Motor energy consumption and carbon emission comprehensive monitoring and analysis method based on equipment state recognition

Abstract

The invention relates to the technical field of motor running state detection, in particular to a motor energy consumption and carbon emission comprehensive monitoring and analysis method based on equipment state recognition. The method comprises the steps that 1, data motor running original data are collected, and the collected running original data comprise the motor rotating speed, a rotating shaft, vibration data and carbon emission factors; step 2, performing missing value complementation on the collected original operation data, and calculating feature correlation among the original operation data to obtain a new data set; and 3, inputting the new data set obtained in the step 2 into the classification decision tree model, and correcting a classification result through an information gain correction value. According to the method, the motor vibration data is combined with the decision tree model and the Pearson's correlation coefficient, and interpolation expansion is carried out after Fourier transform, so that the signal integrity is ensured, and higher precision is provided in energy consumption monitoring and carbon emission calculation.

Description

Technical Field

[0001] The present application relates to the technical field of motor operation status detection, and in particular to a comprehensive monitoring and analysis method for motor energy consumption and carbon emissions based on equipment status identification. Background Art

[0002] Fans, motors, pumps, and compressors are collectively referred to as "industrial motor systems" by the International Energy Agency (IEA). Motors are widely used in modern industry, and industrial motor systems are the main electricity users in China, accounting for more than 50% of all electricity consumption. Therefore, improving motor efficiency and monitoring energy consumption and carbon emissions are of great significance to saving electricity.

[0003] Most existing motor operating status monitoring methods rely on electrical parameters such as current and voltage to monitor energy consumption, which makes it difficult to make a quantitative evaluation of the carbon emissions of new electrical equipment. Summary of the invention

[0004] The technical problem to be solved by the present invention is that the existing motor monitoring accuracy is low and it is difficult to quantitatively evaluate the carbon emissions.

[0005] To this end, the present invention provides a method for comprehensive monitoring and analysis of motor energy consumption and carbon emissions based on equipment status identification.

[0006] The technical solution adopted by the present invention to solve its technical problem is: A comprehensive monitoring and analysis method for motor energy consumption and carbon emissions based on equipment status recognition, comprising: Step 1: Collect the original data of the motor operation, including the motor speed, shaft, vibration data, and carbon emission factor; Step 2: Fill missing values ​​in the collected original running data and calculate the feature correlation between the original running data to obtain a new data set; Step three, input the new data set obtained in step two into the classification decision tree model, and correct the classification result by the information gain correction value.

[0007] Furthermore, in the step 2, missing values ​​of the collected original operation data are filled based on the distance, and then Fourier transform is introduced to process the vibration data, and then the frequency domain data after Fourier transform is expanded using the interpolation method.

[0008] Furthermore, in step 2, the Pearson correlation coefficient is used to evaluate the correlation between each feature and the target variable, and a loss function is constructed based on the correlation between the feature and the target variable. The loss function is reduced by a gradient descent algorithm, that is, the added value is made closer to the true value.

[0009] Furthermore, the loss function is expressed as: ,in is the mapping result of the correlation coefficient of the original data i, is the mapping result of the correlation coefficient of the increasing value, and n is the number of features.

[0010] Furthermore, in the process of using the Pearson correlation coefficient to evaluate the correlation between each feature and the target variable, the distance between each column and the label column is added as an additional parameter to evaluate the correlation of the features.

[0011] Further, , , where Pnew and Porgin are the results of mapping the new and old data through the Pearson correlation coefficient, and d is the distance between each column and the label column as an additional parameter. .

[0012] Furthermore, in step three, the information gain correction value is the product of the information gain and the Pearson correlation coefficient function f(r), and the Pearson correlation coefficient function f(r) converts the Pearson correlation coefficient into a positive number and improves the hierarchical nature of the strength of the interval where the Pearson correlation coefficient is located.

[0013] Further, the Pearson correlation coefficient function is expressed as: .

[0014] Furthermore, the method further includes a step 4, in which the power P of the motor is calculated according to the torque and operating time of the motor, and the carbon emission E of the motor during operation is calculated according to a predetermined carbon emission factor.

[0015] Furthermore, the method further includes step five, in which the operating state of the motor is evaluated based on the classification result obtained in step three and the real-time operating state of the motor.

[0016] The beneficial effect of the present invention is that the present invention discloses a method for energy consumption monitoring and carbon emission calculation based on motor vibration data. By using vibration data and advanced data analysis technology, the energy consumption of the motor under different working conditions can be accurately monitored, and carbon emissions can be automatically estimated. The core is to infer the speed of the motor through the motor vibration data, and then calculate the motor power in combination with the torque and running time, and then determine the energy consumption and carbon emissions. The present invention adopts a decision tree algorithm combined with the Pearson correlation coefficient to further improve the accuracy of data analysis and classification, and provide a scientific basis for energy consumption monitoring and carbon emission calculation.

[0017] First, the vibration data of the motor is collected in real time by a vibration sensor, and the speed change of the motor is deduced based on characteristic data such as vibration frequency. In order to ensure the integrity and accuracy of the vibration signal data, the present invention first uses Fourier transform to convert the vibration signal from the time domain to the frequency domain before analyzing the vibration data, and interpolates and expands the frequency domain signal to complete the subtle changes in the data. This process ensures the complete expression of the signal frequency component, effectively prevents the pseudo-frequency that may be introduced by direct interpolation in the time domain, and ensures the continuity and accuracy of the data after expansion. The present invention introduces vibration data, which can more comprehensively and accurately reflect the actual operating state of the motor, especially in a complex environment where the load fluctuates greatly and the equipment frequently switches working conditions. The vibration data can capture the subtle changes of the motor and significantly improve the monitoring ability of the operating state. And the Pearson correlation coefficient and distance parameter are used to evaluate the correlation between the supplemented data and the original data, optimize the correlation between the supplemented data and the original data in the supplemented data set, and improve the accuracy of the classification structure.

[0018] After obtaining the real-time speed of the motor, the real-time power output of the motor is calculated using a decision tree model by combining the speed, torque and running time. Compared with the existing method of directly measuring power, the present invention uses a decision tree to classify the motor state and combines the Pearson correlation coefficient of each feature when constructing the decision tree, thereby improving the classification accuracy. Specifically, by calculating the information gain of each feature and multiplying it with the corresponding Pearson correlation coefficient function, the sensitivity to changes in the motor state is enhanced, thereby more accurately capturing the details of the motor operation and avoiding the problem of electrical data lag or error in direct power measurement.

[0019] Based on power calculation, the present invention calculates the energy consumption of the motor in different states by accumulating time. Then, according to the pre-set carbon emission factor, the energy consumption of the motor is converted into carbon emissions, thereby helping users to quantitatively evaluate the carbon footprint of the motor and provide a basis for energy conservation and emission reduction.

[0020] In summary, the present invention combines motor vibration data with a decision tree model and Pearson correlation coefficient, and interpolates and expands the data after Fourier transformation to ensure signal integrity, providing higher accuracy in energy consumption monitoring and carbon emission calculation. It is not only suitable for motor status monitoring under complex working conditions, but also provides strong quantitative support and data basis for equipment operation optimization, energy conservation and emission reduction. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0022] Figure 1 It is a schematic diagram of the comprehensive monitoring and analysis method of motor energy consumption and carbon emissions based on equipment status identification in the present invention.

[0023] Figure 2 It is a schematic diagram of reducing the loss function through the gradient descent algorithm in the present invention. DETAILED DESCRIPTION

[0024] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0025] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0026] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0027] A comprehensive monitoring and analysis method for motor energy consumption and carbon emissions based on equipment status recognition includes the following steps: Step 1: Vibration data collection The vibration data of the motor is collected by the vibration sensor as the basic data for motor status monitoring. It is constructed as the input of the model together with the motor speed, shaft, and carbon emission factor. The motor speed and shaft can be directly obtained from the motor's own parameters and sensors installed on the motor; the carbon emission factor can refer to the data in the grid benchmark carbon emission factor file.

[0028] Step 2: Data processing This step interpolates the data collected in step 1 and calculates the feature correlation. The correlation coefficient is used as the loss function when verifying the expanded data set. The interpolated data is dynamically optimized to form a new data set input classification model, while preparing for the expansion of the data set.

[0029] S2.1 Data preprocessing A distance-based missing value completion algorithm is used for input data to ensure data integrity. By introducing Fourier transform to process vibration data and then using interpolation to expand the frequency domain data after Fourier transform, the frequency characteristics of the data can be kept intact, while improving the accuracy of data analysis.

[0030] First, in order to eliminate the impact of dimensions between different columns of data, each column of data needs to be standardized, that is:

[0031] In the above formula, Z represents the result after standardized output, X represents the original data, μ represents the mean of this column of data, and σ represents the standard deviation of this column of data.

[0032] This allows you to map the data in each column to between [-1, 1] and calculate the distances between different rows in order to find the k closest rows:

[0033] In the formula, , Representing two data in different rows and columns respectively, the distance is calculated by subtracting the elements of the corresponding columns of the two rows and then squaring them.

[0034] For example, here k is taken as 5. For missing values, the corresponding column data of the k nearest rows are used to fill in to ensure the integrity of the data. By introducing Fourier transform to process the vibration data, and then using the interpolation method to expand the frequency domain data after Fourier transform, it can ensure that the frequency characteristics of the data are kept intact and improve the accuracy of data analysis.

[0035] The interpolation formula is:

[0036] After Fourier transform, the value of each column exists in the form of (x, y), where x represents the frequency point and y represents the corresponding complex result. For example, after Fourier transformation, the data (0.5, 1.0, 0.3, -0.7, 0.2) is calculated with a sampling frequency of 0.5 seconds. The corresponding x values ​​generated are (0, 0.4, 0.8, 1.2, 1.6), and the corresponding y values ​​are (1.3+0, 0.6-0.2, -0.1+0.5i, 0.4-0.3i, -0.5+0.1i).

[0037] When interpolation is required, first calculate x based on the supplementary range new For example, if there are 100 data points, corresponding to 0 to 99, and we want to add 120 data points, then the first and last data points will remain unchanged, and the difference between them will change from 1 to 99 / 119, so we can determine x. new , x1, x2, y1, y2 in the formula represent x new The values ​​of the points before and after the corresponding point.

[0038] S2.2 Calculate feature correlation and dynamically update the dataset Use the Pearson correlation coefficient to evaluate the correlation between each feature and the target variable, and calculate the correlation between each feature and the target variable. In order to ensure the rationality of the added data, we can calculate the mapping result of the Pearson correlation coefficient of the original data i , and the mapping results of the correlation coefficient of the added value , set the loss function to:

[0039] In this step, the base Pearson correlation coefficient is used to establish the mapping relationship between the original data and the added value. The mapping results of the new and old data through the Pearson correlation coefficient are Pnew and Porgin, respectively. The format is (p1, p2, ... pn), n is the number of features, and the distance between each column and the label column is added to the tail as an additional parameter, namely (d1, d2, ... dn), that is, Pnew and Porgin are changed to (p1+d1, p2+d2, ... pn+dn). In the loss function, the added value will be filtered out when the derivative is calculated, but it will be reflected when the new and old values ​​are subtracted and the change in the loss function is calculated, which will control the convergence of the loss function more accurately. The distance between each column and the label column is as follows:

[0040] Right now , .

[0041] Reference Figure 2 , the loss function is reduced by the gradient descent algorithm, that is, the increased value is closer to the true value.

[0042] The newly added data is recorded as a, and the speed data is recorded as b. The Pearson relationship coefficient between the original data and the speed is recorded as Porgin, and the relationship coefficient between the added data and the speed is recorded as Pnew. The loss function is reduced by the gradient descent algorithm, that is, the added value is made closer to the true value.

[0043]

[0044] In order to facilitate derivation, set the loss function L=(Pnew-Porgin) 2 The goal is to make the objective function as close to zero as possible by changing the newly added data column, that is, to minimize the loss function. Calculate the loss function L and take the derivative of the newly added data a:

[0045]

[0046] The value of a is changed by the gradient descent method, so as to update the loss function and make it converge to verify the correlation between the supplementary data and the original data.

[0047]

[0048] In the above formula, represents an element in data a, η is the learning rate, which is generally taken as 0.001 to 0.01.

[0049] Step 3: Decision tree model construction and training The completed and processed data obtained in step 2 is used to train a classification model based on a decision tree to classify the motor speed state. In the process of building the decision tree, information gain is used to correct the classification result of the decision tree.

[0050] When calculating the information gain of a feature, the Pearson correlation coefficient of the feature is combined, and the formula is as follows:

[0051] This method retains all features and adjusts the weights of the features according to their correlation with the target variable, so that features with higher correlation have greater influence in the decision tree splitting process, thereby improving classification efficiency and accuracy. In the process of building a decision tree, the calculation of information gain is used to measure the contribution of features to the classification task. When introducing the Pearson correlation coefficient, directly using its value (especially negative values) may cause the information gain to be reduced or even become negative. Since the decision tree needs to select features based on the maximum information gain, if the correlation coefficient of a feature is negative, this may cause the information gain to be unreasonably reduced, and then cause the decision tree to select the wrong feature for splitting. Once a feature is used for splitting, it will no longer be used for the division of subsequent nodes, so unreasonable selection will seriously affect the structure of the tree.

[0052]

[0053] In addition, the values ​​of the Pearson correlation coefficient fall into different intervals (such as [0,0.2], [0.2,0.4], etc.), which represent correlations of different strengths. Directly using the original values ​​for calculation will ignore the hierarchy between these intervals. To avoid this, we need to introduce a segmented mapping method to map the Pearson correlation coefficient to a positive range while retaining the hierarchy of its interval strength.

[0054]

[0055] Then the information gain correction value is expressed as .

[0056] For example, before modifying the mapping, the correlation coefficient of the x-axis displacement is -0.0537, and the result after calculating the information gain is 0, because it basically maintains 0 displacement in the low speed state, while the correlation coefficient of the y-axis displacement is -0.0451, and the result after calculating the information gain is 0.007; the actual situation of the motor is observed as follows: due to the influence of the external environment, the machine has a large amplitude up and down, and the y-axis has obvious displacement. When constructing the decision tree, due to the large information gain of the y-axis, the y-axis displacement will be selected as a node, which is due to external influences. The actual displacement of the y-axis should also be close to 0. Selecting the y-axis as a node will cause the normally collected y-axis 0 displacement data and the abnormal y-axis non-0 displacement data to be divided into two categories, which are actually the same category of states. After the mapping conversion, the information gain of the x-axis is 0.650, and the information gain of the y-axis is 0.6470. When the same node is divided, the x-axis displacement will be selected for division, and the x-axis data is not easily disturbed, that is, all 0 displacements will be classified into the same category.

[0057] Step 4: Calculation of motor power, energy consumption, and carbon emissions According to the motor torque and running time, calculate the motor power P. Among them, the torque of the test motor is 48N / m. The power calculation formula is:

[0058] The energy consumption of the motor is calculated based on the motor power P and running time t. Then, the carbon emissions E of the motor during operation are calculated based on the predetermined carbon emission factor. The formula is:

[0059] Step 5: Operation monitoring and evaluation According to the state categories output by the decision tree classification model, taking the motor speed state as an example, the state categories of the motor data output by the decision tree classification model are shutdown state, 500 speed, 1000 speed, and 1500 speed. The running state and power of the motor are monitored in real time, and the corresponding carbon emissions are recorded at the same time. The classification results are used to evaluate the running state of the motor to ensure that it is in a healthy state.

[0060] Based on the carbon emission monitoring data during the operation of the motor, the carbon emission level of the motor is evaluated. If the carbon emission exceeds the standard, optimization suggestions are provided, such as adjusting the operating parameters or regular maintenance, so as to reduce the carbon emission during the operation of the motor.

[0061]

[0062] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A comprehensive monitoring and analysis method for motor energy consumption and carbon emissions based on equipment status recognition, characterized in that: include, Step 1: Collect the original data of the motor operation, including the motor speed, shaft, vibration data, and carbon emission factor; Step 2: Fill missing values ​​in the collected original running data and calculate the feature correlation between the original running data to obtain a new data set; Step three, input the new data set obtained in step two into the classification decision tree model, and correct the classification result by the information gain correction value.

2. The method for comprehensive monitoring and analysis of motor energy consumption and carbon emissions based on equipment status identification according to claim 1 is characterized in that: In the step 2, missing values ​​of the collected original operation data are filled based on the distance, and then Fourier transform is introduced to process the vibration data, and then the frequency domain data after Fourier transform is expanded using the interpolation method.

3. The method for comprehensive monitoring and analysis of motor energy consumption and carbon emissions based on equipment status identification according to claim 1 is characterized in that: In the step 2, the Pearson correlation coefficient is used to evaluate the correlation between each feature and the target variable, and a loss function is constructed based on the correlation between the feature and the target variable. The loss function is reduced by the gradient descent algorithm, that is, the increased value is made closer to the true value.

4. The method for comprehensive monitoring and analysis of motor energy consumption and carbon emissions based on equipment status identification according to claim 3 is characterized in that: The loss function is expressed as: ,in is the mapping result of the correlation coefficient of the original data i, is the mapping result of the correlation coefficient of the increasing value, and n is the number of features.

5. The method for comprehensive monitoring and analysis of motor energy consumption and carbon emissions based on equipment status identification according to claim 4 is characterized in that: In the process of using the Pearson correlation coefficient to evaluate the correlation between each feature and the target variable, the distance between each column and the label column is added as an additional parameter to evaluate the relevance of the feature.

6. The method for comprehensive monitoring and analysis of motor energy consumption and carbon emissions based on equipment status identification according to claim 5 is characterized in that: , , where Pnew and Porgin are the results of mapping the new and old data through the Pearson correlation coefficient, and d is the distance between each column and the label column as an additional parameter. .

7. The method for comprehensive monitoring and analysis of motor energy consumption and carbon emissions based on equipment status identification according to claim 1 is characterized in that: In step three, the information gain correction value is the product of the information gain and the Pearson correlation coefficient function f(r), and the Pearson correlation coefficient function f(r) converts the Pearson correlation coefficient into a positive number and improves the hierarchical nature of the strength of the interval where the Pearson correlation coefficient is located.

8. The method for comprehensive monitoring and analysis of motor energy consumption and carbon emissions based on equipment status identification according to claim 7 is characterized in that: The Pearson correlation coefficient function is expressed as: .

9. The method for comprehensive monitoring and analysis of motor energy consumption and carbon emissions based on equipment status identification according to claim 1 is characterized in that: The method further includes a step 4, in which the power P of the motor is calculated according to the torque and running time of the motor, and the carbon emission E of the motor during operation is calculated according to a predetermined carbon emission factor.

10. The method for comprehensive monitoring and analysis of motor energy consumption and carbon emissions based on equipment status identification according to claim 9 is characterized in that: The method further includes step five, in which the operating state of the motor is evaluated according to the classification result obtained in step three and the real-time operating state of the motor.

Images