EP3449132B1 - Method for detecting an abnormal operating state of a pump unit - Google Patents

Description

The present invention relates to a method for detecting an operating state of a speed-controllable centrifugal pump unit operated at a predefinable target speed for pumping liquid.

The pamphlet DE-A-102014004336 discloses a method for detecting a specific operating state of a speed-controllable centrifugal pump unit operated at a predefinable speed by determining a first hydraulic variable from a mechanical and/or electrical variable. However, this operating condition is not an abnormal operating condition.

Abnormal operating states occur again and again in pump units and can lead to damage or even total failure of the pump unit. Examples of abnormal operating states are dry running, cavitation on the impeller or bearing wear, bearing damage.

In the case of pumps, sealing elements such as e.g. B. Mechanical seals are used, which are intended to prevent the escape of the fluid from the pump or, in the case of glandless pumps, the ingress of particles into the rotor space. The sealing surfaces of the sealing elements are friction surfaces that are lubricated by the pumped medium. In the event of dry running, these friction surfaces are no longer lubricated or not sufficiently lubricated and can therefore wear out quickly. Early detection of a dry run can prevent this.

Conventional methods for dry-running detection are based e.g. B. on the power consumption of the pump set. The power consumption depends on the speed and the volume flow. When running dry, the performance is usually lower at the same speed than when it is filled

Pump chamber, since the rotating impeller must deliver the mechanical energy to the medium to be pumped, ie it is loaded by this medium. In the case of dry running, this medium or the load is missing, or there is not enough medium in the pump chamber. At the same speed, the power consumption in dry running is therefore lower than in normal operation because the impeller is less loaded in dry running. For this reason, a speed-dependent decision threshold between the minimum power in wet operation and the power consumption in dry running is selected in order to detect dry running, which is figure 1 is shown. However, this is not free from disadvantages.

Because at low speeds there is the problem that the wet and dry running curves are close together and the minimum wet running characteristic can also fall below the dry running characteristic. For this reason, reliable dry-running detection with this method is only possible above a certain speed.

Another way of detecting dry running is to briefly accelerate the impeller and determine the energy required for acceleration. If the energy consumption is less than expected, i.e. less than in the case of wet operation, dry running is assumed. The disadvantage here is that the operating point must be changed significantly for the acceleration process in order to be able to carry out an evaluation of the energy consumption.

Cavitation is the phenomenon in which, during operation of the pump unit and pumping a liquid, vapor-filled cavities form on the back of the impeller blades, which quickly collapse and cause the blades to hit. This can destroy the impeller.

Bearing wear or damage to the bearings, especially the radial bearings, is primarily reflected in shaft vibrations. This can be detected using a vibration sensor. However, such a sensor is an additional component that can fail, and besides, the manufacture of the pump unit more complex and difficult. In addition, it makes the production of the pump unit more expensive. It is therefore a concern to dispense with sensors.

It is therefore the object of the present invention to provide an alternative method for detecting an abnormal operating state that does not require a sensor and can be implemented in a pump unit without any effort.

This object is achieved by the method according to claim 1, pump electronics according to claim 19 and a pump assembly according to claim 20. Advantageous developments are given in the dependent claims.

In order to detect an abnormal operating state during operation of the pump unit, it is proposed to apply a periodic excitation signal of a specific frequency to a manipulated variable of the pump unit in such a way that a hydraulic variable of the pump unit, for example the delivery head, is modulated, and then from a mechanical and/or electrical Size of the pump unit as a system response to the excitation signal to calculate an evaluation signal and determine whether an abnormal operating condition is present.

This procedure makes it possible to detect an abnormal operating state in a very simple way. The method can be implemented on a software basis in electronics of the pump. Neither a sensor external to the pump electronics nor additional hardware is required. Furthermore, the method can be used reliably in the entire speed range, since it has very good suppression of measurement noise. In contrast to the methods of the prior art, it is therefore not limited to a specific speed range.

The pump unit can be a centrifugal pump operated by an electric motor, for example a heating pump in a heating system or a coolant pump in a cooling system. In particular, it can be a glandless pump. Here is due to the liquid-lubricated radial and possibly the axial plain bearing, dry running is particularly harmful and must therefore be detected at an early stage.

The modulation of the hydraulic variable is achieved by modulating a manipulated variable, e.g. the speed. It should be noted that "modulate" within the meaning of the invention is generally to be understood as a change, but the type, level and speed of the excitation signal is not restricted in any way. Furthermore, insofar as control of the pump assembly is discussed below, this term also means regulation, since regulation only includes control with feedback of a specific variable.

An abnormal operating state can be detected, for example, by comparing the evaluation signal with a decision threshold. An abnormal operating state can then be concluded if the evaluation signal deviates from a normal range (range permissible during operation). The decision threshold can form a boundary of this normal range, for example a minimum curve or a maximum curve of the normal range, or have a distance to this range in order to avoid erroneous determinations. Depending on the abnormal operating state to be determined, it is therefore necessary to check whether the evaluation signal is above or below the decision threshold. Consequently, the decision threshold depends on the abnormal operating state that is to be detected. In other words, a specific decision threshold is associated with a specific abnormal operating condition. Thus, a first abnormal operating state can be determined by comparing the evaluation signal with a first decision threshold and a second abnormal operating state can be determined by comparing the evaluation signal with another, second decision threshold. Correspondingly, there can also be a third or further decision threshold(s).

It also depends on the abnormal operating state that is to be detected whether the decision threshold defines an upper limit or a lower limit of the permissible normal range. For example, if it defines a lower limit in the case of the first abnormal operating state, then this first abnormal operating state are closed when the evaluation signal is below the first decision threshold. For example, if the decision threshold defines an upper limit in the case of the second abnormal operating state, this second abnormal operating state can be inferred if the evaluation signal is above the second decision threshold.

For example, dry running can be inferred as an abnormal operating state if the evaluation signal is below the decision threshold.

In the simplest case, the decision threshold can be a constant value. Alternatively, the decision threshold can be defined by a curve, preferably a straight line, which defines a mathematical relationship between the evaluation signal and the speed, in particular the actual speed. If the decision threshold is a function of the rotational speed, it can be set according to the rotational speed dependency of normal or faulty operation in such a way that the decision is as insensitive as possible to measurement noise and disturbances. In the case of dry running, the decision threshold, also referred to here as the threshold value curve, suitably lies between the dry running curve and the minimum wet running curve.

According to one embodiment variant, the evaluation signal can be formed from the integral of the product of the system response and a periodic function of the same frequency or a multiple of the frequency of the excitation signal over a predetermined integration period. It should be noted at this point that integral formation within the meaning of the invention is also to be understood as a summation of values, which must necessarily take place when the method is carried out numerically, for example on a microprocessor, due to the discrete values present. Because it is well known to the person skilled in the art that the integral formation in the discrete time domain can be realized by a sum formation.

wobei I(t) das zu berechnende Auswertesignal zum Zeitpunkt t über einen Integrationszeitraum T, X(t) die Systemantwort, S(t) die periodische Funktion, k_l eine positive ganze Zahl und ω die Frequenz des Anregungssignals ist.The evaluation signal can preferably be determined using the following calculation rule: $$I\left(t\right)={\displaystyle \underset{t-T}{\overset{t}{\int }}X\left(t\right)\cdot S\left(t\right)\mathrm{i.e}t}\mathit{with}T=\frac{k_{I}2\pi }{\omega }$$

where I(t) is the evaluation signal to be calculated at time t over an integration period T , X(t) is the system response, S(t) is the periodic function, k _l is a positive integer and ω is the frequency of the excitation signal.

The periodic function can be a sine function. Investigations have shown that this simple function is sufficient for generating the evaluation signal for detecting a dry run. However, depending on the abnormal operating state to be detected, a cosine function, a combination of a sine function and a cosine function, a combination of several sine functions or cosine functions, or a combination of several sine functions and cosine functions can also be used as the periodic function.

The manipulated variable to which the excitation signal is applied is preferably a target speed or a target torque of the pump assembly, i.e. a mechanical variable. The method according to the invention can be implemented particularly easily here, because the speed or the torque is often a controlled variable in a pump unit, i.e. a variable whose magnitude is specified by a setpoint. Speed or torque controls are known per se in pump units. The control of the pump unit then attempts to regulate the setpoint. The periodic excitation of the target speed or the target torque is a simple measure to achieve a modulation of the size. Ultimately, this modulates the mechanical power output by the drive motor.

Alternatively, the manipulated variable can be a current from the pump unit. This is a suitable manipulated variable, especially for pump units electric drive motor is controlled by means of a vector control such as field-oriented control (FOR). On the basis of a motor model, current components id and iq are formed here, which define the current vector rotating with the frequency of the stator field. With the so-called id current, the field with the iq current can influence the torque. Consequently, modulation of the speed or the torque can be achieved indirectly by modulating the current.

The hydraulic variable to be modulated can suitably be the delivery height H or the differential pressure Δp generated by the pump unit. Because the direct effect of applying the excitation signal to the speed or the torque can be seen from the modulation of the delivery head or the differential pressure. Depending on the operating point, this modulation then causes a more or less pronounced modulation of the flow rate of the pump unit.

With regard to the system response, the torque delivered by the pump assembly or the actual speed is preferably used as the mechanical variable. An electrical variable as a system response can be, for example, the electrical power P _el consumed by the electric motor of the pump unit or the current. According to the invention, the change in at least one of these variables as a result of the modulation of the hydraulic variable is regarded as a system response.

It is thus possible to use different pairings between the excited manipulated variable and the system response to be analyzed. For example, the target speed can be modulated and the resulting electrical power consumption can be evaluated. Instead of the power consumption, the output torque or the actual speed can be used for evaluation. And instead of excitation of the setpoint speed, the setpoint torque can be excited and the resulting actual speed, the torque output or the electrical power consumption can be evaluated.

The periodic excitation signal can ideally be a sine signal or a signal containing a sine function. The latter can also be a triangle or sawtooth signal, for example. Except for a weighting factor, the excitation signal preferably corresponds to the periodic function that is used to form the evaluation signal. The weighting factor defines the amplitude of the modulation of the manipulated variable.

The frequency of the excitation signal can be between 0.1 Hz and 100 Hz, preferably between 0.5 Hz and 10 Hz. It should be pointed out that the frequency must be selected depending on the hydraulic system in which the pump unit is operated. There are upper limits to the frequency due to the inertia of the rotor, the impeller and the liquid being pumped. In addition, the excitation frequency and speed controller of the pump unit must be matched to one another. Because if the frequency is too high, the speed controller may not be able to regulate the modulated target speed quickly enough. In this special case, however, the actual speed can be used for the evaluation.

The disadvantage of a frequency that is too low is that the response time for the calculation of the evaluation signal increases in accordance with the duration of the period, so that the method takes longer. If the frequency is too low, the system response can also be weak, so that the information about the abnormal operating state is also only weak in the evaluation signal. For this reason, the excitation frequency should not be too low, for example not less than 0.1 Hz.

The amplitude of the excitation signal can preferably be less than 25% of the speed setpoint. In particular, it can be between 0.1% and 25% of the speed setpoint. With a setpoint speed of 2000 rpm, for example, a speed fluctuation of ±2 rpm to ±500 rpm can be used.

To calculate the evaluation signal, the integral of the product of the system response and the periodic function over a period of time T is calculated.

This integration period T can be at least one period or a multiple of the period of the excitation signal.

The calculation of the evaluation signal or the integration can preferably be carried out during the modulation of the hydraulic variable, in particular the setpoint speed. The calculation therefore does not only take place when the modulation has ended. This prevents only the decaying system response from being analyzed. It is also advantageous if the calculation of the evaluation signal or the integration only begins after a certain period of time, for example after one or a few periods of the excitation signal have elapsed. This ensures that the system response is only analyzed when the system consisting of the pump unit and the connected pipelines has settled. Transient effects therefore do not affect the evaluation of the system response.

It is advantageous if the method according to the invention is carried out continuously while the pump assembly is in operation. As a result, changes in the operating status can be recognized immediately. Alternatively, the method can be carried out at suitable points in time, at time intervals, in particular regularly.

According to another alternative, the method according to the invention can be started by a trigger during pump operation. Such a trigger can be, for example, that another method for detecting an abnormal operating state, as is known in the prior art and can be implemented in an electronic system of the pump parallel to the method according to the invention, detects just such an abnormal operating state. Since this recognition can be imprecise, a verification can be carried out using the method according to the invention. The method according to the invention can thus be activated during operation of the pump assembly if the use of a method known in the prior art for detecting an abnormal operating state detects an abnormal operating state.

If an abnormal operating state is detected using the method according to the invention, an error signal can be output, for example optically, acoustically or as an electronic message, so that a user or service technician or a connected system (e.g. heating system or building management system) is informed of the abnormal operating state. He can then initiate appropriate measures. Alternatively, the pump set can be shut down to prevent further or more serious damage to the pump set or other components of the system.

The mechanical and/or electrical variable as a system response to the modulation can be recorded either at discrete points in time or continuously. The system response is then available as a sequence of values, so that the multiplication with the periodic function and the integration of the product obtained in this way can take place at any time.

According to the invention, pump electronics for controlling and/or regulating the target speed of a pump unit are also proposed, which are set up for carrying out the method described above. A pump assembly having such pump electronics is also proposed. The pump unit can be, for example, a heating pump, a coolant pump or a drinking water pump. The pump assembly is preferably an electric motor-driven centrifugal pump, ideally of wet-running or dry-running design.

The use of the method according to the invention makes it possible to dispense with sensors that are arranged externally to the pump electronics. This structurally simplifies the pump housing and makes its production cheaper. In addition, it is possible to reliably detect an abnormal operating state over a wide speed range.

Figur 1:

Diagramm mit einer Trockenlaufkurve, einer Nasslaufkurve sowie einer dazwischenliegenden Entscheidungsschwelle nach dem Stand der Technik

Figur 2:

Ablaufdiagramm des Verfahrens

Figur 3:

System zur Anwendung des erfindungsgemäßen Verfahrens

Figur 4:

Strukturbild zusammenwirkender Funktionseinheiten zur Ausführung des Verfahrens

Figur 5:

Diagramm mit vermessenen Nass- und Trockenlaufbetriebsfällen

The invention is explained in more detail below using examples and the attached figures. Show it:

Figure 1:

Diagram with a dry-running curve, a wet-running curve and a decision threshold in between according to the prior art

Figure 2:

Flow chart of the procedure

Figure 3:

System for applying the method according to the invention

Figure 4:

Structural diagram of interacting functional units for executing the method

Figure 5:

Diagram with measured wet and dry running cases

figure 1 shows a diagram that illustrates the mode of action of a method for determining a dry run according to the prior art in a pump assembly during operation. The diagram shows the hydraulic power P_hydr over the actual speed n_act. The diagram shows a curve 11 measured in normal, wet-running operation and a curve 9 measured in abnormal, dry-running operation. It becomes clear that the wet-running curve 11 indicates a higher power output than the dry-running curve 9 at the same speed. The power output of the pump set is therefore always lower in the event of dry running, at least unless the pump set is brand new or one that has been dry for a long time. Because in these cases the pump assembly has a higher dry-running curve that even intersects the wet-running curve 11 .

Between the dry-running curve 9 and the wet-running curve 11, a curve is drawn approximately in the middle between the two curves 9, 11, which forms a decision threshold 10. It serves as a reference for the decision as to whether dry running is present or not with regard to the current speed and the current power. This determination can be made by comparison with the decision threshold 10. If the power is below the decision threshold 10, dry running can be assumed. As a rule, instead of the hydraulic power P_hydr in the prior art, the electrical power P _el is determined and used to identify dry running.

At higher speeds, this method provides reliable detection of dry running. However, since the wet-running curve 11 and the dry-running curve 9 are very close to one another at low speeds, reliable dry-running detection in this speed range is not possible with this method.

The method described below for hydraulically determining an abnormal operating state makes use of the dynamic behavior of the system, which is formed from the pump unit 1 and the pipelines connected to it and is analyzed by targeted excitation.

A model of the system in which an embodiment variant of the method according to the invention can be applied is shown figure 3 as a block diagram. A variable-speed centrifugal pump unit 1 is shown there, which is connected to a pipeline system 5 or is integrated into it. The system can be a heating system, for example, and the pump assembly 1 can be a heating pump accordingly. The piping system 5 is then formed by the lines leading to the radiators or heating circuits and leading back from them to a central heating source. For example, water that is driven by the pump unit 1 can circulate in the pipelines 5 as the liquid.

The pump unit 1 consists of a pump unit 2, which forms the hydraulic part of the unit 1, an electric motor drive unit 3, which forms the electromechanical part of the unit 1, and control electronics 4 for controlling and/or regulating the drive unit. The drive unit 3 consists of an electromagnetic part 3a and a mechanical part 3b. The control electronics 4 includes hardware 4b on the one hand and software 4a on the other. The hardware 4b also includes power electronics such as a frequency converter, for example, in order to set a specific speed on the drive unit.

A setpoint speed n ₀ can be predetermined for the control electronics 4 . Although this is shown here as coming from outside the control electronics 4, for example by manual specification, it can alternatively also come from a characteristic curve control or a dynamic, needs-based adjustment of the operating point of the pump unit 1, which itself is part of the control electronics 4, in particular part of its software 4a.

From the current current consumption l _el and the current speed n, the drive unit 3 _calculates the control electronics 4 or its software 4a a voltage U which is specified for the power electronics 4b so that the drive unit 3 has a corresponding electrical power P _el available .

The electromagnetic part 3a of the drive unit 3, which describes the stator, rotor and their electromagnetic coupling, _generates a mechanical torque M actual from the current l _el . This accelerates the rotor and leads to a corresponding speed n actual _of the drive unit 3, which is included in the mechanical part 3b of the model of the drive unit 3. The pump impeller, which _is seated on the rotor shaft, of the hydraulic part 2 of the pump unit 1 is now driven at the speed n actual . As a result, the pump unit 1 generates a differential pressure between the suction side and the pressure side or a delivery height H which generates a more or less large volume flow Q in the pipeline system 5 depending on the pipeline resistance. From the hydraulic power P_hydr and the associated losses, a hydraulic torque M _hyd can be defined, which counteracts the engine torque _M act as a braking torque.

The basic sequence of the method according to the invention is figure 2 shown. The method is carried out when the pump assembly is operating as intended, ie when the pump assembly 1 is connected to a pipework system 5 and is operated at any setpoint speed n ₀ .

• Anregung des Systems, Schritt S2;
• Ermittlung der Systemantwort beispielsweise durch Messung, Schritt S3;
• Berechnung eines Auswertesignals aufgrund der Systemantwort, Schritt S4 und
• Auswertung des Auswertesignals durch Vergleich mit einer Entscheidungsschwelle, Schritt S5.

Based on the specification of the setpoint speed n ₀ in step S1, the method according to the invention comprises the following steps to be carried out one after the other, which can be repeated continuously:

• excitation of the system, step S2;
• Determination of the system response, for example by measurement, step S3;
• Calculation of an evaluation signal based on the system response, step S4 and
• Evaluation of the evaluation signal by comparison with a decision threshold, step S5.

Although the method is described below based on the detection of dry running as an abnormal condition, the method can also be used to detect other faults in the pump unit or in the overall system, provided that in this case the evaluation signal determined is also outside of a normal range that is defined by a corresponding decision threshold is limited.

The system is excited by a manipulated variable, here the steady-state setpoint speed n ₀ , being modulated with an excitation signal f _A (t), so that the new setpoint speed n setpoint _to be set by the pump electronics 4 is the sum of the previously specified setpoint speed n ₀ and the excitation signal f _A (t) gives: $$n_{\mathit{should}}={\mathrm{<figure-callout\; id="0"\; label="n"\; filenames="imgf0001.png,imgf0005.png"\; state="\{\{state\}\}">n</figure-callout>}}_0+{ƒ}_A\left(t\right)$$

mit der Amplitude n₁ und der Frequenz ω = 2πf. The excitation of the speed is purely sinusoidal here, but other modulations are also conceivable. The excitation signal f _A (t) then has the form, for example $${ƒ}_A\left(t\right)=n_{\mathrm{l}}\sin \mathit{\omega t}$$

with amplitude n ₁ and frequency ω = 2 πf.

The amplitude n ₁ is between 0.1% and 25% of the setpoint speed n ₀ and can be set and fixed at the factory, for example at 1%. The excitation frequency f or ω should be dimensioned in such a way that the speed controller can follow the rate of change of the speed sufficiently quickly. In this case, no correction of the controller parameters, e.g Proportional gain, are made. A frequency f of 1 Hz is used in the exemplary embodiment.

The system response that follows as a reaction to the excitation manifests itself simultaneously in various physical parameters of the pump unit. Furthermore, purely mathematical quantities of the models, i.e. the electrical model 4b, electromagnetic model 3a, mechanical model 3b and the hydraulic model 2, also show a reaction to the speed modulation. Nevertheless, it is sufficient to evaluate a single variable, in particular a mechanical or electrical variable of the pump assembly, particularly if the actual speed can follow the modulated setpoint speed. If the actual speed cannot follow the target speed, it is advisable to evaluate two or more variables, e.g. a mechanical variable such as the actual speed and an electrical variable such as the power consumed by the motor.

In the exemplary embodiment, the electrical power P _el consumed is used as the system response X (t) to the speed modulation. This can be measured or determined directly from the measured current and the measured or calculated voltage. Alternatively, the torque or the current drawn can also be used as the system response.

The system response can be determined by sampling at discrete points in time or continuously, so that the system response X (t) is present as a discrete or continuous series of measured values or calculated values. This is from step S3 figure 3 includes. For the sake of simplicity, only the case of the continuous series is treated here.

wobei g ein Skalierungsfaktor sind und k eine positive ganze Zahl ist und ein Vielfaches der Grundfrequenz definiert. Im einfachsten Fall sind g = k = 1 gesetzt. Dies verdeutlicht, dass die periodische Funktion im einfachsten Fall dieselbe periodische Grundstruktur wie das Anregungssignal f_A(t), vgl. Gl. 2, insbesondere dieselbe Frequenz ω bzw. f haben kann, um das erfindungsgemäße Ergebnis zu erreichen.For the dry-running detection in step S5, the evaluation signal I(t) is first determined. This is done by first multiplying the system response X (t) by a periodic function S(t), ie the product of the system response X (t) and this periodic function S(t) is formed. In the present example, the periodic function S(t) is a sine function of the form $$S\left(t\right)=G\cdot \sin \left(k\cdot \mathit{\omega t}\right)$$

where g is a scale factor and k is a positive integer and defines a multiple of the fundamental frequency. In the simplest case g = k = 1 are set. This makes it clear that in the simplest case the periodic function has the same periodic basic structure as the excitation signal f _A (t), see Eq. 2, in particular can have the same frequency ω or f in order to achieve the result according to the invention.

wobei t₀ den Zeitpunkt des Integrationsbeginns angibt. Da zur Integration die Systemantwort für den Integrationszeitraum vorliegen müssen, kann das Integral frühestens nach Ablauf des Integrationszeitraums berechnet werden, d.h. bei t = t₀ + T. Der Zeitpunkt t₀ liegt folglich in der Vergangenheit und stellt nicht die Gegenwart zum Zeitpunkt t dar. Um dies zu verdeutlichen kann die Integration des Integrals von t-T bis t gehen. Um die Gegenwart t nicht mit der Integrationsvariablen zu verwechseln, ist nachfolgend als Integrationsvariable t' gewählt: $$I\left(t\right)={\displaystyle \underset{t-T}{\overset{t}{\int }}X\left(t\prime \right)\cdot S\left(t\prime \right)\mathrm{d}t\prime }$$

The product of the system response X (t) and the function S(t) is then integrated over a period of time T, which corresponds to the period or a multiple k _l of the period of the excitation signal f _A (t). The integral I (t ₀ ) over the product represents the evaluation signal according to the invention and then results in: $$I\left({\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="imgf0001.png,imgf0005.png"\; state="\{\{state\}\}">t</figure-callout>}}_0+T\right)={\displaystyle \underset{t_0}{\overset{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="imgf0001.png,imgf0005.png"\; state="\{\{state\}\}">t</figure-callout>}}_0+T}{\int }}X\left(t\right)\cdot S\left(t\right)\mathrm{i.e}t}={\displaystyle \underset{t_0}{\overset{{\mathrm{<figure-callout\; id="0"\; label="t"\; filenames="imgf0001.png,imgf0005.png"\; state="\{\{state\}\}">t</figure-callout>}}_0+T}{\int }}X\left(t\right)\sin \left(\mathit{\omega t}\right)\mathrm{i.e}t}\mathit{with}T=\frac{{\mathrm{<figure-callout\; id="2"\; label="k\; l"\; filenames="imgf0003.png"\; state="\{\{state\}\}">k</figure-callout>}}_{l}2\pi }{\omega }$$

where t ₀ indicates the time at which integration begins. Since the system response for the integration period must be available for integration, the integral can only be calculated after the integration period has expired, i.e. at t = t ₀ + T. The time t ₀ is therefore in the past and does not represent the present at time t. To illustrate this, the integration of the integral can go from tT to t. In order not to confuse the present t with the integration variable, t' is chosen below as the integration variable: $$I\left(t\right)={\displaystyle \underset{t-T}{\overset{t}{\int }}X\left(t\prime \right)\cdot S\left(t\prime \right)\mathrm{i.e}t\prime }$$

By forming the integral I (t ₀ ), the system response X (t) is evaluated at the excitation frequency ω or a multiple k _l of the excitation frequency ω over one or more periods 2π/ω, see step S4 in 2 . A value is obtained that allows a statement to be made about the operating status.

This value of the evaluation signal is now compared with a decision threshold 10, see step S5 in 2 . An operating state can be derived from the fact that the value is greater or less than the decision threshold.

In the case of dry-running detection, it was recognized that the value of the evaluation signal in normal operation is above a minimum curve 15a that delimits a normal range from below. The normal range is delimited at the top by a maximum curve 15b. A dry run occurs as soon as the evaluation signal I(t) is below the minimum curve 15a. This is examined in step S5 using decision threshold 10 . It makes sense for the decision threshold 10 to be kept at a distance from the minimum curve 15a in order to avoid false triggering.

If, for example in the case of dry running, the value of the evaluation signal I(t) is above the decision threshold 10, then there is no dry running, see No branch and step S8. The method can then be continued with the excitation of the system, step S2. Alternatively, the method can be terminated and reactivated at a later point in time, for example in a time-controlled manner or triggered by another trigger.

The procedure can be similar if it is recognized for another abnormal operation that the signal I(t) in this case lies below another minimum curve or above a maximum curve.

If, for example in the case of dry running, the value of the evaluation signal I(t) is below the decision threshold 10, then an abnormal operating state is present, see Yes branch and step S6. An error message can then be output, step S7, for example as an optical or acoustic signal on the pump unit or another device for monitoring the pump unit, or as an electronic message to a building management system. Alternatively or additionally, the pump unit can be switched off immediately in order to prevent further damage.

The calculation according to equation Eq. 4 or 5 can be carried out numerically in a microprocessor of the pump electronics 4 or by an analog circuit.

A structural diagram with functional units and signals for carrying out the method according to the invention is shown in 4 given. These functional units include a modulation unit 12, the drive motor 3, a system response determination unit 13, a dry-running detection unit 14 and a reaction unit 4a'. The modulation unit 12, the system response determination unit 13, the dry-running detection 14 and the reaction unit 4a' can also be part of the pump electronics 4, in particular its software. You can also be partially formed by hardware components. Thus, the system response determination unit 13 can include a sensor in order to determine the electrical power consumption and/or the reaction unit 4a′ can include a switch in order to switch off the pump unit.

The modulation unit 12 is supplied with the rotational speed setpoint n ₀ . The modulation unit 12 generates the periodic excitation signal f _A (t) in the form _of a sinusoidal signal n ₁ sin(ωt) and adds this to the speed setpoint n ₀ so that a new speed setpoint n ₀ + n ₁ *sin(ωt) which is output from modulation unit 12 . At the same time, the modulation unit 12 outputs the sinusoidal signal S(t)=sin(ωt) from the excitation signal f _A (t) separately and makes it available to the system response determination unit 13 . Alternatively, the excitation signal f _A (t) can also be output directly.

The new speed setpoint n ₀ + n ₁ · sin(ωt) is set for the drive motor 3 using the in 4 Not shown power electronics set 4b. For reasons of simplicity, the path of this signal is not differentiated in more detail here. It corresponds to the usual way when specifying the speed setpoint of an electric motor.

As a result of the speed setting, the drive motor 3 has a specific electrical power consumption P _el , which represents a system response X(t) to the speed modulation. The electrical power consumption P _el is ascertained by the system response determination unit 13 .

The system response determination unit 13 now forms the evaluation signal l(t ₀ ) or l(t) according to one of the equations 4 or 5 from the integral of the product of the system response X (t) = P _el (t) and the periodic function S(t) over a period T of the excitation signal f _A (t). The calculated evaluation signal I(t ₀ ), I(t) is then output by the system response determination unit 13 and made available to the dry-running detection unit 14 . This now determines from the evaluation signal l(t ₀ ), l(t) whether an abnormal operating state is present. For this purpose, the current value of the evaluation signal I(t ₀ ), I(t) is compared with the decision threshold 10 . If the evaluation signal l(t ₀ ), l(t) is below the decision threshold 10, the dry-running detection 14 outputs an error signal. This is provided to the reaction unit 4a′, which reacts to the detected dry run with a predetermined measure, for example displays the error signal, forwards it to a higher-level controller or control system and/or switches off the drive motor 3 .

figure 5 illustrates values of the evaluation signal according to Equation 4 or 5 at different speeds and with dry running on the one hand and wet running on the other hand, with six curves 15 being shown in the case of wet running and only one operating point each in the form of a plus sign being given in the case of dry running. The investigation was carried out at speeds of 800 rpm, 1200 rpm, 1600 rpm, 2000 rpm, 2400 rpm and 2800 rpm. In the case of the wet-running curves 15, the thick black arrow indicates that the delivery flow Q increases upwards, ie the higher Q is, the larger the evaluation signal is. When running dry, Q=0. Nevertheless is in figure 5 a speed-dependent decision threshold 10 is drawn. This is defined here by a linear relationship between the evaluation signal and the speed. It connects points that are approximately in the middle between the minimum operating points (Qmin) of the wet-running curves 15 and the dry-running operating points.

The method presented here for modulating a manipulated variable of the pump unit and analyzing the system response to this makes it possible to easily and reliably make a reliable statement about an abnormal operating state during operation of the pump unit without using a sensor in the entire speed range. It can be easily integrated into the electronics of the pump unit because it can be implemented purely in software. Although dry running is mentioned in the above example, the principle of the invention can be transferred to the detection of other abnormal operating states. The detection of an abnormal operating condition can be used to switch off the pump to protect it from wear or a corresponding signal to a higher-level control, e.g. B. to send a building management system.

Claims (20)

1. Method for detecting an operating state of a rotational-speed-controllable centrifugal pump unit (1) operated at a specifiable rotational speed (n ₀), characterised by the operating state being an abnormal operating state and a periodic excitation signal (f_A(t)) of a certain frequency (f) being applied to a manipulated variable of the pump unit (1) in such a way that a hydraulic variable (H, Δp) of the pump unit (1) is modulated and that, from a mechanical and/or electrical variable (P_el) of the pump unit (1) as a system response (X(t)) to the excitation signal (f_A(t)), an evaluation signal (l(t₀), l(t)) is calculated and this is used to determine whether an abnormal operating state exists.
2. Method according to claim 1, characterised by the evaluation signal (l(t₀), l(t)) being compared to a decision threshold (10) and the abnormal operating state being inferred when the evaluation signal (l(t₀), l(t)) is above or below the decision threshold (10).
3. Method according to claim 2, characterised by dry running being inferred as an abnormal operating state when the evaluation signal (l(t₀), l(t)) is below the decision threshold (10).
4. Method according to claim 2 or 3, characterised by the decision threshold being dependent on the rotational speed.
5. Method according to one of the preceding claims, characterised by the evaluation signal (l(t₀), l(t)) being formed from the integral of the product of the system response (X(t)) and a periodic function (S(t)) of the same or a multiple of the frequency (f) of the excitation signal (f_A(t)) over a specified integration period (T).
6. Method according to claim 5, characterised by the periodic function (S(t)) being a sine function (S₁(t)).
7. Method according to one of the preceding claims, characterised by the manipulated variable being the nominal rotational speed (n₀), a torque value, or a current of the pump unit (1).
8. Method according to one of the preceding claims, characterised by the hydraulic variable (H) being the pumping head (H) or the differential pressure (Δp) of the pump unit (1).
9. Method according to one of the preceding claims, characterised by the mechanical variable being the output torque (M_ist ) or the current rotational speed (n_ist) of the pump unit (1).
10. Method according to one of the preceding claims, characterised by the electrical variable being the electric power consumed by the electric motor of the pump unit (1) (P_el ) or a current draw.
11. Method according to one of the preceding claims, characterised by the excitation signal (f_A(t)) being a sinusoidal signal or a signal containing a sine function.
12. Method according to one of the preceding claims, characterised by the frequency (f) of the excitation signal (f_A(t)) being between 0.1 Hz and 100 Hz.
13. Method according to one of the preceding claims, characterised by the amplitude (n₁) of the excitation signal (f_A(t)) being less than 25% of the target rotational speed (n₀) for speed control of the pump unit (1), in particular between 0.1% and 25% of the target rotational speed (n₀).
14. Method according to claim 5 or according to one of the claims 6 through 13 referring back to claim 5, characterised by the integration period (T) being a period or a multiple (k_l) of the period (2π/ω) of the excitation signal (f_A(t)).
15. Method according to claim 5 or according to one of the claims 6 through 14 referring back to claim 5, characterised by the integration being performed during the modulation of the hydraulic variable (H), in particular the nominal rotational speed (n_soll ).
16. Method according to one of the preceding claims, characterised by it being carried out continuously during the operation of the pump unit (1).
17. Method according to one of the preceding claims, characterised by it being activated during the operation of the pump unit (1) when the use of a method for the detection of an abnormal operating state recognises an abnormal operating state.
18. Method according to one of the preceding claims, characterised by an error signal being output and/or the pump unit (1) being shut off when an abnormal operating state is detected.
19. Pump electronics for the management and/or control of the nominal rotational speed of a pump unit (1), characterised by being configured to carry out the method according to one of the claims 1 through 18.
20. Pump unit having pump electronics according to claim 19.

Images