Sensor-based grain-ferment mixing process production data monitoring system

Abstract

The present invention relates to the field of data monitoring technology, and in particular to a sensor-based grain-ferment mixing process production data monitoring system. The system includes a processor and a memory, and the processor executes a computer program stored in the memory to implement the following steps: according to the initial order value of each characteristic subsequence in the characteristic subsequence set, a predicted temperature data sequence corresponding to the characteristic subsequence is obtained, and the error coefficient corresponding to the initial order value of the characteristic subsequence is obtained according to the predicted temperature data sequence; according to the initial order value of the current target subsequence and the error coefficient corresponding to the initial order value of the characteristic subsequence, the target order value of the current target subsequence is obtained; according to the ARIMA model, the target order value and the current target subsequence, the predicted temperature data at the future moment is obtained, and the temperature monitoring result in the reaction chamber at the current monitoring moment is obtained according to the predicted temperature data. And the present invention can make the obtained temperature monitoring result in the reaction chamber more reliable.

Description

Technical Field

[0001] The present invention relates to the technical field of data monitoring, and in particular to a sensor-based grain-wheat mixing process production data monitoring system. Background Art

[0002] The grain-mash mixing process is a key operation step in the production process of fermented foods, and is usually used in the manufacture of fermented foods such as brewed alcohol (such as white wine), vinegar, and soy sauce. The grain-mash mixing process is generally carried out in a reaction chamber; however, since the temperature in the reaction chamber will directly affect the efficiency and yield of the grain-mash processing, in order to ensure the efficiency and yield of the grain-mash processing, during the grain-mash mixing process production process, it is usually necessary to monitor the temperature in the reaction chamber, and subsequently adjust the temperature in the reaction chamber based on the monitoring results, so as to control the temperature in the reaction chamber within a specified range.

[0003] Normally, the temperature in the reaction chamber is generally collected by sensors. However, due to the time-delay characteristic of temperature transmission, the temperature data collected by the sensor in real time during temperature monitoring is not the actual temperature in the reaction chamber. Therefore, if the temperature monitoring result in the reaction chamber at the corresponding monitoring time is obtained directly based on the temperature data collected by the sensor, the monitoring result will be inaccurate, which will lead to control deviation or inaccurate control when the temperature in the reaction chamber is subsequently controlled. Therefore, in order to obtain accurate monitoring results, the predicted temperature is generally obtained based on the temperature data collected by the sensor and the ARIMA model (autoregressive integrated moving average model). Temperature data is obtained, and the temperature monitoring result in the reaction chamber at the corresponding monitoring time is determined based on the predicted data. However, when predicting based on the temperature data collected by the sensor and the ARIMA model, a fixed order value is usually used or the order value is set to an empirical value. The order value is the order value in the ARIMA model and is also the order value of the autoregressive part. This method of determining the order value will cause the subsequent predicted data to be unreliable or inaccurate, which will in turn cause the subsequent monitoring results determined by the analysis of the predicted data to be inaccurate and unreliable, which will also affect the efficiency and yield of the grain mash processing. Therefore, how to adaptively obtain the order value to improve the accuracy of the prediction is an urgent problem to be solved. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a sensor-based grain-wheat mixing process production data monitoring system, and the technical solution adopted is as follows:

[0005] An embodiment of the present invention provides a sensor-based grain-wheat mixing process production data monitoring system, comprising a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the following steps:

[0006] In the reaction chamber of the grain-wheat mixing process, a sensor is used to obtain a temperature data sequence to be analyzed corresponding to a target monitoring time period, wherein the last monitoring moment in the target monitoring time period is the current monitoring moment;

[0007] A target subsequence set is obtained according to a preset sliding window and the temperature data sequence to be analyzed; a change characteristic value of the target subsequence is obtained according to the total number of peaks in each target subsequence in the target subsequence set; a first target subsequence in the target subsequence set and an initial order value of the first target subsequence are obtained according to the change characteristic value of the target subsequence and the temperature data in the target subsequence;

[0008] According to the initial order value of each characteristic subsequence in the characteristic subsequence set, a predicted temperature data sequence corresponding to the characteristic subsequence is obtained, and according to the predicted temperature data sequence, an error coefficient corresponding to the initial order value of the characteristic subsequence is obtained; the characteristic subsequence set refers to a set constructed by all first target subsequences except the current target subsequence, and the current target subsequence refers to the first target subsequence whose acquisition time is closest to the current monitoring time;

[0009] Obtaining a target order value of the current target subsequence according to the initial order value of the current target subsequence and the error coefficient corresponding to the initial order value of the characteristic subsequence;

[0010] According to the ARIMA model, the target order value and the current target subsequence, the predicted temperature data at a future moment is obtained, and according to the predicted temperature data, the temperature monitoring result in the reaction chamber at the current monitoring moment is obtained.

[0011] Beneficial effects: The present invention first uses a sensor to obtain a temperature data sequence to be analyzed corresponding to a target monitoring time period in a reaction chamber of a grain-wheat mixing process production process; then, a target subsequence set is obtained according to a preset sliding window and the temperature data sequence to be analyzed, and a change characteristic value of the target subsequence is obtained according to the total number of peak values ​​in each target subsequence in the target subsequence set, and a first target subsequence in the target subsequence set and an initial order value of the first target subsequence are obtained according to the change characteristic value of the target subsequence and the temperature data in the target subsequence; then, a predicted temperature data sequence corresponding to the feature subsequence is obtained according to the initial order value of each feature subsequence in the feature subsequence set, and an error coefficient corresponding to the initial order value of the feature subsequence is obtained according to the predicted temperature data sequence; then, a target order value of the current target subsequence is obtained according to the initial order value of the current target subsequence and the error coefficient corresponding to the initial order value of the feature subsequence; finally, the predicted temperature data at a future moment is obtained according to the ARIMA model, the target order value and the current target subsequence, and the temperature monitoring result in the reaction chamber at the current monitoring moment is obtained according to the predicted temperature data. Moreover, the accuracy of the predicted data obtained according to the target order value of the present invention is higher, that is, the predicted data obtained according to the target order value is more accurate and reliable. Therefore, the accuracy and reliability of monitoring the temperature in the reaction chamber can be improved through the obtained predicted temperature data, that is, based on the predicted temperature data, the temperature monitoring results in the reaction chamber can be more reliable and accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0013] Figure 1 The present invention is a flow chart of a sensor-based method for monitoring production data of a grain-wheat mixing process. DETAILED DESCRIPTION

[0014] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the embodiments of the present invention.

[0015] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0016] This embodiment provides a sensor-based grain-wheat mixing process production data monitoring system, including a processor and a memory, wherein the processor executes a computer program stored in the memory to implement a sensor-based grain-wheat mixing process production data monitoring method, such as Figure 1 As shown, the sensor-based grain-wheat mixing process production data monitoring method comprises the following steps:

[0017] Step S001, in a reaction chamber of a grain-wheat mixing process, a sensor is used to obtain a temperature data sequence to be analyzed corresponding to a target monitoring time period.

[0018] Currently, predicted temperature data is usually obtained based on the temperature data collected by the sensor and the ARIMA model, and the temperature monitoring result in the reaction chamber at the corresponding monitoring time is determined based on the predicted data. However, when predicting based on the temperature data collected by the sensor and the ARIMA model, the order value in the ARIMA model usually uses a fixed order value or an empirical value. As the production stage of the grain mash mixing process is different, the temperature in the reaction chamber will vary in different time periods. When the temperature in the reaction chamber has different characteristics in terms of temperature variation at different times, the above-described method of determining the order value will cause the subsequent predicted data to be inconsistent. The problem of reliability or inaccuracy will lead to inaccurate and unreliable monitoring results determined based on the predicted data, which will also affect the efficiency and yield of grain mash processing. In order to make the monitoring results more accurate, this embodiment will adaptively obtain the order value in the prediction model based on the temperature change in the reaction chamber, so that the obtained prediction value is more accurate and reliable, and then based on the obtained prediction value, the purpose of reliably and accurately monitoring the temperature in the reaction chamber is achieved. In addition, for ease of understanding, this embodiment will take any type of grain mash mixing process production process as an example for analysis, such as vinegar, wine, etc., and the reaction chamber that appears subsequently refers to the same reaction chamber.

[0019] Based on the above description, it can be seen that this embodiment is mainly based on the temperature change in the reaction chamber to adaptively obtain the order value in the prediction model. Therefore, this embodiment needs to use a sensor in the reaction chamber of the grain mash mixing process to obtain the temperature data sequence to be analyzed corresponding to the target monitoring time period. The specific acquisition process is:

[0020] First, in the reaction chamber of the grain-mash mixing process, a temperature sensor is used to obtain sensor data at each monitoring moment in the target monitoring time period, wherein the sensor data represents the temperature, and the sensor collects the temperature in the reaction chamber; then, all the collected sensor data are sorted in chronological order, and the sorted sequence is recorded as the temperature data sequence to be analyzed corresponding to the target monitoring time period.

[0021] In specific applications, the implementer needs to set the target monitoring time period and the time interval between adjacent monitoring moments according to actual conditions. For example, in this embodiment, the target monitoring time period can be set to ten minutes before the current monitoring moment, and the last monitoring moment in the target monitoring time period is the current monitoring moment, and the time interval between adjacent monitoring moments is set to 0.5 seconds.

[0022] In addition, the collected data will be transmitted to the grain-mash mixing process production temperature monitoring and control system via wired means, and then predictive analysis will be performed through the data analysis module in the system. The data analysis module can further determine the monitoring results through the prediction results, and then the temperature control module in the reaction chamber can be controlled based on the monitoring results obtained, mainly to control the operating frequency of the heat dissipation equipment and the heating module equipment.

[0023] Therefore, this embodiment obtains the temperature data sequence to be analyzed through the above process.

[0024] Step S002, obtaining a target subsequence set according to a preset sliding window and the temperature data sequence to be analyzed; obtaining a change characteristic value of the target subsequence according to the total number of peaks in each target subsequence in the target subsequence set; obtaining a first target subsequence in the target subsequence set and an initial order value of the first target subsequence according to the change characteristic value of the target subsequence and the temperature data in the target subsequence.

[0025] Based on the above description, it can be known that in the grain mash mixing process, the temperature data in the reaction chamber will change with different grain mash stages. Therefore, when using the ARIMA model to predict the temperature in the reaction chamber, if the prediction accuracy is to be guaranteed, it is necessary to adaptively obtain the order value in the ARIMA model according to the temperature changes at different stages. In order to achieve the purpose of adaptively obtaining the order value according to the temperature changes at different stages, after obtaining the temperature data sequence to be analyzed, the temperature data sequence to be analyzed needs to be divided to obtain a target subsequence set, and the specific acquisition process of the target subsequence set in this embodiment is as follows:

[0026] First, a preset sliding window is obtained, and then the preset sliding window is slid on the temperature data sequence to be analyzed with a preset moving step length to obtain each sliding window located on the temperature data sequence to be analyzed, wherein the first data in the first sliding window on the temperature data sequence to be analyzed is the first temperature data in the temperature data sequence to be analyzed, and the sliding stops immediately when the last data on the temperature data sequence to be analyzed appears in the sliding window for the first time.

[0027] Then, the sequence formed by the data in each sliding window on the temperature data sequence to be analyzed is recorded as the initial subsequence on the temperature data sequence to be analyzed, that is, one sliding window corresponds to one initial subsequence, and the set constructed by all the initial subsequences on the temperature data sequence to be analyzed is recorded as the initial subsequence set, and the bth initial subsequence in the initial subsequence set is the sequence formed by the data in the bth sliding window on the temperature data sequence to be analyzed. The size of the preset sliding window is , a is the length of the preset sliding window, the preset moving step is less than a, and a is required to have opposite parity to the preset moving step; in addition, the first mark value of the cth temperature data in the temperature data sequence to be analyzed is c.

[0028] Therefore, an initial subsequence set is obtained through the above process; and in specific applications, the implementer needs to set the length of the preset sliding window and the preset moving step according to the actual situation and the above requirements. For example, in this embodiment, the length of the preset sliding window, that is, the value of a is set to 7, and the preset moving step is set to 4. If the preset moving step is 4 and the length of the preset sliding window is 7, then in the temperature data sequence to be analyzed, there are usually 3 overlapping data in two adjacent initial subsequences.

[0029] After the initial subsequence set is obtained, the sequence pairs consisting of two adjacent initial subsequences in the initial subsequence set are recorded as initial subsequence pairs, and the set constructed by all the obtained initial subsequence pairs is recorded as the first set, and all the initial subsequence pairs in the first set are recorded as first subsequence pairs, and the total number of the initial subsequence pairs is the total number of the initial subsequences minus 1.

[0030] Then, the similarity characterization value of each first subsequence pair in the first set is obtained; then, in the first set, all first subsequence pairs whose similarity characterization value is less than the preset similarity threshold are obtained, and the set constructed by all first subsequence pairs whose similarity characterization value is less than the preset similarity threshold is recorded as the second set, and all first subsequence pairs in the second set are recorded as second subsequence pairs. In specific applications, the implementer needs to set the preset similarity threshold according to the actual situation. For example, in this embodiment, the preset similarity threshold can be set to 0.3.

[0031] Next, the subsequences to be selected of each second subsequence pair in the second set are obtained, and the temperature data located at the middle position of the subsequence to be selected are recorded as the first data; for example, if any subsequence to be selected is {x1, x2, x3}, then x2 in the subsequence to be selected is the first data.

[0032] Afterwards, on the temperature data sequence to be analyzed, temperature data belonging to the first data is obtained and recorded as segmented temperature data; then, all segmented temperature data on the temperature data sequence to be analyzed are used to divide the temperature data sequence to be analyzed, and a set constructed from all subsequences obtained by the division is recorded as a target subsequence set.

[0033] For example, if the temperature data sequence to be analyzed is {x1, x2, x3, x4, x5, x6, x7, x8, x9}, and if x3 and x6 are segmented temperature data, then using all the segmented temperature data on the temperature data sequence to be analyzed, the subsequences obtained by dividing the temperature data sequence to be analyzed are {x1, x2, x3}, {x4, x5, x6}, and {x7, x8, x9}, that is, the target subsequence set at this time consists of the sequence {x1, x2, x3}, the sequence {x4, x5, x6}, and the sequence {x7, x8, x9}.

[0034] In this embodiment, the method for obtaining the similarity characterization value of the first subsequence pair is as follows: for any first subsequence pair U, firstly, the Pearson correlation coefficient between the two subsequences in the first subsequence pair U is obtained, and the absolute value of the difference between the Pearson correlation coefficient between the two subsequences in the first subsequence pair U and the preset first constant is recorded as the first characterization value of the first subsequence pair U, then the absolute value of the difference between the standard deviations of the two subsequences in the first subsequence pair U is obtained, and recorded as the second characterization value of the first subsequence pair U, and then the value obtained by multiplying the first characterization value by the second characterization value and adding the preset second constant is recorded as the third characterization value, and the value obtained by normalizing the reciprocal of the third characterization value is recorded as the similarity characterization value of the first subsequence pair U; and the specific expression for obtaining the similarity characterization value of the first subsequence pair U is:

[0035]

[0036] Wherein, S is the similarity characterization value of the first subsequence pair U, wnorm() is a weight normalization function, c1 is a preset first constant, c2 is a preset second constant, PER is the Pearson correlation coefficient between the two subsequences in the first subsequence pair U, is the standard deviation of one of the subsequences in the first subsequence pair U, is the standard deviation of the first subsequence to another subsequence in U; ​​in this embodiment, the values ​​of c1 and c2 are set to 1 and 0.01 respectively, and the purpose of c2 is to avoid the denominator being 0.

[0037] In addition, when and The smaller the value of is, the larger the value of S is. The larger the value of S is, the more similar the two subsequences in the first subsequence pair U are. and The larger the value of , the smaller the value of S is, and the smaller the value of S is, the less similar the two subsequences in the first subsequence pair U are.

[0038] In this embodiment, for ease of understanding, a method for obtaining a subsequence to be selected from any second subsequence pair N is taken as an example for description, that is, the process for obtaining a subsequence to be selected from the second subsequence pair N is:

[0039] First, the two subsequences constituting the second subsequence pair N are respectively recorded as sequence N1 and sequence N2, and then the temperature data marked as overlapping are obtained in sequence N1; and the sequence formed by all the temperature data marked as overlapping in sequence N1 is recorded as the subsequence to be selected of the second subsequence pair N; for example, if sequence N1 is {y1, y2, y3, y4, y5, y6, y7, y8}, and y6, y7, and y8 are all marked as overlapping, then the subsequence to be selected of the second subsequence pair N is {y6, y7, y8}.

[0040] And the method for obtaining the temperature data marked as overlapping in the sequence N1 is: for any temperature data c0 in the sequence N1, if there is temperature data with the same first mark value as the temperature data c0 in the sequence N2, then the temperature data c0 is marked as overlapping, that is, since the window moving step is smaller than the window length, there must be overlapping parts between the two subsequences in the subsequence pair.

[0041] Therefore, this embodiment obtains a target subsequence set through the above process, and the temperature data or change rules that are divided into the same target subsequence are relatively similar, while the temperature data or change rules in adjacent target subsequences are relatively different; and after obtaining the target subsequence set, this embodiment immediately obtains the total number of peaks in each target subsequence in the target subsequence set, and then obtains the change characteristic value of each target subsequence according to the total number of peaks in each target subsequence in the target subsequence set; and in this embodiment, the specific method for obtaining the change characteristic value of the target subsequence is:

[0042] For any target subsequence D: the total number of peaks in the target subsequence D is recorded as the quantity representation value, the time interval between the acquisition time of the first temperature data in the target subsequence D and the acquisition time of the last temperature data in the target subsequence D is recorded as the time representation value, and the ratio of the quantity representation value to the time representation value is recorded as the change characteristic value of the target subsequence D; and the change characteristic value of the target subsequence D can reflect the frequency of changes in the temperature data in the target subsequence D, and the greater the frequency of changes in the temperature data in the target subsequence D, the greater the frequency of changes in the target subsequence D.

[0043] In addition, in this embodiment, the AMPD algorithm is used to detect the peak value in the target subsequence. The AMPD (Automatic Multiscale-based Peak Detection) algorithm is an automatic peak detection algorithm based on multi-scale, and the process of using the AMPD algorithm to detect the peak value in the target subsequence is a well-known technology, so it will not be described in detail.

[0044] Therefore, the change characteristic value of each target subsequence can be obtained through the above process. After obtaining the change characteristic value, this embodiment will obtain the first target subsequence in the target subsequence set and the initial order value of the first target subsequence according to the change characteristic value of the target subsequence and the temperature data in the target subsequence. Then, in this embodiment, the specific acquisition process of the first target subsequence in the target subsequence set and the initial order value of the first target subsequence is as follows:

[0045] First, the acquisition time of the first temperature data in each target subsequence is recorded as the representation time of the corresponding target subsequence; then, in the target subsequence set, all target subsequences except the minimum representation time are recorded as the first target subsequence.

[0046] After all the first target subsequences are obtained, the process of obtaining the initial order value of any first target subsequence F will be described as an example, that is, the process of obtaining the initial order value of the first target subsequence F is:

[0047] First, a set constructed from all target subsequences whose representation time is less than the first target subsequence F is recorded as the pre-order set of the first target subsequence F; then, the first index value and the second index value of each target subsequence in the pre-order set are obtained, and the product of the first index value of each target subsequence in the pre-order set and the second index value of the corresponding target subsequence is recorded as the third index value corresponding to the corresponding target subsequence; then, the mean of the third index values ​​corresponding to all target subsequences in the pre-order set is obtained and recorded as the first mean value L1; in addition, the first index value of the f-th target subsequence in the pre-order set refers to the ratio of the change characteristic value of the first target subsequence F to the change characteristic value of the f-th target subsequence, and the second index value of the f-th target subsequence in the pre-order set refers to the ratio of the value range representation value of the first target subsequence F to the value range representation value of the f-th target subsequence, and the value range representation value of the target subsequence refers to the difference between the maximum temperature data and the minimum temperature data in the corresponding target subsequence.

[0048] In this embodiment, the first mean value L1 is obtained by:

[0049]

[0050] Among them, L1 is the first mean, F0 is the number of target subsequences in the pre-order set, is the change characteristic value of the first target subsequence F, is the change characteristic value of the fth target subsequence in the preorder set, is the value range representation value of the first target subsequence F, The value representing the value range of the f-th target subsequence in the preorder set.

[0051] In addition, when and The larger the value of is, the greater the change of the temperature data in the first target subsequence F is compared with the historical data, and the change of the temperature data in the first target subsequence F is greater than the change of the historical temperature data. and The larger the value of is, the larger the value of L1 is. When the value of L1 is larger, a smaller order value should be selected as the initial order value of the first target subsequence F. This can avoid excessive reliance on the temperature data in the previous set in subsequent predictions as much as possible, thereby avoiding overfitting when predicting the temperature data in the first target subsequence L1 as much as possible, and can also make the prediction accuracy and reliability higher. On the contrary, when and The smaller the value of , the smaller the value of L1. The smaller the value of L1, the larger the order value should be selected as the initial order value of the first target subsequence F. In this way, the change details of temperature prediction can be improved according to the long-term advantages of historical data, thereby improving the accuracy of prediction.

[0052] In order to achieve the above-mentioned characteristics of the initial order value, this embodiment will then perform negative correlation mapping on the first mean value A1, and record the value after negative correlation mapping as the first mapping value. The acquisition expression of the first mapping value is: ; Then, the rounded-up value of the first mapping value is used as the initial order value of the first target subsequence F; wherein P0 is the preset order value, exp() is an exponential function with a constant e as the base, and k is the function coefficient.

[0053] In specific applications, the implementer needs to set the preset order value and function coefficient according to the actual situation. For example, in this embodiment, the preset order value and function coefficient are set to 10 and 0.4 respectively.

[0054] Therefore, this embodiment can obtain the initial order value of each first target subsequence through the above process.

[0055] Step S003, obtaining a predicted temperature data sequence corresponding to each feature subsequence according to the initial order value of each feature subsequence in the feature subsequence set, and obtaining an error coefficient corresponding to the initial order value of the feature subsequence according to the predicted temperature data sequence.

[0056] Furthermore, the initial order value obtained as described above may have errors, and the existence of errors may also cause inaccurate prediction values ​​obtained subsequently. In order to reduce the accuracy of the prediction values, this embodiment first obtains a feature subsequence set, and the feature subsequence set refers to a set constructed from all first target subsequences except the current target subsequence, and the current target subsequence refers to the first target subsequence whose acquisition time is closest to the current monitoring moment, that is, the acquisition time corresponding to the last data in the current target subsequence is the current monitoring moment.

[0057] Next, it is necessary to first obtain the error coefficient corresponding to the initial order value of the feature subsequence. Subsequently, the target order value of the current target subsequence will be obtained based on the error coefficient corresponding to the initial order value of the feature subsequence. Before obtaining the error coefficient, it is necessary to first obtain the predicted temperature data sequence corresponding to the feature subsequence based on the initial order value of the feature subsequence. The specific acquisition process of the predicted temperature data sequence corresponding to the feature subsequence is as follows:

[0058] For any feature subsequence H: construct an ARIMA model with an order value equal to the initial order value of the feature subsequence H, and record it as the prediction model of the feature subsequence H, then train the prediction model of the feature subsequence H, and record the trained prediction model of the feature subsequence H as the target prediction model, then input the feature subsequence H into the target prediction model, and record the output sequence as the predicted temperature data sequence corresponding to the feature subsequence H; and the training process of the ARIMA model is a well-known technology, so this embodiment will not be described in detail.

[0059] Therefore, according to the above process, the predicted temperature data sequence corresponding to each characteristic subsequence can be obtained. After the predicted temperature data sequence corresponding to the characteristic subsequence is obtained, this embodiment will obtain the error coefficient corresponding to the initial order value of the characteristic subsequence according to the predicted temperature data sequence corresponding to the characteristic subsequence, and the specific acquisition process is:

[0060] For the characteristic subsequence H: obtain the coefficient of variation corresponding to the predicted temperature data sequence of the characteristic subsequence H and the coefficient of variation corresponding to the characteristic subsequence H, the coefficient of variation corresponding to the sequence refers to the ratio of the standard deviation of the corresponding sequence to the mean of the corresponding sequence, and then the ratio of the coefficient of variation corresponding to the predicted temperature data sequence of the characteristic subsequence H to the coefficient of variation corresponding to the characteristic subsequence H is recorded as the variation ratio; then obtain the mean square error between the predicted temperature data sequence of the characteristic subsequence H and the characteristic subsequence H, and record it as the mean square error characterization value; then subtract the preset first constant from the variation ratio and multiply it by the mean square error characterization value to obtain the value, which is recorded as the error coefficient corresponding to the initial order value of the characteristic subsequence H.

[0061] In addition, the formula for obtaining the error coefficient corresponding to the initial order value of the characteristic subsequence H is:

[0062]

[0063] in, is the error coefficient corresponding to the initial order value of the characteristic subsequence H, V1 is the coefficient of variation corresponding to the predicted temperature data sequence of the characteristic subsequence H, V2 is the coefficient of variation corresponding to the characteristic subsequence H, and Y0 is the mean square error characterization value; and when The larger the value of Y0 is, the The larger the value of The larger the value of , the larger the initial order value of the characteristic subsequence H may be, which may lead to overfitting, that is, the predicted temperature sequence may fluctuate more violently than the real temperature sequence. The smaller the value of Y0, the The smaller the value of The smaller the value of , the smaller the initial order value of the characteristic subsequence H may be. When the initial order value is small, the predicted temperature series may fluctuate more smoothly than the real temperature series, and the prediction may not be able to reflect the temperature change information in a timely and accurate manner.

[0064] Therefore, this embodiment can obtain the error coefficient corresponding to the initial order value of each characteristic subsequence according to the above process.

[0065] Step S004: obtaining a target order value of the current target subsequence according to the initial order value of the current target subsequence and the error coefficient corresponding to the initial order value of the characteristic subsequence.

[0066] Since the initial order value of the current target subsequence obtained above may be too large or too small, this situation will also cause the prediction data obtained subsequently to have low accuracy. In order to further improve the accuracy of the prediction data, this embodiment will correct the initial order value of the current target subsequence based on the error coefficient corresponding to the initial order value of the feature subsequence, so as to obtain the target order value of the current target subsequence, that is, according to the initial order value of the current target subsequence and the error coefficient corresponding to the initial order value of the feature subsequence, the target order value of the current target subsequence is obtained. The specific acquisition process is as follows:

[0067] First, the set of error coefficients corresponding to the initial order values ​​of all feature subsequences in the feature subsequence set is recorded as the error coefficient set, and the value after normalizing the mean of the error coefficient set is recorded as the target error mean, and the mean of the error coefficient set is recorded as the initial error mean; and the formula for normalizing the mean of the error coefficient set is: , where k1 is the normalization function coefficient, in this embodiment the value of k1 is 1.2, and L2 is the mean value of the error coefficient set.

[0068] Then, it is determined whether the initial error mean is greater than a preset first threshold value. If so, it indicates that the initial order value of the current target subsequence needs to be reduced, that is, the value obtained by multiplying the initial order value of the current target subsequence by the preset first constant minus the target error mean is rounded up as the target order value of the current target subsequence; and if it is determined that the initial error mean is less than the preset first threshold value, it indicates that the initial order value of the current target subsequence needs to be increased, that is, the value obtained by multiplying the initial order value of the current target subsequence by the preset first constant plus the target error mean is rounded up as the target order value of the current target subsequence; and if it is determined that the initial error mean is equal to the preset first threshold value, it indicates whether the initial order value of the current target subsequence needs to be changed, that is, the initial order value of the current target subsequence is directly used as the target order value of the current target subsequence.

[0069] Therefore, this embodiment obtains the initial order value of the current target subsequence through the above process; and in specific applications, the implementer needs to set the preset first threshold according to actual conditions, such as setting the preset first threshold to 0 in this embodiment.

[0070] Step S005, obtaining predicted temperature data at a future time according to the ARIMA model, the target order value and the current target subsequence, and obtaining the temperature monitoring result in the reaction chamber at the current monitoring time according to the predicted temperature data.

[0071] After obtaining the target order value, the predicted temperature data at the future moment is obtained according to the ARIMA model, the target order value and the current target subsequence. The specific acquisition process is: first, an ARIMA model with an order value equal to the target order value is constructed and recorded as the first prediction model, and then the first prediction model is trained, and the current target subsequence is input into the trained first prediction model to output the predicted temperature data at the future moment; and in this embodiment, the predicted temperature data at the future moment refers to the predicted temperature data at the (t+w0)th moment, and the tth moment refers to the current monitoring moment; in addition, in this embodiment, the implementer needs to set the value of w0 according to the volume of the reaction chamber, and the setting needs to satisfy that the larger the volume of the reaction chamber, the larger the value of w0, and the smaller the volume of the reaction chamber, the smaller the value of w0, and the larger the volume of the reaction chamber indicates that the temperature transfer delay time may be longer, such as in this embodiment, the value of w0 can be set to 20.

[0072] After obtaining the predicted temperature data at a future moment, the temperature monitoring result in the reaction chamber at the current monitoring moment can be analyzed, specifically: first obtain the preset temperature threshold, and in this embodiment, the preset temperature threshold is set to the optimal reaction temperature in the reaction chamber or the optimal suitable temperature of the grain-mash mixing process, and there are differences in the optimal reaction temperatures in the reaction chamber corresponding to different types of grain-mash mixing processes; then determine whether the predicted temperature data is not equal to the preset temperature threshold. If so, it is determined that the temperature in the reaction chamber at the current monitoring moment does not meet the production requirements; otherwise, it is determined that the temperature in the reaction chamber at the current monitoring moment meets the production requirements.

[0073] And when it is determined that the temperature in the reaction chamber at the current monitoring time does not meet the production requirements, it is necessary to adjust the working frequency of the heat dissipation device or the heating device so that the temperature in the reaction chamber meets the production or reaction requirements, specifically: first, the absolute value of the difference between the predicted temperature data and the preset temperature threshold is recorded as the characteristic difference; if it is determined that the predicted temperature data is greater than the preset temperature threshold, the working power mapping parameter of the heat dissipation device is multiplied by the characteristic difference, recorded as the first target working power, and the working power of the heat dissipation device is adjusted to the first target working power; if it is determined that the predicted temperature data is less than the preset temperature threshold, the working power mapping parameter of the heating device is multiplied by the characteristic difference, recorded as the second target working power, and the working power of the heating device is adjusted to the second target working power. In addition, in this embodiment, the working power mapping parameters of the heat dissipation device and the working power mapping parameters of the heating device are 50 and 130, respectively.

[0074] In summary, in this embodiment, first, in the reaction chamber of the grain-wheat mixing process production, a sensor is used to obtain a temperature data sequence to be analyzed corresponding to the target monitoring time period; then, according to the preset sliding window and the temperature data sequence to be analyzed, a target subsequence set is obtained, and according to the total number of peaks in each target subsequence in the target subsequence set, a change characteristic value of the target subsequence is obtained, and according to the change characteristic value of the target subsequence and the temperature data in the target subsequence, a first target subsequence in the target subsequence set and an initial order value of the first target subsequence are obtained; then, according to the initial order value of each feature subsequence in the feature subsequence set, a predicted temperature data sequence corresponding to the feature subsequence is obtained, and according to the predicted temperature data sequence, an error coefficient corresponding to the initial order value of the feature subsequence is obtained; then, according to the initial order value of the current target subsequence and the error coefficient corresponding to the initial order value of the feature subsequence, a target order value of the current target subsequence is obtained; finally, according to the ARIMA model, the target order value and the current target subsequence, the predicted temperature data at a future moment is obtained, and according to the predicted temperature data, the temperature monitoring result in the reaction chamber at the current monitoring moment is obtained. Moreover, in this embodiment, the accuracy of the predicted data obtained based on the target order value is higher, that is, the accuracy and reliability of the predicted data obtained based on the target order value are higher. Therefore, through the obtained predicted temperature data, the accuracy and reliability of monitoring the temperature in the reaction chamber can be improved, that is, based on the predicted temperature data, the obtained temperature monitoring result in the reaction chamber can be made more reliable.

[0075] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A sensor-based grain-wheat mixing process production data monitoring system, comprising a processor and a memory, characterized in that: The processor executes the computer program stored in the memory to implement the following steps: In the reaction chamber of the grain-wheat mixing process, a sensor is used to obtain a temperature data sequence to be analyzed corresponding to a target monitoring time period, wherein the last monitoring moment in the target monitoring time period is the current monitoring moment; Obtaining a target subsequence set according to a preset sliding window and the temperature data sequence to be analyzed; obtaining a change characteristic value of the target subsequence according to the total number of peaks in each target subsequence in the target subsequence set; Obtaining a first target subsequence in the target subsequence set and an initial order value of the first target subsequence according to the change characteristic value of the target subsequence and the temperature data in the target subsequence; According to the initial order value of each characteristic subsequence in the characteristic subsequence set, a predicted temperature data sequence corresponding to the characteristic subsequence is obtained, and according to the predicted temperature data sequence, an error coefficient corresponding to the initial order value of the characteristic subsequence is obtained; the characteristic subsequence set refers to a set constructed by all first target subsequences except the current target subsequence, and the current target subsequence refers to the first target subsequence whose acquisition time is closest to the current monitoring time; Obtaining a target order value of the current target subsequence according to the initial order value of the current target subsequence and the error coefficient corresponding to the initial order value of the characteristic subsequence; According to the ARIMA model, the target order value and the current target subsequence, the predicted temperature data at a future moment is obtained, and according to the predicted temperature data, the temperature monitoring result in the reaction chamber at the current monitoring moment is obtained.

2. A sensor-based grain-wheat mixing process production data monitoring system as claimed in claim 1, characterized in that: The method for acquiring the target subsequence set includes: Slide the preset sliding window on the temperature data sequence to be analyzed with a preset moving step length to obtain each sliding window on the temperature data sequence to be analyzed, and record the sequence formed by the data in each sliding window on the temperature data sequence to be analyzed as an initial subsequence on the temperature data sequence to be analyzed, and record the set constructed by all initial subsequences as an initial subsequence set, the bth initial subsequence in the initial subsequence set is a sequence formed by the data in the bth sliding window on the temperature data sequence to be analyzed, the preset moving step length is less than the length of the preset sliding window, and the first mark value of the cth temperature data on the temperature data sequence to be analyzed is c; In the initial subsequence set, sequence pairs composed of two adjacent initial subsequences are all recorded as initial subsequence pairs, a set constructed by all the obtained initial subsequence pairs is recorded as a first set, and the initial subsequence pairs in the first set are all recorded as first subsequence pairs; a similarity characterization value of each first subsequence pair in the first set is obtained; a set constructed by all first subsequence pairs in the first set whose similarity characterization values ​​are less than a preset similarity threshold is recorded as a second set, and the first subsequence pairs in the second set are all recorded as second subsequence pairs; Obtaining subsequences to be selected of each second subsequence pair; recording the temperature data at the middle position of the subsequence to be selected as first data, and recording the temperature data located on the temperature data sequence to be analyzed and belonging to the first data as split temperature data; dividing the temperature data sequence to be analyzed using the split temperature data on the temperature data sequence to be analyzed, and recording a set constructed by all subsequences obtained by the division as a target subsequence set; The method for obtaining the subsequence to be selected of the second subsequence pair includes: for any second subsequence pair N: recording two subsequences constituting the second subsequence pair N as sequence N1 and sequence N2 respectively, and obtaining temperature data marked as overlapping in the sequence N1; for any temperature data n0 in the sequence N1, if there is temperature data having the same first mark value as the temperature data n0 in the sequence N2, marking the temperature data n0 as overlapping; and recording a sequence formed by all temperature data marked as overlapping in the sequence N1 as the subsequence to be selected of the second subsequence pair N.

3. A sensor-based grain-wheat mixing process production data monitoring system as claimed in claim 2, characterized in that: The method for obtaining the similarity characterization value of the first subsequence pair includes: For any first subsequence pair U, the absolute value of the difference between the Pearson correlation coefficient between the two subsequences in the first subsequence pair U and the preset first constant is recorded as the first characterization value of the first subsequence pair U, the absolute value of the difference between the standard deviations of the two subsequences in the first subsequence pair U is recorded as the second characterization value of the first subsequence pair U, the value obtained by multiplying the first characterization value by the second characterization value and then adding the preset second constant is recorded as the third characterization value, and the value obtained by normalizing the reciprocal of the third characterization value is recorded as the similarity characterization value of the first subsequence pair U.

4. A sensor-based grain-wheat mixing process production data monitoring system as claimed in claim 1, characterized in that: The method for obtaining the change characteristic value of the target subsequence includes: For any target subsequence D: the total number of peaks in the target subsequence D is recorded as the quantity representation value; the time interval between the acquisition time of the first temperature data in the target subsequence D and the acquisition time of the last temperature data in the target subsequence D is recorded as the time representation value; the ratio of the quantity representation value to the time representation value is recorded as the change characteristic value of the target subsequence D.

5. A sensor-based grain-wheat mixing process production data monitoring system as claimed in claim 1, characterized in that: The method for obtaining the first target subsequence in the target subsequence set and the initial order value of the first target subsequence includes: Recording the acquisition time of the first temperature data in each target subsequence as the representation time of the corresponding target subsequence; in the target subsequence set, recording all target subsequences except the minimum representation time as the first target subsequence; For any first target subsequence F: a set constructed from all target subsequences whose representation time is less than the first target subsequence F is recorded as a pre-order set of the first target subsequence F; a first index value and a second index value of each target subsequence in the pre-order set are obtained, and the product of the first index value of each target subsequence in the pre-order set and the second index value of the corresponding target subsequence is recorded as a third index value corresponding to the corresponding target subsequence; the mean of the third index values ​​corresponding to all target subsequences in the pre-order set is recorded as a first mean value L1; a first mapping value is obtained according to the first mean value L1, and the first mapping value is , taking the rounded value of the first mapping value as the initial order value of the first target subsequence F, wherein P0 is a preset order value, exp() is an exponential function with a constant e as the base, k is a function coefficient, the first index value of the f-th target subsequence in the preamble set refers to the ratio of the change characteristic value of the first target subsequence F to the change characteristic value of the f-th target subsequence, the second index value of the f-th target subsequence in the preamble set refers to the ratio of the value range representation value of the first target subsequence F to the value range representation value of the f-th target subsequence, and the value range representation value of the target subsequence refers to the difference between the maximum temperature data and the minimum temperature data in the corresponding target subsequence.

6. A sensor-based grain-wheat mixing process production data monitoring system as claimed in claim 1, characterized in that: The method for obtaining the predicted temperature data sequence corresponding to the characteristic subsequence includes: For any feature subsequence H: construct an ARIMA model with an order value equal to the initial order value of the feature subsequence H, and record it as the prediction model of the feature subsequence H, train the prediction model of the feature subsequence H, and record the trained prediction model of the feature subsequence H as the target prediction model, input the feature subsequence H into the target prediction model, and record the output sequence as the predicted temperature data sequence corresponding to the feature subsequence H.

7. A sensor-based grain-wheat mixing process production data monitoring system as claimed in claim 1, characterized in that: The method for obtaining the error coefficient corresponding to the initial order value of the characteristic subsequence includes: For any characteristic subsequence H: the ratio of the coefficient of variation corresponding to the predicted temperature data sequence of the characteristic subsequence H to the coefficient of variation corresponding to the characteristic subsequence H is recorded as the variation ratio; the mean square error between the predicted temperature data sequence of the characteristic subsequence H and the characteristic subsequence H is recorded as the mean square error characterization value; the value obtained by subtracting the preset first constant from the variation ratio and then multiplying it by the mean square error characterization value is recorded as the error coefficient corresponding to the initial order value of the characteristic subsequence H.

8. A sensor-based grain-wheat mixing process production data monitoring system as claimed in claim 1, characterized in that: The method for obtaining the target order value of the current target subsequence includes: A set of error coefficients corresponding to the initial order values ​​of all feature subsequences in the feature subsequence set is constructed and recorded as the error coefficient set, a normalized value of the mean of the error coefficient set is recorded as the target error mean, and the mean of the error coefficient set is recorded as the initial error mean; If the initial error mean is greater than a preset first threshold, the integer value of the value obtained by multiplying the initial order value of the current target subsequence by the preset first constant minus the target error mean is used as the target order value of the current target subsequence; if the initial error mean is less than the preset first threshold, the integer value of the value obtained by multiplying the initial order value of the current target subsequence by the preset first constant plus the target error mean is used as the target order value of the current target subsequence; if the initial error mean is equal to the preset first threshold, the initial order value of the current target subsequence is used as the target order value of the current target subsequence.

9. A sensor-based grain-wheat mixing process production data monitoring system as claimed in claim 1, characterized in that: The method for obtaining the predicted temperature data at the future moment includes: An ARIMA model with an order value equal to the target order value is constructed and recorded as a first prediction model. The first prediction model is trained, and the current target subsequence is input into the trained first prediction model to output predicted temperature data at a future moment.

10. A sensor-based grain-wheat mixing process production data monitoring system as claimed in claim 1, characterized in that: The method for obtaining the temperature monitoring result in the reaction chamber at the current monitoring time includes: Obtain a preset temperature threshold; determine whether the predicted temperature data is not equal to the preset temperature threshold, if so, determine that the temperature in the reaction chamber at the current monitoring time does not meet the production requirements, otherwise, determine that the temperature in the reaction chamber at the current monitoring time meets the production requirements.

Images