Technical Summary
Centrifugal pumps face challenges in reliably detecting dry running conditions, especially at low speeds, due to small torque differences between filled and unfilled systems, which can lead to increased wear, thermal overload, and system damage.
Briefly varying the target speed of the centrifugal pump, either by increasing it temporarily or superimposing a periodic signal, to determine changes in torque and motor parameters, allowing for reliable detection of dry running and changes in the pumped medium's properties.
Enables early and reliable detection of dry running and changes in medium properties across all operating points, preventing damage by accurately assessing torque and motor parameters, even at low speeds.
Smart Images
Figure 1.1
Abstract
Description
[0001] KSB SE & Co. KGaA 67227 Frankenthal
[0002] Description
[0003] Method for checking the properties of a pumped medium of a centrifugal pump, in particular for dry running detection
[0004] The invention relates to a method for checking the properties of a pumped medium of a centrifugal pump, in particular for dry running detection.
[0005] Centrifugal pumps are used to pump various fluids. Information regarding the fluid can be important for the proper and, in particular, energy-efficient operation of a centrifugal pump. In particular, changes in the characteristic properties of the fluid, such as changes in temperature or density, can be important for efficient pump control. Therefore, there is a legitimate interest in a suitable method for determining these characteristic parameters of the fluid during pump operation.
[0006] Another important aspect of wet-running centrifugal pumps is dry-running detection. This is because the medium being pumped is regularly used to lubricate the rotating machine parts of the pump unit or to cool the pump. Dry running during pump operation can therefore lead to increased wear or even thermal overload of the pump. Dry running of the pump is also an indication that the pipes in the hydraulic system have run dry, which can cause damage to the system. For this reason, dry-running detection is also important for non-wet-running pumps. To avoid any consequential damage, whether to the pump itself or the hydraulic system, it is therefore desirable to detect such a condition early and reliably, particularly regardless of the operating point.
[0007] If the lines in a hydraulic system are not sufficiently filled with fluid, i.e., the pumped medium, the torque acting on the impeller decreases compared to a filled pipeline. This circumstance can be exploited to reliably detect dry running. For this purpose, the torque is recorded in the laboratory with the valve closed but the line filled for different speeds, and the measured speed or the pair of torque values is stored in the pump as a table. Operation with the valve closed and the resulting zero flow rate corresponds to the "worst-case" scenario with the lowest possible torque in filled pipelines. The stored torque values are the lowest that can occur with a filled pipeline.As soon as the torque during operation is lower than the reference torque stored in the table, dry running is assumed and the pump can issue a corresponding message.
[0008] This method works reliably for medium and high speeds, but at low speeds the torque differences between the filled system with the valve closed and the unfilled system are very small, making a reliable distinction difficult at such an operating point.
[0009] Against this background, it is also sensible to optimize the procedure shown so that dry-running detection is reliably possible for every operating point.
[0010] The problems identified are solved by a method according to the features of claim 1. Advantageous embodiments of the method are the subject of the dependent claims. According to the invention, it is proposed to briefly vary the target speed for the speed-controlled centrifugal pump for the process of checking or detecting dry running. Particularly preferably, the target speed is increased for a short time. When the speed changes, a motor variable characterizing the torque of the pump is then determined and evaluated taking into account at least one reference value. Based on the result of the evaluation, a statement can then be made about a change in the characteristic properties of the pumped medium and / or the presence of dry running can be detected.
[0011] A characteristic property that can be determined using the method according to the invention is, for example, the density and / or temperature of the pumped medium or a change in the density and / or temperature of the pumped medium compared to an initial value. Since the density of the pumped medium and its temperature influence the viscosity of the liquid and thus the required torque of the pump, a detected torque change at a specific speed can be used to make a statement about a deviation or change in the consistency of the pumped medium. This also applies if there is a pressure drop on the suction side or if the pump is running dry, so that the corresponding torque required by the pump motor is reduced.
[0012] For example, the existing torque of the electric motor can be determined as a motor parameter to be recorded. Direct recording of the torque using a torque sensor or, alternatively, deriving the torque from other motor parameters is conceivable here. The torque can also be estimated, in particular using a motor model. It is also conceivable that the electrical power consumption of the motor and / or the drawn motor current are determined or estimated. Since the electrical power consumption or the motor current, in particular the current component forming the motor torque, have a direct influence on the resulting torque or are proportional to it, the method according to the invention can also be carried out by evaluating these operating parameters.As already shown at the beginning, dry-running detection works reliably at medium and higher speeds; only at lower speeds is the torque variation often too small, making a reliable statement impossible. Against this background, it is expedient to only implement the proposed method if the centrifugal pump is currently operating at too low a speed. Only in this case is it necessary to artificially vary or increase the target speed in order to reliably detect dry-running or check the properties of the pumped medium. The execution of the method can therefore be subject to the condition that the currently set target speed is lower than a definable threshold.
[0013] In the simplest case, the speed variation proposed according to the invention can be exhausted in a short-term increase in the target speed, whereby the motor parameter for the currently increased speed is determined and compared with a reference value assigned to the increased target speed. To determine the assigned reference value in advance, the operating parameter, e.g. the torque with a closed valve but a filled pipe, is recorded for different speeds in advance in the laboratory or under real operating conditions, and the measured speed or the pair of torque values is stored in the pump as a table. The stored reference values are the lowest that can occur when the pipe is filled. As soon as the operating parameter, e.g. the torque during operation, is lower than the reference value stored in the table, dry running is assumed and the pump can issue a corresponding message.
[0014] It is particularly preferred if the speed increase sets the target speed to the maximum speed or to a value between 50 - 90% of the maximum speed, preferably to a value between 65 - 85% of the maximum speed.
[0015] Depending on the pump application, however, the speed increase described above may not be permitted. As an alternative, it is therefore proposed to superimpose a time-variable signal on the current target speed, resulting in a defined temporal change or fluctuation in the target speed over a certain period of time, but preferably only with a limited increase in the speed. The dynamic change in the target speed inevitably causes a dynamic change in the motor parameter to be evaluated. For further analysis, a correlation between the time course of the target speed and the time course of the motor operating parameter is examined in particular. Based on the correlation, conclusions can be drawn about a physical property of the pumped medium, in particular a change in this property, and / or the presence of dry running can be detected.Since the system pressure at the suction port and the properties of the pumped medium influence the mass inertia of the pump unit, the existing mass inertia can be determined by evaluating the time curves between the target speed and the reactively adjusted torque or the recorded motor parameter, and based on this, a property of the pumped medium or any dry running can be detected.
[0016] For example, it is preferable to overlay the target speed with a periodic signal, resulting in a periodic variation of the target speed during the observation period. Overlaying it with an alternative time-variant signal is also conceivable.
[0017] Due to mass inertia, there is usually a time delay between the torque curve or the motor parameter compared to the time curve of the target speed. Against this background, the evaluation of the phase shift between the time curve of the target speed and the time curve of the recorded motor parameter is useful. If the viscosity of the pumped medium increases, the phase shift increases; however, if the viscosity decreases, the phase shift inevitably also decreases. In a dry running scenario, the phase shift is minimal due to the lack of pumped medium in the area of the impeller. Alternatively or in addition, the amplitude value of the motor parameter, in particular the torque, can be determined and evaluated while the target speed fluctuates due to the superposition. With higher mass inertia, the torque amplitude or the amplitude of the motor parameter increases in magnitude.It makes sense to compare the determined amplitude value with a reference value or to determine the amplitude ratio to a reference amplitude.
[0018] In order to record changes in the physical properties of the pumped medium or to detect dry running, it can therefore be provided to maintain corresponding reference values for the phase offset or the amplitude ratio or the amplitude itself, whereby a change in the currently determined phase offset and / or the amplitude ratio compared to the reference value can be used to conclude that there has been a change in the pumped medium and, if necessary, that there is dry running.
[0019] The one or more reference values for the process execution can be obtained, for example, during a pump training run, particularly during commissioning or at regular intervals during pump operation. It is also conceivable to determine the reference values through previous measurements and continuously update them.
[0020] In addition to the method according to the invention, the present invention also relates to a centrifugal pump, in particular a heating circulation pump, with an integral or external control / regulation system configured to implement the method according to the invention. Consequently, the centrifugal pump according to the invention offers the same advantages and properties as have already been explained in detail with reference to the method according to the invention. For this reason, a repetitive description is omitted.
[0021] Further advantages and features of the invention will be described below with reference to the exemplary embodiments illustrated by the figures. Figure 1 shows a diagram of the torque of a pump drive unit versus the speed for different pump states.
[0022] Figure 2: an exemplary representation of the temporal torque curve after superposition with a periodic signal for the execution of the method according to the invention, and
[0023] Figure 3: an exemplary comparison of the target speed superimposed with a periodic signal and the resulting torque curve over time.
[0024] Figure 1 shows the torque of a centrifugal pump unit as a function of the set target speed of the pump. The diagram shows two curves. The dotted line represents the torque PQO curve across the entire speed range for a suction line completely filled with fluid, but with the suction line valve closed, so that the pump operates at zero flow. The solid line shows the torque Pory curve across the entire speed range when the pump is running dry.
[0025] Overall, it can be seen that with increasing speed, the deviation between the torque Pory during dry running and the torque PQO when the pump is properly filled increases; the described method therefore works very precisely at medium and high speeds. Particularly in the low speed range, e.g. at the speed marked with reference number 1, the torque differences or differences in power consumption between a filled system with a closed valve and an unfilled system are very small or in some cases barely existent. It is difficult to reliably distinguish whether dry running is actually occurring or whether the valve is simply closed in this speed range. To solve the aforementioned problem, the invention therefore proposes briefly increasing the target speed of the speed-controlled centrifugal pump, e.g. to the speed value marked with reference number 2 in Figure 1.Dry running can be detected very reliably here. As shown in the figure, the speed is increased significantly for dry running detection, in this case by a factor of approximately three. In general, it can be stated that a brief increase in the speed to a value between 50% and 100% of the maximum speed enables the most reliable detection. The advantage of this method is its very simple implementation.
[0026] Since in certain processes a temporary increase in speed to the extent suggested above may not be permitted, the method can alternatively provide for excitation of the target speed with a periodic signal, which not only enables dry running detection, but also allows information on the pumped medium to be obtained.
[0027] As part of parameter identification, it is proposed to apply a periodic signal to the current target speed in order to then observe the phase shift between the target speed curve and the resulting (measured or estimated) torque curve. Alternatively, the amplitude of the measured or estimated torque curve can also be considered.
[0028] In a variable-speed centrifugal pump, the current actual speed and torque are available as estimated values from the motor controller. The phase shift or torque amplitude depend on the inertia of the moving machine parts, e.g., impeller, shaft, etc. (known) and the pumped fluid (to be determined). The process can be divided into two steps.
[0029] Step 1: Recording the phase shifts or torque amplitude under normal conditions. Step 1 takes place in the laboratory. Different operating points (various speed / torque combinations) are approached with a defined reference medium and at a reference temperature, and the target speed is superimposed by a higher-frequency signal (e.g., 10 Hz). For each of these operating points, the phase shift between the target speed and the torque curve, or alternatively, the torque amplitude, is recorded. The resulting characteristic map is stored in the pump.
[0030] Step 2: Identifying different parameters in the field
[0031] While the pump is running, the target speed is overlaid with the same higher-frequency signal as previously recorded in the laboratory at regular intervals or once upon request. The phase shift between the target speed curve and the measured / estimated torque curve, or alternatively the torque amplitude of the torque curve, is recorded. If the phase shift or torque amplitude is significantly lower than the reference value for the phase shift recorded in the laboratory, this can be interpreted as an indication of dry running. Furthermore, the phase shift can provide information about a density or temperature of the medium that deviates from the reference value, or an incorrect mixing ratio.
[0032] Figure 2 shows an example of the simulated torque curve of a centrifugal pump, where the static target speed is superimposed with a periodic signal, resulting in the periodic torque curve designated by reference numeral 10 (corresponds to the laboratory setup). Curve 20 shows the torque curve with an identical test setup, but with the simulated mass inertia of the rotating part of the pump increased by 50%. It is clearly visible that the amplitude of curve 20 is higher. The ratio of the amplitudes, marked here with reference numeral 40, is equal to the ratio of the simulated mass inertias. The pump could detect this deviation of 50% compared to the values recorded in the laboratory and conclude from this a change in mass inertia of 50%. The density of the medium could also have changed equivalent to the mass inertia. The mean value of the torque curves 30 is identical.This means that the information of the mass inertia would not be recognizable at a static target speed (without the application of a periodic signal).
[0033] Figure 3 shows the temporal curves of the target speed and the measured torque. Curve 50 corresponds to the target speed superimposed with a periodic signal. Curve 60 shows the resulting torque curve. Reference symbol 70 marks the resulting phase shift between the two curves, which also increases with increasing mass inertia. As a precaution, please note that in Figure 3, the speed curve 50 and the torque curve 60 have been superimposed to better illustrate the phase shift.
Claims
Patent claims Method for checking the properties of a pumped medium of a centrifugal pump, in particular for dry running detection 1 . Method for checking the properties of a pumped medium of a centrifugal pump, in particular for detecting dry running, using a speed-controlled centrifugal pump, wherein for the check the target speed of the centrifugal pump is briefly varied, in particular increased, and a motor parameter characterizing the torque of the pump is determined and evaluated taking into account a reference value in order to determine a property of the pumped medium and / or to detect the presence of dry running.
2. Method according to claim 1, characterized in that by means of the method a change in the density and / or a change in temperature of the conveying medium compared to a reference value is determined.
3. Method according to one of the preceding claims, characterized in that the torque generated by the pump is determined, in particular estimated, as the motor parameter characterizing the torque of the pump, and / or the electrical power consumption of the motor and / or the motor current consumed is determined or estimated.
4. Method according to one of the preceding claims, characterized in that it is first checked whether the current target speed is below a minimum value and a variation or increase in the target speed only occurs if the current target speed is below the minimum value.
5. Method according to one of the preceding claims, characterized in that the speed variation provides a short-term increase in the target speed, the engine parameter for the currently increased target speed is determined and compared with a reference value assigned to the increased target speed, wherein dry running is detected if the engine parameter is smaller than the reference value or is smaller than the reference value by at least a defined value.
6. Method according to one of the preceding claims 1 to 4, characterized in that the target speed for the check is superimposed with a time-variable signal for a certain period of time, in particular with a periodic signal.
7. Method according to claim 6, characterized in that a property of the pumped medium or the presence of dry running is detected on the basis of a correlation between the resulting target speed curve and the time curve of the motor parameter.
8. The method according to claim 7, characterized in that a phase offset between the time course of the target speed and the time course of the engine parameter is evaluated and preferably compared against a reference value.
9. Method according to one of claims 6 to 8, characterized in that the amplitude of the engine parameter is determined during the superposition of the target speed and is preferably compared with a reference value.
10. Method according to one of the preceding claims, characterized in that the one or more reference values are determined by a training procedure carried out under normal conditions. 11 . Centrifugal pump, in particular a heating circulation pump, with an integral or external control configured to carry out the method according to one of the preceding claims.