Method for improving sensorless flow rate estimation accuracy of pump driven with frequency converter

Abstract

An exemplary method for determining a flow rate of a pump, includes measuring and storing the flow rate (Q.sub.Mean) when the pump is in operation, storing rotational speed (n.sub.Mean) and torque (T.sub.Mean) of the pump during measurement of the flow rate (Q.sub.Mean), determining torque (T.sub.Manufacturer) corresponding to the measured flow rate (Q.sub.Mean), calculating a bias value of the torque (T.sub.Bias) by comparing the torque determined from the characteristic curve (T.sub.Manufacturer) and the stored torque (T.sub.Mean) of the pump, and determining an operating point of the pump based on the bias value of the torque (T.sub.Bias).

Claims

1. A method in connection with a pump driven with a frequency converter, comprising: a) measuring, in a measurement device, and storing a flow rate (Q.sub.Mean) when the pump is in operation; b) storing a used rotational speed (n.sub.Mean) and torque (T.sub.Mean) of the pump during measurement of the flow rate (Q.sub.Mean), which rotational speed and toque are estimated with the frequency converter; and in the frequency converter: c) determining from a characteristic curve that is transformed into stored rotational speed (n.sub.Mean) a predetermined torque (T.sub.Manufacturer) corresponding to the measured flow rate (Q.sub.Mean); d) calculating a bias value of the torque (T.sub.Bias) by comparing the predetermined torque of the characteristic curve (T.sub.Manufacturer) and the stored torque (T.sub.Mean) of the pump; and e) determining an operating point of the pump based on the bias value of the torque (T.sub.Bias), wherein the bias value of the torque is transferred with an affinity law from the rotational speed (n.sub.Mean) in which the bias value of the torque (T.sub.Bias) was initially evaluated to an estimated rotational speed (n.sub.est), and wherein the step of measuring and storing the flow rate comprises storing a time stamp of the measurement, and the step of storing the used rotational speed and torque comprises storing a time stamp of stored events for synchronizing the rotational speed and torque with the stored flow rate, and f) operating the pump with the frequency converter according to the determined operating point.

2. The method according to claim 1, wherein determining the operating point of the pump comprises: estimating torque (T.sub.est) and the rotational speed (n.sub.est), subtracting the bias value of the torque (T.sub.Bias) from the estimated torque (T.sub.est) to obtain the torque (T.sub.UseForQCalculation) value used in a QT calculation.

3. The method according to claim 1, comprising: repeating the steps a) to d) with differing flow rates, and calculating the bias value used in the determination of the operating point from bias values (T.sub.Bias1, T.sub.Bias2, T.sub.Biasn) obtained with the differing flow rates.

4. The method according to claim 3, comprising: calculating the bias value used in the determination of the operating point of the pump as a mean value of the calculated bias values (T.sub.Bias1, T.sub.Bias2, T.sub.Biasn).

5. The method according to claim 3, wherein the bias values obtained with differing flow rates are used to provide linear correction to the estimated torque.

6. The method according to claim 1, comprising: measuring the flow rate with a portable measurement device.

7. The method according to claim 1, wherein the steps of measuring and storing the flow rate comprise: calculating a mean value of the flow rate.

8. The method according to claim 1, wherein the used rotational speed and torque are obtained from the frequency converter.

9. A method of calculating flow rate of a pump, comprising: in a measuring device: a) measuring and storing a flow rate (Q.sub.Mean) when the pump is in operation; and in a frequency converter: b) storing rotational speed (n.sub.Mean) and torque (T.sub.Mean) the pump during measurement of the flow rate (Q.sub.Mean); c) determining a predetermined torque corresponding to the measured flow rate (Q.sub.Mean); d) calculating a bias value of the torque (T.sub.Bias) by comparing the predetermined torque of the characteristic curve (T.sub.Manufacturer) and the stored torque (T.sub.Mean) of the pump; and e) determining an operating point of the pump based on the bias value of the torque (T.sub.Bias), wherein the bias value of the torque is transferred with an affinity law from the rotational speed (n.sub.Mean) in which the bias value of the torque (T.sub.Bias) was initially evaluated to an estimated rotational speed (n.sub.est), and wherein the step of measuring and storing the flow rate comprises storing a time stamp of the measurement, and the step of storing the used rotational speed and torque comprises storing a time stamp of stored events for synchronizing the rotational speed and torque with the stored flow rate.

10. The method of claim 9, wherein the predetermined torque (T.sub.Manufacturer) is obtained from a characteristic curve that is transformed into stored rotational speed (n.sub.Mean).

11. The method according to claim 9, comprising: repeating the steps a) to d) with differing flow rates, and calculating the bias value used in the determination of the operating point from bias values (T.sub.Bias1, T.sub.Bias2, T.sub.Biasn) obtained with the differing flow rates.

12. The method according to claim 11, comprising: calculating the bias value used in the determination of the operating point of the pump as a mean value of the calculated bias values (T.sub.Bias1, T.sub.Bias2, T.sub.Biasn).

13. The method according to claim 11, wherein the bias values obtained with differing flow rates are used to provide linear correction to the estimated torque.

14. The method according to claim 9, comprising: measuring the flow rate with a portable measurement device.

15. The method according to claim 9, wherein the steps of measuring and storing the flow rate comprise: calculating a mean value of the flow rate.

16. The method according to claim 9, wherein the used rotational speed and torque are obtained from the frequency converter.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) In the following, the disclosure will be described in greater detail by means of exemplary embodiments and with reference to the attached drawings, in which

(2) FIG. 1 illustrates an example of characteristic curves of a pump in accordance with a known embodiment;

(3) FIG. 2 illustrates published and measured characteristic curves of a pump in accordance with a known embodiment;

(4) FIG. 3 shows published, measured, and estimated values of the shaft torque as function of the pump flow rate in accordance with an exemplary embodiment of the present disclosure;

(5) FIG. 4 shows estimated and measured flow rates in accordance with an exemplary embodiment of the present disclosure;

(6) FIG. 5 shows estimation errors relating to the flow rates of FIG. 4 in accordance with an exemplary embodiment of the present disclosure;

(7) FIG. 6 shows estimated and measured flow rates near the pump's best efficiency point in accordance with an exemplary embodiment of the present disclosure;

(8) FIG. 7 shows estimation errors relating to the flow rates of FIG. 6 in accordance with an exemplary embodiment of the present disclosure;

(9) FIG. 8 shows estimated and measured flow rates when the pump was driven at a relative flow rate of 70% in accordance with an exemplary embodiment of the present disclosure;

(10) FIG. 9 shows estimation errors relating to the flow rates of FIG. 8 in accordance with an exemplary embodiment of the present disclosure, and

(11) FIG. 10 shows a flow chart for measuring flow rate in accordance with an exemplary embodiment of the present disclosure.

DETAILED DESCRIPTION

(12) An object of the present disclosure is thus to provide a method so as to overcome the above shortcoming in improving the accuracy of the estimation of the operating point of a pump. The object of the disclosure is achieved by a method which is characterized by what is stated in the independent claim. Exemplary embodiments of the disclosure are disclosed in the dependent claims.

(13) The disclosure is based on the idea of using a temporary flow measurement device to measure the flow rate produced by the pump during its normal operation. The measured flow rate together with the estimates obtained from the frequency converter can be used to correct the QT curve.

(14) An advantage of the method of the disclosure is that the estimation of the operating point of the pump can be greatly improved. The method does not specify that the pump is operated in specific operating points, such as against a closed discharge valve. Further, the correction is easy to accomplish and it can be made part of a normal inspection routine of the process.

(15) As already mentioned, FIG. 1 discloses characteristic curves of a pump. These curves or the information needed to construct these curves can be provided by the pump manufacturer. A frequency converter that controls the pump provides estimates for the rotational speed of the pump and the torque provided by a motor to the shaft of the pump. In the following, the disclosure is described with reference to the flowchart of FIG. 10.

(16) FIG. 10 shows a flow chart for measuring flow rate in accordance with an exemplary embodiment of the present disclosure. The measurement can be carried out using a non-intrusive measurement device. This kind of device can be used for measuring the flow rate in a pipe outside of the pipe without disturbing the process itself. The measurement device may be a portable ultrasonic flow meter, for example.

(17) At the same time as the flow rate is measured, the torque T.sub.est provided by the motor and rotational speed n.sub.est of the pump are estimated (101) and stored using the frequency converter. The measurement time should be sufficient, for example five seconds per measurement, so true mean values for the measurements and estimates can be determined. Also, no significant change in the rotational speed or valve positions should occur in this period of time. The determined mean values are later referred to as T.sub.Mean, n.sub.Mean, and Q.sub.VMean. In practice, a longer measurement or several measurements could be carried out, from which a time period with applicable mean values for torque T.sub.Mean, rotational speed n.sub.Mean and flow rate Q.sub.VMean could be selected to determine the QT curve inaccuracy.

(18) For synchronizing the measurement and the estimates, time labels for T.sub.Mean, n.sub.Mean and Q.sub.VMean are stored. After the measurement, flow rates can be transferred to the frequency converter manually by inputting the measurement results via a keyboard or automatically by means of a communications interface in order to correct the characteristics curves of the pump.

(19) The manufacturer gives the pump characteristic curves for a specific rotational speed. These curves are converted (102) to the mean rotational speed acquired in the measurement section. The curves can be converted using the generally known affinity laws

(20) $\begin{array}{cc}{Q}_v=\frac{n}{n_0}{Q}_{v0}& \left(1\right)\\ H={\left(\frac{n}{n_0}\right)}^2{H}_0& \left(2\right)\\ T={\left(\frac{n}{n_0}\right)}^2{T}_0& \left(3\right)\end{array}$
where Q.sub.v, is the flow rate, H is the pump head, T is the pump torque, n is the instantaneous rotational speed (e.g. n.sub.Mean), and the subscript 0 denotes the initial values of pump characteristic curves. Thus, n.sub.0 denotes the nominal rotational speed of the pump.

(21) When the flow rate produced by the pump has been measured, and the estimates for the rotational speed and shaft torque are known for the same time instant or period of the flow rate measurement, accuracy of the pump characteristic curves can be determined. By utilizing the characteristic curves converted for the instantaneous rotational speed, the torque T.sub.Manufacturer corresponding to the mean value of the measured flow rate Q.sub.VMean can be determined. This is compared with the measured value of torque T.sub.Mean in order to determine the bias torque T.sub.Bias for the characteristic curve
T.sub.Bias=T.sub.MeanT.sub.Manufacturer.(4)

(22) This bias torque T.sub.Bias is then used to improve the estimation from torque to flow either by manipulating T.sub.est used for the estimation of the pump flow rate, or by manipulating the published QT curve. In the exemplary embodiment, T.sub.est is manipulated with T.sub.Bias.

(23) During normal operation, the frequency converter gives estimates for the torque T.sub.est. According to the exemplary embodiment, this value is adjusted (103) so that it takes into account the bias in the manufacturer's characteristic curves. Instead of using T.sub.est for the flow rate estimation, variable T.sub.UseForQEstimation is applied for this purpose (104), (105), which can be calculated for instance with

(24) $\begin{array}{cc}{T}_{\mathrm{UseForQEstimation}}=T_{\mathrm{est}}-{T_{\mathrm{Bias}}\left(\frac{n_{\mathrm{est}}}{n_{\mathrm{Mean}}}\right)}^2& \left(5\right)\end{array}$

(25) where n.sub.est is the instantaneous rotational speed given by the frequency converter and n.sub.Mean is the rotational speed where T.sub.bias has been evaluated. In other words, the torque used for the estimation is scaled to the right rotational speed level instead of the curves.

(26) If the flow rate measurement can be carried out for several (i.e., two or more) operating points of the pump, for instance a linear correction can be performed for the pump characteristic curve or for the estimated torque T.sub.est. In an exemplary embodiment of the present disclosure, the linear correction is based on determining the equation for T.sub.Bias as a function of T.sub.est, and by applying it to (5). In the following exemplary equations, T.sub.Bias1 and T.sub.Bias2 have been determined at the mean torques T.sub.Mean1 and T.sub.Mean2, and T.sub.Mean2 is assumed to be larger than T.sub.Mean1

(27) $\begin{array}{cc}{T}_{\mathrm{Bias}}=\frac{T_{\mathrm{Bias}2}-T_{\mathrm{Bias}1}}{T_{\mathrm{Mean}2}-T_{\mathrm{Mean}1}}.Math.\left(T_{\mathrm{est}}-T_{\mathrm{Mean}1}\right)+T_{\mathrm{Bias}1}& \left(6\right)\end{array}$
Equation (6) is basically a first order function for a straight line going through points T.sub.Bias1 and T.sub.Bias2 as T.sub.est changes.

(28) Alternatively, several determined values of T.sub.Bias (i.e., T.sub.Bias1, T.sub.Bias2, . . . T.sub.Biasn) can be applied to calculate a mean value that is used for the correction of T.sub.est

(29) $\begin{array}{cc}{T}_{\mathrm{Bias}}=\frac{1}{n}\underset{i=1}{\overset{n}{.Math.}}{T}_{\mathrm{Bias}n}& \left(7\right)\end{array}$

(30) Besides the manipulation of T.sub.est, T.sub.Bias can be used to adjust the published QT curve of the pump, which initial values are denoted with the subscript 0. Magnitude of T.sub.Bias can be determined similarly as explained above for the exemplary embodiment [equations (4), (6), and (7)]. When T.sub.Bias is known at a certain rotational speed n.sub.Mean, corrected values for the shaft torque of the QT curve (T.sub.0) can be defined with:

(31) $\begin{array}{cc}{T}_0^{}=T_0+T_{\mathrm{Bias}}.Math.{\left(\frac{n_0}{n_{\mathrm{Mean}}}\right)}^2& \left(8\right)\end{array}$

(32) As mentioned, the relationship between the shaft torque and the produced flow rate is presented with a QT curve (FIG. 1). The pump manufacturer can give the characteristic curves only at the nominal rotational speed of the pump. If the pump is operated at a different rotational speed, the QT curve must be converted to the present speed using affinity laws before the flow rate estimation. The affinity laws for torque and flow rate are given in the equations (1) and (3). After this conversion, the flow rate of the pump can be determined from the QT curve. In the case of the exemplary embodiment of the disclosure, a corrected torque value (T.sub.UseForQEstimation) is used to determine the pump flow rate Q.sub.v.

(33) The method was evaluated by utilizing data collected with laboratory measurements. Laboratory measurements were conducted with a Sulzer APP 22-80 centrifugal pump, an 11 kW induction motor, and an ABB ACS 800 series frequency converter. The pump has a radial-flow impeller with a 255 mm impeller, and the internal clearance between the impeller and suction side of the pump can be adjusted without opening the pump. Motor and pump are connected to each other by a Dataflex 22/100 speed and torque measurement shaft which has a torque measurement accuracy of 1 Nm. The pump operating point location was determined with Wika absolute pressure sensors for the head and a pressure difference sensor across the venture tube, which equals the pump flow rate. In addition, a portable ultrasonic flow meter (Omega FD613) was used in the measurements, and its accuracy was verified with a separate measurement sequence.

(34) The pump is located in a process, which includes (e.g., consists of) two water containers, valves, and alternative pipe lines. The shape of the characteristic curve of the process and the resulting operating point location can be modified by adjusting the valves in the pipe lines.

(35) Firstly, characteristic curves of the Sulzer pump were determined for the QT calculation function: since the characteristic curves have been published only for 250 and 266 mm impeller diameters, the QH and QT characteristic curves of the pump had to be determined from the 250 mm curves by affinity equations for the impeller having a constant geometry (Q.sub.vd.sup.3, Hd.sup.2, Td.sup.4)

(36) $\begin{array}{cc}{Q}_V={\left(\frac{d}{d_0}\right)}^3.Math.Q_{V0}& \left(9\right)\\ H={\left(\frac{d}{d_0}\right)}^2{H}_0& \left(10\right)\\ T={\left(\frac{d}{d_0}\right)}^4{T}_0& \left(11\right)\end{array}$

(37) Then, a measurement series was conducted to determine the accuracy of the published pump characteristic curves. The pump was driven at a 1450 rpm rotational speed, and the operating point location was modified with the valves. The pump head, flow rate, rotational speed, and shaft torque were measured in 13 different operating point locations. In addition, the estimates calculated by the frequency converter for the rotational speed and shaft torque were stored for each operating point location. The pump had the same clearance setting as new, meaning that the internal clearances of the pump were in the factory settings during this measurement series. FIG. 2 illustrates published and measured characteristic curves of a pump in accordance with a known embodiment. In FIG. 2, the measurement results are illustrated together with the published characteristics of the pump. In the example results of FIG. 2, the published curves are the lower ones and the measured the upper ones.

(38) As shown in FIG. 2 notable differences between the measured and published characteristics for the pump torque specification as a function of flow rate are notable. In practice, the error of the published torque curve leads to erroneous estimation results for the pump flow rate. For instance, at the shaft torque of 38 Nm, the published characteristic curve would result in a flow rate of 24.7 l/s, when the actual flow rate is 19.3 l/s. In practice, the estimation error for the flow rate is at least 5.4 l/s (19% of the pump's best efficiency flow rate Q.sub.VBEP 28 l/s at 1450 rpm).

(39) To ensure the accuracy of the estimated shaft torque values calculated by the frequency converter, a comparison was carried out for the published, measured, and estimated shaft torques of the pump. The results are shown in FIG. 3.

(40) FIG. 3 shows published, measured, and estimated values of the shaft torque as function of the pump flow rate in accordance with an exemplary embodiment of the present disclosure. The difference of the published and estimated shaft torques is 4.3-6.5 Nm depending on the pump flow rate. On the other hand, the difference between the measured and estimated shaft torques is less than 1 Nm, stating the applicability of the shaft torque estimates in the flow calculation. In the upper plot of FIG. 3, the published values are the leftmost bars, the measured values are the bars in the middle and the estimated values are presented as the rightmost bars at each flow rate.

(41) From the results shown in FIG. 3, a torque difference at the flow rate of 26 l/s was selected as the bias value for the shaft torque. Hence T.sub.Bias=4.86 Nm, which is applied to correct the estimated values of the shaft torque by applying (4). After this, QT estimation was carried out with T.sub.Bias correction for T.sub.est and also without corrections.

(42) FIG. 4 shows estimated and measured flow rates in accordance with an exemplary embodiment of the present disclosure. The resulting flow rates for each measurement are illustrated in FIG. 4. As expected, estimation results can be erroneous without the T.sub.Bias correction. The use of T.sub.Bias for the shaft torque estimates has improved the estimation accuracy of the flow rate. Consequently, the improvement in the estimation accuracy of the pump flow rate also improves the estimation accuracy of the pump head. In FIG. 4, the estimated values without correction are the leftmost bars, the rightmost bars are the estimated values with the correction, and the middle ones are the measured results.

(43) FIG. 5 shows estimation errors relating to the flow rates of FIG. 4 in accordance with an exemplary embodiment of the present disclosure. Estimation errors were determined for both T.sub.Bias cases, and they are shown in FIG. 5. Without the T.sub.Bias correction (leftmost, longer bars), the estimation error of the flow rate is within, for example 26.2 l/s (94% of Q.sub.VBEP) This means that the QT calculation can give unrealistic estimates for the flow rate. With the T.sub.Bias correction (rightmost, shorter bars), the estimation error of the flow rate is within, for example, 7.3 l/s, that is, approximately 26% of Q.sub.VBEP. It should be noted that the estimation accuracy with T.sub.Bias correction is within, for example, 1.2 l/s, when the flow rate ranges, for example, from 11 l/s to 31 l/s (39-111% of Q.sub.VBEP). It can be assumed that the pump is mostly operating within this operating region.

(44) Laboratory measurements were also carried out by changing the rotational speed of the pump, while the valve settings affecting the process curve were untouched. In the first measurement series, the relative flow rate at nominal speed was 100% of the best efficiency flow rate 28 l/s, so the pump was operating at its best efficiency point. Measurements were carried out at rotational speeds ranging, for example, from 1020 rpm to 1620 rpm which correspond to 70% and 112% of the pump's nominal rotational speed, respectively.

(45) FIG. 6 shows estimated and measured flow rates near the pump's best efficiency point in accordance with an exemplary embodiment of the present disclosure. The results of the estimations can be seen in FIG. 6. The estimation without the T.sub.Bias correction gives too high estimates for the flow rate (leftmost bars). The estimation with T.sub.Bias (rightmost bars) gives again significantly more accurate results compared with the estimations without T.sub.Bias. The measured flow rates are shown as the middle bars in each set of results.

(46) FIG. 7 shows estimation errors relating to the flow rates of FIG. 6 in accordance with an exemplary embodiment of the present disclosure. The errors of the estimation methods can be seen in FIG. 7. The estimation error without the T.sub.Bias correction (longer bars) is within, for example, 17.3 l/s (62% of Q.sub.VBEP). With T.sub.Bias, the estimation error (shorter bars) is within, for example, 2.7 l/s (10% of Q.sub.VBEP). This indicates that the estimation with T.sub.Bias improves the accuracy of QT calculation method significantly. It should also be noted that the estimation error of the T.sub.Bias-improved QT calculation is within, for example, 1 l/s at the rotational speeds over 1140 rpm (i.e., flow rate is over 21 l/s in FIG. 7).

(47) In another exemplary embodiment, a measurement series was carried out at the relative flow rate of 70%. FIG. 8 shows estimated and measured flow rates when the pump was driven at a relative flow rate of 70% in accordance with an exemplary embodiment of the present disclosure. As shown in FIG. 8, the T.sub.Bias correction improves once again the estimation accuracy of the flow rate. Flow rates estimated with the correction are shown in the rightmost bars, measured flow rates are in the middle and estimated flow rates without the correction are the leftmost bars.

(48) FIG. 9 shows estimation errors relating to the flow rates of FIG. 8 in accordance with an exemplary embodiment of the present disclosure. As shown in FIG. 9, the estimation error without the T.sub.Bias correction (longer bars) is within, for example, 6.6 l/s (24% of Q.sub.VBEP). With T.sub.Bias the estimation error (shorter bars) is within, for example, 2.3 l/s (8% of Q.sub.VBEP).

(49) The conducted measurements indicate that the estimation with T.sub.Bias improves the QT calculation method significantly, and the T.sub.Bias correction works both with static-speed (FIGS. 4 and 5) and variable-speed pumping applications (FIGS. 6 to 9).

(50) Thus, it will be appreciated by those skilled in the art that the present invention can be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The presently disclosed embodiments are therefore considered in all respects to be illustrative and not restricted. The scope of the invention is indicated by the appended claims rather than the foregoing description and all changes that come within the meaning and range and equivalence thereof are intended to be embraced therein.