Intelligent cladophora organic fertilizer production system based on Internet of Things
Through the intelligent production system of the Internet of Things, combined with timing analysis and wavelet transformation, the temperature prediction problem under the influence of multi-factor coupling is solved, more accurate temperature control is achieved, and the quality of organic fertilizer is ensured.
Patent Information
- Application Number
- CN202510430774.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-08
- Publication Date
- 2025-07-08
AI Technical Summary
The existing temperature predictions are relatively low in the production of organic fertilizers for bristle algae, and the coupling effects of various factors such as humidity, acid-base value and oxygen concentration are not effectively considered, affecting the accuracy of temperature control.
Using an intelligent production system based on the Internet of Things, the monitoring data timing is obtained through the data acquisition module, the mean sequence and trend characteristic values are obtained by the timing analysis module, differential iterative processing is performed and wavelet transformation is performed, the energy intensity and correlation characteristics of the component signal are analyzed, and the temperature prediction and control are used by the ARIMAX model.
It improves the accuracy of temperature prediction, ensures that the temperature is within a reasonable range during the production process, and ensures the quality of organic fertilizers.
Smart Images
Figure CN120276533A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing, and particularly to an intelligent production system for Cladophora organic fertilizer based on the Internet of Things. Background Art
[0002] In the production process of Cladophora organic fertilizer, temperature plays a crucial role in the growth of microalgae and microbial fermentation. The change of temperature directly determines the growth rate, metabolic state of microorganisms and the quality of the final organic fertilizer. Therefore, in the production process of Cladophora organic fertilizer, the monitoring and control of temperature are of great significance for ensuring fermentation efficiency, promoting the decomposition of organic matter, maintaining the activity of microorganisms and improving the composting quality.
[0003] In order to be able to intervene in the change of temperature in time, the temperature can be predicted to obtain the temperature trend in advance. However, the change of temperature is not only affected by a single factor, but there is a complex coupling effect with multiple environmental factors such as humidity, pH value and oxygen concentration. Existing temperature prediction is usually based only on single historical temperature prediction, without considering the influence of other factors on temperature, resulting in low prediction accuracy, thus affecting the accuracy of temperature control in the production process of Cladophora organic fertilizer. Summary of the Invention
[0004] In order to solve the technical problem that in the production process of Cladophora organic fertilizer, the temperature is affected by multiple factors, resulting in low temperature prediction accuracy and affecting the temperature control in the fertilizer production process, the purpose of the present invention is to provide an intelligent production system for Cladophora organic fertilizer based on the Internet of Things. The specific technical solutions adopted are as follows: A data acquisition module, which is used to acquire the time series of monitoring data of different temperature influencing factors and temperature in the historical period during the production process of Cladophora organic fertilizer; A time series analysis module, which is used to obtain a mean sequence according to the data distribution characteristics within a preset sliding window in the monitoring data time series; obtain the change correlation according to the data change trend of the mean sequences of temperature and influencing factors; obtain the trend characteristic value according to the data fluctuation characteristics of the mean sequence; perform differential iteration processing on the monitoring data time series according to the trend characteristic value to obtain a target sequence; A temperature influence analysis module, which is used to perform wavelet transform on the interpolated target sequence to obtain different component signals; obtain the component importance according to the energy intensity characteristics of the component signals; obtain the component correlation characteristic value according to the difference characteristics between the component importance, the component signals corresponding to temperature and influencing factors; obtain the comprehensive correlation characteristic value of the component signals of the influencing factors according to the component correlation characteristic value and the component importance; obtain the final correlation characteristic value of the influencing factors according to the comprehensive correlation characteristic values of all component signals of the influencing factors; A temperature prediction module, which is used to perform temperature prediction based on the final correlation eigenvalue of all target sequences and influencing factors to obtain the predicted temperature at a future moment; and control the temperature in production according to the predicted temperature.
[0005] Further, the step of obtaining the mean sequence according to the data distribution characteristics within a preset sliding window in the monitoring data time series includes: Calculate the data average value within a preset sliding window in the monitoring data time series to obtain a local average value; sort all local average values according to the sliding order of the preset sliding window in the monitoring data time series to obtain the mean sequence.
[0006] Further, the step of obtaining the change correlation according to the data change trend of the mean sequences of temperature and influencing factors includes: Differentiate the mean sequence and calculate the average value of the differentiation result to obtain a differential average value; if the positive and negative natures of the differential average values corresponding to the temperature and influencing factors are the same, the change correlation between the temperature and influencing factors is positively correlated, otherwise the change correlation is negatively correlated.
[0007] Further, the step of obtaining the trend eigenvalue according to the data fluctuation characteristics of the mean sequence includes: Calculate the product of the absolute value of the differential average value of the mean sequence and the standard deviation of the mean sequence to obtain the trend eigenvalue corresponding to the mean sequence.
[0008] Further, the step of performing differential iteration processing on the monitoring data time series according to the trend eigenvalue to obtain the target sequence includes: When the trend eigenvalue is a constant 0, use the monitoring data time series of the influencing factor as the target sequence of the influencing factor; when the trend eigenvalue exceeds the constant 0, differentiate the monitoring data time series and use the differentiation result as the new round of monitoring data time series, and judge whether the trend eigenvalue of the new round of monitoring data time series is less than the trend eigenvalue of the initial monitoring data time series. If it is less, continue to use the differentiation result of the new round of monitoring data time series as the next round of monitoring data time series, and judge whether the trend eigenvalue of the next round of monitoring data time series is less than the previous round of monitoring data time series until the trend eigenvalue of the next round of monitoring data time series is not less than the previous round of monitoring data time series, then stop the differential iteration and use the previous round of monitoring data time series as the target sequence.
[0009] Further, the step of obtaining the component importance according to the energy intensity characteristics of the component signals includes: Calculate the ratio of the energy intensity of any component signal to the total energy intensity of all component signals obtained by the corresponding transformation object to obtain the component importance of the any component signal.
[0010] Further, the step of obtaining the component correlation eigenvalue according to the difference characteristics of the component signals corresponding to the component importance, temperature, and influencing factors includes: Calculating the absolute value of the difference in component importance between any influencing component signal in the influencing factors and any temperature component signal of the temperature to obtain an importance difference value; calculating the dynamic time warping distance between the any influencing component signal and the any temperature component signal and negatively correlating and mapping to obtain a similarity; calculating the sum value of the importance difference value and a preset constant to obtain a difference characterization value; calculating the ratio of the similarity to the difference characterization value to obtain the component correlation eigenvalue.
[0011] Further, the step of obtaining the comprehensive correlation eigenvalue of the component signal of the influencing factor according to the component correlation eigenvalue and the component importance includes: Calculating the average value of the component correlation eigenvalues of any influencing component signal of the influencing factor and the component signals of all temperatures to obtain the average correlation eigenvalue of the any influencing component signal; calculating the product of the component importance of the any influencing component signal and the average correlation eigenvalue to obtain the comprehensive correlation eigenvalue of the any influencing component signal.
[0012] Further, the step of obtaining the final correlation eigenvalue of the influencing factor according to the comprehensive correlation eigenvalues of all component signals of the influencing factor includes: Calculating the average value of the comprehensive correlation eigenvalues of all component signals corresponding to the influencing factor to obtain the final correlation eigenvalue of the influencing factor.
[0013] Further, the step of performing temperature prediction according to all target sequences and the final correlation eigenvalue of the influencing factor to obtain the predicted temperature at a future moment includes: Taking all influencing factors as exogenous variables in the ARIMAX model for predicting temperature; when the change correlation between the influencing factor and the temperature is positively correlated, taking the final correlation eigenvalue as the regression coefficient of the exogenous variable of the influencing factor, and when the change correlation between the influencing factor and the temperature is negatively correlated, taking the opposite number of the final correlation eigenvalue as the regression coefficient of the exogenous variable of the influencing factor; performing temperature prediction through the ARIMAX model according to the target sequences of the influencing factor and the temperature to obtain the predicted temperature at a future moment.
[0014] The present invention has the following beneficial effects: In the present invention, obtaining the mean sequence can smooth the time series of monitoring data, remove the local fluctuation characteristics, and reflect the overall trend characteristics, so as to minimize the trend characteristics in the subsequent process. Obtaining the change correlation can characterize whether the data change trends between the influencing factors and the temperature are the same, thereby improving the accuracy of temperature prediction. Obtaining the trend characteristic value can characterize the degree of trend of the mean sequence. Obtaining the target sequence can weaken the trend characteristics of the time series of monitoring data, thereby accurately obtaining the influence degree of the influencing factors on the temperature and improving the accuracy of temperature prediction. Obtaining the component signal can characterize the frequency characteristics of the target sequence, and analyze the correlation between the influencing factors and the temperature through the frequency characteristics; obtaining the component importance can characterize the importance degree of the component signal. Obtaining the component correlation characteristic value can characterize the correlation degree between any influencing component signal corresponding to the influencing factor and any temperature component signal; obtaining the comprehensive correlation characteristic value can characterize the importance degree of any influencing component of the influencing factor and the correlation with all temperature component signals; obtaining the final correlation characteristic value can characterize the overall correlation between the influencing factor and the temperature. Finally, temperature prediction is performed based on all the target sequences and the final correlation characteristic values of the influencing factors to obtain the predicted temperature at a future moment, improving the accuracy of temperature prediction. Finally, the temperature in the production process is controlled according to the predicted temperature, which can keep the temperature in a reasonable range during the production process and ensure the quality of the organic fertilizer. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.
[0016] Figure 1 It is a block diagram of an intelligent production system for Cladophora organic fertilizer based on the Internet of Things provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0017] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will describe in detail the specific implementation manners, structures, features and effects of an intelligent production system for Cladophora organic fertilizer based on the Internet of Things proposed according to the present invention in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.
[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this invention belongs.
[0019] The following specifically describes the specific solution of an intelligent production system for Cladophora organic fertilizer based on the Internet of Things provided by the present invention in conjunction with the accompanying drawings.
[0020] Please refer to Figure 1 , which shows a block diagram of an intelligent production system for Cladophora organic fertilizer based on the Internet of Things provided by an embodiment of the present invention. The system includes the following modules: The data acquisition module S1 is used to acquire the monitoring data time series of different temperature influencing factors and temperature in the historical period during the production process of Cladophora organic fertilizer.
[0021] In the embodiment of the present invention, the implementation scenario is to predict the temperature during the production process of Cladophora organic fertilizer, so as to be able to obtain the temperature change trend in advance and control the temperature in a timely manner to ensure the quality of the fertilizer. Since the temperature change is not only affected by a single factor, but there is a complex coupling effect with multiple environmental factors such as humidity, pH value, and oxygen concentration; therefore, in the temperature prediction process, not only the historical temperature during the production process should be considered, but also the data changes of the temperature influencing factors should be taken into account; therefore, the existing ARIMAX model can be used for temperature prediction. The ARIMAX multiple time series prediction model is an extension of the ARIMA model, which adds exogenous variables to improve the temperature prediction accuracy. The exogenous variables are other characteristics that affect the change of the prediction object. However, determining the degree of influence on the temperature change is required when adding exogenous variables for prediction to further improve the prediction accuracy. First, the monitoring data time series of different temperature influencing factors and temperature in the historical period during the production process of Cladophora organic fertilizer are acquired. The historical period is the period before the current moment. In the embodiment of the present invention, the temperature influencing factors include: humidity, pH value, and oxygen concentration during the production process of organic fertilizer. Therefore, the monitoring data time series of temperature, humidity, pH value, and oxygen concentration in the historical period during the production process of organic fertilizer are collected through sensors respectively; it should be noted that the collection frequency and collection period of all characteristics are the same, and the implementer can determine them according to the implementation scenario. In order to avoid the influence of noise data, the acquired monitoring feature time series needs to be filtered through the existing Kalman filter algorithm.
[0022] The time series analysis module S2 is used to obtain the mean sequence according to the data distribution characteristics within a preset sliding window in the monitoring data time series; obtain the change correlation according to the data change trend of the mean sequences of temperature and influencing factors; obtain the trend characteristic value according to the data fluctuation characteristics of the mean sequence; and perform differential iteration processing on the monitoring data time series according to the trend characteristic value to obtain the target sequence.
[0023] Due to the influence of microbial activities and environmental factors, there will be trend changes in the time-series data of temperature, humidity, pH value, and oxygen concentration during the production process of Cladophora organic fertilizer. For example, the temperature rises in the initial stage of decomposition and then gradually decreases; the humidity decreases due to water evaporation or microbial utilization but rises when water is replenished; the pH value first decreases and then increases during the degradation of organic matter; the oxygen concentration decreases during aerobic decomposition and rebounds as decomposition weakens or ventilation increases. In the subsequent process, wavelet transform needs to be used to analyze the transformation of the time series of monitoring data, so as to obtain the influence degree of influencing factors on temperature and improve the accuracy of temperature data prediction. Since the purpose of wavelet transform is to decompose different frequency components of the signal, if the data has a strong trend, it will affect the extraction of high-frequency information and reduce the effectiveness of decomposition. At the same time, the purpose of this solution is to predict the temperature during the fertilizer production process based on multi-source data, and the ARIMAX model itself assumes that the data is stationary. The existence of trends in the data will lead to a poor fitting effect of the model. Therefore, it is necessary to minimize the trend of the data with trends through trend feature analysis to improve the prediction accuracy and robustness of the model. Therefore, the possible trends in the time series of monitoring data are first removed to more accurately obtain the influence degree of influencing factors on temperature and improve the accuracy of temperature prediction.
[0024] Furthermore, a mean sequence is obtained according to the data distribution characteristics within a preset sliding window in the time series of monitoring data. Preferably, in the implementation of the present invention, the steps of obtaining the mean sequence include: calculating the average value of the data within the preset sliding window in the time series of monitoring data to obtain a local average value; sorting all local average values according to the sliding order of the preset sliding window in the time series of monitoring data to obtain the mean sequence corresponding to the time series of monitoring data. The length of the preset sliding window is the data length within a preset time range, and the sliding step size each time is 0.8 times the window length, so that there is an overlapping part between adjacent windows during the sliding process. The implementer can determine the length and sliding step size of the preset sliding window according to the implementation scenario. Obtaining the mean sequence can smooth the time series of monitoring data, ignore local fine fluctuations, and more accurately analyze the trend of the time series of monitoring data.
[0025] After obtaining the mean sequence corresponding to the time series of monitoring data, the influence trends of different influencing factors on temperature are different. To initially improve the prediction accuracy, it is necessary to determine the positive and negative correlations between different influencing factors and temperature. Therefore, the change correlation is obtained according to the data change trends of the mean sequences of temperature and influencing factors. Preferably, in the embodiment of the present invention, the steps of obtaining the change correlation include: performing differencing on the mean sequence and calculating the average value of the differencing result to obtain the differencing average value; when the differencing average value is positive, it means that the mean sequence conforms to an increasing trend; when the differencing average value is negative, it means that the mean sequence conforms to a decreasing trend. If the positive and negative natures of the differencing average values corresponding to temperature and the influencing factor are the same, it means that the data change trends of the two are the same, then the change correlation between temperature and the influencing factor is positively correlated; otherwise, the change correlation is negatively correlated. A positive correlation means that when the data of the influencing factor increases, the temperature will increase synchronously; a negative correlation means that the change in the data of the influencing factor will cause the temperature to decrease.
[0026] Furthermore, it is necessary to determine whether there is a trend in the mean sequence and minimize the trend. When the monotonicity and fluctuation amplitude of the mean sequence are weaker, it means that the trend of the mean sequence is smaller. Therefore, the trend characteristic value is obtained according to the data fluctuation characteristics of the mean sequence. Preferably, in the embodiment of the present invention, the steps of obtaining the trend characteristic value include: calculating the product of the absolute value of the differencing average value of the mean sequence and the standard deviation of the mean sequence to obtain the trend characteristic value corresponding to the mean sequence; when the standard deviation of the mean sequence is smaller, it means that the fluctuation characteristic is smaller. At the same time, when the standard deviation is smaller and the differencing average value is closer to 0, it means that the monotonicity of the mean sequence is weaker. Therefore, the closer the trend characteristic value is to 0, the less the mean sequence has a trend characteristic and the less it is necessary to remove the trend. Furthermore, the time series of monitoring data can be processed by differencing iteration according to the trend characteristic value to obtain the target sequence.
[0027] Preferably, in the embodiments of the present invention, the step of obtaining the target sequence includes: when the trend eigenvalue is a constant of 0, it means that there is no trend feature in the mean sequence, and the time series of the monitoring data of this influencing factor can be used as the target sequence of this influencing factor; when the trend eigenvalue exceeds the constant 0, it means that there is a trend, and the trend needs to be weakened as much as possible. The time series of the monitoring data is differenced and the differencing result is used as the new time series of the monitoring data. It is judged whether the trend eigenvalue of the new time series of the monitoring data is less than the trend eigenvalue of the initial time series of the monitoring data. If it is less, it means that the trend of this differencing result is less than that of the initial time series of the monitoring data, and the trend feature still needs to be further removed. Then, the differencing result of the new time series of the monitoring data is continued to be used as the next time series of the monitoring data, and it is judged whether the trend eigenvalue of the next time series of the monitoring data is less than that of the previous time series of the monitoring data. If it is still less, the above differencing iteration step is repeated; the differencing iteration stops until the trend eigenvalue of the next time series of the monitoring data is not less than that of the previous time series of the monitoring data. At this time, it means that the trend feature of the latest differencing result is more than that of the previous differencing result, so the trend of the previous differencing result is the weakest, and the time series of the previous monitoring data is used as the target sequence. The target sequence is the sequence obtained by removing the trend feature after differencing the time series of the monitoring data of this influencing factor. Subsequently, correlation analysis can be carried out according to the target sequences of different influencing factors and temperatures.
[0028] The temperature influence analysis module S3 is used to perform wavelet transform on the interpolated target sequence to obtain different component signals; obtain the component importance according to the energy intensity characteristics of the component signals; obtain the component correlation eigenvalue according to the component importance, the difference characteristics between the temperature and the component signals corresponding to the influencing factors; obtain the comprehensive correlation eigenvalue of the component signals of the influencing factor according to the component correlation eigenvalue and the component importance; obtain the final correlation eigenvalue of the influencing factor according to the comprehensive correlation eigenvalues of all the component signals of the influencing factor.
[0029] Perform wavelet transform on the interpolated target sequence to obtain different component signals; since the trend degrees of the temperature and different influencing factors are different, the number of differencing times is also different. Before performing wavelet transform, the lengths of the target sequences need to be kept consistent. Therefore, the target sequences are interpolated so that the lengths of all target sequences are the same when transformed. In the embodiments of the present invention, the existing spline interpolation method is used to interpolate the target sequences, so that the length of the shorter target sequence after interpolation is the same as that of the longest target sequence, thereby reducing the error in subsequent analysis. Perform wavelet transform on the interpolated target sequence. It should be noted that wavelet transform belongs to the prior art and the specific steps will not be elaborated. The target sequence is decomposed into component signals of different frequencies through wavelet transform.
[0030] Furthermore, the degree of association between the influencing factor and temperature can be judged according to the frequency characteristics of the target sequence. First, the importance of components is obtained based on the energy intensity characteristics of the component signals. Preferably, in the embodiments of the present invention, the step of obtaining the importance of components includes: calculating the ratio of the energy intensity of any component signal to the sum of the energy intensities of all component signals obtained by the corresponding transformation object to obtain the importance of this arbitrary component signal; the energy intensity is the sum of the squares of all amplitudes of this arbitrary component signal; the greater the energy intensity, it means that the energy contribution degree of this arbitrary component signal in the target sequence to which it belongs is greater, the proportion of the principal component information of this arbitrary component signal is greater, and the characteristics of this arbitrary component signal are more important in the subsequent analysis process.
[0031] If the importance of a certain component signal corresponding to the influencing factor is more similar to that of a certain component signal of temperature, and their frequency characteristics are also similar at the same time, then it indicates that the correlation between a certain component signal of the influencing factor and a certain component signal of temperature is greater. Therefore, the component correlation eigenvalue can be obtained according to the importance of components, the difference characteristics of temperature and the component signals corresponding to the influencing factor. Preferably, in the embodiments of the present invention, the step of obtaining the component correlation eigenvalue includes: calculating the absolute value of the difference between the importance of any influencing component signal in the influencing factor and the importance of any temperature component signal of temperature to obtain the importance difference value; when the importance difference value is larger, it means that the difference between the importance of this arbitrary influencing component signal and the importance of this arbitrary temperature component signal is greater. Calculate the dynamic time warping distance between any influencing component signal and any temperature component signal and perform a negative correlation mapping to obtain the similarity; it should be noted that the dynamic time warping distance is obtained by the existing dynamic time warping algorithm. When two sequences are more similar, the dynamic time warping distance is smaller; the greater the similarity, it means that this arbitrary influencing component is more similar to this arbitrary temperature component signal and the frequency characteristics are closer. Calculate the sum value of the importance difference value and a preset constant to obtain the difference characterization value; the preset constant needs to be greater than or equal to zero to avoid the situation where the importance difference value is 0. In the embodiments of the present invention, the preset constant is 1. Calculate the ratio of the similarity to the difference characterization value to obtain the component correlation eigenvalue; when the similarity is greater and the importance difference value is smaller, it means that the correlation between this arbitrary influencing component signal and this arbitrary temperature component signal is greater, and the importance degrees of the component signals are close, the correlation characteristics of the two are more obvious, and the component correlation eigenvalue is larger. When the component correlation eigenvalue is smaller, it means that the degree of association between this arbitrary influencing component and this arbitrary temperature component is smaller. The formula for obtaining the component correlation eigenvalue includes:
[0032] In the formula, represents the component correlation eigenvalue of the a-th influencing component signal and the b-th temperature component signal, n represents the preset constant, represents the importance of the a-th influencing component signal, Represents the component importance degree of the b-th temperature component signal, Represents the importance degree difference value, Represents the difference characterization value, Represents the exponential function with the natural constant as the base, Represents the a-th influencing component signal, Represents the b-th temperature component signal, Represents calculating the dynamic time warping distance, Represents the similarity.
[0033] Furthermore, if the component correlation eigenvalue between any influencing component signal of the influencing factor and all component signals of the temperature is larger, and the component importance degree of the any influencing component signal is larger, it means that the correlation degree between the any influencing component signal and the temperature data is larger, and the any influencing component signal is more important in the influencing factor, and further indicates that the correlation degree between the influencing factor and the temperature data is larger. Therefore, the comprehensive correlation eigenvalue of the component signal of the influencing factor is obtained according to the component correlation eigenvalue and the component importance degree. Preferably, in the embodiment of the present invention, the steps of obtaining the comprehensive correlation eigenvalue include: calculating the average value of the component correlation eigenvalues between any influencing component signal of the influencing factor and all component signals of the temperature to obtain the average correlation eigenvalue of the any influencing component signal; when the average correlation eigenvalue is larger, it means that the correlation between the any influencing component signal and all component signals of the temperature is larger. Calculating the product of the component importance degree of the any influencing component signal and the average correlation eigenvalue to obtain the comprehensive correlation eigenvalue of the any influencing component signal; when the component importance degree of the any influencing component signal is larger, it means that the any influencing component signal can better represent the characteristics of the influencing factor, and at the same time, when the average correlation eigenvalue is larger, the comprehensive correlation eigenvalue is larger, which means that the correlation degree between the influencing factor and the temperature is larger, and the influencing factor can more influence the change of the temperature. The formula for obtaining the comprehensive correlation eigenvalue includes:
[0034] In the formula, Represents the a-th influencing component signal of the influencing factor, Represents the component importance degree of the a-th influencing component signal, B represents the number of component signals of the temperature, Represents the component correlation eigenvalue between the a-th influencing component signal and the b-th temperature component signal, Represents the average correlation eigenvalue.
[0035] After obtaining the comprehensive correlation eigenvalue of any influence component signal of the influencing factor, the final correlation eigenvalue of the influencing factor can be obtained according to the comprehensive correlation eigenvalues of all component signals of the influencing factor; preferably, in the embodiments of the present invention, the steps of obtaining the final correlation eigenvalue include: calculating the average value of the comprehensive correlation eigenvalues of all component signals corresponding to the influencing factor to obtain the final correlation eigenvalue of the influencing factor; when the final correlation eigenvalue is larger, it means that the correlation between the influencing factor and the temperature is greater, and when the influencing factor changes, the temperature will also change to a certain extent. When the final correlation eigenvalue is smaller, it means that the correlation between the two is smaller, and when the influencing factor changes, the degree of influence on the temperature is smaller.
[0036] The temperature prediction module S4 is used to perform temperature prediction according to all target sequences and the final correlation eigenvalue of the influencing factor to obtain the predicted temperature at a future time; control the temperature in production according to the predicted temperature.
[0037] After obtaining the final correlation eigenvalues of all influencing factors, the final correlation eigenvalues can characterize the degree of influence of the influencing factors on the temperature. The larger this value is, the more it can affect the change of the temperature. Therefore, the degree of influence can be used as the regression coefficient of the exogenous variable in the ARIMAX model. Temperature prediction can be performed according to all target sequences and the final correlation eigenvalues of the influencing factors to obtain the predicted temperature at a future time; preferably, in the embodiments of the present invention, the steps of obtaining the predicted temperature include: taking all influencing factors as the exogenous variables in the ARIMAX model for predicting the temperature. When the correlation between the influencing factor and the change of the temperature is positively correlated, it means that the data changes of the two show a positive correlation trend. Therefore, the final correlation eigenvalue is used as the regression coefficient of the exogenous variable of this influencing factor. When the correlation between the influencing factor and the change of the temperature is negatively correlated, it means that the data changes of the two show a negative correlation trend. Therefore, the opposite number of the final correlation eigenvalue is used as the regression coefficient of the exogenous variable of this influencing factor. Then, temperature prediction is performed through the ARIMAX model according to the target sequences of the influencing factors and the temperature to obtain the predicted temperature at a future time. It should be noted that this ARIMAX model belongs to the prior art, and the specific prediction steps will not be elaborated. Since the time series input into this model is the target sequence after differencing the time series of each monitoring data, the differencing analysis step of the monitoring data time series in the subsequent steps of this model can be saved, facilitating the model to perform prediction based on the differenced data, and restoring the predicted data according to the recorded number of differencing times. By obtaining the final correlation eigenvalues of all influencing factors, the accuracy of predicting the temperature using the ARIMAX model is improved.
[0038] Furthermore, after obtaining the predicted temperature at a future time, the implementer can judge the temperature change trend in the fertilizer production process based on the predicted temperature, and then control the temperature during production according to the predicted temperature, so that the temperature during the production process is maintained within a reasonable range to ensure the quality of the fertilizer.
[0039] In summary, the embodiment of the present invention provides an intelligent production system for Cladophora organic fertilizer based on the Internet of Things; obtaining the change correlation based on the mean sequences of temperature and influencing factors; obtaining the trend characteristic value according to the mean sequence and performing differencing on the time series of the monitoring data to obtain the target sequence; performing wavelet transform on the target sequence after interpolation to obtain different component signals; obtaining the component importance according to the energy intensity characteristics of the component signals; obtaining the component correlation characteristic value according to the component importance and the component signals corresponding to the temperature and influencing factors; obtaining the final correlation characteristic value of the influencing factor according to the component correlation characteristic value and the component importance. The present invention predicts the temperature according to all the target sequences and the final correlation characteristic values of the influencing factors; controls the temperature during production according to the predicted temperature, improves the accuracy of temperature prediction, and keeps the temperature during the production process within a reasonable range.
[0040] It should be noted that: the above sequence of the embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0041] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.
Claims
1. An intelligent production system for Cladophora organic fertilizer based on the Internet of Things, characterized in that, The system includes the following modules: A data acquisition module, configured to acquire the monitoring data time series of different temperature influencing factors and temperatures during a historical period in the production process of Cladophora organic fertilizer; A time series analysis module, configured to obtain a mean sequence according to the data distribution characteristics within a preset sliding window in the monitoring data time series; obtain a change correlation according to the data change trend of the mean sequences of temperature and influencing factors; obtain a trend characteristic value according to the data fluctuation characteristics of the mean sequences; Perform a difference iteration process on the monitoring data time series according to the trend characteristic value to obtain a target sequence; A temperature influence analysis module, configured to perform wavelet transform on the target sequence after interpolation to obtain different component signals; obtain a component importance according to the energy intensity characteristics of the component signals; obtain a component correlation characteristic value according to the difference characteristics between the component importance and the component signals corresponding to temperature and influencing factors; obtain a comprehensive correlation characteristic value of the component signals of the influencing factors according to the component correlation characteristic value and the component importance; obtain a final correlation characteristic value of the influencing factors according to the comprehensive correlation characteristic values of all component signals of the influencing factors; A temperature prediction module, configured to perform temperature prediction according to all target sequences and the final correlation characteristic values of influencing factors to obtain a predicted temperature at a future moment; control the temperature in production according to the predicted temperature.
2. The intelligent production system of cladophora organic fertilizer based on the Internet of Things according to claim 1, characterized in that The step of obtaining a mean sequence according to the data distribution characteristics within a preset sliding window in the monitoring data time series includes: Calculate the data average value within a preset sliding window in the monitoring data time series to obtain a local average value; sort all local average values according to the sliding order of the preset sliding window in the monitoring data time series to obtain the mean sequence.
3. The intelligent production system of Cladophora organic fertilizer based on the Internet of Things according to claim 1, wherein, The step of obtaining a change correlation according to the data change trend of the mean sequences of temperature and influencing factors includes: Perform differencing on the mean sequence and calculate the average value of the differencing result to obtain a differencing average value; if the positive and negative natures of the differencing average values corresponding to the temperature and influencing factors are the same, the change correlation between the temperature and influencing factors is positively correlated, otherwise the change correlation is negatively correlated.
4. An intelligent production system for Cladophora organic fertilizer based on the Internet of Things according to claim 3, characterized in that, The step of obtaining a trend characteristic value according to the data fluctuation characteristics of the mean sequence includes: Calculate the product of the absolute value of the differencing average value of the mean sequence and the standard deviation of the mean sequence to obtain the trend characteristic value corresponding to the mean sequence.
5. The intelligent production system of Cladophora organic fertilizer based on the Internet of Things according to claim 1, characterized in that, The step of performing a difference iteration process on the monitoring data time series according to the trend characteristic value to obtain a target sequence includes: When the trend feature value is a constant of 0, the time series of the monitoring data of the influencing factor is used as the target sequence of the influencing factor; when the trend feature value exceeds the constant 0, the time series of the monitoring data is differenced and the differencing result is used as the time series of the new round of monitoring data. Determine whether the trend feature value of the time series of the new round of monitoring data is less than the trend feature value of the initial time series of the monitoring data. If it is less, continue to use the differencing result of the time series of the new round of monitoring data as the time series of the next round of monitoring data, and determine whether the trend feature value of the time series of the next round of monitoring data is less than the time series of the previous round of monitoring data until the trend feature value of the time series of the next round of monitoring data is not less than the time series of the previous round of monitoring data, then stop the differencing iteration, and use the time series of the previous round of monitoring data as the target sequence.
6. The intelligent production system of cladophora organic fertilizer based on the Internet of Things according to claim 1, characterized in that The step of obtaining the component importance according to the energy intensity feature of the component signal includes: Calculate the ratio of the energy intensity of any component signal to the total energy intensity of all component signals obtained from the corresponding transformation object, and obtain the component importance of the any component signal.
7. An intelligent production system for Cladophora organic fertilizer based on the Internet of Things according to claim 1, characterized in that, The step of obtaining the component correlation feature value according to the component importance, the difference feature between the temperature and the component signal corresponding to the influencing factor includes: Calculate the absolute value of the difference between the component importance of any influencing component signal in the influencing factor and any temperature component signal of the temperature to obtain the importance difference value; calculate the dynamic time warping distance between the any influencing component signal and the any temperature component signal and perform a negative correlation mapping to obtain the similarity; calculate the sum value of the importance difference value and a preset constant to obtain the difference characterization value; calculate the ratio of the similarity to the difference characterization value to obtain the component correlation feature value.
8. An intelligent production system for Cladophora organic fertilizer based on the Internet of Things according to claim 7, characterized in that, The step of obtaining the comprehensive correlation feature value of the component signal of the influencing factor according to the component correlation feature value and the component importance includes: Calculate the average value of the component correlation feature values of any influencing component signal of the influencing factor and the component signals of all temperatures to obtain the average correlation feature value of the any influencing component signal; calculate the product of the component importance of the any influencing component signal and the average correlation feature value to obtain the comprehensive correlation feature value of the any influencing component signal.
9. The intelligent production system of Cladophora organic fertilizer based on the Internet of Things according to the claim is characterized in that, The step of obtaining the final correlation feature value of the influencing factor according to the comprehensive correlation feature values of all component signals of the influencing factor includes: Calculate the average value of the comprehensive correlation feature values of all component signals corresponding to the influencing factor to obtain the final correlation feature value of the influencing factor.
10. The intelligent production system for Cladophora organic fertilizer based on the Internet of Things according to claim 3, characterized in that, The step of predicting the temperature at a future moment according to all target sequences and the final correlation feature value of the influencing factor to obtain the predicted temperature at a future moment includes: All influencing factors are used as exogenous variables in the ARIMAX model for predicting temperature; when the correlation between the influencing factor and the change in temperature is positively correlated, the final correlation eigenvalue is used as the regression coefficient of the exogenous variable of the influencing factor, and when the correlation between the influencing factor and the change in temperature is negatively correlated, the opposite of the final correlation eigenvalue is used as the regression coefficient of the exogenous variable of the influencing factor; temperature prediction is carried out through the ARIMAX model based on the target sequences of the influencing factors and temperature to obtain the predicted temperature at future moments.