Regulation and control method and system for raw material ratio and parameter combination in nitrogen-phosphorus-potassium ternary fertilizer production

Abstract

The invention provides a method and a system for regulating and controlling a raw material ratio and parameter combination in production of a nitrogen-phosphorus-potassium ternary fertilizer, and belongs to the technical field of automatic control of fertilizers. The production data is collected in real time through the sensor, the production model is established by using the convolutional neural network and the long-short-term memory network, the optimal parameter combination is calculated by using the particle swarm optimization algorithm, and real-time optimization and self-learning of the production process are realized. By adopting the method and the system for regulating and controlling the raw material ratio and the parameter combination in the production of the nitrogen-phosphorus-potassium ternary fertilizer, the real-time optimization of the production process can be realized, the product quality and the production efficiency can be improved, and the production cost can be reduced.

Description

Technical Field

[0001] The invention relates to the technical field of fertilizer automation control, and in particular to a method and system for regulating raw material ratios and parameter combinations in the production of nitrogen, phosphorus and potassium ternary fertilizers. Background Art

[0002] As a basic fertilizer widely used in agricultural production, the production quality of nitrogen, phosphorus and potassium ternary fertilizer is directly related to the growth and yield of crops. However, the production process of nitrogen, phosphorus and potassium ternary fertilizer involves many complex chemical reactions and fine process parameter control, which makes its production challenging.

[0003] The traditional production method of nitrogen, phosphorus and potassium fertilizers mainly relies on fixed process flow and operator experience. This method lacks flexibility and cannot respond to fluctuations in raw material quality and changes in the production environment in real time. Since the quality of raw materials may vary depending on the supplier, season or batch, traditional production methods cannot ensure the consistency and stability of product quality. At the same time, changes in production environment factors such as temperature, humidity and pressure may also affect product quality, and traditional control methods often cannot adjust these parameters in a timely and effective manner.

[0004] In addition, traditional production methods also have problems of resource waste and low production efficiency. Since it is impossible to accurately control the raw material ratio and process parameters, problems such as raw material waste, increased energy consumption and substandard product quality may occur during the production process, which not only increases production costs but also affects the market competitiveness of enterprises.

[0005] With the development of agricultural modernization and smart agriculture, agricultural production has put forward higher requirements on the quality, performance and production efficiency of fertilizers. The market demand for high-quality and high-efficiency nitrogen, phosphorus and potassium ternary fertilizers is increasing, and traditional production methods can no longer meet these needs. Therefore, there is an urgent need for a technology that can control the raw material ratio and process parameters in the production of nitrogen, phosphorus and potassium ternary fertilizers in real time and intelligently, so as to improve production efficiency, reduce costs, ensure stable product quality and meet market demand. This technology should be able to monitor production data in real time, automatically adjust the raw material ratio and process parameters, and thus achieve precise control of the production process. Summary of the invention

[0006] The purpose of the present invention is to provide a method and system for regulating the raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizers, which can realize real-time optimization of the production process, improve product quality and production efficiency, and reduce production costs.

[0007] To achieve the above object, the present invention provides a method for regulating raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizer, comprising the following steps:

[0008] Step S1, deploying a sensor network at key locations of the production line to collect raw material quality parameters, equipment operating parameters, environmental parameters and product quality indicators in real time;

[0009] Step S2: preprocessing and storing the collected data to form a data set;

[0010] Step S3: Based on the data set obtained in step S2, a nonlinear mapping relationship between raw material quality parameters, equipment operation parameters, environmental parameters and product quality is established using a deep learning model;

[0011] Step S4, using a particle swarm optimization algorithm, taking product quality, production cost and energy consumption as objective functions, to calculate the optimal raw material ratio and process parameter combination;

[0012] Step S5: Send the optimized raw material ratio and process parameters to the execution control module to adjust the raw material input amount and equipment operation parameters in real time;

[0013] Step S6: Real-time detection of the quality indicators of the fertilizer products produced, comparison of the actual detection results with the model prediction results and calculation of the errors, and use of the errors to incrementally train the deep learning model and update the model parameters.

[0014] Preferably, in step S1, a sensor network is deployed at key positions of the production line, specifically: a near-infrared spectroscopy sensor is installed at the raw material feed port, a flow sensor is installed on the conveying pipeline, a humidity sensor is installed in the production environment, a temperature sensor and a pressure sensor are installed on the reactor, and a temperature sensor is installed on the mixer.

[0015] Preferably, in step S2, the preprocessing includes data denoising, missing value filling and outlier processing, and the preprocessed data is stored in a NoSQL database.

[0016] Preferably, in step S3, a nonlinear mapping relationship between production parameters and product quality is established using a deep learning model, and the specific operations are: using a convolutional neural network to extract spatial features of data, and using a long short-term memory network to capture time series features;

[0017] Preferably, in step S4, the fitness function of the particle swarm optimization algorithm comprehensively considers product quality, production cost and energy consumption, and initializes a group of particles, each particle represents a possible combination of raw material ratio and process parameters, evaluates the fitness value of each particle, and updates the speed and position of the particle according to the fitness value, and optimizes iteratively until the termination condition is met, and outputs the optimal raw material ratio and process parameter combination;

[0018] Among them, the fitness function is:

[0019] Fitness=w1×(1-QQmax )+w2×CC max +w3×EE max ;

[0020] Among them, Fitness is the fitness value; Q is the product quality index; Q max is the maximum value of product quality index; C is production cost; C max is the maximum value of production cost; E is energy consumption; E max is the maximum value of energy consumption; w1, w2, w3 are all weight coefficients, and w1+w2+w3=1;

[0021] The particle velocity update formula is:

[0022]

[0023] The particle position update formula is:

[0024]

[0025] in, is the velocity of the i-th particle at the k-th iteration; is the position of the i-th particle at the k-th iteration, i.e., the parameter combination; w is the inertia weight; c1 and c2 are learning factors; r1 and r2 are random numbers between [0,1]; is the best position of the i-th particle in history; g best is the global best position.

[0026] Preferably, in step S5, the execution control module is a PLC controller.

[0027] The present invention also provides a system for realizing the control method of raw material ratio and parameter combination in the production of the above-mentioned nitrogen, phosphorus and potassium ternary fertilizer, comprising:

[0028] Data acquisition module, which is used to deploy sensor networks at key locations on the production line to collect raw material quality parameters, equipment operation parameters, environmental parameters and product quality indicators in real time;

[0029] A data processing module is used to pre-process and store the collected data to form a data set;

[0030] Intelligent control module, which uses deep learning models to establish a nonlinear mapping relationship between production parameters and product quality, and uses particle swarm optimization algorithm to calculate the optimal raw material ratio and process parameter combination with product quality, production cost and energy consumption as objective functions;

[0031] The execution control module is used to adjust the raw material input amount and equipment operation parameters according to the optimized raw material ratio and process parameters output by the intelligent control module;

[0032] The feedback learning module includes a quality detection unit and a model updating unit. The quality detection unit is used to detect the quality indicators of the fertilizer products produced in real time, compare the actual detection results with the model prediction results and calculate the errors. The model updating unit is used to use the errors to perform incremental training on the deep learning model and update the model parameters.

[0033] Preferably, the system also includes an edge computing unit and a cloud computing unit. The edge computing unit is deployed at the production site and is responsible for real-time data processing and some model calculation tasks. The cloud computing unit is responsible for offline training and updating of the model. A secure data transmission channel is established between the edge server and the cloud server, and data and model parameters are synchronized regularly to achieve local rapid response and coordination of high-performance cloud computing. The edge computing unit is responsible for processing certain components or computing tasks of the deep learning model or optimization model. The purpose of this division is to balance real-time performance and computational complexity, adapt to the resource limitations of edge computing devices, and ensure that the system can respond quickly to dynamic changes in the production site.

[0034] Therefore, the present invention adopts the above-mentioned method and system for regulating the raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizers, and the beneficial technical effects are as follows:

[0035] (1) Real-time optimization of the production process: By deploying sensor networks at key locations on the production line, raw material quality parameters, equipment operating parameters, environmental parameters, and product quality indicators are collected in real time. Deep learning models and particle swarm optimization algorithms are used for real-time analysis and optimization, thus achieving dynamic adjustment and real-time optimization of the production process.

[0036] (2) Improve product quality: The present invention can accurately control the raw material ratio and process parameters, thereby significantly improving the product quality of nitrogen, phosphorus and potassium ternary fertilizers. Through real-time monitoring and feedback adjustment, the stability and consistency of product quality are ensured.

[0037] (3) Improve production efficiency: Through automated control and intelligent optimization, human intervention and errors are reduced, and the operating efficiency and stability of the production line are improved. This helps reduce production costs and improve overall production efficiency.

[0038] (4) Reduce production costs: By precisely controlling the use of raw materials and optimizing process parameters, resource waste and energy consumption are reduced, thereby reducing production costs. At the same time, through real-time monitoring and early warning, potential production problems can be discovered and resolved in advance to avoid unnecessary losses.

[0039] (5) Enhanced intelligence: The control method and device of the present invention integrate advanced technologies such as sensor technology, deep learning, and particle swarm optimization algorithm to improve the intelligence level of fertilizer production. This will help promote the development of agricultural modernization and smart agriculture, and provide more efficient and environmentally friendly fertilizer products for agricultural production. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 The overall framework diagram of the control system of raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizer of the present invention;

[0041] Figure 2 This is the structural diagram of the deep learning model;

[0042] Figure 3 Flowchart of the particle swarm optimization algorithm. DETAILED DESCRIPTION

[0043] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0044] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0045] Embodiment 1

[0046] like Figure 1 As shown, it is the overall framework diagram of the control system of raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizer of the present invention, including:

[0047] Data acquisition module, which is used to deploy sensor networks at key locations on the production line to collect raw material quality parameters, equipment operation parameters, environmental parameters and product quality indicators in real time;

[0048] A data processing module is used to pre-process and store the collected data to form a data set;

[0049] Intelligent control module, which uses deep learning models to establish a nonlinear mapping relationship between production parameters and product quality, and uses particle swarm optimization algorithm to calculate the optimal raw material ratio and process parameter combination with product quality, production cost and energy consumption as objective functions;

[0050] The execution control module is used to adjust the raw material input amount and equipment operation parameters according to the optimized raw material ratio and process parameters output by the intelligent control module;

[0051] Feedback learning module: includes a quality detection unit and a model update unit. The quality detection unit is used to detect the quality indicators of the fertilizer products produced in real time, compare the actual detection results with the model prediction results and calculate the errors. The model update unit is used to use the errors to perform incremental training on the deep learning model and update the model parameters.

[0052] The operation method of the above device is as follows:

[0053] Step S1: deploy a sensor network at key locations of the production line to collect raw material quality parameters, equipment operation parameters, environmental parameters and product quality indicators in real time.

[0054] Near infrared spectroscopy (NIR) sensor: installed at the raw material feed port to detect the content of nitrogen, phosphorus and potassium in real time.

[0055] Temperature sensor: installed on reactors, mixers and other equipment to monitor temperature changes.

[0056] Pressure sensor: installed on closed equipment such as reactors to monitor internal pressure.

[0057] Flow sensor: installed on the conveying pipeline to monitor the flow rate of materials.

[0058] Humidity sensor: installed in the production environment to monitor the ambient humidity

[0059] Step S2: pre-process and store the collected data to form a data set.

[0060] Preprocessing includes data denoising, missing value filling and outlier processing. The preprocessed data is stored in a NoSQL database.

[0061] Step S3: Based on the data set obtained in step S2, a deep learning model is used to establish a nonlinear mapping relationship between production parameters and product quality.

[0062] like Figure 2 As shown in the figure, it is a structural diagram of the deep learning model.

[0063] Input layer: Receives preprocessed data, including raw material quality parameters (such as nitrogen, phosphorus, and potassium content), equipment operating parameters (such as temperature, pressure, and speed), and environmental parameters (such as temperature and humidity).

[0064] Convolutional Neural Network (CNN) layer: used to extract spatial features of data, including multiple convolutional layers and pooling layers.

[0065] Long short-term memory network (LSTM) layer: used to capture time series characteristics and process dynamic changes and historical dependencies of parameters.

[0066] Fully connected layer: The features extracted by CNN and LSTM are integrated to generate predictions on product quality.

[0067] Output layer: Outputs the predicted value of product quality.

[0068] Model training: Using historical production data, the model is trained through the back-propagation algorithm and optimizer (such as Adam) to minimize the error between the predicted value and the actual value.

[0069] Step S4: Utilize the particle swarm optimization algorithm to calculate the optimal raw material ratio and process parameter combination with product quality, production cost and energy consumption as the objective function.

[0070] Optimization goal: Minimize production costs and energy consumption while meeting product quality requirements.

[0071] like Figure 3 As shown in the figure, the fitness function of the particle swarm optimization algorithm comprehensively considers product quality, production cost and energy consumption. By initializing a group of particles, each particle represents a possible combination of raw material ratio and process parameters, the fitness value of each particle is evaluated, and the speed and position of the particle are updated according to the fitness value. Through iterative optimization, until the termination condition is met, the optimal raw material ratio and process parameter combination is output;

[0072] Among them, the fitness function is:

[0073] Fitness=w1×(1-QQ max )+w2×CC max +w3×EE max ;

[0074] Among them, Fitness is the fitness value; Q is the product quality index; Q max is the maximum value of product quality index; C is production cost; C max is the maximum value of production cost; E is energy consumption; E max is the maximum value of energy consumption; w1, w2, w3 are all weight coefficients, and w1+w2+w3=1;

[0075] The particle velocity update formula is:

[0076]

[0077] The particle position update formula is:

[0078]

[0079] in, is the velocity of the ith particle at the kth iteration, is the position of the ith particle at the kth iteration, i.e., the parameter combination, w is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers between [0,1], and p i best is the best position of the i-th particle in history, g best is the global best position.

[0080] Step S5: Send the optimized raw material ratio and process parameters to the execution control module to adjust the raw material input amount and equipment operating parameters in real time.

[0081] The execution control module is a PLC controller.

[0082] Step S6: Real-time detection of the quality indicators of the fertilizer products produced, comparison of the actual detection results with the model prediction results and calculation of the errors, incremental training of the deep learning model using the errors, and updating of the model parameters. Through continuous feedback learning, the prediction accuracy and robustness of the model are continuously improved.

[0083] In this embodiment, the deep learning model is trained using the historical production data of a factory over the past year (about 100,000 records).

[0084] The Adam optimizer was used, the learning rate was set to 0.001, and the changes in the loss function were monitored during training to prevent overfitting.

[0085] During the production process, the system performs parameter optimization calculations every 5 minutes.

[0086] The intelligent control module calculates the optimal parameter combination using the PSO algorithm based on real-time data and model prediction results.

[0087] The execution control module receives the optimization results and adjusts the operating parameters of the production equipment in real time.

[0088] The quality detection unit detects product quality (nutrient content, particle size, stability, etc.) in real time and feeds the results back to the model updating unit.

[0089] The quality inspection unit is a comprehensive hardware and software module used to collect and analyze the quality indicators of the produced fertilizer products in real time. Its function is to obtain the key quality data of the finished fertilizer products, compare them with the model prediction results, calculate the errors and feed them back to the model update unit to achieve self-optimization of the model.

[0090] The quality inspection unit includes:

[0091] Near infrared spectrometer (NIR): monitors the accuracy and uniformity of nitrogen (N), phosphorus (P), and potassium (K) content in fertilizers. Through non-contact rapid scanning, it can obtain the nutrient distribution of fertilizers in real time.

[0092] Particle size analyzer: Detects the particle size distribution (uniformity) of fertilizer particles to ensure that product specifications are met. This is achieved using laser particle size analyzer or image analysis technology.

[0093] Moisture detector: detects the moisture content of fertilizer to ensure product stability and adaptability to storage conditions. Infrared moisture meter or rapid evaporation equipment can be used.

[0094] Density / strength tester: used to test the bulk density and mechanical strength of particles to ensure stability during transportation and use.

[0095] Chemical analysis equipment: including wet chemical methods to detect impurity content (such as chlorides, sulfides) or other by-products.

[0096] Online near infrared spectroscopy sensor: for continuous analysis of nutrient content.

[0097] Vision sensors: monitor the physical state of the particle surface, such as shape and color consistency.

[0098] Online moisture sensor: obtain the moisture status of particles in real time.

[0099] Industrial Computer (IPC): Integrates output data from quality inspection equipment and processes real-time signals.

[0100] Data storage and analysis system: Use NoSQL database to store detection data for quick comparison with prediction models.

[0101] Edge computing unit: performs local data processing and preliminary error calculation to reduce the burden on cloud servers.

[0102] The specific work flow of the quality inspection unit is as follows:

[0103] Data collection: Real-time collection of fertilizer quality data such as nutrients, particle size, moisture and strength through sensors or detection equipment.

[0104] Signal processing: pre-process the collected signals (such as noise filtering and data smoothing).

[0105] Compare with model predictions: Compare the detected values ​​with the results predicted by the deep learning model and calculate the error.

[0106] Error feedback: The error results are passed to the model update unit to adjust the model parameters.

[0107] For example, for a certain production line of nitrogen, phosphorus and potassium ternary fertilizer, the goal is to ensure the accuracy of the nutrient ratio of 15-15-15.

[0108] Quality inspection unit equipped with NIR spectrometer:

[0109] Scan the nitrogen, phosphorus and potassium content of fertilizer particles in real time at the discharge port.

[0110] The spectrometer data is processed by an edge computing device and compared with the target values ​​predicted by the model.

[0111] If the test finds that the actual nitrogen content is low (such as 14.5), the model will feedback the production parameters (such as urea feed rate) for adjustment.

[0112] The model update unit performs incremental training on the model based on new data to continuously improve prediction accuracy.

[0113] After one month of trial operation, the product quality qualification rate increased by 5%, the production cost decreased by 8%, and the energy consumption decreased by 10%.

[0114] Embodiment 2

[0115] Based on the first embodiment, the architecture of edge computing and cloud computing is introduced.

[0116] Edge computing deployment: Edge computing servers are deployed at the production site to be responsible for real-time data processing and some model calculation tasks to improve response speed.

[0117] Cloud collaboration: Perform offline training and updates of models in the cloud, and use more powerful computing resources to improve model performance.

[0118] A secure data transmission channel is established between the edge server and the cloud server to synchronize data and model parameters regularly.

[0119] Improved security: Blockchain technology is used to encrypt and verify key data to ensure data security and non-tamperability.

[0120] After one month of trial operation, the system's response speed has been further improved, the real-time performance of parameter adjustment has been enhanced, and production efficiency has been increased by about 15%.

[0121] It is worth noting that the contents not elaborated in detail in the present invention are all prior art and are well known to those skilled in the art.

[0122] Therefore, the present invention adopts the above-mentioned nitrogen, phosphorus and potassium ternary fertilizer production process of raw material ratio and parameter combination control method and system, can realize real-time optimization of the production process, improve product quality and production efficiency, and reduce production costs.

[0123] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for regulating raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizers, characterized in that: The following steps are involved: Step S1, deploying a sensor network at key locations of the production line to collect raw material quality parameters, equipment operating parameters, environmental parameters and product quality indicators in real time; Step S2: preprocessing and storing the collected data to form a data set; Step S3: Based on the data set obtained in step S2, a nonlinear mapping relationship between raw material quality parameters, equipment operation parameters, environmental parameters and product quality is established using a deep learning model; Step S4, using a particle swarm optimization algorithm, taking product quality, production cost and energy consumption as objective functions, to calculate the optimal raw material ratio and process parameter combination; Step S5: Send the optimized raw material ratio and process parameters to the execution control module to adjust the raw material input amount and equipment operation parameters in real time; Step S6: Real-time detection of the quality indicators of the fertilizer products produced, comparison of the actual detection results with the model prediction results and calculation of the errors, and use of the errors to incrementally train the deep learning model and update the model parameters.

2. The method for controlling raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizer according to claim 1, characterized in that: In step S1, a sensor network is deployed at key locations of the production line, specifically: a near-infrared spectroscopy sensor is installed at the raw material feed port, a flow sensor is installed on the conveying pipeline, a humidity sensor is installed in the production environment, a temperature sensor and a pressure sensor are installed on the reactor, and a temperature sensor is installed on the mixer.

3. The method for controlling raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizer according to claim 2, characterized in that: In step S2, preprocessing includes data denoising, missing value filling and outlier processing, and the preprocessed data is stored in a NoSQL database.

4. The method for controlling raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizer according to claim 3, characterized in that: In step S3, a deep learning model is used to establish a nonlinear mapping relationship between production parameters and product quality. The specific operations are: using a convolutional neural network to extract the spatial features of the data, and using a long short-term memory network to capture the time series features.

5. The method for controlling raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizer according to claim 4, characterized in that: In step S4, the fitness function of the particle swarm optimization algorithm comprehensively considers product quality, production cost and energy consumption. By initializing a group of particles, each particle represents a possible combination of raw material ratio and process parameters, the fitness value of each particle is evaluated, and the speed and position of the particle are updated according to the fitness value. Through iterative optimization, until the termination condition is met, the optimal raw material ratio and process parameter combination is output; Among them, the fitness function is: Fitness=w1×(1-QQ max )+w2×CC max +w3×EE max ; Among them, Fitness is the fitness value; Q is the product quality index; Q max is the maximum value of product quality index; C is production cost; C max is the maximum value of production cost; E is energy consumption; E max is the maximum value of energy consumption; w1, w2, w3 are all weight coefficients, and w1+w2+w3=1; The particle velocity update formula is: The particle position update formula is: in, is the velocity of the i-th particle at the k-th iteration; is the position of the i-th particle at the k-th iteration, i.e., the parameter combination; w is the inertia weight; c1 and c2 are learning factors; r1 and r2 are random numbers between [0,1]; is the best position of the i-th particle in history; g best is the global best position.

6. The method for controlling raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizer according to claim 5, characterized in that: In step S5, the execution control module is a PLC controller.

7. A system for implementing the method for controlling the raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizers as described in any one of claims 1 to 6, characterized in that: include: Data acquisition module, which is used to deploy sensor networks at key locations on the production line to collect raw material quality parameters, equipment operation parameters, environmental parameters and product quality indicators in real time; A data processing module is used to pre-process and store the collected data to form a data set; Intelligent control module, which uses deep learning models to establish a nonlinear mapping relationship between production parameters and product quality, and uses particle swarm optimization algorithm to calculate the optimal raw material ratio and process parameter combination with product quality, production cost and energy consumption as objective functions; The execution control module is used to adjust the raw material input amount and equipment operation parameters according to the optimized raw material ratio and process parameters output by the intelligent control module; The feedback learning module includes a quality detection unit and a model updating unit. The quality detection unit is used to detect the quality indicators of the fertilizer products produced in real time, compare the actual detection results with the model prediction results and calculate the errors. The model updating unit is used to use the errors to perform incremental training on the deep learning model and update the model parameters.

8. The control system for raw material ratio and parameter combination in the production of nitrogen, phosphorus and potassium ternary fertilizer according to claim 7, characterized in that: It also includes edge computing units and cloud computing units, and the edge computing units are deployed at the production site.

Images