An electronic belt scale running state diagnosis method based on data analysis

Abstract

The application discloses a kind of based on data analysis electronic belt scale operating condition diagnosis method, comprising: acquisition and storage weighing sensor's code value, tare value and weighing value;Calculate the distribution difference of weighing value data, judge whether to occur partial load;Based on least square fitting weighing value data, judge whether to block material;In not partial load and not block material, based on code value and tare value to weighing sensor state is detected.The application can real-time monitoring, diagnose the operating condition of electronic belt scale.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of operating state detection of intelligent electronic scales, and more particularly to an electronic belt scale operating state diagnosis method based on data analysis. BACKGROUND

[0002] The electronic belt scale is an important process equipment in the tobacco leaf threshing and redrying and silk production line, and plays a crucial role in tobacco quality control. Although the current electronic belt scale has a high degree of automation, it only realizes automatic control and stays at the level of industrial equipment. In maintenance and adjustment, multiple people must work together, important parameters and data lack collection, analysis and storage, and the health state of the device lacks detection and early warning, which is difficult to meet the development needs of the overall intelligentization of the tobacco production line.

[0003] Therefore, how to provide an electronic belt scale operating state diagnosis method based on data analysis is a problem that those skilled in the art need to solve. SUMMARY

[0004] In view of this, the present application provides an electronic belt scale operating state diagnosis method based on data analysis, which can monitor and diagnose the operating state of the electronic belt scale in real time.

[0005] In order to achieve the above purpose, the present application adopts the following technical solutions:

[0006] An electronic belt scale operating state diagnosis method based on data analysis, comprising:

[0007] Collecting and storing the code value, tare value and weighing value of the weighing sensor;

[0008] Calculating the distribution difference of the weighing value data to determine whether the load is unbalanced;

[0009] Fitting the weighing value data based on the least square method to determine whether the material is blocked;

[0010] When the load is not unbalanced and the material is not blocked, the state of the weighing sensor is detected based on the code value and the tare value.

[0011] Preferably, the weighing sensors are symmetrically arranged at both ends below the metering roller, the two weighing sensors below the single metering roller form a weighing sensor module, the single metering roller and the two weighing sensors below it form a scale platform, and the weighing value is the scale platform weighing value, which is the average of the weighing values of the two weighing sensors.

[0012] Preferably, the distribution difference of the weighing value data is judged based on relative entropy, and the average value data distribution of two scale platforms in a period of time is P(X) and Q(X), X represents the time set of the weighing data sampling, and the relative entropy is calculated based on one of the weighing value data distributions as a reference. The specific calculation formula of the relative entropy is:

[0013]

[0014] wherein E = min[(minP(X),minQ(X))]

[0015]

[0016]

[0017] and denote the set of weighing values collected by the two scales during operation, and Δt denotes a minimum constant;

[0018] If the relative entropy approaches the set threshold KL[P(X)||Q(X)]→δ, it indicates that the load is unbiased, otherwise, it is biased, and an alarm information is sent out.

[0019] Preferably, the specific process of judging whether the material is blocked based on the least square method fitting the weighing value data is as follows:

[0020] Suppose the weighing value of the scale downstream of the conveyor belt within a certain time is wherein 1, 2, …, t denote time, and the fitting function is For a minimum value ε, according to the definition of the least square method, the following formula can be established:

[0021]

[0022] Based on the coefficient of determination R 2 Further judge the fitting degree of the fitting function, wherein the calculation formula of the coefficient of determination is:

[0023]

[0024] wherein, is the mean value of the weighing value Y 1 , and e t denotes the residual error between the estimated value and the actual value of each weighing value, and the calculation formula is:

[0025]

[0026] wherein, the weighing value Y 1 of the scale, and all the estimated values

[0027] When and only when R 2 is greater than the preset value, the material blocking is judged, otherwise, it is determined that the fitting function accuracy does not reach the predetermined value, and no further material blocking judgment is performed.

[0028] When the blockage judgment is performed, the specific blockage judgment process is as follows: if the weighing value of the scale table at the downstream of the conveying belt continuously increases, the weighing value of the scale table at the upstream is further judged, if the weighing value of the scale table at the upstream also continuously increases at the next time period, and the data distribution of the two scale tables is similar, the blockage is determined.

[0029] Preferably, the code value of the current weighing sensor module is evaluated at the same weighing value and compared with the historical code value corresponding to the same weighing value stored, if the code value difference is greater than a certain limit, it indicates that the weighing sensor module is faulty, then the code value corresponding to the current tare value is compared with the code value data corresponding to the historical tare value stored, to judge whether the state of the weighing sensor module is normal, so as to locate the faulty weighing sensor module.

[0030] Compared with the prior art, the advantages of the present application are as follows: (1) the problem that the previous scale table cannot utilize historical data is solved, on the basis of real-time detection of the scale table state, a more scientific method is used to diagnose the running state of the scale table in real time, so that the result is more accurate and reliable; (2) the historical data is fully utilized to realize real-time monitoring of the state of the weighing sensor module; (3) compared with the previous use of PLC and other closed-source platforms, the algorithms of the present application are completed on an open-source platform, which has better expandability. BRIEF DESCRIPTION OF DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the provided drawings.

[0032] Figure 1 The attached drawing is a flow chart of an electronic belt scale running state diagnosis method based on data analysis provided by the present application.

[0033] Figure 2 The attached drawing is a partial load schematic diagram in the embodiments provided by the present application.

[0034] Figure 3 The attached drawing is a blockage schematic diagram provided by the present application.

[0035] In the drawing, 1 is a weighing sensor, and 2 is uneven material. DETAILED DESCRIPTION

[0036] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be clearly and completely described below, obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work belong to the scope of protection of the present application.

[0037] The embodiment of the present application discloses a data analysis-based electronic belt scale running state diagnosis method, which firstly judges the load deviation and end blockage, thereby further realizing the state evaluation of the weighing module. Specifically, the method comprises the following steps:

[0038] (a) acquiring the code value, tare value and weighing value of the electronic belt scale and saving them in real time. For the convenience of description, the code value, tare value and weighing value are described as follows: the code value is the original data obtained by converting the analog quantity collected by the weighing sensor through the PLC, and the weight value corresponding to the code value under the empty load is the tare value. The tare value directly reflects the weight of the belt on the measurement section, so a set of tare value data should cover a whole circle of the belt. The weighing value is the total weight value obtained by converting the code value of the measured material weight through the analog-digital conversion when the electronic scale is normally loaded, and the weighing value is the value obtained by subtracting the tare value from the weight measured by the scale table under the load. The relationship among the three can be simply represented as: if the code value of a certain weighing sensor is m, the tare value is w, and the actual load of the sensor is M, then there is a mapping function f(x) such that the following formula is established:

[0039] f(m)=w+M

[0040] The mapping function is determined when the weighing sensor is factory-calibrated and the scale is calibrated, and represents the mapping relationship between the code value and the actual load of the weighing sensor. When M is 0, i.e. in the empty load condition, the code value is calculated to obtain the tare value w through the calibration. When the material passes, the material weight M can be calculated according to the measured f(m) and tare value w. Therefore, the three need to be measured in real time and converted in real time according to the above mutual relationship. The single measurement roller and the two weighing sensors below can be regarded as a scale table, which is called an emergency scale, and the weighing value of the scale table is the average value of the weighing values of the weighing sensors. The two weighing sensors below the single measurement roller are the weighing sensor module.

[0041] (b) calculating the distribution difference of the weighing value data to judge whether the load deviation occurs;

[0042] (c) fitting the weighing value data based on the least square method to judge whether the material is blocked;

[0043] (d) when there is no load deviation and no material blocking, detecting the state of the weighing sensor based on the code value and the tare value.

[0044] There is no order of precedence for overload judgment and material blockage judgment, and they can be performed simultaneously. Alternatively, the overload can be calculated first and then whether there is a material blockage is determined, or whether there is a material blockage is determined first and then whether there is an overload is calculated.

[0045] This embodiment uses linear regression and relative entropy to determine the load imbalance, end-of-line blockage, and lifespan of the weighing module of an electronic belt scale. When the slope of the regression function or the relative entropy meets specified conditions and exceeds a threshold, the belt scale's operating status is determined and promptly fed back to the interactive display. While the load imbalance and end-of-line blockage can be determined simultaneously without conflict, the weighing module lifespan must be determined when neither load imbalance nor end-of-line blockage is present.

[0046] In this embodiment, four load cells are provided, located at the two ends below the two metering rollers, together forming the weighing module of the electronic belt scale. This embodiment is applicable to electronic belt scales with four load cells, two rollers, and direct load bearing.

[0047] Specifically, such as Figure 2 As shown in the figure, uneven material 2 may cause uneven loading. The uneven loading determination process is as follows: Load cell 1 is divided into two groups, one on the left and one on the right, according to the material's orientation. The weight value of each group is the average of the two sensors. If the weighing data from the two groups of load cells 1 are consistent, then the average values ​​of the two groups have similar mathematical distributions. Therefore, only the degree of difference in the data distribution between the two groups of sensors needs to be checked. If the difference is small, the data can be considered consistent. Data consistency is calculated using the preprocessed relative entropy formula. Relative entropy measures the distance between two probability distribution functions. In electronic belt scale platforms, data is a discrete random variable.

[0048] Suppose there are two sets of weight distributions P(X) and Q(X) from a scale over a period of time, where X represents the time set of the weight data sampling. Taking one of the load distributions (such as P(X)) as the reference, it is defined as:

[0049]

[0050] E=min[(minP(X),minQ(x))]

[0051]

[0052]

[0053] Where Δt represents a very small constant, which is introduced to prevent the numerator or denominator from being 0 and affecting the judgment. and represents the set of weighing values ​​collected by the scale during operation. Substituting P(X) and Q(X) into the KL calculation formula can obtain the relative entropy value.

[0054] The difference between P(X) and Q(X) can be obtained (or the degree of distribution consistency). If the difference is smaller, the relative entropy is smaller, that is, the materials on both sides are more balanced, and vice versa, the more serious the load is. When the distribution is completely consistent (in the most ideal case), the relative entropy is 0. Therefore, given a threshold δ, when KL[P(X)||Q(X)]→δ, it can be considered that the two distributions are consistent.

[0055] Specifically, the end blockage judgment process is as follows: Figure 3 During the blockage process, the material will gradually accumulate forward from the end of the conveyor, so the scale table downstream of the conveyor belt will measure the rising trend of the weight earlier than the scale table upstream. Whether the scale table is loaded or not, the trend of the weighing value of the two scales during the blockage period should be consistent. Therefore, the least square method is used to fit the weighing value in a certain period, so as to judge the trend.

[0056] Let the weighing value of the scale table downstream of the conveyor belt in a certain period be Y Where 1, 2, …, t represent time, and the fitting function is Y According to the definition of the least square method, for a very small value ε, the following formula can be established:

[0057]

[0058] The function of the above formula is to simulate the trend of the weighing value in a certain period t using the fitting function. After obtaining the fitting function, the coefficient of determination R 2 is used to further judge the fitting degree of the fitting function, so as to ensure the accuracy of the judgment of blockage or not.

[0059] The calculation formula of the coefficient of determination is:

[0060]

[0061] Where, is the mean value of the weighing value Y 1 , and e t represents the residual error between the estimated value and the actual value of each weighing value, and the calculation formula is:

[0062]

[0063] Where, the weighing value Y 1 of the scale table, and all estimated values

[0064] R 2The greater, the greater the proportion of the error in the total error that explains the sample regression, the better the fitting degree, indicating that the system can grasp the trend of the current scale weight, otherwise the proportion is smaller, the fitting degree is worse, and the system grasps the trend of the current scale weight worse. Therefore, only when the determination function is greater than a certain value (which needs to be determined by experiment), the blocking material is determined, otherwise the fitting function accuracy does not meet the predetermined value, and no further blocking material determination is performed, at this time, on the one hand, the risk of blocking material is prompted, and on the other hand, the above detection is continued to be executed until the fitting function accuracy reaches the predetermined value, triggering further blocking material determination;

[0065] In order to ensure R 2 Under certain conditions, if the scale weight downstream of the conveying belt continues to increase (its derivative continues to be greater than a certain preset value, indicating that the trend of the fitting function f(t) continues to rise), the scale weight upstream of the scale is further judged, if the scale weight upstream of the scale in the next time period is also increasing, and the distribution of the two scale data is similar (the distribution similarity judgment is the same as the load bias judgment, and the relative entropy of the scale weight of the two scales is calculated), it can be determined that the material is blocked, and the "stop" measure is taken to ensure production safety.

[0066] Specifically, the state evaluation of the weighing sensor module is mainly judged from the code value of the weighing sensor, and a two-stage judgment method is used in the evaluation process. The idea of judgment is: after each calibration, the mapping function f(m) of the code value and the actual load will change, that is, the same load before and after calibration will correspond to different code values. Therefore, after calibration, the real-time data generated during operation is used for judgment, and under the premise of no load bias and blocking material characteristics, if the weighing sensor module is fault-free or in good condition, the difference between the code values under the same scale weight at different times should be small or the same. If the difference between the code values under the same scale weight at different times is large, it can be preliminarily determined that the weighing sensor module may have a fault. That is:

[0067] |m 当前值 -m 历史值 |<Δm

[0068] Where Δm is the preset tolerance limit of the code value difference.

[0069] Further, if it is preliminarily determined that a certain weighing sensor module may have a fault, the early tare weight value and code value historical data of the weighing sensor module are called from the database for further analysis. Let the code value set corresponding to the early tare weight value of the weighing sensor module be The code value set corresponding to the current tare weight value Where i represents the position of the tare measurement. Theoretically, the size of the tare weight at each position of the belt is equal, so when KL[m 0 ||m 1When the weight is greater than a certain value, it can be determined that the weight channel is malfunctioning. This two-stage judgment method can effectively reduce the data query frequency, save the IPC computing resources, and improve the system running speed compared with directly comparing the real-time data with the historical data.

[0070] Due to the special structure, the embodiment is also applicable to a direct load bearing type electronic belt scale with more than four weighing sensors and multiple rollers.

[0071] In summary, the present application has the following advantages:

[0072] (1) The present application fills the gap that the past belt scale cannot use historical data. The data on the belt scale is effectively used, a mathematical relationship is established between the data, and a more objective judgment of the state of the scale is achieved.

[0073] (2) Compared with the existing invention, the present application can more accurately locate the cause of the abnormal work of the belt scale. The data on each scale with a clear mathematical relationship can provide more effective reference for the maintenance work of the belt scale.

[0074] In the specification, each embodiment is described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts of each embodiment can be referred to each other. For the device disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the related parts can be referred to the method part.

[0075] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for diagnosing the operating status of an electronic belt scale based on data analysis, characterized in that: include: Collect and store the code value, tare value and weighing value of the weighing sensor; Calculate the distribution difference of weighing value data to determine whether overloading occurs; Based on the least square method to fit the weighing value data, determine whether there is material blockage; When there is no unbalanced load or material blockage, the status of the weighing sensor is detected based on the code value and tare value; Based on relative entropy, we can judge the distribution difference of weighing value data. Assume that the average value data distribution on two weighing platforms within a period of time is P(X) and Q(X), where X represents the time set of weighing data sampling. Taking the distribution of one weighing value data as a reference, the specific calculation formula of relative entropy is: Where E = min[(minP(X), minQ(X))] and It represents the set of weighing values ​​collected by the two scales during operation, and Δt represents a minimum constant; If the relative entropy approaches the set threshold value KL[P(X)||Q(X)]→δ, it means there is no unbalanced load, otherwise it is unbalanced load and an alarm message is issued; Based on the least squares method to fit the weighing value data, the specific process of judging whether there is material blockage is as follows: Assume that the weighing value of the scale downstream of the conveyor belt within a certain period of time is Where 1, 2, ..., t represents time, and the fitting function For a minimum value ε, according to the definition of least squares method, the following formula can be established: Based on the coefficient of determination R 2 Further judge the goodness of fit of the fitting function, where the calculation formula of the coefficient of determination is: in, is the weighing value Y 1 The mean value, e t It represents the residual between the estimated value and the actual value of each weighing value. The calculation formula is: Among them, the weighing value of the weighing platform is Y 1 And all estimates are obtained by least squares method If and only if R 2 If it is greater than the preset value, it will be judged as material blocking. Otherwise, it is determined that the fitting function accuracy does not meet the preset value and no further material blocking judgment will be performed. When judging material blockage, the specific process is as follows: if the weighing value of the scale at the downstream of the conveyor belt continues to increase, the weighing value of the upstream scale is further judged. If the weighing value of the upstream scale is also increasing in the next time period, and the data distribution of the two scales is similar, it is determined to be a material blockage.

2. The method for diagnosing the operating status of an electronic belt scale based on data analysis according to claim 1, characterized in that: The weighing sensors are symmetrically arranged at both ends below the metering roller. The two weighing sensors below the single metering roller are weighing sensor modules. The single metering roller and the two weighing sensors below act as a weighing platform. The weighing value is the weighing value of the weighing platform, and the average weighing value of the two weighing sensors is taken.

3. The method for diagnosing the operating status of an electronic belt scale based on data analysis according to claim 2, characterized in that: Evaluate the code value of the current weighing sensor module under the same weighing value and compare it with the historical code value corresponding to the same stored weighing value. If the code value difference is greater than a certain limit, it indicates that the weighing sensor module has failed. Then, further compare the code value corresponding to the current period's tare value with the code value data corresponding to the stored historical tare value to determine whether the weighing sensor module status is normal, thereby locating the weighing sensor module with the problem.

Images