A method for judging abnormal operation status of water pump based on efficiency analysis

Abstract

The present invention discloses a method for distinguishing abnormal operation status of a water pump based on efficiency analysis. According to the collected current and shaft power data in the operation status of the water pump, the correlation between the current and the shaft power is analyzed, and the relationship between the current and the shaft power is converted into the relationship between the flow rate and the efficiency, so as to illustrate the possible relationship between the current and the efficiency according to the water pump efficiency curve. Subsequently, polynomial regression is performed based on the collected data, and statistical analysis is performed. Finally, it is judged whether the operation of the water pump is abnormal by the 3σ rule of the residual normal distribution, and it is judged whether the water pump has potential faults based on this.

Description

Technical Field

[0001] The present invention relates to a water pump operation state abnormality discrimination technology based on efficiency analysis, and in particular to a water pump operation state abnormality discrimination method based on efficiency analysis. Background Art

[0002] Water pumps are widely and importantly used in industrial production. Once a water pump fails, it will not only cause the equipment to stop operating, but also affect the previous and subsequent production processes, significantly reducing the company's profits. Existing companies usually use post-corrective repairs or regular preventive maintenance to repair equipment, which may indulge in failures or generate unnecessary maintenance costs. If the possible failure of the water pump in the future can be judged and analyzed based on the status information of the water pump before the water pump fails, so as to perform maintenance and repair in advance, then the downtime can be greatly reduced, thereby ensuring the normal production of the entire production line and thus ensuring the company's profits. Summary of the invention

[0003] In order to solve the above problems, the present invention provides a method for distinguishing abnormal operation status of a water pump based on efficiency analysis. According to the collected current and shaft power data in the operation status of the water pump, the correlation between the current and the shaft power is analyzed, and the relationship between the current and the shaft power is converted into a relationship between flow and efficiency, so as to illustrate the possible relationship between the current and the efficiency according to the water pump efficiency curve. Subsequently, polynomial regression is performed based on the collected data, and statistical analysis is performed. Finally, it is judged whether there is abnormality in the operation of the water pump by the 3σ law of the residual normal distribution, thereby judging whether there is a potential fault in the water pump.

[0004] A method for determining abnormal operation status of a water pump based on efficiency analysis in the present invention comprises the following steps:

[0005] Step S1: obtaining two state data of the instantaneous current during the operation of the water pump and the water pump shaft power at the corresponding moment through a sensor installed on the water pump;

[0006] Step S2: Analyze the mechanism of the water pump and obtain the performance curve fitting type;

[0007] Step S3: fitting the collected state monitoring data using a polynomial;

[0008] Step S4: Use the normal distribution 3σ criterion to perform abnormal point analysis and detection of the operating status.

[0009] In the above scheme, step S2 includes the following steps:

[0010] Step S21: Calculate the output power of the water pump motor under the instantaneous current in step S1. The calculation formula is P is the output power of the motor, U is the line voltage, and I represents the current flowing through the motor. is the power factor, which is determined by the type of load;

[0011] Step S22: The efficiency of the water pump reaches its maximum value when it is close to the rated power, and its efficiency is lower when it is lower than the rated power or higher than the rated power. After considering the efficiency of the water pump itself, the formula in step S21 is rewritten as P a represents the shaft power of the water pump, U is the line voltage, I is the current flowing through the motor, is the power factor, η represents the efficiency of the pump itself;

[0012] Step S23: Taking the motor current as the independent variable, the total efficiency of the water pump is calculated as the dependent variable according to the equation on the left side of the formula in step S22, and the distribution of the water pump efficiency η-current I of different water pump motors at different times can be obtained.

[0013] In the above scheme, in step S21, U is 380V when used for industrial electricity. is 0.8.

[0014] In the above scheme, step S3 includes:

[0015] Step S31: Use Matlab toolbox to fit the data with a cubic polynomial. The fitting criterion is the least squares fitting. The calculation target of the least squares method is N is the number of samples, f(x i ) is the regression value of the fitted function, y i is the actual value, so as to find the cubic curve that minimizes the sum of squared residuals;

[0016] Step S32: Calculate the regression statistical index determination coefficient R 2 And the sample standard deviation σ of the residual is calculated as R 2 represents the coefficient of determination, y represents the actual value of the sample, represents the regression function value, represents the mean of the sample values, σ represents the sample standard deviation of the residual, and n represents the number of samples.

[0017] In the above scheme, the random error of the polynomial linear regression in step S4 satisfies N(0, σ 2) distribution, that is, a normal distribution with a mean of 0 and a standard deviation of σ. The outliers can be tested according to the 3σ rule of the normal distribution, including the outliers in the training samples and the outliers in the new input data. After the regression curve is established, the original input data is tested. If the difference with the fitting curve exceeds 3σ, the data is considered to be outside the confidence interval of the random error and is an outlier. Subsequently, the outliers need to be removed and refitted until there are no outliers in the training data. At this time, the regression model has been established.

[0018] In the above scheme, after the regression model is established in step S4, if there is new data, it is input into the model for detection. If the difference with the regression value is greater than 3σ, it is considered that it is not caused by random error but an outlier.

[0019] The advantages and beneficial effects of the present invention are:

[0020] 1. From the perspective of water pump efficiency analysis, the present invention analyzes the intrinsic relationship between the two parameters of current and power, finds the performance curve of the water pump as a bridge in the fitting process, and thus determines the fitting method and the degree of the fitting polynomial;

[0021] 2. The present invention can establish an abnormality detection model for a water pump from sample data with a small data dimension and a small amount of data, and can also address the problem of a small amount of fault data in the collected samples;

[0022] 3. The present invention can establish a one-to-one corresponding abnormality detection model for each water pump;

[0023] 4. The present invention can quickly and conveniently establish a water pump operation abnormality detection model and has strong engineering practice capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0025] Figure 1 It is the polynomial fitting result of the water pump performance curve in the specific implementation method.

[0026] Figure 2 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0027] The specific implementation of the present invention is further described below in conjunction with the accompanying drawings and embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0028] like Figure 1 and Figure 2 As shown, the present invention is a method for determining abnormal operation status of a water pump based on efficiency analysis, comprising the following steps:

[0029] Step S1: obtaining two state data of the instantaneous current during the operation of the water pump and the water pump shaft power at the corresponding moment through a sensor installed on the water pump motor;

[0030] Specifically, this embodiment collects data on the motor current and shaft power of a water pump during the actual operation of an energy station. The water pump application types include primary pumps, secondary pumps, etc. Taking a secondary pump No. 02 cooling pump as an example, its status data at the time of operation is collected.

[0031] Step S2: Analyze the mechanism of the water pump and obtain the performance curve fitting type;

[0032] Calculate the output power of the water pump motor under the instantaneous current in step S1, and the calculation formula is:

[0033]

[0034] Among them, P is the output power of the motor, U is the line voltage, which is 380V for industrial electricity, and I represents the current flowing through the motor. It is the power factor, which is determined by the load type. For the water pump motors used in engineering practice, an empirical value of 0.8 is generally taken.

[0035] Theoretically, the shaft power of the water pump should be proportional to the current, but the efficiency of the water pump is different in different working conditions. From the water pump performance curve, it can be seen that the efficiency of the water pump reaches its maximum value when it is close to the rated power, and its efficiency is lower when it is lower than the rated power and higher than the rated power.

[0036] Although the performance curves of most water pumps are not exactly the same, the trends of the curves are consistent. Therefore, after considering the efficiency of the water pump itself, formula (1) can be rewritten as

[0037]

[0038] Among them, P a represents the shaft power of the water pump, U is the line voltage, I is the current flowing through the motor, is the power factor, and η represents the efficiency of the pump itself.

[0039] When the motor current is large, the pump power is large, and the pump can pump more liquid to do work, so there is a positive correlation between the flow rate and the motor current. At this time, the change trend of the efficiency η-flow rate Q curve, one of the performance curves of the pump, can also be considered as the change trend of the efficiency η-current I. Therefore, if the motor current is used as the independent variable and the total efficiency of the pump is calculated as the dependent variable according to the equation on the left side of formula (2), the distribution of the pump efficiency η-current I of different pump motors at different times can be obtained.

[0040] At the same time, in Zhao Linming's document "Analytical Expression Method of Water Pump Performance Curve" in the journal "Farmland Water Conservancy and Small Hydropower", the author successfully used cubic polynomials to fit the performance curves of two models of water pumps, and found that using cubic curves to fit the performance curves of water pumps, especially the relationship between efficiency and flow, can achieve better results. Therefore, based on the characteristic that the efficiency in the performance curve is positively correlated with the flow, the present invention also uses cubic polynomials to fit the data points.

[0041] Step S3: fitting the collected state monitoring data using a polynomial;

[0042] The Matlab toolbox is used to fit the data to a cubic polynomial. The fitting criterion is the least squares fitting. The calculation objectives of the least squares method are as follows:

[0043]

[0044] Where N is the number of samples, f(x i ) is the regression value of the fitted function, y i is the actual value, so as to find the cubic curve that minimizes the sum of squared residuals;

[0045] Then calculate the regression statistical indicator determination coefficient R 2 And the sample standard deviation σ of the residual is calculated as

[0046]

[0047]

[0048] In the formula, R 2 represents the coefficient of determination, y represents the actual value of the sample, represents the regression function value, represents the mean of the sample values, σ represents the sample standard deviation of the residual, and n represents the number of samples.

[0049] The Matlab toolbox is used for fitting, and the fitting results are as follows Figure 1As shown, the vertical axis represents the overall efficiency of the pump multiplied by the efficiency of each link, and the horizontal axis represents the current of the motor. According to 3σ as the judgment criterion, the points with a difference of more than 3σ from the fitting curve are deleted as abnormal points and retrained in a cycle. After all the training is completed (a total of two regression fittings are performed in the present invention), the model fitting finally obtained the polynomial: -5.013×10 -8 x 3 +3.382×10 -5 x 2 +0.0094x-0.6216, the sample standard deviation of the residual is 0.0463, and the coefficient of determination R 2 =0.9707, which means that the regression model can represent 97% of the data variation.

[0050] Step S4: Use the normal distribution 3σ criterion to perform abnormal point analysis and detection of the operating status;

[0051] Theoretically, the water pump efficiency calculated by formula (2) from the actual measured data should fall on the water pump efficiency regression curve. However, in the actual production process, due to water pump equipment failures such as unstable voltage, bearing damage, rotor dynamic imbalance and inaccurate measurement factors, the calculated efficiency will deviate from the ideal regression curve. Since the efficiency of the water pump is calculated from the current and voltage values, excessive fluctuations in efficiency can be considered to be caused by the mismatch between the current current and power values. At this time, the water pump equipment must have experienced abnormal operation or a part of the equipment has failed. Therefore, it is necessary to determine whether the fluctuation exceeds the specified range through calculation.

[0052] According to mathematical statistics, the random error of polynomial linear regression satisfies N(0, σ 2 ) distribution, that is, a normal distribution with a mean of 0 and a standard deviation of σ. At the same time, the errors here can be considered to be caused by random, countless, independent, and multiple factors. Therefore, the outliers can be tested according to the 3σ rule of the normal distribution, including outliers in the training samples and outliers in the newly input data. After the regression curve is established, the previously input data is tested. If the difference with the fitting curve exceeds 3σ, the data is considered to be outside the confidence interval of the random error and is an outlier. Subsequently, the outliers need to be removed and refitted until there are no outliers in the training data. At this point, the regression model has been established. If there is new data, it can be input into the model for detection. If the difference with the regression value is greater than 3σ, it can be considered that it is not caused by random errors, but outliers. In the present invention, an outlier may represent a fault in the equipment, or it may be caused by non-fault reasons such as excessive starting current when the pump motor is started.

[0053] When new test data is re-entered, it is also judged according to the 3σ criterion. If the data is considered an abnormal point by the model, it outputs 1, and if it is considered to be within the normal range, it outputs 0. In this model, 7 new time data are input for testing, and the result is that 6 data are within the normal range, and one data is considered an abnormal point by the regression model, which may be caused by a fault in the equipment.

[0054] The present invention analyzes the correlation between current and shaft power based on the collected current and shaft power data in the running state of the water pump, and converts the relationship between current and shaft power into the relationship between flow and efficiency, thereby illustrating the possible relationship between current and efficiency based on the water pump efficiency curve. Subsequently, polynomial regression is performed based on the collected data, and statistical analysis is performed. Finally, the residual normal distribution 3σ rule is used to determine whether the water pump operation is abnormal, and based on this, it is determined whether the water pump has potential faults. Its advantages are:

[0055] 1. From the perspective of water pump efficiency analysis, the present invention analyzes the intrinsic relationship between the two parameters of current and power, finds the performance curve of the water pump as a bridge in the fitting process, and thus determines the fitting method and the degree of the fitting polynomial;

[0056] 2. The present invention can establish an abnormality detection model for a water pump from sample data with a small data dimension and a small amount of data, and can also address the problem of a small amount of fault data in the collected samples;

[0057] 3. The present invention can establish a one-to-one corresponding abnormality detection model for each water pump;

[0058] 4. The present invention can quickly and conveniently establish a water pump operation abnormality detection model and has strong engineering practice capabilities.

[0059] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for determining abnormal operation status of a water pump based on performance analysis, characterized in that: The steps include: Step S1: obtaining two state data of the instantaneous current during the operation of the water pump and the water pump shaft power at the corresponding moment through a sensor installed on the water pump; Step S2: Analyze the mechanism of the water pump and obtain the performance curve fitting type; The step S2 comprises the following steps: Step S21: Calculate the output power of the water pump motor under the instantaneous current in step S1. The calculation formula is P is the output power of the motor, U is the line voltage, and I represents the current flowing through the motor. is the power factor, which is determined by the type of load; Step S22: The efficiency of the water pump reaches its maximum value when it is close to the rated power, and its efficiency is lower when it is lower than the rated power or higher than the rated power. After considering the efficiency of the water pump itself, the formula in step S21 is rewritten as P a represents the shaft power of the water pump, U is the line voltage, I is the current flowing through the motor, is the power factor, η represents the efficiency of the pump itself; Step S23: taking the motor current as the independent variable, and calculating the total efficiency of the water pump as the dependent variable according to the equation on the left side of the formula in step S22, the distribution of the water pump efficiency η-current I of different water pump motors at different times can be obtained; Step S3: fitting the collected state monitoring data using a polynomial; The step S3 comprises: Step S31: Use Matlab toolbox to fit the data with a cubic polynomial. The fitting criterion is the least squares fitting, and the calculation target of the least squares method is min. N is the number of samples, f(x i ) is the regression value of the fitted function, y i is the actual value, so as to find the cubic curve that minimizes the sum of squared residuals; Step S32: Calculate the regression statistical index determination coefficient R 2 And the sample standard deviation σ of the residual is calculated as R 2 represents the coefficient of determination, y represents the actual value of the sample, represents the regression function value, represents the mean of the sample values, σ represents the sample standard deviation of the residual, and n represents the number of samples; Step S4: Use the normal distribution 3σ criterion to perform abnormal point analysis and detection of the operating status.

2. A method for determining abnormal operation status of a water pump based on efficiency analysis according to claim 1, characterized in that: In step S21, U is 380V when used for industrial electricity. is 0.

8.

3. The method for determining abnormal operation status of a water pump based on efficiency analysis according to claim 1 is characterized in that: The random error of the polynomial linear regression in step S4 satisfies N(0, σ 2 ) distribution, that is, a normal distribution with a mean of 0 and a standard deviation of σ. The outliers can be tested according to the 3σ rule of the normal distribution, including the outliers in the training samples and the outliers in the new input data. After the regression curve is established, the original input data is tested. If the difference with the fitting curve exceeds 3σ, the data is considered to be outside the confidence interval of the random error and is an outlier. Subsequently, the outliers need to be removed and refitted until there are no outliers in the training data. At this time, the regression model has been established.

4. A method for determining abnormal operation status of a water pump based on efficiency analysis according to claim 3, characterized in that: After the regression model is established in step S4, if there is new data, it is input into the model for detection. If the difference with the regression value is greater than 3σ, it is considered that it is not caused by random error but an abnormal point.

Images