Feces and straw turning and composting control method and system

Abstract

The invention belongs to the technical field of organic fertilizer preparation, and provides a manure straw turning and composting control method and system. The method comprises the following steps: receiving a fertilizer expected preparation time length and a fertilizer property parameter range combination input by a user, and screening according to the fertilizer expected preparation time length to obtain a corresponding pile-turning water-logged compost process; performing parameter measurement on the mixed raw material of the excrement and the straws to be used to obtain raw material parameters; inputting the expected preparation duration of the fertilizer, the turning composting process, the raw material parameters and the fertility parameter range combination into the process parameter optimization model, and predicting to obtain an optimal process parameter combination; and preparing the organic fertilizer from the to-be-used mixed raw materials according to the optimal process parameter combination. According to the process parameter optimization model, the optimal process parameter combination can be obtained through automatic analysis according to the expected fertilizer preparation duration and the fertilizer property parameter range combination required by the user, so that the personalized preparation requirements of the user are met.

Description

Technical Field

[0001] The invention relates to the technical field of organic fertilizer preparation, and in particular to a manure straw composting control method and system. Background Art

[0002] Turning and composting of manure and straw is an environmentally friendly technology that converts agricultural organic waste into fertilizer. Compost volume, material addition ratio, water addition ratio and number of turning the compost are the main process parameters in turning and composting of manure and straw. By adjusting these process parameters, the preparation effect and preparation time can be optimized. At the same time, in reality, users may have different urgency for turning and composting of manure and straw. For example, some users need to complete the turning and composting of manure and straw in the shortest possible time. However, the process parameters such as compost volume, material addition ratio, water content and number of turning the compost in the existing preparation technology are determined based on empirical values, and it is difficult to support the personalized needs of users for the preparation time of organic fertilizers, resulting in failure to meet user needs. This technical problem needs to be improved. Summary of the invention

[0003] In order to solve at least one of the above technical problems, the present invention specifically provides a method, system, electronic equipment, computer storage medium and computer program product for controlling manure and straw composting.

[0004] The present invention provides a method for controlling composting of manure and straw, comprising the following steps: Receive a combination of expected fertilizer preparation time and fertilizer property parameter range input by a user, and select and obtain a corresponding composting process according to the expected fertilizer preparation time; Parameters of the mixed raw material of manure, sewage and straw to be used are measured to obtain raw material parameters; The expected preparation time of the fertilizer, the compost turning and fermentation process, the raw material parameters and the fertilizer property parameter range are combined and input into a process parameter optimization model based on Transformer, and the process parameter optimization model outputs a predicted optimal process parameter combination, wherein the optimal process parameter combination includes compost volume, material addition ratio, water addition ratio and compost turning times; The mixed raw materials to be used are used to prepare organic fertilizer according to the optimal process parameter combination, including composting and fermentation, and compost turning treatment.

[0005] Furthermore, the receiving of the expected fertilizer preparation time and the fertilizer property parameter range combination input by the user includes: Receiving organic fertilizer preparation demand information input by a user, and parsing the organic fertilizer preparation demand information; If the analysis result includes the expected preparation time of the fertilizer and the range of the fertilizer property parameter combination, the expected preparation time of the fertilizer and the range of the fertilizer property parameter combination will be confirmed and transmitted; If the parsing result does not include the expected preparation time of the fertilizer and the combination of the fertilizer property parameter range, a semantic analysis plug-in is called to perform semantic analysis on the organic fertilizer preparation demand information to obtain the expected preparation time of the fertilizer and the combination of the fertilizer property parameter range.

[0006] Furthermore, the process parameter optimization model is trained in the following manner: Collecting multiple sets of local preparation history data of organic fertilizer prepared by composting of manure and straw, constructing a first training data set and a first test data set based on the multiple sets of preparation history data, and using the first training data set to train the process parameter optimization model once; After one training is completed, the process parameter optimization model that has been trained once is tested using the first test data set, and a comprehensive test deviation is calculated, and the number of collaborative nodes is determined according to the comprehensive test deviation; The process parameter optimization model is copied according to the number of collaborative nodes to obtain a corresponding number of model copies, and each model copy is published on the alliance chain to the corresponding collaborative nodes; Receive the second key model parameters and the second training data set fed back by each of the collaborative nodes, fuse the local first key model parameters and each of the second key model parameters to obtain target key model parameters, and import the target key model parameters into the local trained process parameter optimization model; A set number of training data are randomly selected from each of the second training data sets to form a third training data set and a second test data set. The process parameter optimization model that has been trained once is trained for a second time using the third training data set, and the process parameter optimization model after the second training is tested for a second time using the second test data set. If the test result meets the standard, the training is terminated.

[0007] Furthermore, the number of collaborative nodes is calculated based on the following formula: (1) (2) In the above formula, is the number of collaborative nodes, To preset the number of collaborative nodes, is the average value of the test deviations obtained after testing each test data in the first test data set, namely, the comprehensive test deviation, is the total number of test data in the first test data set; is the normalization function, is the ceiling function, is the adjustment value; is the weight coefficient, and ; is the maximum mode value of the test deviation obtained after testing each test data in the first test data set, The number of most common values.

[0008] Furthermore, the set quantity is determined by: Calculate the maximum deviation amplitude of each test deviation obtained in one test, wherein the maximum deviation amplitude is the difference between the maximum value and the minimum value of the test deviation; The set quantity is determined based on the maximum deviation amplitude and in accordance with a positive correlation.

[0009] Furthermore, the first key model parameters and each of the second key model parameters include at least model architecture parameters, self-attention mechanism parameters, feedforward neural network parameters, LayerNorm parameters, and position encoding parameters.

[0010] The present invention also provides a manure straw composting control system, the system comprising a processing module and a storage module; the processing module is electrically connected to the storage module; the processing module calls the executable computer code stored in the storage module to implement the following steps: Receive a combination of expected fertilizer preparation time and fertilizer property parameter range input by a user, and select and obtain a corresponding composting process according to the expected fertilizer preparation time; Parameters of the mixed raw material of manure, sewage and straw to be used are measured to obtain raw material parameters; The expected preparation time of the fertilizer, the compost turning and fermentation process, the raw material parameters and the fertilizer property parameter range are combined and input into a process parameter optimization model based on Transformer, and the process parameter optimization model outputs a predicted optimal process parameter combination, wherein the optimal process parameter combination includes compost volume, material addition ratio, water addition ratio and compost turning times; The mixed raw materials to be used are used to prepare organic fertilizer according to the optimal process parameter combination, including composting and fermentation, and compost turning treatment.

[0011] The present invention also provides an electronic device, comprising: a memory storing executable program code; a processor coupled to the memory; the processor calls the executable program code stored in the memory to execute the method as described in any of the preceding items.

[0012] The present invention also provides a computer storage medium, on which a computer program is stored. When the computer program is executed by a processor, any of the above methods is executed.

[0013] The present invention also provides a computer program product, which is used to be called and executed by a processor of an electronic device to implement any of the methods described above.

[0014] The beneficial effect of the present invention is that the process parameter optimization model in the scheme of the present invention can automatically analyze and obtain the optimal process parameter combination according to the expected preparation time of the fertilizer required by the user and the combination of the fertilizer property parameter range, so that the organic fertilizer that meets the fertilizer property parameter range combination can be prepared within the expected preparation time, thereby meeting the personalized preparation needs of the user, which is beneficial to improving the economic benefits of the organic fertilizer preparation manufacturer. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 The present invention discloses a flow chart of a method for controlling manure and straw composting.

[0017] Figure 2 The invention is a schematic structural diagram of a manure and straw composting control system disclosed in an embodiment of the invention. DETAILED DESCRIPTION

[0018] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0019] See also Figure 1 As shown in the flow chart, the embodiment of the present invention provides a method for controlling manure straw composting, comprising the following steps: Receive a combination of expected fertilizer preparation time and fertilizer property parameter range input by a user, and select and obtain a corresponding composting process according to the expected fertilizer preparation time; Parameters of the mixed raw material of manure, sewage and straw to be used are measured to obtain raw material parameters; The expected preparation time of the fertilizer, the compost turning and fermentation process, the raw material parameters and the fertilizer property parameter range are combined and input into a process parameter optimization model based on Transformer, and the process parameter optimization model outputs a predicted optimal process parameter combination, wherein the optimal process parameter combination includes compost volume, material addition ratio, water addition ratio and compost turning times; The mixed raw materials to be used are used to prepare organic fertilizer according to the optimal process parameter combination, including composting and fermentation, and compost turning treatment.

[0020] For the technical problems mentioned in the background technology, the scheme of the present invention can be based on the expected preparation time of the fertilizer expected by the user and the combination of fertilizer parameter ranges, and automatically determine the optimal process parameter combination that can complete the preparation of organic fertilizer within the expected preparation time of the fertilizer, and can ensure that the fertilizer index of the prepared organic fertilizer is within the combination of fertilizer parameter ranges. Specifically, the system receives the expected preparation time of the fertilizer input by the user, for example, 70 days. Since the preparation time required for different composting processes is different, it is necessary to screen the corresponding composting process according to the expected preparation time of the fertilizer, such as strip composting (30-60 days), film composting (10-15 days), NCS intelligent film composting (7-25 days), etc. It should be noted that the costs of different composting processes are different, and the rapid preparation method equipped by the organic fertilizer preparation factory should also be retained for standby. If each batch of organic fertilizer preparation orders adopts, for example, NCS intelligent film composting, it is obviously not economical enough. Therefore, the present invention still sets up a compost turning and composting process with a suitable duration according to the expected preparation duration of the fertilizer. And also receives a fertilizer parameter range combination input by a user, the fertilizer parameter range combination includes multiple fertilizer index ranges, mainly including the mass fraction range of organic matter, the mass fraction range of total nutrients, etc.

[0021] At the same time, the parameters of the mixed raw materials of manure and straw (i.e. the mixture of manure and straw after mixing) to be used for the preparation of organic fertilizer are measured to obtain the raw material parameters. Among them, the raw material parameters mainly include the carbon-nitrogen ratio (C / N ratio) and the percentage of moisture content. Carbon-nitrogen ratio: The appropriate carbon-nitrogen ratio can promote the metabolism of microorganisms and accelerate the decomposition of organic matter; Moisture content percentage: The moisture content in the mixed raw materials of composting should generally be controlled between 50% and 70%. Too low moisture is not conducive to the growth of microorganisms. When the moisture content exceeds 70%, it will block the gaps in the pile, affect ventilation, and cause anaerobic fermentation.

[0022] Next, the expected preparation time of the fertilizer, composting process, raw material parameters, and fertilizer parameter range combinations obtained above are input into the Transformer-based process parameter optimization model. The process parameter optimization model processes the above input data comprehensively and in depth to obtain the predicted optimal process parameter combination, which includes compost volume, material addition ratio, water addition ratio, and compost turning times.

[0023] Finally, the organic fertilizer preparation process is carried out according to the optimal process parameter combination determined by the process parameter optimization model, that is, the mixed raw materials to be used are subjected to the main preparation processes such as composting and fermentation, compost turning, etc., so that the organic fertilizer that meets the fertilizer parameter range combination can be prepared within the expected preparation time of the fertilizer.

[0024] Therefore, the process parameter optimization model in the scheme of the present invention can automatically analyze and obtain the optimal process parameter combination according to the expected preparation time of the fertilizer required by the user and the combination of the fertilizer parameter range, so that the organic fertilizer that meets the fertilizer parameter range combination can be prepared within the expected preparation time, thereby meeting the personalized preparation needs of the user, which is beneficial to improving the economic benefits of the organic fertilizer preparation manufacturer.

[0025] Furthermore, the receiving of the expected fertilizer preparation time and the fertilizer property parameter range combination input by the user includes: Receiving organic fertilizer preparation demand information input by a user, and parsing the organic fertilizer preparation demand information; If the analysis result includes the expected preparation time of the fertilizer and the range of the fertilizer property parameter combination, the expected preparation time of the fertilizer and the range of the fertilizer property parameter combination will be confirmed and transmitted; If the parsing result does not include the expected preparation time of the fertilizer and the combination of the fertilizer property parameter range, a semantic analysis plug-in is called to perform semantic analysis on the organic fertilizer preparation demand information to obtain the expected preparation time of the fertilizer and the combination of the fertilizer property parameter range.

[0026] In this example, the solution of the present invention can be used for the order receiving system of the organic fertilizer preparation manufacturer, such as an order receiving webpage, a dedicated APP, etc. The user can input the organic fertilizer preparation demand information to the order receiving webpage, the dedicated APP, etc., and the organic fertilizer preparation demand information is used to characterize the user's relevant requirements for the organic fertilizer to be purchased. At the same time, the user can directly input the expected preparation time of the fertilizer and the combination of the fertilizer parameter range; or input a description text that can indirectly reflect their needs, such as the user of the organic fertilizer, the expected use time, etc. For the former, the system can directly parse out the expected preparation time of the fertilizer and the combination of the fertilizer parameter range, and confirm and transmit it to the subsequent module for processing; and for the latter, it is necessary to call the semantic analysis plug-in to perform semantic analysis on the organic fertilizer preparation demand information, thereby indirectly obtaining the expected preparation time of the fertilizer and the combination of the fertilizer parameter range, and then confirm and transmit it to the subsequent module for processing.

[0027] Therefore, the scheme of the present invention can quickly determine the organic fertilizer preparation demand regardless of whether the user inputs clear or vague organic fertilizer preparation demand information, thereby reducing the input difficulty for the user.

[0028] It should be noted that when the semantic analysis plug-in performs semantic analysis on the organic fertilizer preparation demand information, it can extract the expected fertilizer preparation time and a corresponding set of demand keywords, and then crawl network data based on the set of demand keywords or ask questions to the commercial big model to obtain the fertilizer parameter range combination.

[0029] Furthermore, the process parameter optimization model is trained in the following manner: Collecting multiple sets of local preparation history data of organic fertilizer prepared by composting of manure and straw, constructing a first training data set and a first test data set based on the multiple sets of preparation history data, and using the first training data set to train the process parameter optimization model once; After one training is completed, the process parameter optimization model that has been trained once is tested using the first test data set, and a comprehensive test deviation is calculated, and the number of collaborative nodes is determined according to the comprehensive test deviation; The process parameter optimization model is copied according to the number of collaborative nodes to obtain a corresponding number of model copies, and each model copy is published on the alliance chain to the corresponding collaborative nodes; Receive the second key model parameters and the second training data set fed back by each of the collaborative nodes, fuse the local first key model parameters and each of the second key model parameters to obtain target key model parameters, and import the target key model parameters into the local trained process parameter optimization model; A set number of training data are randomly selected from each of the second training data sets to form a third training data set and a second test data set. The process parameter optimization model that has been trained once is trained for a second time using the third training data set, and the process parameter optimization model after the second training is tested for a second time using the second test data set. If the test result meets the standard, the training is terminated.

[0030] In this example, due to various factors, the production history data of an organic fertilizer production plant in one or one region is usually insufficient, which makes it difficult to fully train the process parameter optimization model. In response to this technical problem, the present invention constructs a distributed training method based on the alliance chain, that is, each collaborative node on the alliance chain collects training data on its own to participate in the collaborative training of the process parameter optimization model, and then obtains the optimal key model parameters based on the key model parameters obtained by the training of each collaborative node, thereby solving the above-mentioned technical problem of insufficient training data. Specifically: First, the local node collects multiple sets of local preparation history data for preparing organic fertilizer through composting of manure and straw. Each piece of preparation history data contains composting process, raw material parameters, fertilizer parameter range combination (the three are used as training / test data) and corresponding preparation time (as label data of training / test data). The first training data set and the first test data set are constructed from these preparation history data, and the process parameter optimization model is trained once using the first training data set. After the training is completed, the trained process parameter optimization model is tested using the first test data set, and the comprehensive test deviation is calculated. The number of collaborative nodes is determined based on the comprehensive test deviation.

[0031] Then, the process parameter optimization model is copied according to the number of collaborative nodes determined in the previous step, and each model copy is published one by one on the alliance chain to the corresponding collaborative nodes. Each collaborative node collects local training data and trains and tests the model copy until the training meets the standard, and then feeds back the second key model parameters of the model copy that meets the standard and the second training data set used for training (which can only contain training data or both training data and test data) to the local node through the alliance chain. The first key model parameters of the model master obtained by the one-time training of the local node and the second key model parameters of each model copy are fused to obtain the target key model parameters, and the target key model parameters are imported into the local process parameter optimization model that has been trained once.

[0032] Finally, a set number of training data are randomly selected from each of the second training data sets fed back above to form a third training data set and a second test data set. The process parameter optimization model that has been trained once is trained twice using the third training data set. After the second training is completed, the process parameter optimization model after the second training is tested again using the second test data set. If the test result meets the standard, the training is terminated. If the test result does not meet the standard, a copy of the model after the second training can be generated and published on the alliance chain, and the above process is repeated until the test result meets the standard.

[0033] Therefore, this embodiment of the present invention can train the process parameter optimization model without relying entirely on local training data, but can be collaboratively trained with the help of multiple other collaborative nodes on the alliance chain, thereby solving the problem of insufficient local training data; moreover, since the training data used by the collaborative training nodes on the alliance chain is different from that of the local nodes, this also improves the diversity of training data, which is conducive to improving the training effect of the model.

[0034] Furthermore, the number of collaborative nodes is calculated based on the following formula: (1) (2) In the above formula, is the number of collaborative nodes, is the preset number of collaborative nodes (for example, 5 or 10). is the average value of the test deviations obtained after testing each test data in the first test data set, namely, the comprehensive test deviation, is the total number of test data in the first test data set; is the normalization function (used to remove The unit is time length), is the ceiling function, is the adjustment value; is the weight coefficient, and ; is the maximum mode value of the test deviation obtained after testing each test data in the first test data set, The number of most common values.

[0035] In this example, the process parameter optimization model after training is tested once using each test data in the first test data set, and the test deviation of each test is recorded, that is, the absolute value of the deviation between the predicted preparation time and the preparation time in the label data corresponding to the test data. The average value of all test deviations is taken as the comprehensive test deviation, and the maximum number of these test deviations is also calculated, that is, the test deviation with the most occurrences. It should be noted that the test deviation with the most occurrences can also be a group of test deviations with the largest number of clusters, and the cluster center of this group of test deviations is taken as the above-mentioned maximum number of values.

[0036] Furthermore, the set quantity is determined by: Calculate the maximum deviation amplitude of each test deviation obtained in one test, wherein the maximum deviation amplitude is the difference between the maximum value and the minimum value of the test deviation; The set quantity is determined based on the maximum deviation amplitude and in accordance with a positive correlation.

[0037] In this example, after the process parameter optimization model after a training is completed is tested once using each test data in the first test data set, it is also necessary to calculate the deviation between the maximum and minimum values ​​of the test deviation of each test recorded, that is, the maximum deviation amplitude. When the maximum deviation amplitude is larger, it indicates that the diversity of the first training data set and the first test data set used by the local node is worse, that is, the composting process, raw material parameters, fertilizer parameter range combination and preparation time involved in each preparation history data of the organic fertilizer preparation factory are relatively close, which leads to the lack of robustness of the process parameter optimization model after a training, and the larger the maximum deviation amplitude obtained by the test, the more collaborative nodes need to be selected to perform collaborative training on the process parameter optimization model. On the contrary, it is necessary to select a smaller number of collaborative nodes to perform collaborative training on the process parameter optimization model, which can reduce the difficulty of training and improve the training speed.

[0038] Furthermore, the first key model parameters and each of the second key model parameters include at least model architecture parameters, self-attention mechanism parameters, feedforward neural network parameters, LayerNorm parameters, and position encoding parameters.

[0039] In this example, the training of the Transformer-based process parameter optimization model is mainly used to determine the model architecture parameters, self-attention mechanism parameters, feedforward neural network parameters, LayerNorm parameters, and position encoding parameters of the model. These parameters determine the ability of the process parameter optimization model to perform accurate analysis and prediction. The specific description of each parameter is as follows: 1. Model architecture parameters: Number of encoder and decoder layers: The Transformer model consists of multiple encoder and decoder layers, each of which contains a self-attention mechanism and a feed-forward neural network. The number of layers affects the depth and learning ability of the model.

[0040] Hidden layer dimension (d_model): is the size of the hidden layer in the model, that is, the dimension of each word vector.

[0041] Attention heads: The number of heads in the multi-head attention mechanism, which allows the model to learn information in different representation subspaces in parallel.

[0042] 2. Self-attention mechanism parameters: Weight matrices for query (Q), key (K), and value (V): These are the parameters in the self-attention mechanism and are used to calculate the attention scores and weights.

[0043] Output weight matrix and bias: The output after the self-attention mechanism needs to pass through a linear layer, which also has its own weight matrix and bias.

[0044] 3. Feedforward neural network parameters: weight matrices and biases of two linear layers: FFN consists of two linear layers, the first one maps the dimension from d_model to a larger dimension d_ff, and the second one maps the dimension from d_ff back to d_model.

[0045] 4. Position encoding parameters: Positional encodings: Transformer models require positional encodings to capture the order information of words in a sequence. These encodings can be fixed or learned.

[0046] In addition, the local node performs fusion processing on the local first key model parameters and the second key model parameters of other collaborative nodes, which may be average processing of the above parameters.

[0047] like Figure 2 As shown in the structural schematic diagram, the embodiment of the present invention also discloses a manure straw composting control system, the system includes a processing module and a storage module; the processing module is electrically connected to the storage module; the processing module calls the executable computer code stored in the storage module to implement the following steps: Receive a combination of expected fertilizer preparation time and fertilizer property parameter range input by a user, and select and obtain a corresponding composting process according to the expected fertilizer preparation time; Parameters of the mixed raw material of manure, sewage and straw to be used are measured to obtain raw material parameters; The expected preparation time of the fertilizer, the compost turning and fermentation process, the raw material parameters and the fertilizer property parameter range are combined and input into a process parameter optimization model based on Transformer, and the process parameter optimization model outputs a predicted optimal process parameter combination, wherein the optimal process parameter combination includes compost volume, material addition ratio, water addition ratio and compost turning times; The mixed raw materials to be used are used to prepare organic fertilizer according to the optimal process parameter combination, including composting and fermentation, and compost turning treatment.

[0048] An embodiment of the present invention further discloses an electronic device, comprising: a memory storing executable program code; a processor coupled to the memory; the processor calls the executable program code stored in the memory to execute the method described in the above embodiment.

[0049] The embodiment of the present invention further discloses a computer storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method described in the above embodiment is executed.

[0050] An embodiment of the present invention further discloses a computer program product, which is used to be called and executed by a processor of an electronic device to implement any of the above methods.

[0051] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable load balancing device, so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0052] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0053] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for controlling composting of manure and straw, characterized in that: The steps include: Receive a combination of expected fertilizer preparation time and fertilizer property parameter range input by a user, and select and obtain a corresponding composting process according to the expected fertilizer preparation time; Parameters of the mixed raw material of manure, sewage and straw to be used are measured to obtain raw material parameters; The expected preparation time of the fertilizer, the compost turning and fermentation process, the raw material parameters and the fertilizer property parameter range are combined and input into a process parameter optimization model based on Transformer, and the process parameter optimization model outputs a predicted optimal process parameter combination, wherein the optimal process parameter combination includes compost volume, material addition ratio, water addition ratio and compost turning times; The mixed raw materials to be used are used to prepare organic fertilizer according to the optimal process parameter combination, including composting and fermentation, and compost turning treatment.

2. A method for controlling composting of manure and straw according to claim 1, characterized in that: Receive the expected fertilizer preparation time and fertilizer property parameter range combination input by the user, including: Receiving organic fertilizer preparation demand information input by a user, and parsing the organic fertilizer preparation demand information; If the analysis result includes the expected preparation time of the fertilizer and the range of the fertilizer property parameter combination, the expected preparation time of the fertilizer and the range of the fertilizer property parameter combination will be confirmed and transmitted; If the parsing result does not include the expected preparation time of the fertilizer and the combination of the fertilizer property parameter range, a semantic analysis plug-in is called to perform semantic analysis on the organic fertilizer preparation demand information to obtain the expected preparation time of the fertilizer and the combination of the fertilizer property parameter range.

3. A method for controlling composting of manure and straw according to claim 2, characterized in that: The process parameter optimization model is trained in the following way: Collecting multiple sets of local preparation history data of organic fertilizer prepared by composting of manure and straw, constructing a first training data set and a first test data set based on the multiple sets of preparation history data, and using the first training data set to train the process parameter optimization model once; After one training is completed, the process parameter optimization model that has been trained once is tested using the first test data set, and a comprehensive test deviation is calculated, and the number of collaborative nodes is determined according to the comprehensive test deviation; The process parameter optimization model is copied according to the number of collaborative nodes to obtain a corresponding number of model copies, and each model copy is published on the alliance chain to the corresponding collaborative nodes; Receiving the second key model parameters and the second training data set fed back by each of the collaborative nodes, fusing the local first key model parameters and each of the second key model parameters to obtain target key model parameters, and importing the target key model parameters into the local trained process parameter optimization model; A set number of training data are randomly selected from each of the second training data sets to form a third training data set and a second test data set. The process parameter optimization model that has been trained once is trained for a second time using the third training data set, and the process parameter optimization model after the second training is tested for a second time using the second test data set. If the test result meets the standard, the training is terminated.

4. A method for controlling composting of manure and straw according to claim 3, characterized in that: The number of collaborative nodes is calculated based on the following formula: (1) (2) In the above formula, is the number of collaborative nodes, To preset the number of collaborative nodes, is the average value of the test deviations obtained after testing each test data in the first test data set, namely, the comprehensive test deviation, is the total number of test data in the first test data set; is the normalization function, is the ceiling function, is the adjustment value; is the weight coefficient, and ; is the maximum mode value of the test deviation obtained after testing each test data in the first test data set, The number of most common values.

5. A method for controlling composting of manure and straw according to claim 4, characterized in that: The set quantity is determined as follows: Calculate the maximum deviation amplitude of each test deviation obtained in one test, wherein the maximum deviation amplitude is the difference between the maximum value and the minimum value of the test deviation; The set quantity is determined based on the maximum deviation amplitude and in accordance with a positive correlation.

6. A method for controlling composting of manure and straw according to claim 3, characterized in that: The first key model parameters and each of the second key model parameters include at least model architecture parameters, self-attention mechanism parameters, feedforward neural network parameters, LayerNorm parameters, and position encoding parameters.

7. A manure straw composting control system, the system comprising a processing module and a storage module; the processing module is electrically connected to the storage module; characterized in that: The processing module calls the executable computer code stored in the storage module to implement the following steps: Receive a combination of expected fertilizer preparation time and fertilizer property parameter range input by a user, and select and obtain a corresponding composting process according to the expected fertilizer preparation time; Parameters of the mixed raw material of manure, sewage and straw to be used are measured to obtain raw material parameters; The expected preparation time of the fertilizer, the compost turning and fermentation process, the raw material parameters and the fertilizer property parameter range are combined and input into a process parameter optimization model based on Transformer, and the process parameter optimization model outputs a predicted optimal process parameter combination, wherein the optimal process parameter combination includes compost volume, material addition ratio, water addition ratio and compost turning times; The mixed raw materials to be used are used to prepare organic fertilizer according to the optimal process parameter combination, including composting and fermentation, and compost turning treatment.

8. An electronic device comprising: A memory storing executable program code; A processor coupled to the memory; characterized in that: the processor calls the executable program code stored in the memory to execute the method according to any one of claims 1-6.

9. A computer storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is executed.

10. A computer program product, characterized in that: The computer program product is used to be called and executed by a processor of an electronic device to implement the method according to any one of claims 1 to 6.

Images