A dynamic production working condition motor energy-saving amount measuring and calculating system and method

Abstract

The application provides a motor energy-saving amount measuring and calculating system and method under dynamic production conditions, the system comprises a data acquisition module, a data cleaning module, a working condition identification and feature extraction module, a model construction module and a real-time calculation module; the method comprises the following sub-steps: S1: collecting real-time data of new and old motors and performing data cleaning; S2: obtaining working condition feature vectors and time period power consumption; S3: constructing a new and old motor association model; S4: obtaining the energy-saving amount of the new motor; the method matches the working condition feature vector of the new motor with the working condition feature vector of the old motor closest in time, gets rid of the interference of external factors such as production plan adjustment and seasonal change on working condition comparison, ensures that the energy-saving amount calculation is not affected by working condition fluctuations, dynamically fits the model parameters of the real-time working condition features and energy consumption data of the old motor, automatically corrects the model as the motor ages, avoids the model deviation caused by equipment parameter errors and motor aging, and reduces the energy-saving amount misjudgment caused by model failure.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of motor energy saving, and in particular to a motor energy saving measurement and calculation system and method under dynamic production conditions. BACKGROUND

[0002] The motor system is the main power consumption equipment in the industrial field, and its energy saving reconstruction is the key direction of contract energy management. Under the double carbon target and strict energy efficiency constraints, the demand for motor system energy saving reconstruction in high energy consumption industries such as steel, chemical industry and building materials continues to grow. The contract energy management mode becomes the first choice for enterprises because it can help enterprises realize zero investment or low investment reconstruction, and the core of this mode is energy saving benefit sharing. Therefore, accurate measurement and verification of energy saving amount is crucial. The production load, process flow, environmental conditions and other factors of the motor system affect the energy consumption. Accurate and fair calculation of energy consumption effect under complex and variable conditions is a technical difficulty and the basis for customer trust.

[0003] The existing motor energy saving amount calculation methods mainly include direct measurement method and indirect calculation method. The direct measurement method directly, continuously and accurately calculates the input power of the motor system before and after the reconstruction under similar conditions, compares the differences between the two, and determines the energy saving rate. The indirect calculation method estimates the energy saving efficiency through theoretical calculation or device parameters, establishes a detailed physical model of the motor and its driving load, inputs the actual operating parameters (flow, pressure, speed, etc.), and simulates the energy consumption before and after the reconstruction.

[0004] The existing motor energy saving amount calculation methods have the following disadvantages:

[0005] The direct measurement method relies on similar condition definition and adjustment, and is easily disturbed by external factors such as production plan adjustment and seasonal change. The pre-set energy saving benchmark is difficult to adapt to new production scenarios.

[0006] The indirect calculation method is complex in modeling process, requires professional knowledge and accurate device parameters, and parameter errors will cause result deviation. With the increase of motor service life, the efficiency of the motor will decrease, and the model error will increase. SUMMARY

[0007] The purpose of the present application is to solve the problems in the prior art and provide a motor energy saving amount measurement and calculation method and system under dynamic production conditions.

[0008] To achieve the above purpose, the present application adopts the following technical solutions:

[0009] A motor energy saving amount measurement and calculation system under dynamic production conditions, the system comprises a data acquisition module, a data cleaning module, a working condition recognition and feature extraction module, a model construction module and a real-time calculation module.

[0010] The data acquisition module is used for acquiring real-time data of new and old motors, including rotating speed, output torque, input current, output power, cumulative power consumption, etc.

[0011] The data cleaning module is used for cleaning data of the acquired new and old motors, including eliminating abnormal data, eliminating noise, etc.

[0012] The working condition recognition and feature extraction module is used for extracting working condition feature vectors and power consumption of time periods.

[0013] The model construction module is used for constructing a mapping model of working condition features of new and old motors and a correlation model of working condition feature vectors and power consumption.

[0014] The real-time calculation module is used for calculating energy-saving amount of new motors.

[0015] The application further provides a motor energy-saving amount measuring and calculating method under dynamic production working conditions, comprising the following steps:

[0016] S1: acquiring real-time data of new and old motors and cleaning data;

[0017] In advance, among a plurality of same production lines, one production line uses an old motor as a reference motor, and the remaining production lines are reformed to use energy-saving motors as new motors;

[0018] S11: periodically acquiring real-time data of new and old motors according to a set acquisition period;

[0019] The data acquisition module sets an acquisition period in advance, and a thing networking acquisition gateway periodically acquires real-time data of new and old motors according to the set acquisition period, wherein the real-time data includes rotating speed, output torque, input current, output power, cumulative power consumption and corresponding time stamps of the new and old motors, etc.

[0020] S12: cleaning data of the acquired real-time data of new and old motors;

[0021] The data cleaning includes eliminating abnormal data, eliminating noise, etc.

[0022] The data cleaning module sets value intervals of rotating speed, output torque, input current and output power according to actual rated parameters and production scenes of the new and old motors in advance; the data cleaning module compares the acquired real-time data of the new and old motors with corresponding value intervals one by one, and if there is data outside the corresponding value intervals, the data is judged as abnormal data and the abnormal data is eliminated.

[0023] The abnormal judgment method for cumulative power consumption is as follows:

[0024] The data cleaning module obtains the rated maximum power of the new and old motor in advance, multiplies the rated maximum power by the collection period to obtain the power growth threshold in a unit period, extracts the current period cumulative power consumption and the last collection period cumulative power consumption from the cumulative power consumption data collected in step S11, subtracts the last collection period cumulative power consumption from the current period cumulative power consumption to obtain the actual power consumption increment in a unit period, and determines that it is abnormal data if the actual power consumption increment < 0;

[0025] The actual power consumption increment is compared with the power growth threshold in a unit period, and if it is greater than the power growth threshold in a unit period, it is also determined to be abnormal data, and the abnormal data is removed.

[0026] The Kalman filtering algorithm is used to eliminate the Gaussian noise in the real-time data of the new and old motor after removing the abnormal data.

[0027] S2: Obtain the working condition feature vector and the period power consumption;

[0028] S21: Screen the initial candidate working condition period;

[0029] The time threshold is stored in advance in the working condition recognition and feature extraction module ;

[0030] The real-time data of the new and old motor after data cleaning is arranged in chronological order according to the time stamp corresponding to the data, and continuous time intervals are extracted therefrom; each time interval is marked, the starting time is , the ending time is , and the duration of the time interval is , ;

[0031] The duration is compared with the time threshold , and the time segment of is screened out as the candidate working condition period;

[0032] S22: Data integrity and running stability check;

[0033] The candidate working condition period that passes the data integrity and running stability check is retained;

[0034] S23: Extract the working condition feature vector and the period power consumption;

[0035] The average values of the speed, output torque, input current and output power in each retained candidate working condition period are combined into a column vector as the working condition feature vector in the candidate working condition period;

[0036] Let the speed be R, the output torque be T, the input current be I, the output power be P, and the working condition feature vector be EV, ;

[0037] Specifically, respectively are the average value of the rotation speed, the average value of the output torque, the average value of the input current, and the average value of the output power in the candidate working condition period;

[0038] For each candidate working condition period to be retained, the cumulative power consumption at the starting time and the ending time is extracted, the cumulative power consumption at the ending time is subtracted from the cumulative power consumption at the starting time , and the power consumption in the candidate working condition period is obtained, denoted as .

[0039] S3: Construct a new-old motor association model;

[0040] S31: Construct a mapping model of new-old motor working condition characteristics;

[0041] Artificially query the production plan, filter out new-old motors driving the same type of equipment and producing the same type of products in the production line, and determine that the new-old motors are in the same production working condition;

[0042] Map the working condition characteristic vectors of the new-old motors in the same production working condition, and based on the principle of the nearest time matching, match the working condition characteristic vectors of the old motors closest in time to each working condition characteristic vector of the new motor to form a one-to-one mapping data pair;

[0043] Establish a mapping model of new-old motor working condition characteristics, and the mapping model of new-old motor working condition characteristics is as follows:

[0044] ;

[0045] ;

[0046] ;

[0047] wherein, , are the working condition characteristic vectors of the new motor and the corresponding matched old motor, is a parameter to be fitted;

[0048] Substitute the mapping data pair as a training sample into the mapping model of new-old motor working condition characteristics, and obtain the parameter value group with the smallest total error through a numerical fitting algorithm, that is, the parameter value of ;

[0049] S32: Construct an association model of working condition characteristic vectors and power consumption;

[0050] Based on the principle of motor power loss (copper loss, iron loss, mechanical loss, stray loss, etc.), the correlation model of old motor working condition characteristic vector and power consumption is constructed, and the model is as follows:

[0051] ;

[0052] Wherein, the values of R, T, I, P are the corresponding values in the working condition characteristic vector of the old motor in the production working condition period;

[0053] is the actual power consumption per unit time of the old motor, i.e. input power, , is the power consumption in the production working condition, is the duration of the corresponding production working condition;

[0054] is the parameter to be fitted;

[0055] The working condition characteristic vector, actual power consumption and duration of each production working condition are a data sample, which constitutes a data sample set; the sample data set is divided into training set and test set according to the proportion;

[0056] Substitute the data in the training set into the correlation model of old motor working condition characteristic vector and power consumption to obtain the parameter group that minimizes the total error, which is The corresponding parameter value; the total error is the cumulative error of all training samples in the data sample set; and the result is verified by the data in the test set, and the model is continuously optimized.

[0057] S4: obtain the energy saving of new motor;

[0058] For the collected real-time data of new motor, denoted as rotating speed R n,i , output torque T n,i , input current I n,i , output power P n,i , cumulative power consumption E n,i , time stamp t n,i ; the real-time calculation module subtracts the time stamps of adjacent two times of collection to obtain the time interval , subtracts the cumulative power consumption of adjacent two times of collection to obtain the difference of cumulative power consumption ;

[0059] The Kalman filter is used to smooth the real-time data of new motor, eliminating the Gaussian noise of real-time data of new motor; the smoothed rotating speed , output torque , input current , output power are obtained and substituted into the mapping model of new and old motor working conditions to obtain the rotating speed of the corresponding old motor , output torque , input current , output power ;

[0060] Put the rotating speed , output torque , input current , output power Into the working condition characteristic vector and the correlation model of power consumption, the actual power consumption of the old motor in unit time under the same production working condition is obtained ;

[0061] Put the input power Into the following formula, the energy saving amount E of the new motor is obtained s,i ;

[0062] ;

[0063] Wherein, The difference between the cumulative power consumption of the new motor in the adjacent two times, The time interval between the adjacent two times, The actual power consumption of the old motor in the adjacent two times.

[0064] Compared with the prior art, the beneficial effects of the present application are:

[0065] The method compares the old and new motors in real time, maps the working condition characteristics, retains the old motor production line as a dynamic reference, does not depend on the preset similar working condition, matches the same product / working condition through production plan, matches the time closest working condition characteristic vector of the old motor for each working condition characteristic vector of the new motor, and then calibrates the data deviation of different motors through the linear model, so as to get rid of the interference of production plan adjustment, seasonal change and other external factors on the working condition comparison, even if the production scene changes dynamically, the real-time and accurate old motor reference can be found, and the energy saving amount calculation is not affected by the working condition fluctuation;

[0066] The method is based on the core loss principle of motor copper loss (changes with the square of current) and iron loss (changes linearly with rotating speed), constructs a simplified energy consumption correlation model, reduces the modeling difficulty, and does not need professional complex debugging; meanwhile, the model parameters are dynamically fitted with the real-time working condition characteristics and energy consumption data of the old motor, instead of relying on fixed equipment parameters, which can automatically correct the model with the aging of the motor, avoid the model deviation caused by equipment parameter error and motor aging, keep the high precision of energy consumption calculation for a long time, and reduce the energy saving amount misjudgment caused by model failure. BRIEF DESCRIPTION OF DRAWINGS

[0067] Figure 1 It is the step flow chart of the motor energy saving amount measuring and calculating method in the dynamic production working condition. DETAILED DESCRIPTION

[0068] In order to make the purpose, structure, characteristics and functions of the present application more clear, the following detailed description is given in conjunction with the embodiments.

[0069] As shown in Figure 1 A motor energy saving amount measuring system under dynamic production conditions, the system comprises a data acquisition module, a data cleaning module, a working condition recognition and feature extraction module, a model construction module, a real-time calculation module;

[0070] The data acquisition module is used for acquiring real-time data of new and old motors, including speed, output torque, input current, output power, cumulative power consumption, etc.

[0071] The data cleaning module is used for cleaning the data of the collected new and old motors, including rejecting abnormal data and eliminating noise, etc.

[0072] The working condition recognition and feature extraction module is used for extracting working condition feature vectors and time period power consumption;

[0073] The model construction module is used for constructing a mapping model of new and old motor working condition features and a correlation model of working condition feature vectors and power consumption.

[0074] The real-time calculation module is used for calculating the energy saving amount of the new motor.

[0075] The system contains a data acquisition module, a data cleaning module, a working condition recognition and feature extraction module, a real-time calculation module and other parts, which can measure the energy saving amount through automatic data acquisition and real-time analysis on site, reduce the complexity of manual operation, improve the convenience of use, and is suitable for different scales and types of production lines.

[0076] The present application also provides a motor energy saving amount measuring method under dynamic production conditions, comprising the following steps:

[0077] S1: Collecting real-time data of new and old motors and cleaning the data;

[0078] In advance, among a plurality of same production lines, one production line uses an old motor as a reference motor, and the remaining production lines are reformed to use energy-saving motors as new motors;

[0079] S11: Collecting real-time data of new and old motors according to a set collection period;

[0080] The data acquisition module sets a collection period in advance, and the Internet of Things collection gateway collects real-time data of new and old motors according to the set collection period, the real-time data including speed, output torque, input current, output power, cumulative power consumption and corresponding time stamp of new and old motors, etc.

[0081] S12: data cleaning on the collected real-time data of the new and old motors;

[0082] The data cleaning includes eliminating abnormal data, eliminating noise, etc.

[0083] The data cleaning module sets the numerical interval of the speed, output torque, input current, and output power in advance according to the actual rated parameters of the new and old motors and the production scene, such as the numerical interval of the input current being [0, 50A] and the like. The data cleaning module compares the collected real-time data of the new and old motors with the corresponding numerical interval one by one. If there is data outside the corresponding numerical interval, it is determined that this data is abnormal data, and the abnormal data is eliminated.

[0084] Further, the abnormal judgment method of the cumulative power consumption is as follows:

[0085] The data cleaning module obtains the rated maximum power of the new and old motors in advance, multiplies the rated maximum power by the collection period to obtain the power growth threshold in the unit period, extracts the current period cumulative power consumption and the last collection period cumulative power consumption from the cumulative power consumption data collected in step S11, subtracts the last collection period cumulative power consumption from the current period cumulative power consumption to obtain the actual power consumption increment in the unit period, and if the actual power consumption increment <0, it is determined as abnormal data.

[0086] The actual power consumption increment is compared with the power growth threshold in the unit period. If it is greater than the power growth threshold in the unit period, it is also determined as abnormal data, and the abnormal data is eliminated.

[0087] The Gaussian noise in the real-time data of the new and old motors after eliminating the abnormal data is eliminated by the Kalman filtering algorithm.

[0088] S2: obtain the working condition feature vector and the period power consumption;

[0089] S21: screen the initial candidate working condition period;

[0090] The time threshold value is stored in advance in the working condition recognition and feature extraction module ;

[0091] The real-time data of the new and old motors after data cleaning is arranged in chronological order according to the time stamp corresponding to the data, and a continuous time interval is intercepted therefrom. Each time interval is marked, the starting time is , the ending time is , and the duration of the time interval is , ;

[0092] The duration is compared with the time threshold value The comparison is performed to screen out time segments as candidate working condition periods;

[0093] Specifically, the method of intercepting continuous time intervals is as follows:

[0094] The timestamp of the first sorted new and old motor real-time data is taken as the starting time of the first continuous time interval , and each subsequent new and old motor real-time data is checked in time sequence according to the time interval from the previous new and old motor real-time data. If the interval is equal to the preset collection period, the data is included in the current interval. If the time interval is not equal to the collection period (such as missing data leading to a long interval, or shutdown leading to data interruption), the timestamp of the previous new and old motor real-time data is taken as the end time of the current time interval , and the current time interval is completed. At the same time, the timestamp of the data is taken as the starting time of the new interval , and the above verification steps are repeated to continue to determine the next continuous interval.

[0095] S22: Perform data integrity and running stability verification;

[0096] The data integrity verification method is as follows:

[0097] For all candidate working condition periods, all new and old motor real-time data included in the period are checked one by one. If for any timestamp in the period, any data of speed, output torque, input current, or output power is missing, it is determined that the data integrity verification of the candidate working condition period fails, and the candidate working condition period is removed. Otherwise, the period is retained;

[0098] The running stability verification method is as follows:

[0099] A fixed proportion is stored in advance in the working condition recognition and feature extraction module ;

[0100] For each candidate working condition period retained after data integrity verification, the average values of speed, output torque, input current, and output power are calculated respectively by the average value formula;

[0101] All speeds, output torques, input currents, and output powers are sorted by the sort component to obtain the maximum and minimum values of the speed, output torque, input current, and output power in each candidate working condition period. The maximum value minus the minimum value of each of the above data in the candidate working condition period is obtained as the fluctuation amplitude;

[0102] The fluctuation amplitude is compared with the product of the average value of each item of data and the fixed proportion . If the fluctuation amplitude of any item of data is greater than the product of the average value and the fixed proportion, the candidate working condition period is removed. If the product of the two is greater than 1, the candidate working condition period is rejected; otherwise, the candidate working condition period is retained.

[0103] S23: Extract the working condition feature vector and the power consumption in the period;

[0104] The average values of the rotation speed, output torque, input current, and output power in each retained candidate working condition period are combined into a column vector as the working condition feature vector in the candidate working condition period;

[0105] Let the rotation speed be R, the output torque be T, the input current be I, the output power be P, and the working condition feature vector be EV, ;

[0106] For each retained candidate working condition period, the cumulative power consumption from the start time to the end time is extracted, the cumulative power consumption from the start time is subtracted from the cumulative power consumption from the end time , and the power consumption in the candidate working condition period is obtained, denoted as .

[0107] S3: Construct a new-old motor association model;

[0108] S31: Construct a mapping model of new-old motor working condition features;

[0109] Artificially query the production plan, filter out new-old motors that drive the same type of equipment and produce the same type of products in the production line, and determine that the new-old motors are in the same production working condition;

[0110] Map the working condition feature vectors of the new-old motors in the same production working condition, and based on the principle of the nearest time match, match the working condition feature vector of the nearest old motor for each working condition feature vector of the new motor to form a one-to-one mapping data pair;

[0111] Establish a mapping model of new-old motor working condition features, and the mapping model of new-old motor working condition features is as follows:

[0112] ;

[0113] ;

[0114] ;

[0115] wherein, , are the working condition feature vectors of the new motor and the corresponding matched old motor, is a parameter to be fitted;

[0116] The mapping data pairs are substituted into the mapping model of the working condition characteristics of the new and old motors, and the parameter value group with the minimum total error is obtained through a numerical fitting algorithm, i.e. the parameter value of the mapping model of the working condition characteristics of the new and old motors;

[0117] Specifically, the calculation method of the parameter value group with the minimum error is as follows:

[0118] The mapping model of the working condition characteristics of the new and old motors is decomposed into four independent dimension equations:

[0119]

[0120]

[0121]

[0122]

[0123] In the equations, the left side of the equal sign is the actual value of the old motor, and the right side is the predicted value of the corresponding dimension of the old motor. The error is obtained by subtracting the predicted value from the actual value, which corresponds to the four error functions. The mapping data pairs are proportionally divided into a training set and a test set. The mapping data pairs in the training set are substituted into the four error functions one by one, and the partial derivatives of the error functions are calculated through a least squares method or other numerical fitting algorithms, and are set to 0. The parameter value group with the minimum total error is obtained by solving, i.e. the parameter value of the mapping model of the working condition characteristics of the new and old motors; the total error is the cumulative error of all mapping data pairs; and the model is verified through a verification set and is continuously optimized.

[0124] S32: Constructing a working condition characteristic vector and power consumption correlation model;

[0125] Based on the motor power loss principle (copper loss, iron loss, mechanical loss, and stray loss), a working condition characteristic vector and power consumption correlation model of the old motor is constructed, and the model is as follows:

[0126]

[0127] wherein the values of R, T, I, and P are the corresponding values in the working condition characteristic vector of the old motor during the production working condition period;

[0128] is the actual power consumption per unit time of the old motor, i.e., the input power, , is the power consumption during the production working condition, is the duration of the corresponding production working condition;

[0129] is a parameter to be fitted;

[0130] ​​​​​The working condition characteristic vector, actual power consumption and duration of each production condition are a data sample, and constitute a data sample set; the sample data set is divided into a training set and a test set according to a proportion;

[0131] The data in the training set is substituted into the correlation model of the working condition characteristic vector and power consumption of the old motor, the predicted power consumption of the old motor per unit time is on the right side of the equal sign in the model, the corresponding error is obtained by subtracting the predicted power consumption from the actual power consumption, an error function is constructed, the partial derivative of the error function is solved by a numerical fitting algorithm such as least squares method, and is set to 0 to obtain the parameter group that minimizes the total error, which is The corresponding parameter value; the total error is the cumulative error of all training samples in the data sample set; and the result is verified by the data in the test set, and the model is continuously optimized.

[0132] The method flexibly identifies the motor running conditions under different production lines and different working conditions through the working condition recognition and feature extraction module, and dynamically adjusts and adapts to various production environments through the establishment of the working condition characteristic mapping model of the new and old motors, so as to ensure accurate energy saving calculation under multiple production lines and different working conditions.

[0133] S4: obtaining the energy saving amount of the new motor;

[0134] For the collected real-time data of the new motor, denoted as rotating speed R n,i , output torque T n,i , input current I n,i , output power P n,i , cumulative power consumption E n,i , and time stamp t n,i ; the real-time calculation module subtracts the time stamps collected at adjacent times to obtain the time interval , and subtracts the cumulative power consumptions collected at adjacent times to obtain the difference between the cumulative power consumptions.

[0135] The Kalman filter is used to smooth the real-time data of the new motor to eliminate Gaussian noise of the real-time data of the new motor; the smoothed rotating speed , output torque , input current , and output power are obtained and substituted into the mapping model of the working condition characteristics of the new and old motors to obtain the rotating speed , output torque , input current , and output power of the corresponding old motor.

[0136] The rotating speed , output torque , input current , and output power Substitute into the working condition characteristic vector and the correlation model of power consumption, obtain the equivalent input power of the old motor under the same production working condition ;

[0137] Substitute input power Into the following formula, obtain the energy saving amount E of the new motor s,i ;

[0138] .

[0139] Through range checking, energy consumption trend checking, abnormal data is eliminated, Kalman filter is used to eliminate Gaussian noise; then through the characteristic data of the new and old motors under the same working condition, a linear mapping model is fitted, and the system data deviation of different motors is corrected; the noise and deviation are eliminated from the data source, the data consistency of different brands and models of motors is ensured, the calculation error of energy saving amount caused by data problems is avoided, and the result reliability is improved.

[0140] Through real-time calculation module, the collected new motor data is analyzed in real time, the energy saving amount is continuously calculated in combination with the dynamic working condition characteristic model, instead of relying on batch processing mode of fixed time period, the real-time and adaptability of energy saving amount measurement are improved, which is helpful to provide more timely energy saving feedback and optimization suggestions.

[0141] The present application has been described by the above-mentioned related embodiments, however, the above-mentioned embodiments are only examples for implementing the present application. It must be pointed out that the disclosed embodiments do not limit the scope of the present application. On the contrary, changes and modifications made without departing from the spirit and scope of the present application are all within the scope of the patent protection of the present application.

Claims

1. A method for measuring and calculating the energy saving of a motor in a dynamic production process, characterized in that: It comprises the following steps: S1: Collecting real-time data of new and old motors and performing data cleaning; In advance, among a plurality of same production lines, one production line uses an old motor as a reference motor, and the rest of the production lines are reformed to use energy-saving motors as new motors; S11: Collecting real-time data of new and old motors according to a set collection period; The data collection module sets a collection period in advance, and collects real-time data of new and old motors according to the set collection period through the Internet of Things collection gateway, wherein the real-time data includes the rotating speed, output torque, input current, output power, cumulative power consumption and corresponding time stamp of the new and old motors; S12: Cleaning the collected real-time data of new and old motors; The data cleaning includes eliminating abnormal data and eliminating noise; S2: Obtaining a working condition characteristic vector and a time period power consumption; S21: Screening an initial candidate working condition period; S22: Performing data integrity and running stability verification; S23: Extracting a working condition characteristic vector and a time period power consumption; S3: Constructing a new and old motor association model; S31: Constructing a mapping model of new and old motor working condition characteristics; For new and old motors under the same production working condition, a working condition characteristic vector mapping is performed, and based on the time nearest matching principle, for each working condition characteristic vector of the new motor, a working condition characteristic vector of the old motor with the nearest time is matched to form a one-to-one mapping data pair; A mapping model of new and old motor working condition characteristics is established, and the mapping model of new and old motor working condition characteristics is as follows: ； ； ； wherein, , are the working condition feature vectors of the new electric machine and the corresponding matching old electric machine, respectively, , , , , , , , are the parameters to be fitted; The mapping data pairs are taken as training samples, and are substituted into the mapping model of the working condition characteristics of the new and old motors to obtain a parameter value group with minimum total error through a numerical fitting algorithm, i.e. 、 、 、 、 、 、 、 The total error is the cumulative error of all the mapping data pairs. S32: Constructing an association model of working condition characteristic vector and power consumption; An association model of old motor working condition characteristic vector and power consumption is constructed, and the model is as follows: ； Wherein, the values of R, T, I and P are the corresponding values in the working condition characteristic vector of the old motor in the production working condition period; actual power consumption of the old motor per unit time, i.e. input power, duration of the corresponding production condition;​​​​ a1, a2, b1, b2, c1, c2, d1, d2 and e are parameters to be fitted; S4: Obtaining the energy-saving amount of the new motor; The collected real-time data of the new motor is smoothed, the smoothed data is substituted into the mapping model of new and old motor working condition characteristics, the data of the corresponding old motor is obtained, the data of the old motor is substituted into the association model of working condition characteristic vector and power consumption, the equivalent input power of the old motor under the same production working condition is obtained, and the energy-saving amount of the new motor is obtained based on the equivalent input power; For the collected real-time data of the new motor, denoted as rotating speed R n,i , output torque T n,i , input current I n,i , output power P n,i , cumulative power consumption E n,i , time stamp t n,i ; the real-time calculation module subtracts the time stamps of the adjacent two times of collection to obtain the time interval , subtracts the cumulative power consumption of the adjacent two times of collection to obtain the difference of the cumulative power consumption ; Kalman filtering is used to smooth the real-time data of the new motor, eliminating Gaussian noise in the real-time data; the smoothed speed is then obtained. Output torque Input current Output power Then, by substituting these values ​​into the mapping model of the operating characteristics of the old and new motors, the corresponding speed of the old motor can be obtained. Output torque Input current Output power ; The rotational speed , the output torque , the input current , the output power is substituted into the correlation model between the working condition characteristic vector and the power consumption to obtain the actual power consumption of the old motor in unit time under the same production working condition, i.e. the input power P oi ; The input power P oi Substituting the following formula, the energy saving E of the new motor is obtained s,i ; 。 2. The dynamic production working condition motor energy-saving amount measuring method according to claim 1, wherein: The specific content of step S12 is as follows: The data cleaning module sets the value range of rotating speed, output torque, input current and output power in advance according to the actual rated parameters of new and old motors and production scenes; the collected real-time data of new and old motors are compared one by one with the corresponding value range, if there is data outside the corresponding value range, it is judged that this data is abnormal data, and the abnormal data is eliminated; The abnormal judgment method of cumulative power consumption is as follows: The data cleaning module obtains the rated maximum power of the new and old motors in advance, multiplies the rated maximum power by the collection period to obtain the power growth threshold in a unit period, extracts the current period cumulative power consumption and the last collection period cumulative power consumption from the cumulative power consumption data collected in step S11, subtracts the last collection period cumulative power consumption from the current period cumulative power consumption to obtain the actual power consumption increment in a unit period, and determines that the actual power consumption increment is abnormal data if the actual power consumption increment is less than 0; The actual power consumption increment is compared with the power growth threshold in a unit period, and if the actual power consumption increment is greater than the power growth threshold in a unit period, it is also determined that the actual power consumption increment is abnormal data, and the abnormal data is removed. Gaussian noise in the real-time data of the new and old motors after removing the abnormal data is eliminated by a Kalman filtering algorithm.

3. The dynamic production working condition motor energy saving measurement method according to claim 1, characterized in that: Step S2 includes the following specific contents: S21: screening initial candidate working condition period; The working condition recognition and feature extraction module stores a time threshold value in advance ​ The real-time data of the new and old motors after data cleaning is arranged according to the time stamps of the data in chronological order, and a continuous time interval is intercepted from the data; each time interval is marked, the starting time is , the ending time is , and the duration of the time interval is , = - ; duration threshold comparison, screening out ≥ time segment as a candidate working condition period S22: performing data integrity and running stability verification; S23: extracting working condition feature vector and period power consumption; S24: retaining candidate working condition period passing the data integrity and running stability verification; S25: extracting working condition feature vector and period power consumption; Let the rotating speed be R, the output torque be T, the input current be I, the output power be P, and the working condition characteristic vector be EV, EV=[ , , , ] ; For each candidate working condition period reserved, the cumulative power consumption with the starting time and the ending time is extracted, the cumulative power consumption with the ending time is subtracted from the cumulative power consumption with the starting time , and the power consumption in the candidate working condition period is obtained, denoted as .

4. The dynamic production working condition motor energy saving measurement method according to claim 1, characterized in that: Step S3 includes the following specific contents: S31: constructing a mapping model of new and old motor working condition features; S32: constructing a correlation model of working condition feature vector and power consumption; S33: constructing a data sample set; S34: dividing the sample data set into a training set and a test set according to a proportion; S35: substituting the data in the training set into the correlation model of old motor working condition feature vector and power consumption, and obtaining the corresponding error by subtracting the predicted power consumption from the actual power consumption to construct an error function, and obtaining the parameter group that minimizes the total error by solving the partial derivative of the error function and setting it to 0, wherein the total error is the cumulative error of all training samples in the data sample set, and the parameter group is a1, a2, b1, b2, c1, c2, d1, d2, and e corresponding parameter values; and verifying the result by the data in the test set to continuously optimize the model.

5. A dynamic production working condition motor energy saving measurement system for implementing the method according to any one of claims 1-4, characterized in that: The system includes a data collection module, a data cleaning module, a working condition identification and feature extraction module, a model construction module, and a real-time calculation module; The data collection module is used to collect real-time data of new and old motors, including speed, output torque, input current, output power, and cumulative power consumption; The data cleaning module is used to clean the collected data of new and old motors. The working condition recognition and feature extraction module is configured to extract a working condition feature vector and a power consumption of a time period; The model construction module is configured to construct a mapping model of working condition features of the new and old motors and a correlation model of the working condition feature vector and the power consumption; The real-time calculation module is configured to calculate energy saving of the new motor.

Images