A real-time production beat calculation method and system based on waveform feature matching

Abstract

The present invention discloses a real-time production beat calculation method and system based on waveform feature matching, which relates to the field of industrial big data processing. The method is as follows: S1: real-time acquisition of raw current data; S2: preprocessing the raw current data using a cubic spline interpolation method, and resampling the data to obtain standard current data with a sampling time interval of t; S3: extracting standard current waveform features, and establishing a historical standard current waveform feature library for storing the waveform features of historical standard current data and the equipment operating status corresponding to the waveform features of the historical standard current data; S4: matching the standard current waveform features with the waveform features of the historical standard current data, judging the equipment operating status, and obtaining the number of processed waveforms; S5: counting the number of processed waveforms and the number of workpieces processed by the equipment within a time period to calculate the production beat; the present invention improves the calculation efficiency of the production beat by processing the data collected by the cubic spline interpolation method at equal time intervals.

Description

Technical Field

[0001] The present invention relates to the field of industrial big data processing, and in particular to a real-time production rhythm calculation method and system based on waveform feature matching. Background Art

[0002] Currently, most factories operate on assembly lines. The efficiency of an entire production line depends not on the fastest workstation but on the slowest. A bottleneck at any workstation can negatively impact the efficiency of the entire line. To eliminate the efficiency losses caused by imbalanced workstations and the wasted waiting time at various workstations, real-time monitoring of production rhythms is essential, allowing for adjustments to workloads and maximized production efficiency.

[0003] However, the quantitative measurement of production tact time related to traditional production lines is generally achieved manually. Engineers use a handheld stopwatch to test the tact time of each workstation on the production line. Based on the detected tact time, they analyze the problems with the production system's production tact time. This method is inefficient and results in the inability to promptly resolve tact time anomalies in the production system.

[0004] For the application scenario of multi-channel parallel sampling of equipment data in smart factories, due to network congestion, electromagnetic interference and other reasons, the collection system cannot achieve uniform sampling. The time intervals of the equipment operation data collected by the system are not completely equal, which makes data processing and analysis complicated.

[0005] In the prior art, the equipment status is monitored by collecting analog information of the equipment, but the prior art has the disadvantage that the calculation efficiency of the production rhythm is low. Summary of the Invention

[0006] In order to solve the problem of low calculation efficiency of production rhythm in the prior art, the present invention proposes a real-time production rhythm calculation method and system based on waveform feature matching to improve the calculation efficiency of production rhythm.

[0007] To solve the above technical problems, the technical solution adopted by the present invention is: a real-time production rhythm calculation method based on waveform feature matching, the steps are as follows: S1: real-time collection of original current data of production equipment operation; S2: preprocessing the original current data using the cubic spline interpolation method, and converting the original current data with non-equal time intervals into preprocessed current data with equal time intervals; resampling the preprocessed current data to obtain standard current data with a sampling time interval of t; S3: extracting standard current waveform features based on the standard current data, and establishing a historical standard current waveform feature library, which is used to store historical standard current data waveform features and the equipment operating status corresponding to the historical standard current data waveform features; S4: matching the standard current waveform features with the historical standard current data waveform features, judging the corresponding equipment operating status, and obtaining the number of processed waveforms based on the equipment operating status; S5: counting the number of processed waveforms and the number of workpieces processed by the equipment within a time period to calculate the production rhythm.

[0008] The working principle of the present invention is as follows:

[0009] Based on the cubic spline interpolation method, the parallel sampled original current data with non-uniform time intervals are interpolated and reconstructed to obtain original current data with equal time intervals. The reconstructed original current data are analyzed to realize real-time online calculation of the production station's rhythm and achieve real-time optimization of workshop production.

[0010] Preferably, the raw current data includes the serial number of each production device, the device name, the address of the collection point, the timestamp, the data identifier, and the current value.

[0011] Furthermore, the specific steps for collecting the original current data of the production equipment are as follows: the equipment operating status parameters are collected, and the hexadecimal message is encapsulated using the ModBus protocol. The response information is {register address, operation code, data bit, check bit}. The message needs to be parsed when it is transmitted to the server, and the data is extracted and converted into real values. The parsed original current data is pushed to the streaming message queue; the data format of the original current data is [device number, device name, collection point address, timestamp, data identifier, current value].

[0012] Furthermore, the specific steps of the preprocessing are as follows: for the original current data in the streaming message queue, a sampling time window is established according to the historical data, so that the length of a sampling time window includes several production waveforms, the current value and time stamp in the original current data are extracted in the sampling time window, a current time series signal is formed, and the signal is reconstructed by interpolation using the cubic spline interpolation method; the original current data includes n+1 different time nodes t0, t1, t2, ..., t nThe current sequence data, the corresponding current values are i0, i1, i2, ..., i n ; Wherein, p=t0<t1<t2<...<t n =q, p is the starting time node of the original current data, q is the last time node of the original current data; construct a cubic spline interpolation function S(t) and make it meet the following conditions: S(t) has a second-order continuous derivative on [p, q]; S(t k1 )=y k1 , k1=0,1,2,...,n;In each subinterval [t k2 , t k2+1 ], k2=0, 1, 2, ..., n-1, S k2 (t) is a cubic polynomial; S k2 (t) is expressed as follows:

[0013]

[0014] The mathematical expression of the cubic function S(t) corresponding to n intervals is:

[0015]

[0016] Each segment of the cubic function S(t) has four coefficients: A, B, C, and D. Based on the connection requirements between each segment of the cubic spline function and the boundary constraints, the equations required to solve the coefficients A, B, C, and D are generated:

[0017]

[0018] Solve S(t) to get the fitting curve function of the device current within the sampling time window. Reconstruct and resample the data through time t to obtain the standard current data with a sampling time interval of t.

[0019] Furthermore, the method for establishing a historical standard current waveform feature library is as follows:

[0020] Through the standard current data of the equipment processing process, the characteristic values of the current data of each state are extracted to form a characteristic vector. The characteristic vector includes peak, trough, amplitude, skewness, kurtosis, wavelet transform coefficient, wavelet variance, processing time, and processing state category; and the characteristic vector of the current waveform of a single processing time under each standard processing state is used to establish a historical standard current waveform feature library.

[0021] Furthermore, the waveform characteristics of the historical standard current data include the waveform characteristics of the processing state when processing several types of products; the equipment operation state corresponding to the waveform characteristics of the historical standard current data includes the waveform characteristics of the equipment shutdown state, the waveform characteristics of the equipment standby state, and the waveform characteristics of the equipment fault state.

[0022] Furthermore, the method for extracting the characteristic value of the current data of each state is as follows:

[0023] The characteristic values are extracted using wavelet analysis; the characteristic values include peaks, troughs, amplitude, skewness, kurtosis, wavelet transform coefficients, and wavelet variance; square integrable function is a basic wavelet or wavelet mother function, and satisfies:

[0024]

[0025] After stretching and translating, a cluster of functions is obtained;

[0026]

[0027] Where a is the expansion factor and b is the translation factor; is The generated continuous wavelet function family or wavelet basis function depends on a, b; and a, b∈R, a≠0; the wavelet transform is as follows:

[0028]

[0029] W f (a, b) are wavelet transform coefficients, which contain the time-frequency domain variation characteristics of the current waveform. The square value of the wavelet coefficient is integrated over the b domain to obtain the wavelet variance. The wavelet variance reflects the distribution of the energy of the signal fluctuation with the scale a, and the expression is as follows:

[0030]

[0031] According to W f (a, b)<f(t), Extract the characteristic value of the real-time current data, where f(t) represents the signal of the standard current waveform.

[0032] Furthermore, the method for matching waveform features in S4 is as follows:

[0033] A long-short time variable sliding window is used to obtain standard current data; the long-short time variable sliding window includes a long time window and a short time window; the processing time of a single workpiece is T, the long time window obtains standard current data of the processing time of several workpieces, and the short time window obtains standard current data of the processing time of a single workpiece; the starting point of the long time window is the same as the starting point of the short time window; after the short time window obtains the real-time standard current data, waveform matching is performed according to the characteristic value, and after the characteristic value of the real-time standard current waveform of the short time window is extracted, it is compared with the characteristic value of the waveform feature of the historical standard current data in the standard current waveform feature library, and the processing waveform with the smallest difference in characteristic value with the real-time standard current waveform is selected to obtain the matching waveform of the real-time waveform feature and mark the state category. ; If the characteristic value of the real-time standard current waveform cannot match the characteristic value of the historical standard current data waveform characteristic, it is considered that the real-time standard current waveform is a non-processing waveform, and the short-time window slides to the next time point and continues to perform waveform matching; if the characteristic value of the real-time standard current waveform matches the characteristic value of the historical standard current data waveform characteristic, the short-time window slides forward for a time period of T, and the processing time T of a single workpiece becomes the processing time of the matched processing waveform, and the long-time window slides forward so that its starting point is consistent with the starting point of the short-time window; if the short-time window slides to the end of the long-time window and the real-time standard current waveform is a non-processing waveform, the short-time window slides forward for a time period of T, and the long-time window slides forward so that the starting point of the long-time window is consistent with the starting point of the segment time window.

[0034] Furthermore, the calculation method of the production cycle is as follows:

[0035] Calculate the production rhythm of the production station. The calculation formula of the production rhythm is: production rhythm = time required to produce workpieces / number of workpieces produced; the number of workpieces produced is the number of workpieces processed by the equipment in the time required to produce workpieces; the number of workpieces produced is obtained by the number of processing waveforms identified by waveform matching.

[0036] A real-time production rhythm calculation system based on waveform feature matching is applicable to the above-mentioned real-time production rhythm calculation method based on waveform feature matching, comprising: a data acquisition module, a data preprocessing module, a historical standard current waveform feature library, a waveform matching module, and a real-time calculation and display module; the data acquisition module is used to collect the original current data of the production equipment in real time; the data preprocessing module: uses the cubic spline interpolation method to preprocess the original current data, and converts the original current data with non-equal time intervals into preprocessed current data with equal time intervals; resamples the preprocessed current data to obtain standard current data with a sampling time interval of t; the historical standard current waveform feature library is used to store the waveform features of historical standard current data and the equipment operating status corresponding to the waveform features of the historical standard current data; the waveform matching module is used to match the standard current waveform features in the standard current data with the waveform features of the historical standard current data; the real-time calculation and display module is used to count the number of processing waveforms in a time period and calculate the production rhythm; and the production rhythm monitoring results of the equipment are displayed in real time.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. Use the cubic spline interpolation method to process the data into equal time intervals to reduce the complexity of data processing and analysis.

[0039] 2. By comparing the collected current data with the current data in the standard current waveform feature library, the operating status of the equipment can be judged, and equipment problems can be discovered in time when the equipment is operating abnormally.

[0040] 3. Through the calculation and real-time display of the production rhythm, the production status of the production line can be monitored in real time, and each workstation of the production line can be adjusted to improve the production efficiency of the production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Schematic diagram of the flow of the real-time production beat calculation method based on waveform feature matching.

[0042] Figure 2 This is a workflow diagram of the real-time production rhythm calculation system based on waveform feature matching. DETAILED DESCRIPTION

[0043] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] Example 1

[0045] A real-time production beat calculation method based on waveform feature matching, the steps are as follows:

[0046] S1: Real-time collection of raw current data of production equipment operation; S2: Preprocessing the raw current data using the cubic spline interpolation method, converting raw current data with non-equal time intervals into preprocessed current data with equal time intervals; resampling the preprocessed current data to obtain standard current data with a sampling time interval of t; S3: Extracting standard current waveform features based on the standard current data, and establishing a historical standard current waveform feature library, which is used to store the waveform features of historical standard current data and the equipment operating status corresponding to the waveform features of historical standard current data; S4: Matching the standard current waveform features with the waveform features of historical standard current data, judging the corresponding equipment operating status, and obtaining the number of processed waveforms based on the equipment operating status; S5: Counting the number of processed waveforms and the number of workpieces processed by the equipment within a time period, and calculating the production rhythm.

[0047] The working principle of the present invention is as follows:

[0048] Based on the cubic spline interpolation method, the parallel sampled original current data with non-uniform time intervals are interpolated and reconstructed to obtain original current data with equal time intervals. The reconstructed original current data are analyzed to realize real-time online calculation of the production station's rhythm and achieve real-time optimization of workshop production.

[0049] Preferably, the raw current data includes the serial number of each production equipment, the equipment name, the address of the collection point, the timestamp, the data identifier, and the current value.

[0050] More specifically, the steps for collecting raw current data from production equipment are as follows:

[0051] The device operating status parameters are collected and encapsulated in hexadecimal messages using the ModBus protocol. The response information is {register address, operation code, data bit, check bit}. The message needs to be parsed when transmitted to the server, and the data is extracted and converted into real values. The parsed raw current data is pushed to the streaming message queue; the data format of the raw current data is [device number, device name, collection point address, timestamp, data identifier, current value].

[0052] More specifically, the specific steps of the pre-processing are as follows:

[0053] For the original current data in the streaming message queue, a sampling time window is established based on historical data, so that the length of a sampling time window includes several production waveforms. The current value and timestamp in the original current data are extracted in the sampling time window to form a current time series signal and reconstructed using the cubic spline interpolation method; the original current data includes n+1 different time nodes t0, t1, t2, ..., t nThe current sequence data, the corresponding current values are i0, i1, i2, ..., i n ; Wherein, p=t0<t1<t2<...<t n =q, p is the starting time node of the original current data, q is the last time node of the original current data; construct a cubic spline interpolation function S(t) and make it meet the following conditions: S(t) has a second-order continuous derivative on [p, q]; S(t k1 )=y k1 , k1=0,1,2,...,n;In each subinterval [t k2 , t k2+1 ], k2=0, 1, 2, ..., n-1, S k2 (t) is a cubic polynomial; S k2 (t) is expressed as follows:

[0054]

[0055] The mathematical expression of the cubic function S(t) corresponding to n intervals is:

[0056]

[0057] Each segment of the cubic function S(t) has four coefficients: A, B, C, and D. Based on the connection requirements between each segment of the cubic spline function and the boundary constraints, the equations required to solve the coefficients A, B, C, and D are generated:

[0058]

[0059] Solve S(t) to get the fitting curve function of the device current within the sampling time window. Reconstruct and resample the data through time t to obtain the standard current data with a sampling time interval of t.

[0060] More specifically, the method for establishing a historical standard current waveform feature library is as follows:

[0061] Through the standard current data of the equipment processing process, the characteristic values of the current data of each state are extracted to form a characteristic vector. The characteristic vector includes peak, trough, amplitude, skewness, kurtosis, wavelet transform coefficient, wavelet variance, processing time, and processing state category; and the characteristic vector of the current waveform of a single processing time under each standard processing state is used to establish a historical standard current waveform feature library.

[0062] In this embodiment, the waveform characteristics of the historical standard current data include the waveform characteristics of the processing state when processing several types of products; the equipment operation state corresponding to the waveform characteristics of the historical standard current data includes the waveform characteristics of the equipment shutdown state, the waveform characteristics of the equipment standby state, and the waveform characteristics of the equipment fault state.

[0063] More specifically, the method for extracting the characteristic values of the current data of each state is as follows:

[0064] The characteristic values are extracted using wavelet analysis; the characteristic values include peaks, troughs, amplitude, skewness, kurtosis, wavelet transform coefficients, and wavelet variance; square integrable function is a basic wavelet or wavelet mother function, and satisfies:

[0065]

[0066] After stretching and translation, a cluster of functions is obtained.

[0067]

[0068] Where a is the expansion factor and b is the translation factor; is The generated continuous wavelet function family or wavelet basis function depends on a, b; and a, b∈R, a≠0; the wavelet transform is as follows:

[0069]

[0070] W f (a, b) are wavelet transform coefficients, which contain the time-frequency domain variation characteristics of the current waveform. The square value of the wavelet coefficient is integrated over the b domain to obtain the wavelet variance. The wavelet variance reflects the distribution of the energy of the signal fluctuation with the scale a, and the expression is as follows:

[0071]

[0072] According to W f (a, b)<f(t), dt extracts the characteristic value of real-time current data, where f(t) represents the signal of the standard current waveform.

[0073] More specifically, the method for matching waveform features in S4 is as follows:

[0074] A long-short time variable sliding window is used to obtain standard current data; the long-short time variable sliding window includes a long time window and a short time window; the processing time of a single workpiece is T, the long time window obtains standard current data of the processing time of several workpieces, and the short time window obtains standard current data of the processing time of a single workpiece; the starting point of the long time window is the same as the starting point of the short time window; after the short time window obtains the real-time standard current data, waveform matching is performed according to the characteristic value, and after the characteristic value of the real-time standard current waveform of the short time window is extracted, it is compared with the characteristic value of the waveform feature of the historical standard current data in the standard current waveform feature library, and the processing waveform with the smallest difference in characteristic value with the real-time standard current waveform is selected to obtain the matching waveform of the real-time waveform feature and mark the state category. ; If the characteristic value of the real-time standard current waveform cannot match the characteristic value of the historical standard current data waveform characteristic, it is considered that the real-time standard current waveform is a non-processing waveform, and the short-time window slides to the next time point and continues to perform waveform matching; if the characteristic value of the real-time standard current waveform matches the characteristic value of the historical standard current data waveform characteristic, the short-time window slides forward for a time period of T, and the processing time T of a single workpiece becomes the processing time of the matched processing waveform, and the long-time window slides forward so that its starting point is consistent with the starting point of the short-time window; if the short-time window slides to the end of the long-time window and the real-time standard current waveform is a non-processing waveform, the short-time window slides forward for a time period of T, and the long-time window slides forward so that the starting point of the long-time window is consistent with the starting point of the segment time window.

[0075] More specifically, the calculation method of the production cycle is as follows:

[0076] Calculate the production rhythm of the production station. The calculation formula of the production rhythm is: production rhythm = time required to produce workpieces / number of workpieces produced; the number of workpieces produced is the number of workpieces processed by the equipment in the time required to produce workpieces; the number of workpieces produced is obtained by the number of processing waveforms identified by waveform matching.

[0077] Example 2

[0078] A real-time production rhythm calculation system based on waveform feature matching is applicable to the real-time production rhythm calculation method based on waveform feature matching described in Example 1, including: a data acquisition module, a data preprocessing module, a historical standard current waveform feature library, a waveform matching module, and a real-time calculation and display module; the data acquisition module is used to collect the original current data of the production equipment in real time; the data preprocessing module: uses a cubic spline interpolation method to preprocess the original current data, and converts the original current data with non-equal time intervals into preprocessed current data with equal time intervals; resamples the preprocessed current data to obtain standard current data with a sampling time interval of t; the historical standard current waveform feature library is used to store the waveform features of historical standard current data and the equipment operating status corresponding to the waveform features of the historical standard current data; the waveform matching module is used to match the standard current waveform features in the standard current data with the waveform features of the historical standard current data; the real-time calculation and display module is used to count the number of processing waveforms in a time period and calculate the production rhythm; and the production rhythm monitoring results of the equipment are displayed in real time.

[0079] Example 3

[0080] A real-time production rhythm calculation system based on waveform feature matching includes: a data acquisition module, a data preprocessing module, a historical standard current waveform feature library, a waveform matching module, and a real-time calculation and display module.

[0081] The data acquisition module includes sensors, parsers, and databases installed on each production equipment; the sensors collect raw current data on the production equipment and transmit the raw current data to the parser for parsing and processing, the parsed raw current data is transmitted to the streaming message queue of the corresponding equipment and stored in the storage unit for storage, and the parsed raw current data is transmitted to the data preprocessing module; the raw current data includes the number of each production equipment, the equipment name, the address of the collection point, the timestamp, the data identifier, and the current value.

[0082] The data preprocessing module obtains the parsed raw current data from the streaming message queue, performs preprocessing, and uses the cubic spline interpolation method to perform spline interpolation on the parsed raw current data to obtain an original current fitting function, which fits the current value at each acquisition time point; the system obtains standard current data with an acquisition time interval of 4 seconds by taking the time t=4 seconds in the current fitting function.

[0083] The data preprocessing process is as follows: establish a sampling time window, obtain the parsed raw current data containing several production and processing waveform times from the streaming message queue, and the data content format is [device number, device name, acquisition point address, timestamp, data identifier, current value]; extract the current value and timestamp from the data through the time window to form a non-uniform time interval current time series signal; for the non-uniform time interval current time series signal, it consists of n+1 different time nodes t0, t1, t2, ..., t n The current sequence data consists of the corresponding current values i0, i1, i2, ..., i n ; Wherein, p=t0<t1<t2<...<t n =q, p is the starting time node of the original current data, and q is the last time node of the original current data; construct a cubic spline interpolation function S(t); solve S(t) to obtain the fitting curve function of the device current in the time window, reconstruct the data through time t=4s, and obtain standard current data sampled at an equal time interval of 4s.

[0084] The waveform feature extraction module extracts the standard current waveform features of the processing time of a single workpiece under each standard processing state based on the standard current data, and constructs a historical standard current waveform feature library.

[0085] The feature extraction process is as follows: obtain the standard current waveform of a single processing time under each standard processing state, including the current waveforms of the processing state, equipment shutdown state, equipment standby state, and equipment fault state when processing several products; extract the features of each standard current waveform, including peak, trough, amplitude, skewness, kurtosis, wavelet transform coefficient, and wavelet variance; the wavelet transform coefficient is given by The wavelet variance is obtained by Obtain; extract the characteristic values of each state to form a characteristic vector, wherein the characteristic vector includes peak, trough, amplitude, skewness, kurtosis, wavelet transform coefficient, wavelet variance, processing time, and processing state category; establish a historical standard current waveform feature library.

[0086] The waveform matching module uses a long-short time sliding window to obtain standard current waveform data of the message queue, and performs waveform matching to obtain the current waveform category.

[0087] The steps of waveform matching are as follows: using a long and short time variable sliding window to obtain standard current data; the long and short time variable sliding window includes a long time window and a short time window; the processing time of a single workpiece is T, the long time window obtains the standard current data of the processing time of several workpieces, and the short time window obtains the standard current data of the processing time of a single workpiece; the starting point of the long time window is the same as the starting point of the short time window; after the short time window obtains the real-time standard current data, waveform matching is performed according to the characteristic value, and after the characteristic value of the real-time standard current waveform of the short time window is extracted, it is compared with the characteristic value of the waveform feature of the historical standard current data in the standard current waveform feature library, and the processing waveform with the smallest difference in characteristic value with the real-time standard current waveform is selected to obtain the matching waveform of the real-time waveform feature and mark it. Record the state category; if the characteristic value of the real-time standard current waveform cannot match the characteristic value of the historical standard current data waveform characteristic, it is considered that the real-time standard current waveform is a non-processing waveform, and the short-time window slides to the next time point and continues to perform waveform matching; if the characteristic value of the real-time standard current waveform matches the characteristic value of the historical standard current data waveform characteristic, the short-time window slides forward for a duration of T, and the single workpiece processing time T becomes the processing time of the matched processing waveform, and the long-time window slides forward so that its starting point is consistent with the starting point of the short-time window; if the short-time window slides to the end of the long-time window and the real-time standard current waveform is a non-processing waveform, the short-time window slides forward for a duration of T, and the long-time window slides forward so that the starting point of the long-time window is consistent with the starting point of the segment time window.

[0088] The real-time calculation and display module calculates the production rhythm and displays it in real time. The calculation method of the production rhythm is as follows:

[0089] Calculate the production rhythm of the production station. The calculation formula of the production rhythm is: production rhythm = number of production workpieces / time required to produce workpieces; the number of production workpieces is the number of workpieces processed by the equipment in the time required to produce workpieces; the number of production workpieces is obtained by the number of processing waveforms identified by waveform matching.

[0090] Obviously, the above embodiments of the present invention are merely examples for the purpose of illustrating the present invention, and are not intended to limit the embodiments of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A real-time production beat calculation method based on waveform feature matching, characterized in that: Here are the steps: S1: Real-time collection of raw current data of production equipment operation; S2: using the cubic spline interpolation method to preprocess the raw current data, converting the raw current data with non-uniform time intervals into preprocessed current data with equal time intervals; Resampling the preprocessed current data to obtain standard current data with a sampling time interval of t; S3: extracting standard current waveform features based on standard current data, and establishing a historical standard current waveform feature library, wherein the historical standard current waveform feature library is used to store the historical standard current data waveform features and the device operating status corresponding to the historical standard current data waveform features; S4: Match the standard current waveform characteristics with the historical standard current data waveform characteristics, determine the corresponding equipment operating status, and obtain the number of processed waveforms based on the equipment operating status; S5: Count the number of processing waveforms and the number of workpieces processed by the equipment within a time period to calculate the production cycle.

2. The method for calculating real-time production rhythm based on waveform feature matching according to claim 1, characterized in that: The original current data includes the serial number of each production equipment, the equipment name, the address of the collection point, the timestamp, the data identifier, and the current value.

3. The method for calculating real-time production rhythm based on waveform feature matching according to claim 2, characterized in that: The specific steps for collecting the raw current data of production equipment operation are as follows: The device's operating status parameters are collected and encapsulated in hexadecimal format using the ModBus protocol. The response information is {register address, operation code, data bit, check bit}. The message is transmitted to the server for parsing, the data is extracted and converted into real values, and the parsed raw current data is pushed to the streaming message queue. The data format of the original current data is [device number, device name, address of the acquisition point, timestamp, data identifier, current value].

4. The method for calculating real-time production rhythm based on waveform feature matching according to claim 3 is characterized in that: The specific steps of the pretreatment are as follows: For the raw current data in the streaming message queue, a sampling time window is established based on historical data. The length of a sampling time window includes several production waveforms. The current value and timestamp in the raw current data are extracted in the sampling time window to form a current time series signal, which is then reconstructed using the cubic spline interpolation method. The raw current data includes n+1 different time nodes t0, t1, t2, ..., t n The current sequence data, the corresponding current values are i0, i1, i2, ..., i n ; Where p = t0 <t1<t2<...<t n =q, p is the starting time node of the original current data, q is the last time node of the original current data; construct a cubic spline interpolation function S(t) and make it meet the following conditions: S(t) has a second-order continuous derivative on [p, q; S(t k1 )=y k1 ,k1=0,1,2,...,n;in each subinterval [t k2 , t k2+1 ], k2=0,1,2,...,n-1, S k2 (t) is a cubic polynomial; S k2 (t) is expressed as follows: The mathematical expression of the cubic function S(t) corresponding to n intervals is: Each cubic function S(t) has four coefficients A, B, C, and D; Based on the connection requirements between each piecewise function of the cubic spline function and the boundary constraints, the equations required to solve the coefficients A, B, C, and D are generated: Solve S(t) to get the fitting curve function of the device current within the sampling time window. Reconstruct and resample the data through time t to obtain the standard current data with a sampling time interval of t.

5. The method for calculating real-time production rhythm based on waveform feature matching according to claim 4, characterized in that: The method for establishing a historical standard current waveform feature library is as follows: Through the standard current data of the equipment processing process, the characteristic values of the current data of each state are extracted to form a characteristic vector. The characteristic vector includes peak, trough, amplitude, skewness, kurtosis, wavelet transform coefficient, wavelet variance, processing time, and processing state category; and the characteristic vector of the current waveform of a single processing time under each standard processing state is used to establish a historical standard current waveform feature library.

6. The method for calculating real-time production rhythm based on waveform feature matching according to claim 4, characterized in that: The waveform characteristics of the historical standard current data include the waveform characteristics of the processing state when processing several types of products; the equipment operation state corresponding to the waveform characteristics of the historical standard current data includes the waveform characteristics of the equipment shutdown state, the waveform characteristics of the equipment standby state, and the waveform characteristics of the equipment fault state.

7. The method for calculating real-time production rhythm based on waveform feature matching according to claim 6, characterized in that: The method for extracting the characteristic values of the current data of each state is as follows: Extracting eigenvalues using wavelet analysis; the eigenvalues include peaks, troughs, amplitude, skewness, kurtosis, wavelet transform coefficients, and wavelet variance; Square integrable function is a basic wavelet or wavelet mother function, and satisfies: After stretching and translating, a cluster of functions is obtained; Where a is the expansion factor and b is the translation factor; is The generated continuous wavelet function family or wavelet basis function depends on a, b; and a, b∈R, a≠0; The wavelet transform is as follows: W f (a, b) are wavelet transform coefficients, including the time-frequency domain variation characteristics of the current waveform; Integrating the square of the wavelet coefficients over the b domain yields the wavelet variance, which reflects the distribution of the energy of the signal fluctuation with scale a, as expressed in the following way: According to W f (a,b) <f(), Extract the characteristic value of the real-time current data, where f(t) represents the signal of the standard current waveform.

8. The method for calculating real-time production rhythm based on waveform feature matching according to claim 7, characterized in that: The method for matching waveform features in S4 is as follows: A long-short time variable sliding window is used to obtain standard current data; the long-short time variable sliding window includes a long time window and a short time window; The processing time of a single workpiece is T. The long time window obtains the standard current data of the processing time of several workpieces, and the short time window obtains the standard current data of the processing time of a single workpiece. The starting point of the long time window is the same as the starting point of the short time window. After acquiring real-time standard current data in a short time window, waveform matching is performed based on the characteristic value. After extracting the characteristic value of the real-time standard current waveform in the short time window, it is compared with the characteristic value of the waveform characteristics of historical standard current data in the standard current waveform characteristic library. The processed waveform with the smallest difference in characteristic value with the real-time standard current waveform is selected to obtain the matching waveform of the real-time waveform characteristics and mark the state category; If the characteristic value of the real-time standard current waveform cannot match the characteristic value of the historical standard current data waveform, the real-time standard current waveform is considered to be a non-processed waveform, and the short time window slides to the next time point and continues the waveform matching; If the characteristic value of the real-time standard current waveform matches the characteristic value of the historical standard current data waveform, the short-time window slides forward by T time, and the processing time of a single workpiece T becomes the processing time of the matching processing waveform, and the long-time window slides forward so that its starting point is consistent with the starting point of the short-time window; If the short time window slides to the end of the long time window and the real-time standard current waveform is a non-processed waveform, the short time window slides forward for T time and the long time window slides forward so that the starting point of the long time window is consistent with the starting point of the short time window.

9. The method for calculating real-time production rhythm based on waveform feature matching according to claim 8, characterized in that: The calculation method of the production cycle is as follows: Calculate the production rhythm of the production station. The calculation formula of the production rhythm is: production rhythm = time required to produce workpieces / number of workpieces produced; the number of workpieces produced is the number of workpieces processed by the equipment in the time required to produce workpieces; the number of workpieces produced is obtained by the number of processing waveforms identified by waveform matching.

10. A real-time production takt calculation system based on waveform feature matching, applicable to the real-time production takt calculation method based on waveform feature matching according to any one of claims 1 to 9, characterized in that: include: Data acquisition module, data preprocessing module, historical standard current waveform feature library, waveform matching module, real-time calculation and display module; The data acquisition module is used to collect the original current data of the production equipment in real time; The data preprocessing module: uses a cubic spline interpolation method to preprocess the original current data, and converts the original current data with unequal time intervals into preprocessed current data with equal time intervals; Resampling the preprocessed current data to obtain standard current data with a sampling time interval of t; A historical standard current waveform feature library is used to store the historical standard current data waveform features and the equipment operating status corresponding to the historical standard current data waveform features; The waveform matching module is used to match the standard current waveform characteristics in the standard current data with the waveform characteristics of the historical standard current data; The real-time calculation and display module is used to count the number of processing waveforms within a time period and calculate the production rhythm; and to display the production rhythm monitoring results of the equipment in real time.

Images