Adjustment control method and system of intelligent mulberry leaf feed fermentation device

Abstract

The invention relates to the technical field of feed production, in particular to an adjustment control method and system of an intelligent mulberry leaf feed fermentation device. On the basis of the fermentation inhibition degree, ideally iterating supplementary adjustment control of the initial oxygen concentration content to obtain a first adjustment control scheme; collecting time sequence oxygen content distribution data of fermentation of the mulberry leaf feed through an oxygen detection device, and calculating oxygen content variation distribution deviation under stirring air power according to variation of the time sequence oxygen content distribution data so as to adjust and control the stirring speed to obtain a second adjustment control scheme; real-time pH acid-base signals of mulberry leaf feed fermentation are obtained through a pH detection device, normal pH acid-base gradient dislocation fragment calculation is carried out on the real-time pH acid-base signals so as to control bacterium liquid spraying, and a third adjustment control scheme is obtained. Various important parameters in the intelligent mulberry leaf feed fermentation process can be precisely adjusted and controlled, and the fermentation efficiency and the fermentation stability can be greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of feed production, and in particular to a regulating and controlling method and system of an intelligent mulberry leaf feed fermentation device. Background Art

[0002] With the continuous development of agricultural animal husbandry, feed quality and efficiency have become key issues. Moreover, the fermentation of feed raw materials is of great significance for improving their nutritional value and reducing harmful substances. Therefore, a more intelligent and efficient feed fermentation device is needed to meet these needs. As a feed raw material with potential nutritional value, the fermentation process of mulberry leaf feed also needs to be controlled and optimized. Existing mulberry leaf feed fermentation devices usually have functions such as automated stirring, pH value and oxygen content monitoring, and sample extraction to improve fermentation efficiency and ensure the quality and safety of feed.

[0003] However, the existing mulberry leaf feed fermentation equipment lacks the support of efficient and accurate regulation and control systems, which means that manual intervention is often required during the fermentation process of the equipment. There are problems such as difficulty in accurately controlling process parameters, low fermentation efficiency, and complicated operation. In addition, the existing equipment is still unable to accurately control the rationality of the stirring rate, oxygen distribution, oxygen supplementation, and pH acid-base abnormalities during the fermentation process, making it difficult to meet the needs of modern breeding for efficient and intelligent management. Summary of the invention

[0004] The present invention overcomes the deficiencies of the prior art and provides a regulating and controlling method and system for an intelligent mulberry leaf feed fermentation device.

[0005] To achieve the above object, the technical solution adopted by the present invention is: The first aspect of the present invention provides a regulation and control method for an intelligent mulberry leaf feed fermentation device, comprising the following steps: S102: Obtaining the fermentation inhibition degree of the auxiliary fermentation material on the mulberry leaf fermentation raw material, and ideally iterating the supplementary regulation and control of the initial oxygen content based on the fermentation inhibition degree to obtain a first regulation and control scheme; S104: collecting the time-series oxygen concentration distribution data of the mulberry leaf feed fermentation through the oxygen detection device, and calculating the variational distribution deviation of the oxygen concentration under the stirring air power according to the time-series oxygen concentration distribution data to adjust and control the stirring rate, and obtain a second adjustment control scheme; S106: obtaining a real-time pH acid-base signal of mulberry leaf feed fermentation through a pH detection device, performing normalized pH acid-base gradient dislocation fragment calculation on the real-time pH acid-base signal to control bacterial liquid spraying, and obtaining a third regulation control scheme; S108: Sampling and testing to obtain the fermented mulberry leaf feed sample after adjustment and control, and mapping the indicators of the multi-dimensional sampling test data and performing test analysis based on the unqualified fermentation image data to obtain the test results.

[0006] More specifically, the step S102 includes the following steps: Obtaining the target demand of the mulberry leaf feed fermentation device, obtaining the preset oxygen control strategy of the intelligent mulberry leaf feed fermentation device for the target demand, and extracting the initial oxygen content input by the intelligent mulberry leaf feed fermentation device through the preset oxygen control strategy; The ideal oxygen content is preset according to the target demand. If the initial oxygen content is lower than the ideal oxygen content, the historical fermentation data of the mulberry leaf fermentation raw material and each auxiliary fermentation material are obtained according to the target demand, and the fermentation inhibition degree of each auxiliary fermentation material on the mulberry leaf fermentation raw material is calculated based on the weighted and matched distribution of the historical fermentation data; A penalty factor constraint function is preset based on the degree of fermentation inhibition, and the Lagrangian algorithm is introduced. The initial oxygen concentration is calculated in the Lagrangian algorithm based on the preset oxygen control strategy to obtain the Lagrangian functions of different oxygen control solutions. The penalty factor constraint function and the ideal oxygen content are calculated by minimizing the weighted calculation using a trade-off algorithm to obtain an iterative central path, and the iterative central path is tracked by executing different oxygen control strategies through a Lagrangian function. During the tracking process, the preset oxygen control strategy is continuously updated and iterated to approach the optimal solution for the intelligent mulberry leaf feed fermentation device to achieve the ideal oxygen content, and the current iteration step is output; According to the ideal oxygen content, an iteration step threshold is preset. If the current iteration step does not exceed the iteration step threshold, the tracking iteration continues. If the current iteration step exceeds the iteration step threshold, the tracking iteration operation is stopped, and finally a series of optimal oxygen control solutions are generated. A series of optimal oxygen control solutions are deployed on the control terminal of the intelligent mulberry leaf feed fermentation device to perform oxygen supplementation adjustment and control on the initial oxygen content to obtain the first adjustment and control scheme.

[0007] More specifically, the method of obtaining the historical fermentation data of the mulberry leaf fermentation raw material and each auxiliary fermentation material according to the target demand, and calculating the fermentation inhibition degree of each auxiliary fermentation material on the mulberry leaf fermentation raw material based on the weighted and matched distribution of the historical fermentation data, specifically includes the following steps: Acquire mulberry leaf fermentation raw materials and several auxiliary materials according to target requirements, and simultaneously acquire the actual dosage of mulberry leaf fermentation raw materials and each auxiliary fermentation material; The historical fermentation data of the mulberry leaf fermentation raw material when the actual dosage is output within a preset time period is extracted through the fermentation log, which is defined as the dependent historical fermentation data, and the historical fermentation data of each auxiliary fermentation material when the actual dosage is output within a preset time period is extracted, which is defined as the covariant historical fermentation data; A logistic regression model is introduced to perform regression estimation on each dependent historical fermentation data and each covariant historical fermentation data to obtain several logistic regression functions, and the propensity score of each covariant historical fermentation data leading to the treatment of each dependent historical fermentation data is determined according to the logistic regression function; Based on the propensity score, the current covariate distribution when the covariate historical fermentation data causes the dependent historical fermentation data to be processed is obtained, and each covariate historical fermentation data is bundled with the corresponding dependent historical fermentation data that causes the processing to be received into a weighted control group to obtain a plurality of weighted control groups; A trade-off algorithm is introduced to weigh each weighted control group and generate the mean of the weighted control group. Based on the mean of the weighted control group, each covariate historical fermentation data is individually matched with the dependent historical fermentation data to generate a covariate matching distribution. The registration covariate matching distribution and the current covariate distribution are used to obtain the registration balance difference, and the fermentation inhibition degree of each auxiliary fermentation material on the mulberry leaf fermentation raw material is determined according to the registration balance difference.

[0008] More specifically, the step S104 includes the following steps: The oxygen detection device is used to detect oxygen in the intelligent mulberry leaf feed fermentation device after the first adjustment and control scheme is executed, so as to obtain the time-series oxygen concentration distribution data of the mulberry leaf feed fermentation, and at the same time obtain the predetermined detection area range of the oxygen detection device; Constructing a bubble space domain of an intelligent mulberry leaf feed fermentation device, fitting the time-series oxygen concentration distribution data into the bubble space domain one by one, and after the fitting is completed, dividing the bubble space domain into M sub-bubble domains based on a predetermined detection area range, generating M oxygen content bubble domains, and obtaining the bubble volume change value of each oxygen content bubble domain as the time series transitions; If the bubble volume change value is greater than the preset bubble volume change value, the oxygen content bubble field is calibrated as a type I bubble field in the bubble space field; if it is less than the preset bubble volume change value, the oxygen content bubble field is calibrated as a type II bubble field. The change layout of the type I bubble field and the type II bubble field is visualized to obtain the time series change pattern of oxygen concentration; Obtain the time-series stirring rate corresponding to the time-series oxygen concentration distribution data generated by the intelligent mulberry leaf feed fermentation device, obtain the aerodynamic knowledge graph based on the big data network, identify the time-series stirring rate through the aerodynamic knowledge graph, and output the reference aerodynamic vector coefficient generated inside the device during the stirring process; According to the time series variation pattern of oxygen concentration, a plurality of random oxygen variation distribution samples are constructed. Based on the reference aerodynamic vector coefficient, a variational extrapolation of the oxygen concentration variational distribution deviation is performed on each random oxygen variation distribution sample. According to the oxygen concentration variational distribution deviation, the stirring rate is adjusted and controlled to obtain a second adjustment control scheme.

[0009] More specifically, the method constructs a plurality of random oxygen change distribution samples according to the time series change pattern of oxygen concentration, performs variational calculation of oxygen concentration variation distribution deviation for each random oxygen change distribution sample based on the reference aerodynamic vector coefficient, and adjusts and controls the stirring rate according to the oxygen concentration variation distribution deviation to obtain the second adjustment control scheme, which specifically includes the following steps: Constructing a number of random oxygen change distribution samples resulting from the time-series change pattern of oxygen concentration caused by the time-series stirring rate, and calculating the probability mass function of each random oxygen change distribution sample based on the reference aerodynamic vector coefficient; The Bayesian variational inference network is introduced to estimate the marginal probability lower bound of each probability mass function. By calculating the gradient of the marginal probability lower bound for each random oxygen change distribution sample, the distribution state of the corresponding random oxygen change distribution sample is updated based on the gradient to obtain the actual oxygen concentration variation distribution. An ideal oxygen concentration variational distribution that eliminates the time-series variation pattern of oxygen content is obtained, the distribution difference between the actual oxygen concentration variational distribution and the ideal oxygen concentration variational distribution is calculated, and the oxygen concentration variational distribution deviation is obtained. Based on the oxygen concentration variational distribution deviation, the stirring rate of the intelligent mulberry leaf feed fermentation device is adjusted and controlled to obtain a second adjustment and control scheme.

[0010] More specifically, the step S106 includes the following steps: The pH detection device is used to detect the pH of the intelligent mulberry leaf feed fermentation device after the second adjustment control scheme is executed, and the real-time pH acid-base signal of the mulberry leaf feed fermentation in the target time period is obtained, and the wavelet transform algorithm is introduced to extract the features of the real-time pH acid-base signal to obtain a number of real-time pH acid-base frequency values; Obtaining a reference pH acid-base interval for mulberry leaf feed fermentation, setting an abnormal pH acid-base frequency value range according to the reference pH acid-base interval, extracting only the real-time pH acid-base frequency value within the abnormal pH acid-base frequency value range and marking it as an abnormal pH acid-base frequency value, and obtaining a plurality of abnormal pH acid-base frequency values; Obtaining a preset detection interval of a pH detection device within a target time period, establishing a wavelet basis variation coefficient describing the pH time series variation based on the preset detection interval, and performing a time series gradient estimation on all abnormal pH acid-base frequency values ​​based on the wavelet basis variation coefficient to construct an abnormal pH acid-base gradient diagram; According to the target requirements, the normal pH frequency range of mulberry leaf feed fermentation is obtained, and a normal pH acid-base gradient diagram is established based on the normal pH frequency range. The concept dislocation algorithm is introduced to calculate the gradient dislocation between the abnormal pH acid-base gradient diagram and the normal pH acid-base gradient diagram, and the pH acid-base gradient dislocation fragment is obtained. The pH acid-base gap of the abnormal pH acid-base gradient diagram compared with the normal pH acid-base gradient diagram is determined according to the pH acid-base gradient dislocation fragment, and the bacterial liquid spraying of the intelligent mulberry leaf feed fermentation device is controlled based on the pH acid-base gradient gap to obtain a third regulation and control scheme.

[0011] More specifically, the step S108 includes the following steps: By sampling and extracting mulberry leaf feed fermentation samples after executing the first adjustment control scheme, the second adjustment control scheme and the third adjustment control scheme, and using biological technology to detect the mulberry leaf feed fermentation samples, a number of multi-dimensional sampling detection data are obtained, which are defined as actual sampling detection data; Based on the big data network, unqualified fermentation image data of mulberry leaf feed and several multi-dimensional sampling detection data when unqualified fermentation occurs are obtained, which are defined as reference sampling detection data. The LBP algorithm is introduced to extract features of the unqualified fermentation image data, and the LBP value of the unqualified fermentation feature is generated. Based on the LBP value, an unqualified fermentation feature mapping space of the reference sampling detection data is constructed; Calculate the similarity between each actual sampling test data and each reference sampling test data one by one, generate sampling test similarity clusters, and construct a Laplacian matrix of multi-dimensional comparison mapping of sampling test data based on the sampling test similarity clusters; Acquire a sampling detection index according to target requirements, calculate and obtain the eigenvalue of the dimension corresponding to the sampling detection index in the Laplace matrix and the eigenvector of each eigenvalue based on the sampling detection index, and if the eigenvalue is greater than a preset eigenvalue, extract the eigenvector corresponding to the eigenvalue and mark it as a quasi-mapping eigenvector; stacking one or more quasi-mapping feature vectors in the unqualified fermentation feature mapping space to generate a multidimensional sampling detection index map, and obtaining an area value of the multidimensional sampling detection index map; If the area value is less than the preset area value, the mulberry leaf feed fermentation sample is calibrated as a normal sample; if the area value is less than the preset area value, the mulberry leaf feed fermentation sample is calibrated as an abnormal sample, and the test result is obtained.

[0012] The second aspect of the present invention provides a regulation and control system for an intelligent mulberry leaf feed fermentation device, the regulation and control system comprising a memory and a processor, the memory storing a regulation and control method program for an intelligent mulberry leaf feed fermentation device, and when the regulation and control method program is executed by the processor, any one of the regulation and control method steps is implemented.

[0013] The present invention solves the technical defects existing in the background technology, and the beneficial technical effects of the present invention are: The fermentation inhibition degree of auxiliary fermentation materials for mulberry leaf fermentation raw materials is obtained, and the supplementary regulation and control of the initial oxygen concentration content based on the idealized iteration of the fermentation inhibition degree is obtained to obtain the first regulation and control scheme; the time-series oxygen content distribution data of mulberry leaf feed fermentation is collected by an oxygen detection device, and the variational distribution deviation of the oxygen content under the stirring air power is calculated according to the time-series oxygen content distribution data to adjust and control the stirring rate to obtain the second regulation and control scheme; the real-time pH acid-base signal of mulberry leaf feed fermentation is obtained by a pH detection device, and the pH acid-base gradient dislocation fragment calculation of the real-time pH acid-base signal is normalized to control the bacterial liquid spraying to obtain the third regulation and control scheme; sampling detection obtains the mulberry leaf feed fermentation sample after regulation and control, and the multidimensional sampling detection data is mapped and detected and analyzed based on the unqualified fermentation image data to obtain the detection result. The present invention can regulate and control the aerobic supplementation, stirring rate, pH acid-base gradient change and whether the final fermentation product is normal for the intelligent mulberry leaf feed, accurately adjust the important parameters in the fermentation process, provide the best growth environment for the microorganisms of mulberry leaf fermentation, promote the rapid progress of the mulberry leaf fermentation process, and greatly improve the fermentation efficiency and fermentation stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0015] Figure 1 A first method flow chart of a regulation and control method of an intelligent mulberry leaf feed fermentation device is shown; Figure 2 A second method flow chart of a regulation and control method of an intelligent mulberry leaf feed fermentation device is shown; Figure 3 The system framework diagram of a regulating and controlling system of an intelligent mulberry leaf feed fermentation device is shown. DETAILED DESCRIPTION

[0016] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0017] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0018] The first aspect of the present invention provides a method for regulating and controlling an intelligent mulberry leaf feed fermentation device. Figure 1 As shown, the following steps are included: S102: Obtaining the fermentation inhibition degree of the auxiliary fermentation material on the mulberry leaf fermentation raw material, and ideally iterating the supplementary regulation and control of the initial oxygen content based on the fermentation inhibition degree to obtain a first regulation and control scheme; S104: collecting the time-series oxygen concentration distribution data of the mulberry leaf feed fermentation through the oxygen detection device, and calculating the variational distribution deviation of the oxygen concentration under the stirring air power according to the time-series oxygen concentration distribution data to adjust and control the stirring rate, and obtain a second adjustment control scheme; S106: obtaining a real-time pH acid-base signal of mulberry leaf feed fermentation through a pH detection device, performing normalized pH acid-base gradient dislocation fragment calculation on the real-time pH acid-base signal to control bacterial liquid spraying, and obtaining a third regulation control scheme; S108: Sampling and testing to obtain the fermented mulberry leaf feed sample after adjustment and control, and mapping the indicators of the multi-dimensional sampling test data and performing test analysis based on the unqualified fermentation image data to obtain the test results.

[0019] More specifically, the step S102 includes the following steps: Obtaining the target demand of the mulberry leaf feed fermentation device, obtaining the preset oxygen control strategy of the intelligent mulberry leaf feed fermentation device for the target demand, and extracting the initial oxygen content input by the intelligent mulberry leaf feed fermentation device through the preset oxygen control strategy; The ideal oxygen content is preset according to the target demand. If the initial oxygen content is lower than the ideal oxygen content, the historical fermentation data of the mulberry leaf fermentation raw material and each auxiliary fermentation material are obtained according to the target demand, and the fermentation inhibition degree of each auxiliary fermentation material on the mulberry leaf fermentation raw material is calculated based on the weighted and matched distribution of the historical fermentation data; A penalty factor constraint function is preset based on the degree of fermentation inhibition, and the Lagrangian algorithm is introduced. The initial oxygen concentration is calculated in the Lagrangian algorithm based on the preset oxygen control strategy to obtain the Lagrangian functions of different oxygen control solutions. The penalty factor constraint function and the ideal oxygen content are calculated by minimizing the weighted calculation using a trade-off algorithm to obtain an iterative central path, and the iterative central path is tracked by executing different oxygen control strategies through a Lagrangian function. During the tracking process, the preset oxygen control strategy is continuously updated and iterated to approach the optimal solution for the intelligent mulberry leaf feed fermentation device to achieve the ideal oxygen content, and the current iteration step is output; According to the ideal oxygen content, an iteration step threshold is preset. If the current iteration step does not exceed the iteration step threshold, the tracking iteration continues. If the current iteration step exceeds the iteration step threshold, the tracking iteration operation is stopped, and finally a series of optimal oxygen control solutions are generated. A series of optimal oxygen control solutions are deployed on the control terminal of the intelligent mulberry leaf feed fermentation device to perform oxygen supplementation adjustment and control on the initial oxygen content to obtain the first adjustment and control scheme.

[0020] It should be noted that the fermentation process of some mulberry leaf feeds needs to be carried out under aerobic conditions. The appropriate oxygen content can promote the growth of beneficial microorganisms such as lactic acid bacteria and yeasts, avoid nutritional loss and feed mildew, thereby improving the fermentation quality of mulberry leaf feeds. However, if excessive oxygen is supplemented, the original environment of mulberry leaf feed fermentation inside the device may be destroyed, which will not only inhibit the stable growth and proliferation of beneficial microorganisms, but also cause the loss of nutrients, which is not conducive to the fermentation and production of high-quality mulberry leaf feeds. Therefore, it is particularly important to accurately control the oxygen supplementation of the intelligent mulberry leaf feed fermentation device. Therefore, this method uses the degree of fermentation inhibition of each auxiliary fermentation material on the mulberry leaf fermentation raw material as the oxygen demand regulation constraint, and under the premise of this oxygen demand regulation constraint, the initial oxygen content of the device input is optimized to the ideal oxygen content. It is iterative, thereby achieving oxygen supplementation suitable for different mulberry leaf feed fermentation raw materials and environmental conditions, and finally tending to the idealized fermentation oxygen supplementation effect, replacing the traditional large error step of supplementing oxygen based on artificial experience, so that the oxygen content regulation and control inside the mulberry leaf feed fermentation device is more accurate and efficient, reducing the supplementation error phenomenon, and improving the reproduction and growth quality of beneficial microorganisms and the nutrient locking performance.

[0021] It should be noted that for the constrained iteration of the initial oxygen content, this method first sets a penalty factor constraint function based on the degree of fermentation inhibition of each auxiliary fermentation material on the mulberry leaf fermentation raw material. The penalty factor constraint function can avoid the search process from crossing the obstacle boundary item of the fermentation inhibition consideration, so that the optimal oxygen supplement is more inclined to search within the feasible domain of the inhibitory influence between the raw materials, so that the final supplemented oxygen content is more in line with the mulberry leaf feed fermentation scenarios of different raw materials and environments. Then, the Lagrangian functions of different oxygen control solutions required for the initial oxygen concentration are calculated, and the ideal oxygen content solution is gradually approached by optimizing oxygen supplementation on these Lagrangian functions. Then, the weighted calculation penalty factor constraint function and the ideal oxygen concentration are minimized to generate an iterative central path, which defines the iterative direction of the regulation and control system for the initial oxygen content toward the optimal solution of the ideal oxygen content. The path is within the feasible domain of the inhibitory influence between the raw materials, and can avoid touching the boundary of the fermentation inhibition consideration to the greatest extent, thereby improving the accuracy and reliability of each oxygen content supplementation solution provided to the regulation and control system. Then, the iterative central path is tracked by the Lagrangian function to perform different oxygen control strategies. In addition, the preset oxygen control strategy is continuously updated and iterated during the tracking process to approximate the optimal solution of the intelligent mulberry leaf feed fermentation device to achieve the ideal oxygen concentration content, and the current iteration step size is output. The iteration step size determines the update amount of the optimal solution approaching the ideal oxygen content in each iteration, which can ensure that the oxygen content supplement satisfies the real-time fermentation inhibition degree and environmental changes to the greatest extent, so that the oxygen supplementation in the fermentation process is more real-time optimal and credible.

[0022] More specifically, the historical fermentation data of the mulberry leaf fermentation raw material and each auxiliary fermentation material are obtained according to the target demand, and the fermentation inhibition degree of each auxiliary fermentation material on the mulberry leaf fermentation raw material is calculated based on the weighted and matched distribution of the historical fermentation data, such as Figure 2 As shown, the specific steps include: S202: obtaining mulberry leaf fermentation raw materials and several auxiliary materials according to target requirements, and obtaining actual dosages of the mulberry leaf fermentation raw materials and each auxiliary fermentation material; S204: extracting historical fermentation data of the mulberry leaf fermentation raw material when the actual dosage is output within a preset time period through the fermentation log, which is defined as dependent historical fermentation data, and extracting historical fermentation data of each auxiliary fermentation material when the actual dosage is output within a preset time period, which is defined as covariant historical fermentation data; S206: Introducing a logistic regression model to perform regression estimation on each dependent historical fermentation data and each covariant historical fermentation data, obtaining a number of logistic regression functions, and determining the propensity score of each covariant historical fermentation data causing each dependent historical fermentation data to accept treatment according to the logistic regression function; S208: obtaining the current covariate distribution when the covariate historical fermentation data causes the dependent historical fermentation data to be processed based on the propensity score, and bundling each covariate historical fermentation data with the corresponding dependent historical fermentation data causing the processing to be processed into a weighted control group, to obtain a plurality of weighted control groups; S210: introducing a trade-off algorithm to weight each weighted control group, generating a weighted control group mean, performing individual matching of each covariate historical fermentation data with the dependent historical fermentation data based on the weighted control group mean, and generating a covariate matching distribution; S212: registering the covariate matching distribution with the current covariate distribution to obtain a registration balance difference, and determining the fermentation inhibition degree of each auxiliary fermentation material on the mulberry leaf fermentation raw material according to the registration balance difference.

[0023] It should be noted that the raw material for mulberry leaf fermentation is mulberry leaf, and the auxiliary fermentation materials include corn and soybean meal, etc. Some auxiliary fermentation materials contain certain nutrients such as protein, sugar and fat. The addition of these auxiliary fermentation materials may make the concentration of certain specific nutrients in mulberry leaf too high or too low, which will affect the fermentation process of mulberry leaf itself. For example, too high protein concentration may cause excessive growth of certain microorganisms and inhibit the activity of other microorganisms. And different auxiliary fermentation materials may carry different microbial communities, which may cause the microbial community of the auxiliary fermentation material to compete with the microbial community of the mulberry leaf fermentation raw material. Some microorganisms may have an inhibitory effect on key enzymes or metabolites in mulberry leaf fermentation, thereby affecting the fermentation quality of mulberry leaf feed. Therefore, when adjusting and controlling various parameters in the fermentation process, the degree of inhibition of the fermentation of mulberry leaf by the auxiliary fermentation material should be fully considered. In this regard, the method first obtains the historical fermentation data of the mulberry leaf fermentation raw material and the auxiliary fermentation material respectively. Since there may be more than one auxiliary fermentation material, it is regarded as a covariate, and the historical fermentation data of the mulberry leaf fermentation raw material is inhibited by the auxiliary fermentation material, so it is regarded as a dependent variable. Then, the logistic regression function of the two is estimated through logistic regression model regression, which reflects the tendency probability of a mulberry leaf fermentation raw material receiving the inhibition treatment of the other auxiliary fermentation materials, that is, the propensity score. The propensity score can integrate the covariate effects of different auxiliary fermentation materials into a single indicator, provide a basis for subsequent matching, and reduce the selection bias and confounding bias of fermentation inhibition. Then each weighted control group is weighed and weighted, and the generated weighted control group mean can match each covariate historical fermentation data with the corresponding dependent historical fermentation data that causes the treatment, so that the distribution of covariates that cause the fermentation inhibition of auxiliary fermentation materials should be balanced, simulating the distribution effect of random fermentation inhibition in randomized controlled trials, and improving the reliability of causal inference. The registration balance difference between the covariate matching distribution and the current covariate distribution is a causal measurement value of the fermentation inhibition of mulberry leaf fermentation raw materials caused by the auxiliary fermentation materials, and finally the regulation control system can quickly determine the degree of fermentation inhibition.

[0024] It should be noted that, in summary, this method can quickly calculate the fermentation inhibition tendency of mulberry leaves by the auxiliary fermentation materials put into the device in real time based on the input historical fermentation data, thereby providing a reliable constraint basis for the oxygen supplementation of subsequent devices, effectively improving the accuracy of oxygen supplementation, and achieving the effect of aerobic rationality of fermentation under different materials and environments.

[0025] More specifically, the step S104 includes the following steps: The oxygen detection device is used to detect oxygen in the intelligent mulberry leaf feed fermentation device after the first adjustment and control scheme is executed, so as to obtain the time-series oxygen concentration distribution data of the mulberry leaf feed fermentation, and at the same time obtain the predetermined detection area range of the oxygen detection device; Constructing a bubble space domain of an intelligent mulberry leaf feed fermentation device, fitting the time-series oxygen concentration distribution data into the bubble space domain one by one, and after the fitting is completed, dividing the bubble space domain into M sub-bubble domains based on a predetermined detection area range, generating M oxygen content bubble domains, and obtaining the bubble volume change value of each oxygen content bubble domain as the time series transitions; If the bubble volume change value is greater than the preset bubble volume change value, the oxygen content bubble field is calibrated as a type I bubble field in the bubble space field; if it is less than the preset bubble volume change value, the oxygen content bubble field is calibrated as a type II bubble field. The change layout of the type I bubble field and the type II bubble field is visualized to obtain the time series change pattern of oxygen concentration; Obtain the time-series stirring rate corresponding to the time-series oxygen concentration distribution data generated by the intelligent mulberry leaf feed fermentation device, obtain the aerodynamic knowledge graph based on the big data network, identify the time-series stirring rate through the aerodynamic knowledge graph, and output the reference aerodynamic vector coefficient generated inside the device during the stirring process; According to the time series variation pattern of oxygen concentration, a plurality of random oxygen variation distribution samples are constructed. Based on the reference aerodynamic vector coefficient, a variational extrapolation of the oxygen concentration variational distribution deviation is performed on each random oxygen variation distribution sample. According to the oxygen concentration variational distribution deviation, the stirring rate is adjusted and controlled to obtain a second adjustment control scheme.

[0026] It should be noted that the stirring rate plays a key role in the fermentation process of mulberry leaf feed. When the stirring rate is improper (too fast or too slow), it is difficult to adjust the ideal aerobic or ideal anaerobic environment inside the fermentation of mulberry leaf feed, which can easily lead to uneven oxygen distribution in the mulberry leaf pile, and there may be more oxygen residue, causing hypoxia or excess oxygen growth of microorganisms. At the same time, it is easy to generate too much heat, resulting in excessive local temperature, inhibiting the growth of fermentation microorganisms and affecting the fermentation effect. It can be seen that the precise regulation and control of the stirring rate of the device is crucial. In this regard, the method fits the time-series oxygen concentration distribution data of mulberry leaf feed fermentation through a bubble graph to form a time-series oxygen concentration change display of bubble volume change, that is, the oxygen concentration time-series change pattern, and then analyzes and calculates the random oxygen distribution of the oxygen concentration time-series change pattern based on the aerodynamic force formed inside the device during the stirring process, so that the regulation control system can accurately adjust the stirring rate of mulberry leaf feed fermentation according to the random distribution trend of oxygen inside the device and follow aerodynamics, thereby further making the random distribution of oxygen in the mulberry leaf pile during the fermentation process more reasonable and stable, and can more reliably provide an ideal aerobic or anaerobic growth environment for microorganisms, effectively improving the fermentation quality of mulberry leaf feed.

[0027] It should be noted that for the display of the time-series oxygen concentration change, this method uses the form of a bubble chart to efficiently fit the detected time-series oxygen concentration data, and the bubble chart has the special chart that changes its own volume according to the size of the fitted data to display the data, making the real-time display of the time-series data faster and more reliable. Compared with the traditional real-time data display method, it can save unnecessary calculation steps and improve the efficiency of data display. And based on the established detection area range of the oxygen detection device, the oxygen distribution area inside the device is divided, that is, the sub-bubble area, so that the unreasonable oxygen concentration display is more detailed, and the fermentation area with unreasonable oxygen concentration can be clearly identified at a glance. If the bubble volume change value is greater than the preset bubble volume change value, it means that the oxygen concentration change amplitude and frequency of the fermentation area are high, which is obviously caused by unreasonable oxygen distribution; if it is less than the preset bubble volume change value, it means that the oxygen concentration change amplitude and frequency of the fermentation area are relatively stable, and usually present a reasonable distribution. Since the time-series oxygen concentration distribution data is acquired in real time, the time-series change pattern of oxygen concentration presents a time-series state that continuously and real-time alternately displays the changes in bubbles, which enables the regulating and control system to have certain real-time benefits in displaying the changes in oxygen distribution inside the device, and effectively improves the stirring accuracy and control precision of the mulberry leaf feed fermentation device.

[0028] More specifically, the method constructs a plurality of random oxygen change distribution samples according to the time series change pattern of oxygen concentration, performs variational calculation of oxygen concentration variation distribution deviation for each random oxygen change distribution sample based on the reference aerodynamic vector coefficient, and adjusts and controls the stirring rate according to the oxygen concentration variation distribution deviation to obtain the second adjustment control scheme, which specifically includes the following steps: Constructing a number of random oxygen change distribution samples resulting from the time-series change pattern of oxygen concentration caused by the time-series stirring rate, and calculating the probability mass function of each random oxygen change distribution sample based on the reference aerodynamic vector coefficient; The Bayesian variational inference network is introduced to estimate the marginal probability lower bound of each probability mass function. By calculating the gradient of the marginal probability lower bound for each random oxygen change distribution sample, the distribution state of the corresponding random oxygen change distribution sample is updated based on the gradient to obtain the actual oxygen concentration variation distribution. An ideal oxygen concentration variational distribution that eliminates the time-series variation pattern of oxygen content is obtained, the distribution difference between the actual oxygen concentration variational distribution and the ideal oxygen concentration variational distribution is calculated, and the oxygen concentration variational distribution deviation is obtained. Based on the oxygen concentration variational distribution deviation, the stirring rate of the intelligent mulberry leaf feed fermentation device is adjusted and controlled to obtain a second adjustment and control scheme.

[0029] It should be noted that since the stirring rate forms a certain aerodynamic force inside the device, it will drive the random movement distribution of oxygen. Therefore, this method uses the reference aerodynamic vector coefficient as a premise to calculate the probability mass function of the time-series stirring rate causing the time-series change pattern of oxygen concentration. This probability mass function reflects the distribution probability of the stirring rate causing the oxygen concentration to move, which is the premise for inferring the random distribution of oxygen concentration. Then, the Bayesian variational inference network is introduced to estimate the marginal probability lower bound of each probability mass function. Through the marginal probability lower bound, the random movement analysis of oxygen concentration of the time-series oxygen concentration distribution data can be made closer to the true posterior distribution of aerodynamic disturbances, so that the actual distribution trend of the random movement of the actual oxygen concentration driven by the stirring inside the device can be calculated, that is, the actual oxygen concentration variational distribution, which provides a reliable control basis for the subsequent precise adjustment of the stirring rate and improves the accuracy of the stirring rate of the device.

[0030] More specifically, the step S106 includes the following steps: The pH detection device is used to detect the pH of the intelligent mulberry leaf feed fermentation device after the second adjustment control scheme is executed, and the real-time pH acid-base signal of the mulberry leaf feed fermentation in the target time period is obtained, and the wavelet transform algorithm is introduced to extract the features of the real-time pH acid-base signal to obtain a number of real-time pH acid-base frequency values; Obtaining a reference pH acid-base interval for mulberry leaf feed fermentation, setting an abnormal pH acid-base frequency value range according to the reference pH acid-base interval, extracting only the real-time pH acid-base frequency value within the abnormal pH acid-base frequency value range and marking it as an abnormal pH acid-base frequency value, and obtaining a plurality of abnormal pH acid-base frequency values; Obtaining a preset detection interval of a pH detection device within a target time period, establishing a wavelet basis variation coefficient describing the pH time series variation based on the preset detection interval, and performing a time series gradient estimation on all abnormal pH acid-base frequency values ​​based on the wavelet basis variation coefficient to construct an abnormal pH acid-base gradient diagram; According to the target requirements, the normal pH frequency range of mulberry leaf feed fermentation is obtained, and a normal pH acid-base gradient diagram is established based on the normal pH frequency range. The concept dislocation algorithm is introduced to calculate the gradient dislocation between the abnormal pH acid-base gradient diagram and the normal pH acid-base gradient diagram, and the pH acid-base gradient dislocation fragment is obtained. The pH acid-base gap of the abnormal pH acid-base gradient diagram compared with the normal pH acid-base gradient diagram is determined according to the pH acid-base gradient dislocation fragment, and the bacterial liquid spraying of the intelligent mulberry leaf feed fermentation device is controlled based on the pH acid-base gradient gap to obtain a third regulation and control scheme.

[0031] It should be noted that, generally speaking, as the fermentation of mulberry leaf feed continues, it will develop towards the initial, middle and late fermentation stages. In this process, the suitable pH value for the growth of lactic acid bacteria is between 5.5-6.5. Therefore, the pH value in this process should be acidic and decrease in a certain acid-base gradient. If the pH value is too high (too alkaline), it will destroy the enzyme activity in the microbial cells, affect its material metabolism and energy conversion process, and cause the microorganisms to grow slowly or even stop growing. For example, when the pH value reaches above 8, the growth of lactic acid bacteria will be significantly inhibited, and the fermentation speed will slow down, thereby extending the fermentation cycle of mulberry leaf feed. Therefore, it is necessary to accurately detect abnormal gradient changes in pH to ensure that the microorganisms are always in a suitable acid-base environment during the fermentation process. To this end, the method converts the real-time pH acid-base signal obtained by the pH detection device into a series of real-time pH acid-base frequency values. When there is an abnormal pH acid-base gradient in the device, some of these real-time pH acid-base frequency values ​​are specific frequency values ​​of abnormal pH acid-base. Therefore, by extracting these abnormal real-time pH acid-base frequency values, the abnormal pH acid-base gradient inside the device can be determined, wherein the reference pH acid-base range is a standard range of the optimal pH acid-base value for relatively standardized mulberry leaf feed fermentation. The gradient misalignment between the abnormal pH acid-base gradient diagram and the normal pH acid-base gradient diagram is the stage gradient error of the internal pH acid-base of the mulberry leaf feed compared with the normal pH acid-base as the fermentation stage progresses. These gradient errors may be discretely present in different fermentation stages and are therefore in the form of fragments. Finally, based on these pH acid-base gradient misalignment fragments, the pH acid-base gap of the abnormal pH acid-base gradient diagram compared with the normal pH acid-base gradient diagram can be further determined, thereby achieving precise adjustment and control of the bacterial liquid spraying of the intelligent mulberry leaf feed fermentation device, so that the device can reasonably optimize and improve the internal pH environment of the mulberry leaf feed fermentation process, ensure that the microorganisms grow and reproduce in a suitable fermentation pH environment, and improve the fermentation efficiency and quality of the mulberry leaf feed.

[0032] More specifically, the step S108 includes the following steps: By sampling and extracting mulberry leaf feed fermentation samples after executing the first adjustment control scheme, the second adjustment control scheme and the third adjustment control scheme, and using biological technology to detect the mulberry leaf feed fermentation samples, a number of multi-dimensional sampling detection data are obtained, which are defined as actual sampling detection data; Based on the big data network, unqualified fermentation image data of mulberry leaf feed and several multi-dimensional sampling detection data when unqualified fermentation occurs are obtained, which are defined as reference sampling detection data. The LBP algorithm is introduced to extract features of the unqualified fermentation image data, and the LBP value of the unqualified fermentation feature is generated. Based on the LBP value, an unqualified fermentation feature mapping space of the reference sampling detection data is constructed; Calculate the similarity between each actual sampling test data and each reference sampling test data one by one, generate sampling test similarity clusters, and construct a Laplacian matrix of multi-dimensional comparison mapping of sampling test data based on the sampling test similarity clusters; Acquire a sampling detection index according to target requirements, calculate and obtain the eigenvalue of the dimension corresponding to the sampling detection index in the Laplace matrix and the eigenvector of each eigenvalue based on the sampling detection index, and if the eigenvalue is greater than a preset eigenvalue, extract the eigenvector corresponding to the eigenvalue and mark it as a quasi-mapping eigenvector; stacking one or more quasi-mapping feature vectors in the unqualified fermentation feature mapping space to generate a multidimensional sampling detection index map, and obtaining an area value of the multidimensional sampling detection index map; If the area value is less than the preset area value, the mulberry leaf feed fermentation sample is calibrated as a normal sample; if the area value is less than the preset area value, the mulberry leaf feed fermentation sample is calibrated as an abnormal sample, and the test result is obtained.

[0033] It should be noted that traditional detection of fermented mulberry leaf feed samples usually uses simple image feature comparison. If abnormal features are identified, it is judged as unqualified fermented mulberry leaf feed. However, this detection method has certain drawbacks and is prone to large comparison errors, resulting in unreliable detection results. And with the in-depth study of mulberry leaf feed fermentation, the indicators that need to be detected are increasing, such as sampling and detection data of different dimensions such as physical and chemical indicators, microbial indicators, toxin indicators, nutrient component indicators, and animal feeding test indicators after fermentation. It is difficult for traditional detection methods to comprehensively display the detection quality of these data. And because these indicators will cause certain unqualified expressions in the color, morphology and properties of the fermented samples to a certain extent, for example, excessive mold content may cause mildew on the surface of the feed. In view of this, this method first uses the unqualified fermentation image data of mulberry leaf feed as the mapping expression base of the detection data, so as to maximize the comprehensive display of the degree to which the detection quality tends to the unqualified fermentation image. Since it is a comprehensive mapping display of multi-dimensional sampling test data, it is necessary to establish a mapping space that can reveal the unreasonable degree of detection quality to express these sampling test data. Therefore, this method uses the LBP value of the unqualified fermentation feature to spatially construct several multi-dimensional sampling test data when unqualified fermentation occurs. The purpose of using several multi-dimensional sampling test data when unqualified fermentation occurs is to make the tendency of unqualified fermentation image features in the detection quality mapped by the actual sampling test data more clearly expressed, which can further improve the accurate comprehensive display of sampling test data in different dimensions.

[0034] It should be noted that the similarity between each actual sampling test data and each reference sampling test data is calculated to make the unqualified fermentation image trend mapping of the actual sampling test data more accurate, and the eigenvalues ​​and eigenvectors of the Laplace matrix are calculated to extract the low-dimensional structural information of the actual sampling test data. The quasi-mapping eigenvector extracted based on the sampling test index can display the local structure of the sampling test index data obtained in different dimensions, thereby ensuring the globality of the comprehensive expression of the actual test quality, avoiding the missing of indicator information in dimensions such as microbial indicators, toxin indicators or nutrient components indicators, and finally the mapping of all sampling test index data will form a multidimensional sampling test index map on the regulating and control system to reflect the actual fermentation quality of mulberry leaf feed after multi-dimensional sampling and detection; if the area value of the multidimensional sampling test index map is less than the preset area value, it means that the actual fermentation quality has not fluctuated significantly and is good overall, so the mulberry leaf feed fermentation sample is calibrated as a normal sample; otherwise, it means that the actual fermentation quality is poor, and the test results do not meet the actual application of mulberry leaf feed, so the mulberry leaf feed fermentation sample is calibrated as an abnormal sample. This method can accurately and reliably perform multi-dimensional detection on mulberry leaf feed samples fermented after adjustment and control by the device, solving the problems of large detection errors and inability to comprehensively display sampling detection indicators in different dimensions caused by traditional methods, thereby improving the detection accuracy of mulberry leaf feed quality.

[0035] The second aspect of the present invention provides a regulation and control system for an intelligent mulberry leaf feed fermentation device, such as Figure 3 As shown, the regulating and controlling system includes a memory 31 and a processor 32. The memory 31 stores a regulating and controlling method program of an intelligent mulberry leaf feed fermentation device. When the regulating and controlling method program is executed by the processor 32, any one of the regulating and controlling method steps is implemented.

[0036] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A regulation and control method for an intelligent mulberry leaf feed fermentation device, characterized in that: The following steps are involved: S102: Obtaining the fermentation inhibition degree of the auxiliary fermentation material on the mulberry leaf fermentation raw material, and ideally iterating the supplementary regulation and control of the initial oxygen content based on the fermentation inhibition degree to obtain a first regulation and control scheme; S104: collecting the time-series oxygen concentration distribution data of the mulberry leaf feed fermentation through the oxygen detection device, and calculating the variational distribution deviation of the oxygen concentration under the stirring air power according to the time-series oxygen concentration distribution data to adjust and control the stirring rate, and obtain a second adjustment control scheme; S106: obtaining a real-time pH acid-base signal of mulberry leaf feed fermentation through a pH detection device, performing normalized pH acid-base gradient dislocation fragment calculation on the real-time pH acid-base signal to control bacterial liquid spraying, and obtaining a third regulation control scheme; S108: Sampling and testing to obtain the fermented mulberry leaf feed sample after adjustment and control, and mapping the indicators of the multi-dimensional sampling test data and performing test analysis based on the unqualified fermentation image data to obtain the test results.

2. The regulating and controlling method of an intelligent mulberry leaf feed fermentation device according to claim 1 is characterized in that: The step S102 specifically includes the following steps: Obtaining the target demand of the mulberry leaf feed fermentation device, obtaining the preset oxygen control strategy of the intelligent mulberry leaf feed fermentation device for the target demand, and extracting the initial oxygen content input by the intelligent mulberry leaf feed fermentation device through the preset oxygen control strategy; The ideal oxygen content is preset according to the target demand. If the initial oxygen content is lower than the ideal oxygen content, the historical fermentation data of the mulberry leaf fermentation raw material and each auxiliary fermentation material are obtained according to the target demand, and the fermentation inhibition degree of each auxiliary fermentation material on the mulberry leaf fermentation raw material is calculated based on the weighted and matched distribution of the historical fermentation data; A penalty factor constraint function is preset based on the degree of fermentation inhibition, and the Lagrangian algorithm is introduced. The initial oxygen concentration is calculated in the Lagrangian algorithm based on the preset oxygen control strategy to obtain the Lagrangian functions of different oxygen control solutions. The penalty factor constraint function and the ideal oxygen content are calculated by minimizing the weighted calculation using a trade-off algorithm to obtain an iterative central path, and the iterative central path is tracked by executing different oxygen control strategies through a Lagrangian function. During the tracking process, the preset oxygen control strategy is continuously updated and iterated to approach the optimal solution for the intelligent mulberry leaf feed fermentation device to achieve the ideal oxygen content, and the current iteration step is output; According to the ideal oxygen content, an iteration step threshold is preset. If the current iteration step does not exceed the iteration step threshold, the tracking iteration continues. If the current iteration step exceeds the iteration step threshold, the tracking iteration operation is stopped, and finally a series of optimal oxygen control solutions are generated. A series of optimal oxygen control solutions are deployed on the control terminal of the intelligent mulberry leaf feed fermentation device to perform oxygen supplementation adjustment and control on the initial oxygen content to obtain the first adjustment and control scheme.

3. The regulating and controlling method of an intelligent mulberry leaf feed fermentation device according to claim 2 is characterized in that: The method of obtaining the historical fermentation data of the mulberry leaf fermentation raw material and each auxiliary fermentation material according to the target demand, and calculating the fermentation inhibition degree of each auxiliary fermentation material on the mulberry leaf fermentation raw material based on the weighted and matched distribution of the historical fermentation data, specifically includes the following steps: Acquire mulberry leaf fermentation raw materials and several auxiliary materials according to target requirements, and simultaneously acquire the actual dosage of mulberry leaf fermentation raw materials and each auxiliary fermentation material; The historical fermentation data of the mulberry leaf fermentation raw material when the actual dosage is output within a preset time period is extracted through the fermentation log, which is defined as the dependent historical fermentation data, and the historical fermentation data of each auxiliary fermentation material when the actual dosage is output within a preset time period is extracted, which is defined as the covariant historical fermentation data; A logistic regression model is introduced to perform regression estimation on each dependent historical fermentation data and each covariant historical fermentation data to obtain several logistic regression functions, and the propensity score of each covariant historical fermentation data leading to the treatment of each dependent historical fermentation data is determined according to the logistic regression function; Based on the propensity score, the current covariate distribution when the covariate historical fermentation data causes the dependent historical fermentation data to be processed is obtained, and each covariate historical fermentation data is bundled with the corresponding dependent historical fermentation data that causes the processing to be received into a weighted control group to obtain a plurality of weighted control groups; A trade-off algorithm is introduced to weigh each weighted control group and generate the mean of the weighted control group. Based on the mean of the weighted control group, each covariate historical fermentation data is individually matched with the dependent historical fermentation data to generate a covariate matching distribution. The registration covariate matching distribution and the current covariate distribution are used to obtain the registration balance difference, and the fermentation inhibition degree of each auxiliary fermentation material on the mulberry leaf fermentation raw material is determined according to the registration balance difference.

4. The regulating and controlling method of an intelligent mulberry leaf feed fermentation device according to claim 1 is characterized in that: The step S104 specifically includes the following steps: The oxygen detection device is used to detect oxygen in the intelligent mulberry leaf feed fermentation device after the first adjustment and control scheme is executed, so as to obtain the time-series oxygen concentration distribution data of the mulberry leaf feed fermentation, and at the same time obtain the predetermined detection area range of the oxygen detection device; Constructing a bubble space domain of an intelligent mulberry leaf feed fermentation device, fitting the time-series oxygen concentration distribution data into the bubble space domain one by one, and after the fitting is completed, dividing the bubble space domain into M sub-bubble domains based on a predetermined detection area range, generating M oxygen content bubble domains, and obtaining the bubble volume change value of each oxygen content bubble domain as the time series transitions; If the bubble volume change value is greater than the preset bubble volume change value, the oxygen content bubble field is calibrated as a type I bubble field in the bubble space field; if it is less than the preset bubble volume change value, the oxygen content bubble field is calibrated as a type II bubble field. The change layout of the type I bubble field and the type II bubble field is visualized to obtain the time series change pattern of oxygen concentration; Obtain the time-series stirring rate corresponding to the time-series oxygen concentration distribution data generated by the intelligent mulberry leaf feed fermentation device, obtain the aerodynamic knowledge graph based on the big data network, identify the time-series stirring rate through the aerodynamic knowledge graph, and output the reference aerodynamic vector coefficient generated inside the device during the stirring process; According to the time series change pattern of oxygen concentration, multiple random oxygen change distribution samples are constructed, and the variational distribution deviation of oxygen concentration is variationally extrapolated for each random oxygen change distribution sample based on the reference aerodynamic vector coefficient. The stirring rate is adjusted and controlled according to the variational distribution deviation of oxygen concentration to obtain a second adjustment control scheme.

5. The regulating and controlling method of the intelligent mulberry leaf feed fermentation device according to claim 4 is characterized in that: The method constructs a plurality of random oxygen change distribution samples according to the time series change pattern of oxygen concentration, performs variational calculation of oxygen concentration variation distribution deviation for each random oxygen change distribution sample based on the reference aerodynamic vector coefficient, and adjusts and controls the stirring rate according to the oxygen concentration variation distribution deviation to obtain a second adjustment control scheme, which specifically includes the following steps: Constructing a number of random oxygen change distribution samples resulting from the time-series change pattern of oxygen concentration caused by the time-series stirring rate, and calculating the probability mass function of each random oxygen change distribution sample based on the reference aerodynamic vector coefficient; The Bayesian variational inference network is introduced to estimate the marginal probability lower bound of each probability mass function. By calculating the gradient of the marginal probability lower bound for each random oxygen change distribution sample, the distribution state of the corresponding random oxygen change distribution sample is updated based on the gradient to obtain the actual oxygen concentration variation distribution. An ideal oxygen concentration variational distribution that eliminates the time-series variation pattern of oxygen content is obtained, the distribution difference between the actual oxygen concentration variational distribution and the ideal oxygen concentration variational distribution is calculated, and the oxygen concentration variational distribution deviation is obtained. Based on the oxygen concentration variational distribution deviation, the stirring rate of the intelligent mulberry leaf feed fermentation device is adjusted and controlled to obtain a second adjustment and control scheme.

6. The regulating and controlling method of an intelligent mulberry leaf feed fermentation device according to claim 1 is characterized in that: The step S106 specifically includes the following steps: The pH detection device is used to detect the pH of the intelligent mulberry leaf feed fermentation device after the second adjustment control scheme is executed, and the real-time pH acid-base signal of the mulberry leaf feed fermentation in the target time period is obtained, and the wavelet transform algorithm is introduced to extract the features of the real-time pH acid-base signal to obtain a number of real-time pH acid-base frequency values; Obtaining a reference pH acid-base interval for mulberry leaf feed fermentation, setting an abnormal pH acid-base frequency value range according to the reference pH acid-base interval, extracting only the real-time pH acid-base frequency value within the abnormal pH acid-base frequency value range and marking it as an abnormal pH acid-base frequency value, and obtaining a plurality of abnormal pH acid-base frequency values; Obtaining a preset detection interval of a pH detection device within a target time period, establishing a wavelet basis variation coefficient describing the pH time series variation based on the preset detection interval, and performing a time series gradient estimation on all abnormal pH acid-base frequency values ​​based on the wavelet basis variation coefficient to construct an abnormal pH acid-base gradient diagram; According to the target requirements, the normal pH frequency range of mulberry leaf feed fermentation is obtained, and a normal pH acid-base gradient diagram is established based on the normal pH frequency range. The concept dislocation algorithm is introduced to calculate the gradient dislocation between the abnormal pH acid-base gradient diagram and the normal pH acid-base gradient diagram, and the pH acid-base gradient dislocation fragment is obtained. The pH acid-base gap of the abnormal pH acid-base gradient diagram compared with the normal pH acid-base gradient diagram is determined according to the pH acid-base gradient dislocation fragment, and the bacterial liquid spraying of the intelligent mulberry leaf feed fermentation device is controlled based on the pH acid-base gradient gap to obtain a third regulation and control scheme.

7. The regulating and controlling method of an intelligent mulberry leaf feed fermentation device according to claim 1 is characterized in that: The step S108 specifically includes the following steps: By sampling and extracting mulberry leaf feed fermentation samples after executing the first adjustment control scheme, the second adjustment control scheme and the third adjustment control scheme, and using biological technology to detect the mulberry leaf feed fermentation samples, a number of multi-dimensional sampling detection data are obtained, which are defined as actual sampling detection data; Based on the big data network, unqualified fermentation image data of mulberry leaf feed and several multi-dimensional sampling detection data when unqualified fermentation occurs are obtained, which are defined as reference sampling detection data. The LBP algorithm is introduced to extract features of the unqualified fermentation image data, and the LBP value of the unqualified fermentation feature is generated. Based on the LBP value, an unqualified fermentation feature mapping space of the reference sampling detection data is constructed; Calculate the similarity between each actual sampling test data and each reference sampling test data one by one, generate sampling test similarity clusters, and construct a Laplacian matrix of multi-dimensional comparison mapping of sampling test data based on the sampling test similarity clusters; Acquire a sampling detection index according to target requirements, calculate and obtain the eigenvalue of the dimension corresponding to the sampling detection index in the Laplace matrix and the eigenvector of each eigenvalue based on the sampling detection index, and if the eigenvalue is greater than a preset eigenvalue, extract the eigenvector corresponding to the eigenvalue and mark it as a quasi-mapping eigenvector; stacking one or more quasi-mapping feature vectors in the unqualified fermentation feature mapping space to generate a multidimensional sampling detection index map, and obtaining an area value of the multidimensional sampling detection index map; If the area value is less than the preset area value, the mulberry leaf feed fermentation sample is calibrated as a normal sample; if the area value is less than the preset area value, the mulberry leaf feed fermentation sample is calibrated as an abnormal sample, and the test result is obtained.

8. A regulation and control system for an intelligent mulberry leaf feed fermentation device, characterized in that: The regulating and controlling system includes a memory and a processor. The memory stores a regulating and controlling method program for an intelligent mulberry leaf feed fermentation device. When the regulating and controlling method program is executed by the processor, the regulating and controlling method steps as described in any one of claims 1 to 7 are implemented.

Images