A coated controlled-release fertilizer and preparation method thereof

Abstract

The present application discloses a coated controlled-release fertilizer and a preparation method thereof, and relates to the technical field of fertilizer preparation. In the process of using a pressure pump to atomize a coating liquid to obtain an atomized coating liquid, a time queue of the coating liquid atomization pressure collected by a pressure sensor is obtained, and a data analysis and processing technology based on deep learning is used to scan and sample the time queue of the coating liquid atomization pressure and perform local time series correlation, thereby automatically obtaining the recommended value of the atomization speed at the next time point according to the propagation characteristics of the local time series correlation characteristics of the atomization pressure of each coating liquid in the entire time domain. In this way, the atomization speed can be dynamically adjusted according to the real-time changing pressure data, rather than relying on a static set value. In this way, it is ensured that pressure changes can be responded to at any time during the production process, thereby improving the intelligence level of the atomization process.

Description

Technical Field

[0001] The present application relates to the technical field of fertilizer preparation, and more specifically, to a coated controlled-release fertilizer and a preparation method thereof. Background Art

[0002] Coated controlled-release fertilizers refer to fertilizers that are released in the soil at a controlled rate by forming one or more layers of coating material on the surface of fertilizer particles, thereby achieving the effect of extending the release period of the fertilizer. This technology can reduce the number of fertilizations, improve fertilizer utilization, and reduce environmental pollution. Chinese patent CN103951503B proposes a method for preparing coated controlled-release fertilizers, which is prepared by mixing polyisocyanates with polyols to generate water-soluble polyurethane, adding turpentine and polyvinyl alcohol to form a coating liquid, and spraying the coating liquid on the surface of preheated fertilizer particles, and forming a coating after drying.

[0003] After the coating liquid is formed, it needs to be atomized to evenly disperse the coating liquid into fine droplets, so that these droplets can more easily and evenly adhere to the surface of the fertilizer particles, thereby ensuring that each fertilizer particle can obtain a consistent coating layer. The control of the atomization speed is one of the key factors to achieve this goal. However, the traditional method of adjusting the atomization speed relies on manual experience. This manual adjustment method is prone to introduce subjective errors, resulting in unstable atomization speed during the production process. In addition, during the atomization process, changes in pressure will also affect the atomization speed. The traditional method usually controls the atomization speed based on a static set value, without taking into account the real-time changing pressure data, and therefore cannot be adjusted dynamically.

[0004] Therefore, an optimized preparation scheme for coated controlled-release fertilizers is desired. Summary of the invention

[0005] The present application provides a coated controlled-release fertilizer and a preparation method thereof, which can dynamically adjust the atomization speed according to the real-time changing pressure data, rather than relying on a static set value, so as to ensure that the pressure changes can be responded to at any time during the production process, thereby improving the intelligence level of the atomization process.

[0006] In a first aspect, a method for preparing a coated controlled-release fertilizer is provided, comprising: preparing a water-soluble polyurethane; mixing the water-soluble polyurethane, turpentine and polyvinyl alcohol in a predetermined ratio and placing the mixture in a reactor for stirring and reacting to obtain a coating liquid; atomizing the coating liquid using a pressure pump to obtain an atomized coating liquid; heating fertilizer particles to 60° C., spraying the atomized coating liquid on the surface of the heated fertilizer particles using a spray device, and performing hot air drying on the fertilizer particles sprayed with the coating liquid to obtain a first coated fertilizer; sending the first coated fertilizer into a circulating fluidized bed for further coating treatment to obtain the coated controlled-release fertilizer;

[0007] The method of using a pressure pump to atomize the coating liquid to obtain an atomized coating liquid comprises:

[0008] Obtaining a time sequence of coating liquid atomization pressure collected by a pressure sensor;

[0009] Scanning and sampling the time queue of the coating liquid atomization pressure to obtain a set of local time series of the coating liquid atomization pressure, and then performing sequence encoding to obtain a set of local time series correlation features of the coating liquid atomization pressure;

[0010] The node potential energy of the set of local time series correlation characteristics of the coating liquid atomization pressure is significantly propagated to obtain the time series significant propagation characteristics of the coating liquid atomization pressure, including: extracting the current local time series correlation characteristics of the coating liquid atomization pressure and the set of historical local time series correlation characteristics of the coating liquid atomization pressure from the set of local time series correlation characteristics of the coating liquid atomization pressure; performing feature propagation of energy space-time transfer on the set of historical local time series correlation characteristics of the coating liquid atomization pressure and the current local time series correlation characteristics of the coating liquid atomization pressure to obtain the time series significant propagation characteristics of the coating liquid atomization pressure; and obtaining an atomization speed control instruction based on the time series significant propagation characteristics of the coating liquid atomization pressure.

[0011] In a possible implementation, scanning and sampling the time queue of the coating liquid atomization pressure to obtain a set of local time series of the coating liquid atomization pressure, and then performing sequence encoding to obtain a set of local time series correlation features of the coating liquid atomization pressure, including:

[0012] Based on a predetermined time window and a predetermined step length, scanning and sampling the time queue of the coating liquid atomization pressure to obtain a set of local time series of the coating liquid atomization pressure;

[0013] The local time series of each coating liquid atomization pressure in the set of local time series of the coating liquid atomization pressure are respectively input into the sequence encoder based on the Bi-LSTM model to obtain a set of local time series correlation feature vectors of the coating liquid atomization pressure as the set of local time series correlation features of the coating liquid atomization pressure.

[0014] In a possible implementation, extracting a set of current local time series correlation features of coating liquid atomization pressure and a set of historical local time series correlation features of coating liquid atomization pressure from the set of local time series correlation features of coating liquid atomization pressure includes:

[0015] A current local timing correlation feature vector of the coating liquid atomization pressure is extracted from the set of local timing correlation feature vectors of the coating liquid atomization pressure as the current local timing correlation feature of the coating liquid atomization pressure, and other local timing correlation feature vectors of the coating liquid atomization pressure in the set of local timing correlation feature vectors of the coating liquid atomization pressure are defined as historical local timing correlation feature vectors of the coating liquid atomization pressure, so as to obtain a set of historical local timing correlation feature vectors of the coating liquid atomization pressure as the set of historical local timing correlation features of the coating liquid atomization pressure.

[0016] In a possible implementation, the set of the historical coating liquid atomization pressure local time series correlation features and the current coating liquid atomization pressure local time series correlation features are subjected to feature propagation of energy space-time transfer to obtain the coating liquid atomization pressure time series significant propagation features, including:

[0017] Calculating the coating liquid atomization pressure static energy value of each historical coating liquid atomization pressure local time series correlation characteristic vector in the set of the historical coating liquid atomization pressure local time series correlation characteristic vectors;

[0018] Performing spatiotemporal modulation on the static energy values ​​of the coating liquid atomization pressure of each historical local time series associated characteristic vector of the coating liquid atomization pressure to obtain the spatiotemporal transfer value of the dynamic energy of the coating liquid atomization pressure of each historical local time series associated characteristic vector of the coating liquid atomization pressure relative to the current local time series associated characteristic vector of the coating liquid atomization pressure;

[0019] Inputting the coating liquid atomization pressure dynamic energy spatiotemporal transfer values ​​of the local time series associated characteristic vectors of the historical coating liquid atomization pressure into the energy transfer gating module based on the mask function to obtain a set of coating liquid atomization pressure dynamic energy spatiotemporal transfer weights;

[0020] Based on the set of spatiotemporal transfer weights of the dynamic energy of the coating liquid atomization pressure, a weighted sum of the set of local time series associated characteristic vectors of the historical coating liquid atomization pressure is calculated to obtain a historical coating liquid atomization pressure time series transfer aggregation modulation vector;

[0021] The positional sum of the historical coating liquid atomization pressure time series transmission aggregation modulation vector and the current coating liquid atomization pressure local time series correlation feature vector is calculated to obtain the full time domain coating liquid atomization pressure time series significant propagation feature vector as the coating liquid atomization pressure time series significant propagation feature.

[0022] In a possible implementation, calculating the coating liquid atomization pressure static energy value of each historical coating liquid atomization pressure local time series correlation feature vector in the set of historical coating liquid atomization pressure local time series correlation feature vectors includes:

[0023] Extracting the maximum value of the local time series correlation characteristic vector of the historical coating liquid atomization pressure to obtain the maximum value of the local time series correlation characteristic of the historical coating liquid atomization pressure;

[0024] Respectively calculating the mean and variance of the local time series correlation characteristic vector of the historical coating liquid atomization pressure to obtain the mean of the local time series correlation characteristic of the historical coating liquid atomization pressure and the variance of the local time series correlation characteristic of the historical coating liquid atomization pressure; subtracting the local time series correlation characteristic vector of the historical coating liquid atomization pressure from the local time series correlation characteristic mean of the historical coating liquid atomization pressure by position, calculating the fourth power of the characteristic value of each position in the local time series difference vector of the historical coating liquid atomization pressure to obtain the local time series difference modulation vector of the historical coating liquid atomization pressure; calculating the expected value of the local time series difference modulation vector of the historical coating liquid atomization pressure, and dividing the obtained expected value by the square of the variance of the local time series correlation characteristic of the historical coating liquid atomization pressure to obtain the local time series energy value of the historical coating liquid atomization pressure;

[0025] The maximum value of the local time series correlation characteristic of the historical coating liquid atomization pressure is added to the local time series energy value of the historical coating liquid atomization pressure to obtain the static energy value of the coating liquid atomization pressure.

[0026] In a possible implementation, the coating liquid atomization pressure static energy value of each historical coating liquid atomization pressure local time series correlation characteristic vector is subjected to spatiotemporal modulation to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value of each historical coating liquid atomization pressure local time series correlation characteristic vector relative to the current coating liquid atomization pressure local time series correlation characteristic vector, including:

[0027] Determine the coating liquid atomization pressure energy spatial attenuation factor based on the spatial span coefficient between the local time series correlation characteristic vectors of the respective historical coating liquid atomization pressures and the local time series correlation characteristic vector of the current coating liquid atomization pressure;

[0028] Determine the coating liquid atomization pressure energy time attenuation factor based on the time span coefficient between the local time series correlation characteristic vectors of the respective historical coating liquid atomization pressures and the local time series correlation characteristic vector of the current coating liquid atomization pressure;

[0029] Based on the spatial attenuation factor of the coating liquid atomization pressure energy and the temporal attenuation factor of the coating liquid atomization pressure energy, the static energy value of the coating liquid atomization pressure of each historical local time series correlation characteristic vector of the coating liquid atomization pressure is temporally and spatially modulated to obtain the spatiotemporal transfer value of the dynamic energy of the coating liquid atomization pressure of each historical local time series correlation characteristic vector of the coating liquid atomization pressure relative to the current local time series correlation characteristic vector of the coating liquid atomization pressure.

[0030] In a possible implementation, based on the spatial attenuation factor of the coating liquid atomization pressure energy and the temporal attenuation factor of the coating liquid atomization pressure energy, the static energy value of the coating liquid atomization pressure of each historical coating liquid atomization pressure local time series correlation characteristic vector is temporally modulated to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value of the coating liquid atomization pressure of each historical coating liquid atomization pressure local time series correlation characteristic vector relative to the current coating liquid atomization pressure local time series correlation characteristic vector, including: multiplying the coating liquid atomization pressure energy spatial attenuation factor and the coating liquid atomization pressure energy temporal attenuation factor of the historical coating liquid atomization pressure local time series correlation characteristic vector with the first modulation parameter and the second modulation parameter respectively, and adding the obtained coating liquid atomization pressure energy spatial attenuation modulation factor and coating liquid atomization pressure energy temporal attenuation modulation factor to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value.

[0031] In a possible implementation, the dynamic energy spatiotemporal transfer value of the coating liquid atomization pressure of each historical local time series correlation characteristic vector of the coating liquid atomization pressure is input into an energy transfer gating module based on a mask function to obtain a set of dynamic energy spatiotemporal transfer weights of the coating liquid atomization pressure, including: in response to the dynamic energy spatiotemporal transfer value of the coating liquid atomization pressure being greater than a predetermined threshold, the dynamic energy spatiotemporal transfer value of the coating liquid atomization pressure is input into a sigmoid function to obtain a value as the dynamic energy spatiotemporal transfer weight of the coating liquid atomization pressure.

[0032] In one possible implementation, based on the significant propagation characteristics of the full-time-domain coating liquid atomization pressure time series, an atomization speed control instruction is obtained, including: inputting the significant propagation characteristic vector of the full-time-domain coating liquid atomization pressure time series into a decoder-based atomization speed control module to obtain the atomization speed control instruction, and the atomization speed control instruction includes a decoded value of the recommended value of the atomization speed at the next time point.

[0033] In a second aspect, a coated controlled-release fertilizer is provided, wherein the coated controlled-release fertilizer is prepared by the method for preparing the coated controlled-release fertilizer according to the first aspect.

[0034] The present application provides a coated controlled-release fertilizer and a preparation method thereof, in which a pressure pump is used to atomize the coating liquid to obtain the atomized coating liquid, by obtaining the time queue of the coating liquid atomization pressure collected by the pressure sensor, and using the data analysis and processing technology based on deep learning to scan and sample the coating liquid atomization pressure and locally correlate the time series, so as to automatically obtain the recommended value of the atomization speed at the next time point according to the propagation characteristics of the local time series correlation characteristics of the atomization pressure of each coating liquid in the entire time domain. In this way, the atomization speed can be dynamically adjusted according to the real-time changing pressure data, rather than relying on the static set value. In this way, it is ensured that pressure changes can be responded to at any time during the production process, thereby improving the intelligence level of the atomization process. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0036] Figure 1 This is a schematic flow chart of a method for preparing a coated controlled-release fertilizer according to an embodiment of the present application.

[0037] Figure 2 This is a schematic flow chart of the method for preparing the coated controlled-release fertilizer according to an embodiment of the present application, in which a pressure pump is used to atomize the coating liquid to obtain the atomized coating liquid.

[0038] Figure 3 In the method for preparing the coated controlled-release fertilizer according to the embodiment of the present application, a pressure pump is used to atomize the coating liquid to obtain a data flow diagram of the atomized coating liquid.

[0039] Figure 4 In the method for preparing the coated controlled-release fertilizer of an embodiment of the present application, a schematic flow chart of performing significant propagation of node potential energy on a set of local time series correlation characteristics of the coating liquid atomization pressure to obtain significant time series propagation characteristics of the coating liquid atomization pressure is provided.

[0040] Figure 5 In the method for preparing the coated controlled-release fertilizer of an embodiment of the present application, a schematic flow chart of the significant propagation characteristics of the coating liquid atomization pressure time series is obtained by performing characteristic propagation of energy space-time transfer on the set of the historical coating liquid atomization pressure local time series correlation characteristics and the current coating liquid atomization pressure local time series correlation characteristics. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0042] Based on this, the present application proposes a method for preparing a coated controlled-release fertilizer. Figure 1 Schematic flow chart of the method for preparing the coated controlled-release fertilizer of the embodiment of the present application. Figure 1 As shown, the preparation method of the coated controlled-release fertilizer includes: S1, preparing water-soluble polyurethane; mixing the water-soluble polyurethane, turpentine and polyvinyl alcohol in a predetermined ratio and placing the mixture in a reactor for stirring reaction to obtain a coating liquid; S2, atomizing the coating liquid using a pressure pump to obtain an atomized coating liquid; S3, heating the fertilizer particles to 60°C, spraying the atomized coating liquid on the surface of the heated fertilizer particles using a spray device, and performing hot air drying on the fertilizer particles sprayed with the coating liquid to obtain the first coated fertilizer; S4, sending the first coated fertilizer into the circulating fluidized bed to continue the coating treatment to obtain the coated controlled-release fertilizer.

[0043] Accordingly, in the process of atomizing the coating liquid using a pressure pump to obtain the atomized coating liquid, the technical concept of the present application is to obtain the time queue of the coating liquid atomization pressure collected by the pressure sensor, and use the data analysis and processing technology based on deep learning to scan and sample the coating liquid atomization pressure and locally correlate the time series, so as to automatically obtain the recommended value of the atomization speed at the next time point according to the propagation characteristics of the local time series correlation characteristics of the atomization pressure of each coating liquid in the entire time domain. In this way, the atomization speed can be dynamically adjusted according to the real-time changing pressure data, rather than relying on the static set value. In this way, it is ensured that the pressure changes can be responded to at any time during the production process, thereby improving the intelligence level of the atomization process.

[0044] Figure 2 This is a schematic flow chart of the method for preparing the coated controlled-release fertilizer according to an embodiment of the present application, in which a pressure pump is used to atomize the coating liquid to obtain the atomized coating liquid. Figure 3 The method for preparing the coated controlled-release fertilizer of the embodiment of the present application is to use a pressure pump to atomize the coating liquid to obtain a data flow diagram of the atomized coating liquid. Figure 2 and Figure 3As shown, a pressure pump is used to atomize the coating liquid to obtain an atomized coating liquid, including: S21, obtaining a time queue of the coating liquid atomization pressure collected by a pressure sensor; S22, scanning and sampling the time queue of the coating liquid atomization pressure to obtain a set of local time series of the coating liquid atomization pressure, and then performing sequence encoding to obtain a set of local time series correlation features of the coating liquid atomization pressure; S23, performing node potential energy significant propagation on the set of local time series correlation features of the coating liquid atomization pressure to obtain a time series significant propagation feature of the coating liquid atomization pressure; S24, obtaining an atomization speed control instruction based on the time series significant propagation feature of the coating liquid atomization pressure.

[0045] In the preparation method of the above-mentioned coated controlled-release fertilizer, the S21 obtains the time queue of the coating liquid atomization pressure collected by the pressure sensor. Optionally, in one embodiment of the present application, the time queue of the coating liquid atomization pressure collected by the pressure sensor is obtained, including: installing a pressure sensor at a key part of the atomization equipment to monitor the state of the coating liquid under high pressure. When the coating liquid is atomized using a pressure pump, the pressure sensor continuously collects the pressure data of the coating liquid during the atomization process to form a series of pressure readings at time points. The above-mentioned collected pressure data are arranged in chronological order to form a time queue. Each element in this queue is a pressure value corresponding to a time point. In this way, we can obtain a time queue of the coating liquid atomization pressure collected by the pressure sensor, which is further used as the basic data for subsequent analysis and processing. These data will be used for scanning sampling and sequence encoding to capture the local timing characteristics of the atomization pressure, and finally dynamically adjust the atomization speed to improve the intelligent level of production.

[0046] In the above-mentioned method for preparing coated controlled-release fertilizers, the S22 scans and samples the time queue of the coating liquid atomization pressure to obtain a set of local time series of the coating liquid atomization pressure, and then performs sequence encoding to obtain a set of local time series correlation characteristics of the coating liquid atomization pressure. It should be understood that, considering that the time series characteristics and dynamic changes of the atomization pressure in different local times in the coating liquid atomization pressure are different, based on this, in order to analyze and understand the local change pattern of the atomization pressure in different time ranges in a more fine-grained manner, in the technical solution of the present application, based on a predetermined time window and a predetermined step size, the time queue of the coating liquid atomization pressure is scanned and sampled to obtain a set of local time series of the coating liquid atomization pressure.

[0047] Then, considering that the local time series of each coating liquid atomization pressure in the set of local time series of the coating liquid atomization pressure have mutual correlation and influence in different time spans, therefore, in order to better capture and characterize the temporal variation characteristics and trends of the atomization pressure in different time scales, in the technical solution of the present application, the local time series of each coating liquid atomization pressure in the set of local time series of the coating liquid atomization pressure are respectively input into the sequence encoder based on the Bi-LSTM model to obtain a set of local time series correlation feature vectors of the coating liquid atomization pressure.

[0048] Optionally, in one embodiment of the present application, the S22, scans and samples the time queue of the coating liquid atomization pressure to obtain a set of local time series of the coating liquid atomization pressure, and then performs sequence encoding to obtain a set of local time series correlation features of the coating liquid atomization pressure, including: based on a predetermined time window and a predetermined step size, scans and samples the time queue of the coating liquid atomization pressure to obtain a set of local time series of the coating liquid atomization pressure; and inputs each local time series of the coating liquid atomization pressure in the set of local time series of the coating liquid atomization pressure into a sequence encoder based on the Bi-LSTM model to obtain a set of local time series correlation feature vectors of the coating liquid atomization pressure as a set of local time series correlation features of the coating liquid atomization pressure.

[0049] In the preparation method of the above-mentioned coated controlled-release fertilizer, the S23 performs node potential energy significant propagation on the set of local time-series correlation characteristics of the coating liquid atomization pressure to obtain the coating liquid atomization pressure time-series significant propagation characteristics. It should be understood that considering that the change of the coating liquid atomization pressure at different time points is not only related to its current state, but also closely related to the previous state. Therefore, in order to analyze and extract the time and space relationship contained in the characteristics, so as to better understand and predict the dynamic changes in the atomization process, in the technical solution of the present application, the set of local time-series correlation feature vectors of the coating liquid atomization pressure is subjected to node potential energy significant propagation to obtain the full-time domain coating liquid atomization pressure time-series significant propagation feature vector as the coating liquid atomization pressure time-series significant propagation feature.

[0050] Optionally, in one embodiment of the present application, Figure 4 In the method for preparing the coated controlled-release fertilizer of the embodiment of the present application, a schematic flow chart of performing significant propagation of node potential energy on the set of local time series correlation features of the coating liquid atomization pressure to obtain significant propagation features of the coating liquid atomization pressure time series. Figure 4As shown, the set of local time series correlation characteristics of the coating liquid atomization pressure is subjected to significant propagation of node potential energy to obtain significant time series propagation characteristics of the coating liquid atomization pressure, including: S231, extracting the current local time series correlation characteristics of the coating liquid atomization pressure and the set of historical local time series correlation characteristics of the coating liquid atomization pressure from the set of local time series correlation characteristics of the coating liquid atomization pressure; S232, performing feature propagation of energy spatiotemporal transfer on the set of historical local time series correlation characteristics of the coating liquid atomization pressure and the current local time series correlation characteristics of the coating liquid atomization pressure to obtain significant time series propagation characteristics of the coating liquid atomization pressure.

[0051] Optionally, in one embodiment of the present application, the S231 extracts a set of current coating liquid atomization pressure local timing association features and historical coating liquid atomization pressure local timing association features from the set of coating liquid atomization pressure local timing association features, including: extracting a current coating liquid atomization pressure local timing association feature vector from the set of coating liquid atomization pressure local timing association feature vectors as the current coating liquid atomization pressure local timing association feature, and defining other coating liquid atomization pressure local timing association feature vectors in the set of coating liquid atomization pressure local timing association feature vectors as historical coating liquid atomization pressure local timing association feature vectors, so as to obtain a set of historical coating liquid atomization pressure local timing association feature vectors as the set of historical coating liquid atomization pressure local timing association features.

[0052] Optionally, in one embodiment of the present application, Figure 5In the method for preparing the coated controlled-release fertilizer of an embodiment of the present application, a schematic flow chart of the significant propagation characteristics of the coating liquid atomization pressure time series is obtained by performing characteristic propagation of energy space-time transfer on the set of the historical coating liquid atomization pressure local time series correlation characteristics and the current coating liquid atomization pressure local time series correlation characteristics. The S232, performs characteristic propagation of energy spatiotemporal transfer on the set of historical coating liquid atomization pressure local time series correlation features and the current coating liquid atomization pressure local time series correlation features to obtain the coating liquid atomization pressure time series significant propagation features, including: S2321, calculates the coating liquid atomization pressure static energy value of each historical coating liquid atomization pressure local time series correlation feature vector in the set of historical coating liquid atomization pressure local time series correlation feature vectors; S2322, performs spatiotemporal modulation on the coating liquid atomization pressure static energy value of each historical coating liquid atomization pressure local time series correlation feature vector to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value of each historical coating liquid atomization pressure local time series correlation feature vector relative to the current coating liquid atomization pressure local time series correlation feature vector. ; S2323, input the dynamic energy spatiotemporal transfer value of the coating liquid atomization pressure of each historical coating liquid atomization pressure local time series associated characteristic vector into the energy transfer gating module based on the mask function to obtain a set of dynamic energy spatiotemporal transfer weights of the coating liquid atomization pressure; S2324, based on the set of dynamic energy spatiotemporal transfer weights of the coating liquid atomization pressure, calculate the weighted sum of the set of historical coating liquid atomization pressure local time series associated characteristic vectors to obtain the historical coating liquid atomization pressure time series transfer aggregate modulation vector; S2325, calculate the positional sum of the historical coating liquid atomization pressure time series transfer aggregate modulation vector and the current coating liquid atomization pressure local time series associated characteristic vector to obtain the full time domain coating liquid atomization pressure time series significant propagation characteristic vector as the coating liquid atomization pressure time series significant propagation characteristic.

[0053] Optionally, in one embodiment of the present application, S2321, calculating the static energy value of the coating liquid atomization pressure of each historical coating liquid atomization pressure local time series correlation feature vector in the set of the historical coating liquid atomization pressure local time series correlation feature vectors, including: extracting the maximum value of the historical coating liquid atomization pressure local time series correlation feature vector to obtain the maximum value of the historical coating liquid atomization pressure local time series correlation feature; respectively calculating the mean and variance of the historical coating liquid atomization pressure local time series correlation feature vector to obtain the mean value of the historical coating liquid atomization pressure local time series correlation feature and the variance of the historical coating liquid atomization pressure local time series correlation feature; and converting the historical coating liquid atomization pressure local time series correlation feature vector into the static energy value of the coating liquid atomization pressure local time series correlation feature vector. The quantity is subtracted from the mean value of the local time series correlation characteristic of the historical coating liquid atomization pressure by position, and the fourth power of the characteristic value of each position in the local time series difference vector of the historical coating liquid atomization pressure is calculated to obtain the local time series difference modulation vector of the historical coating liquid atomization pressure; the expected value of the local time series difference modulation vector of the historical coating liquid atomization pressure is calculated, and the expected value is divided by the square of the variance of the local time series correlation characteristic of the historical coating liquid atomization pressure to obtain the local time series energy value of the historical coating liquid atomization pressure; the maximum value of the local time series correlation characteristic of the historical coating liquid atomization pressure is added to the local time series energy value of the historical coating liquid atomization pressure to obtain the static energy value of the coating liquid atomization pressure.

[0054] Optionally, in one embodiment of the present application, S2322, the static energy value of the coating liquid atomization pressure of each historical coating liquid atomization pressure local time series correlation characteristic vector is subjected to spatiotemporal modulation to obtain the spatiotemporal transfer value of the dynamic energy of the coating liquid atomization pressure of each historical coating liquid atomization pressure local time series correlation characteristic vector relative to the current coating liquid atomization pressure local time series correlation characteristic vector, including: determining the coating liquid atomization pressure energy spatial attenuation factor based on the spatial span coefficient between the each historical coating liquid atomization pressure local time series correlation characteristic vector and the current coating liquid atomization pressure local time series correlation characteristic vector; based on the each The time span coefficient between the historical local time series correlation characteristic vector of the coating liquid atomization pressure and the current local time series correlation characteristic vector of the coating liquid atomization pressure is used to determine the time attenuation factor of the coating liquid atomization pressure energy; based on the spatial attenuation factor of the coating liquid atomization pressure energy and the time attenuation factor of the coating liquid atomization pressure energy, the static energy value of the coating liquid atomization pressure of each historical local time series correlation characteristic vector of the coating liquid atomization pressure is temporally and spatially modulated to obtain the spatiotemporal transfer value of the dynamic energy of the coating liquid atomization pressure of each historical local time series correlation characteristic vector of the coating liquid atomization pressure relative to the current local time series correlation characteristic vector of the coating liquid atomization pressure.

[0055] Optionally, in one embodiment of the present application, based on the spatial attenuation factor of the coating liquid atomization pressure energy and the temporal attenuation factor of the coating liquid atomization pressure energy, the static energy value of the coating liquid atomization pressure of each historical coating liquid atomization pressure local time series correlation characteristic vector is temporally modulated to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value of the coating liquid atomization pressure of each historical coating liquid atomization pressure local time series correlation characteristic vector relative to the current coating liquid atomization pressure local time series correlation characteristic vector, including: multiplying the coating liquid atomization pressure energy spatial attenuation factor and the coating liquid atomization pressure energy temporal attenuation factor of the historical coating liquid atomization pressure local time series correlation characteristic vector with the first modulation parameter and the second modulation parameter respectively, and adding the obtained coating liquid atomization pressure energy spatial attenuation modulation factor and coating liquid atomization pressure energy temporal attenuation modulation factor to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value.

[0056] Optionally, in one embodiment of the present application, the dynamic energy spatiotemporal transfer values ​​of the coating liquid atomization pressure of each historical local time series associated characteristic vector of the coating liquid atomization pressure are input into an energy transfer gating module based on a mask function to obtain a set of dynamic energy spatiotemporal transfer weights of the coating liquid atomization pressure, including: in response to the dynamic energy spatiotemporal transfer value of the coating liquid atomization pressure being greater than a predetermined threshold, the dynamic energy spatiotemporal transfer value of the coating liquid atomization pressure is input into a sigmoid function to obtain a value as the dynamic energy spatiotemporal transfer weight of the coating liquid atomization pressure.

[0057] Optionally, in one embodiment of the present application, based on the spatial span coefficient between the local time series correlation characteristic vectors of the historical coating liquid atomization pressures and the local time series correlation characteristic vectors of the current coating liquid atomization pressures, the spatial attenuation factor of the coating liquid atomization pressure energy is determined, including: calculating the square of the spatial span value between the local time series correlation characteristic vectors of the historical coating liquid atomization pressures and the local time series correlation characteristic vectors of the current coating liquid atomization pressures to obtain the spatial span coefficient of the coating liquid atomization pressure; dividing the static energy value of the coating liquid atomization pressure corresponding to the local time series correlation characteristic vectors of the historical coating liquid atomization pressures by the spatial span coefficient of the coating liquid atomization pressure to obtain the spatial attenuation factor of the coating liquid atomization pressure energy.

[0058] Optionally, in one embodiment of the present application, based on the time span coefficient between the local time series correlation characteristic vectors of each historical coating liquid atomization pressure and the local time series correlation characteristic vector of the current coating liquid atomization pressure, determining the coating liquid atomization pressure energy time attenuation factor includes: subtracting the coating liquid atomization pressure timestamp of the current coating liquid atomization pressure local time series correlation characteristic vector from the coating liquid atomization pressure timestamp of the historical coating liquid atomization pressure local time series correlation characteristic vector, and rounding down to obtain the coating liquid atomization pressure time span coefficient. degree value; dividing the coating liquid atomization pressure time span value by the time attenuation parameter to obtain the coating liquid atomization pressure time series span energy attenuation coefficient; taking the coating liquid atomization pressure time series span energy attenuation coefficient as the exponential power, calculating the natural exponential function value with the natural constant e as the base to obtain the coating liquid atomization pressure time span coefficient; dividing the coating liquid atomization pressure static energy value corresponding to the local time series correlation eigenvector of the historical coating liquid atomization pressure by the coating liquid atomization pressure time span coefficient to obtain the coating liquid atomization pressure energy time attenuation factor.

[0059] In summary, in the embodiment of the present application, the node potential energy is significantly propagated for the set of local time series correlation features of the coating liquid atomization pressure, which can be expressed by the formula:

[0060] ;

[0061] ;

[0062] ;

[0063] ;

[0064] ;

[0065] ;

[0066] ;

[0067] in, is the set of local time series correlation feature vectors of the coating liquid atomization pressure, are the first, second, ..., and third in the set of local time series correlation feature vectors of the coating liquid atomization pressure. ,..., The local time series correlation feature vector of the coating liquid atomization pressure, It is The local time series correlation feature vector of the coating liquid atomization pressure The eigenvalues ​​at the positions, To calculate the expected value, and They are The mean and variance of For extraction The maximum value in yes The corresponding static energy value of the coating liquid atomization pressure, express and The space span value between yes and The spatial attenuation factor of the coating liquid atomization pressure energy between and Distribution represents and The timestamp of the local time series correlation feature vector of the coating liquid atomization pressure, To round down, is the time decay parameter, represents the exponential function value with the natural constant e as the base, yes and The coating liquid atomization pressure energy time decay factor between and are the first modulation parameter and the second modulation parameter respectively, yes and The dynamic energy space-time transfer value of the coating liquid atomization pressure between For masking, yes function, is a predetermined threshold, is the number of vectors in the set of local time series correlation feature vectors of the coating liquid atomization pressure, It is the significant propagation characteristic vector of the full-time domain coating liquid atomization pressure time series.

[0068] In detail, first extract the current coating liquid atomization pressure local time series associated feature vector as the current node feature vector from the set of the coating liquid atomization pressure local time series associated feature vector, and the set of the historical coating liquid atomization pressure local time series associated feature vector as the set of historical node feature vectors. That is, the current node feature vector represents the atomization pressure feature at the current time point, and the historical node feature vector set contains the feature information at multiple previous time points, so that the dynamic change pattern of the atomization pressure over time can be captured. Then, calculate the static energy value of each historical node feature vector to obtain each coating liquid atomization pressure static energy value. In particular, the static energy value is obtained based on the maximum value, mean and variance of each historical node feature vector, which provides a quantitative feature vector for measuring the degree of inherent properties in space, and provides a reference standard for the decay of potential energy over time and space, which helps the model to identify the importance of each node in space and time. Next, based on the space span coefficient and time span coefficient between each historical node feature vector and the current node feature vector, the energy space attenuation factor and energy time attenuation factor are determined to obtain the coating liquid atomization pressure energy space attenuation factor and the coating liquid atomization pressure energy time attenuation factor. In particular, the space span coefficient and the time span coefficient quantify the spatial and temporal differences between nodes, respectively. The attenuation factor uses these coefficients to adjust the weight of information transfer between nodes. That is, this design enables the model to adjust the intensity of information flow according to the distance between nodes. By introducing the attenuation factor of time and space, the model enhances the response to locality and time sensitivity, making the information transfer process closer to the actual physical phenomenon. Furthermore, based on the coating liquid atomization pressure energy space attenuation factor and the coating liquid atomization pressure energy time attenuation factor, the coating liquid atomization pressure static energy value of each historical node feature vector is spatiotemporally modulated to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value. That is, the propagation of node features is more in line with the actual physical process through spatiotemporal modulation, thereby enhancing the ability to capture the dynamic changes between each feature. Then, the dynamic energy spatiotemporal transfer value of the coating liquid atomization pressure of each historical node feature vector is input into the energy transfer gating module based on the mask function to screen and strengthen important spatiotemporal transfer information, and the importance of the feature is adaptively adjusted. Finally, the masked factors are used as weights to perform weighted fusion on the features of each historical node to integrate the information of each historical node, and the historical coating liquid atomization pressure time series transfer aggregation modulation vector is obtained, and the historical modulation vector and the current coating liquid atomization pressure local time series correlation feature vector are added by position to obtain the full time domain coating liquid atomization pressure time series significant propagation feature vector that integrates the integrated time information.

[0069] In the preparation method of the above-mentioned coated controlled-release fertilizer, the S24 obtains an atomization speed control instruction based on the significant propagation characteristics of the coating liquid atomization pressure sequence. Optionally, in one embodiment of the present application, based on the significant propagation characteristics of the coating liquid atomization pressure sequence in the full time domain, the atomization speed control instruction is obtained, including: inputting the significant propagation characteristic vector of the coating liquid atomization pressure sequence in the full time domain into the atomization speed control module based on the decoder to obtain the atomization speed control instruction, and the atomization speed control instruction includes the decoded value of the recommended value of the atomization speed at the next time point. That is, the coating liquid atomization pressure sequence significant propagation characteristics obtained by the significant propagation of the coating liquid atomization pressure in the full time domain using the set of the local time series associated characteristic vectors of the coating liquid atomization pressure are decoded and processed, so as to automatically obtain the recommended value of the atomization speed at the next time point. In this way, the atomization speed can be dynamically adjusted according to the real-time changing pressure data, rather than relying on the static set value. In this way, it is ensured that the pressure changes can be responded to at any time during the production process, thereby improving the intelligence level of the atomization process.

[0070] That is, considering that each local time-series correlation feature vector of the coating liquid atomization pressure in the set of local time-series correlation feature vectors of the coating liquid atomization pressure respectively represents the short-range-long-range bidirectional time-series correlation feature of the coating liquid atomization pressure in the local time domain, when it is input into the characteristic message propagation network based on the node potential energy spatiotemporal transfer characteristics, the difference in the time-series propagation weight based on the difference in the potential energy of the source time-series node time-series distribution transfer will make the characteristic manifold of the full-time domain coating liquid atomization pressure time-series significant propagation feature vector have sparse expression and uneven distribution in the high-dimensional feature space, thereby affecting the accuracy of the atomization speed control instruction obtained through the decoder-based atomization speed control module.

[0071] Preferably, the full-time domain coating liquid atomization pressure time series significant propagation feature vector is input into a decoder-based atomization speed control module to obtain an atomization speed control instruction, including:

[0072] Performing L random sampling on the significant propagation characteristic vector of the full-time domain coating liquid atomization pressure time series to obtain L coating liquid atomization pressure random disturbance characteristic values;

[0073] Arrange the L random perturbation eigenvalues ​​of the coating liquid atomization pressure in order from large to small to obtain a random perturbation ordered vector of the coating liquid atomization pressure;

[0074] Calculating the inner product of the significant propagation characteristic vector of the coating liquid atomization pressure time series in the whole time domain to obtain the coating liquid atomization pressure energy modulation characteristic value;

[0075] Based on the coating liquid atomization pressure energy modulation characteristic value, the coating liquid atomization pressure random perturbation ordered vector is semantically scaled to obtain a semantically enhanced coating liquid atomization pressure random perturbation ordered vector;

[0076] Calculating the product of the random disturbance ordered vector of the semantically enhanced coating liquid atomization pressure and the transposed vector of the significant propagation characteristic vector of the coating liquid atomization pressure in the full time domain to obtain the coating liquid atomization pressure optimal sample prompt orthogonal modulation matrix;

[0077] Calculating the matrix product between the orthogonal modulation matrix of the best sample of the coating liquid atomization pressure and the significant propagation characteristic vector of the time series of the coating liquid atomization pressure in the full time domain to obtain the optimized significant propagation characteristic vector of the time series of the coating liquid atomization pressure in the full time domain;

[0078] The optimized full-time domain coating liquid atomization pressure time series significant propagation feature vector is input into the decoder-based atomization speed control module to obtain an atomization speed control instruction.

[0079] The above optimization process is expressed as follows:

[0080] ;

[0081] ;

[0082] ;

[0083] ;

[0084] ;

[0085] ;

[0086] ;

[0087] ;

[0088] in, Indicates The random perturbation eigenvalue of the coating liquid atomization pressure, represents L random samplings, represents the Gaussian density probability function, represents the random disturbance characteristic value of the coating liquid atomization pressure, represents the significant propagation characteristic vector of the full-time-domain coating liquid atomization pressure time series, The first one represents the significant propagation characteristic vector of the full-time coating liquid atomization pressure time series eigenvalues, It means to calculate the average value of the vector. represents the length of the significant propagation characteristic vector of the full-time-domain coating liquid atomization pressure time series, and They represent the mean value and variance of the significant propagation characteristic vector of the full-time-domain coating liquid atomization pressure time series, It indicates that the random disturbance eigenvalues ​​of the coating liquid atomization pressure are arranged in order from large to small. represents the ordered vector of random disturbance of the coating liquid atomization pressure, represents the characteristic value of the coating liquid atomization pressure energy modulation, represents the square of the vector's two-norm, It means point multiplication by position. represents matrix multiplication, represents the transpose of a vector, The best example of coating liquid atomization pressure indicates the orthogonal modulation matrix, Represents the optimized significant propagation eigenvector of the full-domain coating liquid atomization pressure time series.

[0089] Accordingly, by performing L random sampling on the significant propagation feature vector of the coating liquid atomization pressure time series in the entire time domain to obtain a set of new feature values ​​based on the original features but containing slight changes, randomness is introduced to explore different aspects of the feature space, and by simulating the noise or variation in the actual data, the model is better generalized to unseen data. Then, these coating liquid atomization pressure random perturbation feature values ​​are arranged in order from large to small to form a coating liquid atomization pressure random perturbation ordered vector to identify which perturbation directions may be the most important for the current problem. Such an ordering not only helps the algorithm focus on the feature differences that can best distinguish different categories, but also helps to reveal hidden data structures or patterns. Subsequently, the inner product of the significant propagation eigenvector of the coating liquid atomization pressure time series in the entire time domain is calculated to obtain the coating liquid atomization pressure energy modulation eigenvalue, which reflects the overall strength or "energy" of the original eigenvector and provides a benchmark for measuring the importance of the feature. Using this coating liquid atomization pressure energy modulation eigenvalue as a reference point, the previously formed coating liquid atomization pressure random perturbation ordered vector is semantically scaled to construct a semantically enhanced coating liquid atomization pressure random perturbation ordered vector. In this way, those disturbances that appear to be more important in the original feature space are amplified, thereby enhancing the model's sensitivity to changes in key features, while also retaining relatively small but potentially useful disturbance information.

[0090] Then, the product of the semantically enhanced ordered vector of random disturbance of coating liquid atomization pressure and the transpose of the significant propagation feature vector of coating liquid atomization pressure in the whole time domain is calculated to obtain the orthogonal modulation matrix of the best sample prompt of coating liquid atomization pressure. The purpose is to find the interaction mode between the features under two different perspectives. The formed matrix can be regarded as a conversion tool for capturing and representing the complex linear relationship between features. Finally, the best sample prompt orthogonal modulation matrix of coating liquid atomization pressure constructed above is applied to the original significant propagation feature vector of coating liquid atomization pressure in the whole time domain to complete the final feature structure modulation, so that the features are more in line with the requirements of the objective function, reduce redundant information, and maximize the retention of useful information, so as to improve learning efficiency and prediction performance.

[0091] In summary, the feature manifold optimization process of the significant propagation feature vector of the full-time-domain coating liquid atomization pressure time series is explained, which extracts the best prompt sample based on the fine-grained target retrieval of random perturbations, and then guides the adversarial robustness learning based on the best prompt sample to complete the structure of the high-dimensional manifold of the feature vector in the feature space, so as to improve the detail and structure of the feature expression, thereby improving the accuracy of the atomization speed control instruction obtained by the decoder-based atomization speed control module. In this way, the atomization speed can be dynamically adjusted according to the real-time changing pressure data instead of relying on the static set value. In this way, it is ensured that the pressure changes can be responded to at any time during the production process, thereby improving the intelligence level of the atomization process.

[0092] In summary, the preparation method of the coated controlled-release fertilizer according to the embodiment of the present application is explained, which uses a pressure pump to atomize the coating liquid to obtain the atomized coating liquid, and obtains the time queue of the coating liquid atomization pressure collected by the pressure sensor, and uses the data analysis and processing technology based on deep learning to scan and sample the coating liquid atomization pressure and locally correlate the time series, so as to automatically obtain the recommended value of the atomization speed at the next time point according to the propagation characteristics of the local time series correlation characteristics of the atomization pressure of each coating liquid in the entire time domain. In this way, the atomization speed can be dynamically adjusted according to the real-time changing pressure data, rather than relying on a static set value. In this way, it is ensured that pressure changes can be responded to at any time during the production process, thereby improving the intelligence level of the atomization process.

[0093] Specifically, a coated controlled-release fertilizer is also provided, and the coated controlled-release fertilizer is prepared by the aforementioned method for preparing the coated controlled-release fertilizer.

[0094] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0095] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the module division is only a logical function division, and there may be other division methods in actual implementation. The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0096] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0097] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0098] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units stated in the device claim can also be implemented by one unit through software or hardware.

[0099] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for preparing a coated controlled-release fertilizer, comprising: Preparation of water-soluble polyurethane; The water-soluble polyurethane, turpentine and polyvinyl alcohol are mixed in a predetermined ratio and placed in a reaction kettle for stirring reaction to obtain a coating liquid; A pressure pump is used to atomize the coating liquid to obtain an atomized coating liquid; after heating the fertilizer particles to 60° C., a spray device is used to spray the atomized coating liquid on the surface of the heated fertilizer particles, and the fertilizer particles sprayed with the coating liquid are subjected to hot air drying to obtain a first coated fertilizer; the first coated fertilizer is sent to a circulating fluidized bed for further coating treatment to obtain the coated controlled-release fertilizer; The method is characterized in that the coating liquid is atomized by using a pressure pump to obtain an atomized coating liquid, comprising: Obtaining a time sequence of coating liquid atomization pressure collected by a pressure sensor; Scanning and sampling the time queue of the coating liquid atomization pressure to obtain a set of local time series of the coating liquid atomization pressure, and then performing sequence encoding to obtain a set of local time series correlation features of the coating liquid atomization pressure; The node potential energy is significantly propagated on the set of local time series correlation features of the coating liquid atomization pressure to obtain the significant time series propagation features of the coating liquid atomization pressure, including: extracting the current local time series correlation features of the coating liquid atomization pressure and the set of historical local time series correlation features of the coating liquid atomization pressure from the set of local time series correlation features of the coating liquid atomization pressure; performing feature propagation of energy space-time transfer on the set of historical local time series correlation features of the coating liquid atomization pressure and the current local time series correlation features of the coating liquid atomization pressure to obtain the significant time series propagation features of the coating liquid atomization pressure; Based on the significant propagation characteristics of the coating liquid atomization pressure time series, an atomization speed control instruction is obtained; The time queue of the coating liquid atomization pressure is scanned and sampled to obtain a set of local time series of the coating liquid atomization pressure, and then sequence encoding is performed to obtain a set of local time series correlation features of the coating liquid atomization pressure, including: Based on a predetermined time window and a predetermined step length, scanning and sampling the time queue of the coating liquid atomization pressure to obtain a set of local time series of the coating liquid atomization pressure; Inputting each local time series of the coating liquid atomization pressure in the set of local time series of the coating liquid atomization pressure into a sequence encoder based on the Bi-LSTM model to obtain a set of local time series correlation feature vectors of the coating liquid atomization pressure as a set of local time series correlation features of the coating liquid atomization pressure; Extracting a set of current local time series correlation features of coating liquid atomization pressure and a set of historical local time series correlation features of coating liquid atomization pressure from the set of local time series correlation features of coating liquid atomization pressure, including: Extracting a current coating liquid atomization pressure local time series correlation feature vector from the set of coating liquid atomization pressure local time series correlation feature vectors as the current coating liquid atomization pressure local time series correlation feature, and defining other coating liquid atomization pressure local time series correlation feature vectors in the set of coating liquid atomization pressure local time series correlation feature vectors as historical coating liquid atomization pressure local time series correlation feature vectors, so as to obtain a set of historical coating liquid atomization pressure local time series correlation feature vectors as the set of historical coating liquid atomization pressure local time series correlation features; The set of the historical coating liquid atomization pressure local time series correlation features and the current coating liquid atomization pressure local time series correlation features are subjected to characteristic propagation of energy space-time transfer to obtain the coating liquid atomization pressure time series significant propagation features, including: Calculating the coating liquid atomization pressure static energy value of each historical coating liquid atomization pressure local time series correlation characteristic vector in the set of the historical coating liquid atomization pressure local time series correlation characteristic vectors; Performing spatiotemporal modulation on the static energy values ​​of the coating liquid atomization pressure of each historical local time series associated characteristic vector of the coating liquid atomization pressure to obtain the spatiotemporal transfer value of the dynamic energy of the coating liquid atomization pressure of each historical local time series associated characteristic vector of the coating liquid atomization pressure relative to the current local time series associated characteristic vector of the coating liquid atomization pressure; Inputting the coating liquid atomization pressure dynamic energy spatiotemporal transfer values ​​of the local time series associated characteristic vectors of the historical coating liquid atomization pressure into the energy transfer gating module based on the mask function to obtain a set of coating liquid atomization pressure dynamic energy spatiotemporal transfer weights; Based on the set of spatiotemporal transfer weights of the dynamic energy of the coating liquid atomization pressure, a weighted sum of the set of local time series associated characteristic vectors of the historical coating liquid atomization pressure is calculated to obtain a historical coating liquid atomization pressure time series transfer aggregation modulation vector; Calculate the positional sum of the historical coating liquid atomization pressure time series transmission aggregation modulation vector and the current coating liquid atomization pressure local time series correlation feature vector to obtain the full time domain coating liquid atomization pressure time series significant propagation feature vector as the coating liquid atomization pressure time series significant propagation feature; Calculating the coating liquid atomization pressure static energy value of each historical coating liquid atomization pressure local time series correlation characteristic vector in the set of the historical coating liquid atomization pressure local time series correlation characteristic vectors, comprising: Extracting the maximum value of the local time series correlation feature vector of the historical coating liquid atomization pressure to obtain the maximum value of the local time series correlation feature of the historical coating liquid atomization pressure; Respectively calculating the mean and variance of the local time series correlation characteristic vector of the historical coating liquid atomization pressure to obtain the mean of the local time series correlation characteristic of the historical coating liquid atomization pressure and the variance of the local time series correlation characteristic of the historical coating liquid atomization pressure; subtracting the local time series correlation characteristic vector of the historical coating liquid atomization pressure from the local time series correlation characteristic mean of the historical coating liquid atomization pressure by position, calculating the fourth power of the characteristic value of each position in the local time series difference vector of the historical coating liquid atomization pressure to obtain the local time series difference modulation vector of the historical coating liquid atomization pressure; calculating the expected value of the local time series difference modulation vector of the historical coating liquid atomization pressure, and dividing the obtained expected value by the square of the variance of the local time series correlation characteristic of the historical coating liquid atomization pressure to obtain the local time series energy value of the historical coating liquid atomization pressure; The maximum value of the local time series correlation characteristic of the historical coating liquid atomization pressure is added to the local time series energy value of the historical coating liquid atomization pressure to obtain the static energy value of the coating liquid atomization pressure; The coating liquid atomization pressure static energy value of each historical coating liquid atomization pressure local time series correlation characteristic vector is subjected to spatiotemporal modulation to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value of each historical coating liquid atomization pressure local time series correlation characteristic vector relative to the current coating liquid atomization pressure local time series correlation characteristic vector, including: Determine the coating liquid atomization pressure energy spatial attenuation factor based on the spatial span coefficient between the local time series correlation characteristic vectors of the respective historical coating liquid atomization pressures and the local time series correlation characteristic vector of the current coating liquid atomization pressure; Determine the coating liquid atomization pressure energy time attenuation factor based on the time span coefficient between the local time series correlation characteristic vectors of the respective historical coating liquid atomization pressures and the local time series correlation characteristic vector of the current coating liquid atomization pressure; Based on the coating liquid atomization pressure energy spatial attenuation factor and the coating liquid atomization pressure energy temporal attenuation factor, the coating liquid atomization pressure static energy value of each historical coating liquid atomization pressure local time series associated characteristic vector is temporally modulated to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value of each historical coating liquid atomization pressure local time series associated characteristic vector relative to the current coating liquid atomization pressure local time series associated characteristic vector; Based on the coating liquid atomization pressure energy spatial attenuation factor and the coating liquid atomization pressure energy temporal attenuation factor, the coating liquid atomization pressure static energy value of each historical coating liquid atomization pressure local time series associated characteristic vector is temporally modulated to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value of each historical coating liquid atomization pressure local time series associated characteristic vector relative to the current coating liquid atomization pressure local time series associated characteristic vector, including: multiplying the coating liquid atomization pressure energy spatial attenuation factor and the coating liquid atomization pressure energy temporal attenuation factor of the historical coating liquid atomization pressure local time series associated characteristic vector with the first modulation parameter and the second modulation parameter respectively, and adding the obtained coating liquid atomization pressure energy spatial attenuation modulation factor and the coating liquid atomization pressure energy temporal attenuation modulation factor to obtain the coating liquid atomization pressure dynamic energy spatiotemporal transfer value; Inputting the coating liquid atomization pressure dynamic energy spatiotemporal transfer value of each historical coating liquid atomization pressure local time series associated characteristic vector into an energy transfer gating module based on a mask function to obtain a set of coating liquid atomization pressure dynamic energy spatiotemporal transfer weights, including: in response to the coating liquid atomization pressure dynamic energy spatiotemporal transfer value being greater than a predetermined threshold, inputting the coating liquid atomization pressure dynamic energy spatiotemporal transfer value into a sigmoid function to obtain a value as the coating liquid atomization pressure dynamic energy spatiotemporal transfer weight; Based on the significant propagation characteristics of the full-time-domain coating liquid atomization pressure time series, an atomization speed control instruction is obtained, including: inputting the significant propagation characteristic vector of the full-time-domain coating liquid atomization pressure time series into an atomization speed control module based on a decoder to obtain the atomization speed control instruction, wherein the atomization speed control instruction includes a decoded value of a recommended atomization speed value at a next time point; Inputting the significant propagation characteristic vector of the full-time domain coating liquid atomization pressure time series into the decoder-based atomization speed control module to obtain an atomization speed control instruction, including: Performing L random sampling on the significant propagation characteristic vector of the full-time domain coating liquid atomization pressure time series to obtain L coating liquid atomization pressure random disturbance characteristic values; Arrange the L random perturbation eigenvalues ​​of the coating liquid atomization pressure in order from large to small to obtain a random perturbation ordered vector of the coating liquid atomization pressure; Calculating the inner product of the significant propagation characteristic vector of the coating liquid atomization pressure time series in the whole time domain to obtain the coating liquid atomization pressure energy modulation characteristic value; Based on the coating liquid atomization pressure energy modulation characteristic value, the coating liquid atomization pressure random perturbation ordered vector is semantically scaled to obtain a semantically enhanced coating liquid atomization pressure random perturbation ordered vector; Calculating the product of the random disturbance ordered vector of the semantically enhanced coating liquid atomization pressure and the transposed vector of the significant propagation characteristic vector of the coating liquid atomization pressure in the full time domain to obtain the coating liquid atomization pressure optimal sample prompt orthogonal modulation matrix; Calculating the matrix product between the orthogonal modulation matrix of the best sample of the coating liquid atomization pressure and the significant propagation characteristic vector of the time series of the coating liquid atomization pressure in the full time domain to obtain the optimized significant propagation characteristic vector of the time series of the coating liquid atomization pressure in the full time domain; The optimized full-time domain coating liquid atomization pressure time series significant propagation feature vector is input into the decoder-based atomization speed control module to obtain an atomization speed control instruction.

2. A coated controlled-release fertilizer, characterized in that: The coated controlled-release fertilizer is prepared by the method for preparing the coated controlled-release fertilizer according to claim 1.

Images