Fermentation tank fermentation process control service system and method based on cloud computing

Abstract

The present invention belongs to the technical field of fermentation, and discloses a fermentation process control service system and method for a fermenter based on cloud computing; the system comprises: a device end collects fermentation parameters in real time during the fermentation process, and sends the fermentation parameters to a cloud computing platform; the fermentation parameters include fermentation characteristic parameters and fermentation control parameters; the cloud computing platform analyzes the fermentation characteristic parameters to obtain optimal fermentation control parameters; the cloud computing platform identifies deviation parameters in the fermentation control parameters collected in real time; the cloud computing platform determines whether to generate an early warning instruction or an adjustment instruction according to the deviation parameters; if an adjustment instruction is generated, the cloud computing platform calculates the adjustment amount corresponding to each deviation parameter, and sends the adjustment amount to the corresponding device end; the present invention realizes intelligent monitoring and control in the fermentation process through the cloud computing platform, can effectively control the fermentation quality and output, improve the fermentation efficiency, and thus improve the automation level and stability of the fermentation process.

Description

Technical Field

[0001] The present invention relates to the field of fermentation technology, and more specifically, to a fermentation tank fermentation process management and control service system and method based on cloud computing. Background Art

[0002] In modern biotechnology and pharmaceutical industries, the fermentation process is a key link in the production of important biological products (such as drugs, enzymes and biofuels). Its efficiency and quality directly affect the economic and market competitiveness of the products. Traditional fermentation tank management relies more on manual monitoring and experience accumulation, and faces problems such as insufficient real-time data, slow response speed, and low process control accuracy. These deficiencies not only lead to reduced production efficiency, but may also cause unstable product quality.

[0003] Of course, there are also intelligent fermentation tank management methods. For example, the Chinese patent with publication number CN116463210A discloses a method and control device for automatically controlling the temperature in a fermentation tank; including: a parameter acquisition module, a PID control module, a first plate heat exchanger, a second plate heat exchanger, and a third plate heat exchanger. In a stable fermentation state, a low-power plate heat exchanger controls the rise and fall of the temperature. When encountering abnormal conditions such as extreme heating, a high-power plate heat exchanger is used to quickly cool down, which can control the temperature in the fermentation tank more flexibly and accurately, saving the cost of use; in addition, the PID controller that controls the high-power plate heat exchanger can set multi-level parameters, and can automatically switch parameters to adjust the control mode for different tank environment parameters, further achieving the effect of precise temperature control; in addition, this invention does not require manual intervention, so while reducing manual errors, it can also improve the stability of the production process and increase the final output.

[0004] However, the above technologies mainly adopt the cascade PID control method, and the PID parameters are set using empirical parameters. It is not possible to perform adaptive optimization according to the fermentation characteristics of different stages, and the flexibility is insufficient. Moreover, the PID controller is usually based on the linear system assumption and is suitable for the control of linear relationships. It has poor control effect when dealing with complex and highly nonlinear fermentation processes. In addition, the above technologies can only meet the temperature control requirements. However, the fermentation process usually needs to consider multiple objectives at the same time (such as temperature, pH, stirring speed, etc.). It is difficult to coordinate multi-objective requirements using PID control.

[0005] In view of this, the present invention proposes a fermentation process control service system and method for a fermenter based on cloud computing to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned purpose, the present invention provides the following technical solution: a fermentation process control service method of a fermenter based on cloud computing, applied in a cloud computing platform, comprising:

[0007] The fermentation parameters collected in real time by the receiving device during the fermentation process; the fermentation parameters include fermentation characteristic parameters and fermentation control parameters;

[0008] Analyze the fermentation characteristic parameters to obtain the optimal fermentation control parameters;

[0009] comparing the optimal fermentation control parameters with the fermentation control parameters collected in real time, and identifying deviation parameters in the fermentation control parameters collected in real time;

[0010] According to the deviation parameters, generate warning instructions or adjustment instructions;

[0011] If an adjustment instruction is generated, the adjustment amount corresponding to each deviation parameter is calculated and sent to the corresponding device end; if an early warning instruction is generated, the early warning instruction is sent to the corresponding device end.

[0012] Furthermore, the step of obtaining the optimal fermentation control parameters includes:

[0013] Step 1: Construct n parameter sets, set different numerical labels for each parameter set, and mark them as parameter labels, where n is an integer greater than 1;

[0014] Step 2: Preset the population size Z and the number threshold T;

[0015] Step 3: Initialize the population. The position of each individual in the initialization population is defined in a one-dimensional search space. The position of the individual corresponds to the parameter label one by one. The number of iterations t of the initialization population is 0.

[0016] Step 4: Determine the fitness function;

[0017] Step 5: Calculate the candidate position of each individual in the next iteration and mark it as the candidate position;

[0018] Step 6: Determine the behavior performed by each individual and update the candidate position of each individual;

[0019] Step 7: Calculate the dynamic boundary, generate the reverse solution corresponding to the current position of each individual and add it to the population, using the reverse solution as the reverse position;

[0020] Step 8: Calculate the fitness of each individual's current position, updated candidate position, and reverse position, and determine whether to replace the position;

[0021] Step 9: If the number of iterations t is less than the number threshold T, then set the number of iterations , and return to step 5; if the number of iterations t is greater than or equal to the number threshold T, proceed to step 10;

[0022] Step 10: Calculate the fitness corresponding to the current position of each individual, sort all fitnesses from large to small, obtain the current position of the individual corresponding to the fitness at the front, obtain the parameter set corresponding to the corresponding parameter label according to the obtained individual current position, and use it as the optimal fermentation control parameter;

[0023] The steps for obtaining the optimal fermentation control parameters at each device are the same.

[0024] Furthermore, the fermentation characteristic parameters include substrate parameters and microbial parameters; the fermentation control parameters include fermentation liquid temperature, fermentation liquid pH, stirring speed and dissolved oxygen concentration;

[0025] In the step 1, a control range is obtained, and the control range includes a fermentation liquid temperature range, a fermentation liquid pH range, a stirring speed range, and a dissolved oxygen concentration range; a value is randomly selected from each range in the control range to construct a parameter set, and a total of n parameter sets are constructed;

[0026] In step 3, the population is initialized , including Z individuals, is the Zth individual;

[0027] The parameter labels and the range of the one-dimensional search space are ; The expression of each individual's initial position is: ; In the formula, is the initial position of the ith individual, is the random coefficient of the ith individual, , ;

[0028] In step 4, the fitness function is expressed as: ; In the formula, For fitness, It is an indicator of fermentation products;

[0029] Methods for obtaining fermentation product indicators include:

[0030] The parameter set corresponding to the parameter label corresponding to the current position of the individual and the fermentation characteristic parameters are used as analysis data, and the analysis data are sequentially input into the trained prediction model to predict the corresponding fermentation yield, product quality and fermentation cycle; wherein the prediction model includes a yield prediction model, a quality prediction model and a cycle prediction model, the output of the yield prediction model is the fermentation yield, the output of the quality prediction model is the product quality, and the output of the cycle prediction model is the fermentation cycle;

[0031] The predicted fermentation yield, product quality and fermentation cycle are weighted and summed to calculate the fermentation product index.

[0032] Furthermore, in step 5, the method of exploring the candidate positions includes migrating to unknown positions, migrating to known positions, and migrating to other individual positions; wherein the unknown positions are positions that have not been explored by the individual; the known positions are positions that have been explored by the individual but currently do not have an individual; and the other individual positions are positions that have been explored by the individual and currently have an individual;

[0033] Preset judgment coefficient m, ; For each individual, two The random numbers between are and ; The calculation method of the candidate position includes:

[0034] like ,but ;

[0035] like ,but ; ; ; ; ;

[0036] like ,but ;

[0037] In the formula, is the candidate position of the ith individual, , , , Both A random number between is the current position of a random individual, is the behavior coefficient, is the position coefficient, is the influence coefficient, is the current position of the i-th individual, The candidate position of the i-th individual in the last iteration, for Random numbers within.

[0038] Furthermore, in step 6, the execution behavior includes following behavior and competing behavior;

[0039] The switching coefficient q is preset; the method of updating the candidate position of each individual includes:

[0040] Calculate the fitness of each individual's current position, and mark the current position of the individual with the largest fitness as the best position ;

[0041] The behavior coefficient Compare with the switching factor q;

[0042] like ,but ; ;

[0043] like , then calculate the success probability , success probability The expression is: ; Preset probability threshold , ; The probability of success With probability threshold For comparison, if ,but ;like ,but ; ; ;

[0044] In the formula, is the candidate position of the i-th individual after update, is the following coefficient, is the competition coefficient, e is the natural constant, for A random number between .

[0045] Furthermore, in step 7, the dynamic boundary is , obtain the parameter label corresponding to the current position of the individual with the smallest fitness in the population and the parameter label corresponding to the current position of the individual with the largest fitness, and compare them. is a parameter label with a small value, It is the parameter label with large value;

[0046] The methods for generating the inverse solution include:

[0047] ;

[0048] In the formula, is the reverse solution of the ith individual, for Random numbers in ;

[0049] If the generated inverse solution is greater than or less than , it is marked as a transcendental solution, a random number function is used to randomly generate a value within the dynamic boundary, and the randomly generated value is assigned to the transcendental solution;

[0050] In step 8, the fitness corresponding to the current position, updated candidate position and reverse position of each individual is sorted from large to small; if the fitness ranked first is the fitness corresponding to the current position, the current position of the corresponding individual is not replaced; if the fitness ranked first is not the fitness corresponding to the current position, the current position of the corresponding individual is replaced with the position corresponding to the fitness ranked first.

[0051] Furthermore, the method for identifying deviation parameters in the fermentation control parameters collected in real time includes:

[0052] Collect x groups of historical control parameters, where the historical control parameters are fermentation control parameters collected at historical moments; obtain the optimal fermentation control parameters corresponding to the x groups of historical control parameters at corresponding moments, and mark them as historical optimal parameters, where the historical optimal parameters correspond to the historical control parameters one by one; subtract the corresponding parameters in the corresponding historical control parameters from each parameter in each group of historical optimal parameters and take the absolute value, obtain the parameter difference of each parameter in the fermentation control parameters, and mark them as historical difference; calculate the difference mean of each parameter according to the historical difference of each parameter in the fermentation control parameters, and then calculate the difference standard deviation of each parameter;

[0053] The fermentation control parameters collected in real time are marked as real-time control parameters; each parameter in the real-time control parameters is subtracted from the corresponding parameter in the optimal fermentation control parameters and the absolute value is taken to obtain the parameter difference of each parameter in the fermentation control parameters, and the difference is marked as the real-time difference; the parameters whose real-time difference is greater than or equal to the corresponding difference standard deviation are marked as deviation parameters, and the parameters whose real-time difference is less than the corresponding difference standard deviation are not marked.

[0054] Furthermore, the step of determining whether to generate a warning instruction or an adjustment instruction includes:

[0055] Step a1: Establish fuzzy sets and divide each deviation parameter into multiple fuzzy sets;

[0056] Step a2: convert the real-time difference corresponding to each deviation parameter into the membership degree of each corresponding fuzzy set through fuzzification technology;

[0057] Step a3: define fuzzy rules;

[0058] Step a4: matching the fuzzified real-time difference with the fuzzy rule, performing fuzzy reasoning, and obtaining the fuzzy reasoning result, the fuzzy reasoning result is the membership degree of generating the adjustment instruction and the membership degree of generating the warning instruction;

[0059] Step a5: Defuzzify the fuzzy reasoning result to obtain the defuzzified result, which is the generated instruction; the defuzzification operation is: according to the fuzzy reasoning result, respectively obtain the membership of the generated adjustment instruction and the membership of the generated warning instruction and compare them, and take the instruction corresponding to the membership with a larger value as the generated instruction.

[0060] Further, the corresponding parameter in the optimal fermentation control parameter is subtracted from each deviation parameter to obtain the adjustment amount corresponding to each deviation parameter; after the device end receives the adjustment amount corresponding to each deviation parameter, the device corresponding to the deviation parameter in the adjustment device is controlled to adjust each deviation parameter, and when the value of each deviation parameter reaches the value of the corresponding parameter in the optimal fermentation control parameter, the control of the adjustment device is stopped; wherein the adjustment device includes a temperature control device, an acid and alkali adding device, a stirring device and an oxygen supply device;

[0061] While sending the warning instruction to the corresponding device end, the fuzzified real-time difference is also sent to the corresponding device end.

[0062] Furthermore, the method further comprises: if the deviation parameter is the fermentation liquid temperature or the stirring speed, recalculating the adjustment amount corresponding to the dissolved oxygen concentration;

[0063] The fermentation liquid temperature marked as the deviation parameter is marked as the temperature deviation parameter, and the stirring speed marked as the deviation parameter is marked as the speed deviation parameter; the adjustment amount corresponding to the temperature deviation parameter is marked as the temperature adjustment amount, and the adjustment amount corresponding to the speed deviation parameter is marked as the speed adjustment amount; the influence values ​​of the temperature adjustment amount and the speed adjustment amount on the dissolved oxygen concentration are calculated respectively, and marked as the temperature influence value and the speed influence value respectively; the temperature influence value is added to the speed influence value to obtain the actual value of the dissolved oxygen concentration, and the dissolved oxygen concentration in the optimal fermentation control parameter is subtracted from the actual value of the dissolved oxygen concentration to recalculate the adjustment amount corresponding to the dissolved oxygen concentration;

[0064] The calculation method of temperature impact value includes:

[0065] Preset a temperature adjustment threshold, and compare the temperature adjustment amount with the temperature adjustment threshold;

[0066] If the temperature adjustment amount is less than the temperature adjustment threshold, the temperature adjustment amount is substituted into the pre-built influence function to calculate the corresponding temperature influence value;

[0067] If the temperature adjustment amount is greater than or equal to the temperature adjustment threshold, the temperature adjustment amount, dissolved oxygen concentration and fermentation characteristic parameters are used as test data, and the test data are input into the trained change prediction model to predict the corresponding dissolved oxygen concentration change value. The training process of the change prediction model is consistent with the training process of the yield prediction model, and both are deep neural network models; the dissolved oxygen concentration change value is the value after the dissolved oxygen concentration changes under the influence of the temperature adjustment amount; the dissolved oxygen concentration change value is subtracted from the dissolved oxygen concentration in the real-time control parameter to obtain the temperature influence value;

[0068] The calculation method of speed influence value is the same as that of temperature influence value.

[0069] The fermentation process control service system of a fermentation tank based on cloud computing implements the fermentation process control service method of a fermentation tank based on cloud computing and is applied to a cloud computing platform, including:

[0070] The parameter acquisition module is used to receive the fermentation parameters collected in real time by the equipment end during the fermentation process; the fermentation parameters include fermentation characteristic parameters and fermentation control parameters;

[0071] Parameter analysis module, used to analyze fermentation characteristic parameters and obtain optimal fermentation control parameters;

[0072] A deviation identification module, used for comparing the optimal fermentation control parameters with the fermentation control parameters collected in real time, and identifying the deviation parameters in the fermentation control parameters collected in real time;

[0073] Deviation judgment module, used to judge and generate warning instructions or adjustment instructions according to deviation parameters;

[0074] The deviation adjustment module is used to calculate the adjustment amount corresponding to each deviation parameter and send it to the corresponding device end if an adjustment instruction is generated; if an early warning instruction is generated, the early warning instruction is sent to the corresponding device end.

[0075] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the cloud computing-based fermentation tank fermentation process management and control service method when executing the computer program.

[0076] A computer-readable storage medium stores a computer program, and when the computer program is executed, the cloud computing-based fermentation tank fermentation process management and control service method is implemented.

[0077] The technical effects and advantages of the fermentation process control service system and method of the fermentation tank based on cloud computing of the present invention are as follows:

[0078] 1. By collecting and analyzing fermentation parameters in real time, considering multiple target requirements at the same time, dynamically calculating the optimal fermentation control parameters, automatically identifying the deviation parameters in the fermentation control parameters collected in real time, and using fuzzy logic control methods to judge and generate early warning instructions or adjustment instructions, and perform corresponding early warning or deviation adjustment operations; intelligent monitoring and control of the fermentation process is realized through the cloud computing platform, and the fermentation control parameters can be adaptively optimized according to the fermentation characteristic parameters at different stages; and deep learning technology is used in the optimization process, which can handle complex and highly nonlinear fermentation process control; it can effectively control the fermentation quality and fermentation output, while saving labor costs and improving fermentation efficiency, thereby effectively improving the automation level and stability of the fermentation process in the fermenter.

[0079] 2. Taking into account the influence of fermentation liquid temperature and stirring speed on dissolved oxygen concentration, the adjustment threshold is set, the adjustment range of fermentation liquid temperature and stirring speed is identified, and the temperature influence value and speed influence value are calculated for fermentation liquid temperature and stirring speed in different adjustment ranges through a strategy combining linear and nonlinear models. The adjustment of dissolved oxygen concentration can simultaneously respond to changes in fermentation liquid temperature and stirring speed, realize fine adjustment of dissolved oxygen concentration, and realize rapid response adjustment in small range changes and accurate numerical prediction in large range changes, which not only improves the prediction accuracy, but also simplifies the calculation and analysis process; it is conducive to improving and optimizing the intelligent control effect of the entire fermentation process, thereby improving the fermentation efficiency and product quality in the fermentation process of the fermenter. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 This is a schematic diagram of a fermentation process control service system for a fermenter based on cloud computing according to Example 1 of the present invention;

[0081] Figure 2 This is an example diagram of the interaction process between the cloud computing platform and the device end according to Embodiment 1 of the present invention;

[0082] Figure 3 This is a schematic diagram of a fermentation process control service system for a fermenter based on cloud computing according to Example 2 of the present invention;

[0083] Figure 4 This is a flow chart of a fermentation process control service method for a fermenter based on cloud computing according to Example 3 of the present invention;

[0084] Figure 5 This is a schematic diagram of an electronic device according to Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0085] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0086] Example 1

[0087] See also Figure 1 As shown, the fermentation process control service system of the fermentation tank based on cloud computing described in this embodiment includes a parameter acquisition module, a parameter analysis module, a deviation identification module, a deviation judgment module and a deviation adjustment module; each module is connected by wired and / or wireless means to realize data transmission between modules;

[0088] The parameter collection module is used to collect fermentation parameters in real time during the fermentation process on the equipment side and send them to the cloud computing platform; the fermentation parameters include fermentation characteristic parameters and fermentation control parameters.

[0089] The device end is a fermentation tank, and one fermentation tank is one device end; the cloud computing platform is a comprehensive cloud computing service platform.

[0090] Fermentation characteristic parameters include substrate parameters and microbial parameters;

[0091] Substrate parameters include substrate type and substrate concentration; substrate type includes carbon source (such as sugar, starch, etc.), nitrogen source (such as protein powder, ammonia water, etc.); substrate concentration includes starch concentration, ammonia water concentration, etc.; microbial parameters include at least inoculation size (initial concentration of microbial population), microbial species (such as yeast, bacteria, fungi, etc.); fermentation characteristic parameters, except for microbial species, are obtained through an online biodetector installed in the fermentation tank, which integrates a variety of sensors, such as conductivity sensors, optical sensors, electrochemical sensors, etc.; the microbial species are confirmed by staff through experimental analysis.

[0092] The fermentation control parameters include at least the fermentation liquid temperature, the fermentation liquid pH, the stirring speed and the dissolved oxygen concentration; the fermentation liquid is the liquid produced during the fermentation process, which usually contains microorganisms and their metabolites; the stirring speed is the speed at which the stirring equipment rotates per unit time; the dissolved oxygen concentration is the concentration of oxygen dissolved in the fermentation liquid; wherein, the fermentation liquid temperature, the fermentation liquid pH and the dissolved oxygen concentration are also obtained through the online biological detector, which also integrates the temperature sensor, the pH sensor, the dissolved oxygen sensor, etc.; the stirring speed is obtained by the speed sensor (such as a photoelectric sensor, a Hall effect sensor, etc.) installed on the driving shaft of the stirring equipment.

[0093] It should be noted that different substrate types will be utilized by microorganisms through different metabolic pathways to produce different metabolites, which will affect the pH of the fermentation broth, dissolved oxygen concentration, etc.; different substrate concentrations will affect the fermentation rate and microbial growth rate. Too high a substrate concentration will lead to an inhibitory effect and affect the activity of the microorganisms, while too low a substrate concentration will not be able to meet the growth needs of the microorganisms; the inoculation amount will affect the start-up speed of the fermentation and the speed of the overall fermentation process. Too low an inoculation amount may lead to a slow fermentation process, while too high an inoculation amount may lead to a rapid depletion of nutrients, affecting the fermentation effect; different types of microorganisms have different requirements for fermentation conditions. For example, some yeasts grow well at high temperatures, while some bacteria perform better at lower temperatures; therefore, by collecting fermentation characteristic parameters in real time, it is convenient to conduct subsequent real-time analysis of the optimal fermentation control parameters and to manage and control them in a timely manner to improve fermentation efficiency and product quality.

[0094] The parameter analysis module is used to analyze the fermentation characteristic parameters on the cloud computing platform to obtain the optimal fermentation control parameters.

[0095] The steps to obtain the optimal fermentation control parameters include:

[0096] Step 1: Construct n parameter sets, set different numerical labels for each parameter set, and mark them as parameter labels, where n is an integer greater than 1;

[0097] Step 2: Preset the population size Z and the number threshold T;

[0098] Step 3: Initialize the population. The position of each individual in the initialization population is defined in a one-dimensional search space. The position of the individual corresponds to the parameter label one by one. The number of iterations t of the initialization population is 0.

[0099] Step 4: Determine the fitness function;

[0100] Step 5: Calculate the candidate position of each individual in the next iteration and mark it as the candidate position;

[0101] Step 6: Determine the behavior performed by each individual and update the candidate position of each individual;

[0102] Step 7: Calculate the dynamic boundary, generate the reverse solution corresponding to the current position of each individual and add it to the population, using the reverse solution as the reverse position;

[0103] Step 8: Calculate the fitness of each individual's current position, updated candidate position, and reverse position, and determine whether to replace the position;

[0104] Step 9: If the number of iterations t is less than the number threshold T, then set the number of iterations , that is, add one to the value of the number of iterations and assign it to the number of iterations, and return to step 5; if the number of iterations t is greater than or equal to the number threshold T, go to step 10;

[0105] Step 10: Calculate the fitness corresponding to the current position of each individual, sort all fitnesses from large to small, obtain the current position of the individual corresponding to the fitness at the front, obtain the parameter set corresponding to the corresponding parameter label according to the obtained individual current position, and use it as the optimal fermentation control parameter.

[0106] In the above step 1, a control range is obtained, and the control range includes at least a fermentation liquid temperature range, a fermentation liquid pH range, a stirring speed range and a dissolved oxygen concentration range; wherein the fermentation liquid temperature range, the fermentation liquid pH range and the dissolved oxygen concentration range are all preset by technicians in this field according to actual conditions; the stirring speed range is obtained from the technical manual of the stirring equipment; a value is randomly selected from each range in the control range to construct a parameter set, and a total of n parameter sets are constructed, and the n parameter sets are all inconsistent.

[0107] In the above step 2, the population size Z is determined by a technician in this field. During the fermentation process of the historical fermentation tank, multiple groups of different fermentation characteristic parameters are collected. Under the conditions of multiple groups of different fermentation characteristic parameters, multiple different population sizes are set for the same fermentation characteristic parameters. Multiple analyses are performed according to steps 1 to 10. After the same number of iterations, the corresponding parameter labels are obtained, and the population size with the parameter label closest to the actual parameter label is used as the population size corresponding to the group of fermentation characteristic parameters; the actual parameter label is a digital label of the parameter set that best matches the group of fermentation characteristic parameters, and the actual parameter label is obtained by a technician in this field based on experiments and actual experience; and the population size corresponding to each group of fermentation characteristic parameters is obtained in this way, and the average of the multiple population sizes is used as the preset population size Z.

[0108] The number threshold T is determined by a technician in this field. During the fermentation process in the historical fermentation tank, multiple groups of different fermentation characteristic parameters are collected. Under the conditions of multiple groups of different fermentation characteristic parameters, the same fermentation characteristic parameters are analyzed multiple times according to steps 1 to 10 to obtain multiple parameter labels, wherein the number of iterations in each analysis process is different, and the population size is the same and is Z; the number of iterations that is closest to the actual parameter label is used as the number of iterations corresponding to the group of fermentation characteristic parameters; and the number of iterations corresponding to each group of fermentation characteristic parameters is obtained in this way, and the average of multiple iteration numbers is used as the number threshold T.

[0109] It should be understood that the population size Z determines the breadth of the search. A larger population size can explore more solution spaces, while the number threshold T determines the termination condition and can control the convergence speed.

[0110] In step 3 above, the population is initialized , including Z individuals, is the Zth individual.

[0111] The parameter labels and the range of the one-dimensional search space are ; The expression of each individual's initial position is: ; In the formula, is the initial position of the ith individual, is the random coefficient of the ith individual, , .

[0112] In step 4 above, the fitness function is expressed as: ; In the formula, For fitness, It is an indicator of fermentation products.

[0113] Methods for obtaining fermentation product indicators include:

[0114] The parameter set corresponding to the parameter label corresponding to the individual's current position and the fermentation characteristic parameters are used as analysis data, and the analysis data are input into the trained prediction model in sequence to predict the corresponding fermentation yield, product quality and fermentation cycle; wherein the prediction model includes a yield prediction model, a quality prediction model and a cycle prediction model, the output of the yield prediction model is the fermentation yield, the output of the quality prediction model is the product quality, and the output of the cycle prediction model is the fermentation cycle; the fermentation yield is the amount of product obtained after the fermentation process is completed, usually measured in volume, weight or concentration, reflecting the production efficiency and raw material conversion rate; product quality is the characteristics and standards of the product generated during the fermentation process, including taste, aroma, color, purity, nutritional content, etc., which affects the acceptance and applicability of the product; the fermentation cycle is the time required from the start to the completion of fermentation, and the length of the fermentation cycle directly affects the production efficiency.

[0115] The predicted fermentation yield, product quality and fermentation cycle are weighted and summed to calculate the fermentation product index; the expression of the fermentation product index is: ; In the formula, is the fermentation yield, For product quality, For the fermentation cycle, , , All of them are preset weight coefficients; the specific values ​​of the weight coefficients can be set according to actual conditions. The weight coefficients reflect the influence of fermentation yield, product quality and fermentation cycle on the fermentation product indicators. Technical personnel in this field can preset corresponding weight coefficients according to the actual influence of fermentation yield, product quality and fermentation cycle on the fermentation product indicators, so as to accurately evaluate the comprehensive performance of the fermentation process.

[0116] Among them, under the premise of not affecting the product quality and fermentation cycle, the greater the fermentation output, the greater the production capacity and economic benefits, and the greater the fermentation product index; under the premise of not affecting the fermentation output and fermentation cycle, the greater the product quality, the better the product can meet market demand, meet standards and gain consumer recognition, and the greater the fermentation product index; under the premise of not affecting the fermentation output and product quality, the shorter the fermentation cycle, the more it can improve production efficiency and the speed of responding to market demand, and therefore the greater the fermentation product index.

[0117] The specific training process of the yield prediction model includes:

[0118] Preliminarily collect the fermentation yields corresponding to a group of analytical data, where a is an integer greater than 1, and convert the analytical data and the corresponding fermentation yields into a corresponding set of feature vectors; the fermentation yields corresponding to the analytical data are collected by a technician in the field of technology in the fermentation process of the historical fermentation tank, and a group of different analytical data are collected, and fermentation experiments are carried out in sequence under the conditions of each group of analytical data, and the fermentation yields of the fermentation products are measured and calculated after each fermentation experiment is completed, and the fermentation yields corresponding to each group of analytical data are obtained, and the corresponding fermentation yields are set for the group of analytical data respectively;

[0119] Each set of feature vectors is used as the input of the yield prediction model. The yield prediction model uses a set of fermentation yields corresponding to each set of analytical data as output, and the actual fermentation yield corresponding to each set of analytical data as the prediction target. The actual fermentation yield is the fermentation yield corresponding to the analytical data collected in advance. Minimizing the sum of the prediction errors of all analytical data is used as the training target. The calculation formula of the prediction error is: ,in is the prediction error, K is the group number of the eigenvector corresponding to the analysis data, is the fermentation yield corresponding to the Kth group of analysis data, is the actual fermentation yield corresponding to the Kth group of analysis data; the yield prediction model is trained until the sum of the prediction errors reaches convergence and the training is stopped.

[0120] The above-mentioned yield prediction model is specifically a deep neural network model; the specific training process of the quality prediction model and the cycle prediction model is consistent with the training process of the yield prediction model, and both are deep neural network models.

[0121] In the above step 5, the method of exploring candidate locations includes migrating to unknown locations, migrating to known locations, and migrating to other individual locations; among which, the unknown location is a location that has not been explored by the individual; the known location is a location that has been explored by the individual but the individual does not currently exist; and the other individual locations are locations that have been explored by the individual and the individual currently exists.

[0122] Preset judgment coefficient m, ; For each individual, two The random numbers between are and ; The calculation method of the candidate position includes:

[0123] like ,but ; It shows that the method of exploring candidate locations is to migrate to unknown locations;

[0124] like ,but ; ; ; ; ; It shows that the method of exploring candidate positions is to migrate to other individual positions, for Random numbers within

[0125] like ,but ; It shows that the method of exploring candidate locations is to migrate to known locations;

[0126] In the formula, is the candidate position of the ith individual, , , , Both A random number between is the current position of a random individual, is the behavior coefficient, is the position coefficient, is the influence coefficient, is the current position of the i-th individual, The candidate position of the i-th individual in the previous iteration.

[0127] In the above step 6, the execution behavior includes following behavior and competing behavior; wherein, following behavior is the behavior of the remaining individuals following the best individual, and the best individual is the individual with the largest corresponding fitness; as the number of iterations increases, the fitness of the remaining individuals gradually approaches the best individual, so they will compete with the best individual, that is, competing behavior occurs.

[0128] The switching coefficient q is preset, and the switching coefficient q is preset by a person skilled in the art according to actual conditions; the method for updating the candidate position of each individual includes:

[0129] Calculate the fitness of each individual's current position, and mark the current position of the individual with the largest fitness as the best position .

[0130] The behavior coefficient Compare with the switching factor q;

[0131] like ,but ; ; It indicates that the behavior performed by the individual is a following behavior;

[0132] like , then calculate the success probability , success probability The expression is: ; Preset probability threshold , probability threshold It is preset by those skilled in the art according to the actual situation. ; The probability of success With probability threshold For comparison, if ,but ;like ,but ; ; ; It indicates that the behavior performed by the individual is competitive behavior;

[0133] In the formula, is the candidate position of the i-th individual after update, is the following coefficient, is the competition coefficient, e is the natural constant, for A random number between .

[0134] In step 7 above, the dynamic boundary is , obtain the parameter label corresponding to the current position of the individual with the smallest fitness in the population and the parameter label corresponding to the current position of the individual with the largest fitness, and compare them. is a parameter label with a small value, Labels for parameters with large values.

[0135] The methods for generating the inverse solution include:

[0136] ;

[0137] In the formula, is the reverse solution of the ith individual, for The random number in .

[0138] If the generated inverse solution is greater than or less than , it is marked as a transcendental solution, a random number function is used to randomly generate a value within the dynamic boundary, and the randomly generated value is assigned to the transcendental solution.

[0139] In the above step 8, the fitness corresponding to the current position, updated candidate position and reverse position of each individual is sorted from large to small; if the fitness ranked first is the fitness corresponding to the current position, the current position of the corresponding individual is not replaced; if the fitness ranked first is not the fitness corresponding to the current position, the current position of the corresponding individual is replaced with the position corresponding to the fitness ranked first.

[0140] The steps for obtaining the optimal fermentation control parameters at each device are the same.

[0141] It should be noted that the reason for obtaining the optimal fermentation control parameters according to steps 1 to 10 is that steps 1 to 10 can effectively integrate the advantages of swarm intelligence optimization to ensure comprehensive and flexible exploration and optimization in a complex solution space; by constructing a parameter set and dynamically updating individual positions and fitness evaluations, efficient response to multi-objective optimization is ensured, and ultimately the optimal fermentation control parameters obtained have higher accuracy and adaptability, thereby improving the efficiency of the fermentation process and product quality, reducing the need for manual intervention, and enhancing the stability and economic benefits of the overall fermentation process.

[0142] The deviation identification module is used in the cloud computing platform to compare the optimal fermentation control parameters with the fermentation control parameters collected in real time, and to identify the deviation parameters in the fermentation control parameters collected in real time.

[0143] The method for identifying deviation parameters in fermentation control parameters collected in real time includes:

[0144] Collect x groups of historical control parameters, where the historical control parameters are fermentation control parameters collected at historical moments; obtain the optimal fermentation control parameters corresponding to the x groups of historical control parameters at corresponding moments, and mark them as historical optimal parameters, where the historical optimal parameters correspond to the historical control parameters one by one; subtract the corresponding parameters in the corresponding historical control parameters from each parameter in each group of historical optimal parameters and take the absolute value, obtain the parameter difference of each parameter in the fermentation control parameters, and mark them as historical differences; calculate the mean difference of each parameter based on the historical difference of each parameter in the fermentation control parameters, and then calculate the standard deviation of the difference of each parameter; the expression of the mean difference is: ; In the formula, is the mean difference of the Yth parameter in the fermentation control parameters, is the yth historical difference corresponding to the Yth parameter in the fermentation control parameters, , And is an integer, different values ​​of Y correspond to different parameters in the fermentation control parameters. For example, Y=1 represents the fermentation liquid temperature, Y=2 represents the fermentation liquid pH, Y=3 represents the stirring speed, and Y=4 represents the dissolved oxygen concentration; the expression of the difference standard deviation is: ; In the formula, is the standard deviation of the difference of the Yth parameter in the fermentation control parameters;

[0145] The fermentation control parameters collected in real time are marked as real-time control parameters; each parameter in the real-time control parameters is subtracted from the corresponding parameter in the optimal fermentation control parameters and the absolute value is taken to obtain the parameter difference of each parameter in the fermentation control parameters, and the difference is marked as the real-time difference; the parameters whose real-time difference is greater than or equal to the corresponding difference standard deviation are marked as deviation parameters, and the parameters whose real-time difference is less than the corresponding difference standard deviation are not marked.

[0146] The deviation judgment module is used in the cloud computing platform to judge and generate early warning instructions or adjustment instructions based on deviation parameters.

[0147] The steps of determining whether to generate a warning instruction or an adjustment instruction include:

[0148] Step a1: Establish fuzzy sets and divide each deviation parameter into multiple fuzzy sets; for example, the fuzzy sets corresponding to the fermentation liquid temperature are slight temperature deviation, moderate temperature deviation, severe temperature deviation, etc., and the fuzzy sets corresponding to the stirring speed are slight speed deviation, moderate speed deviation, severe speed deviation, etc.;

[0149] Step a2: The real-time difference corresponding to each deviation parameter is converted into the membership of each corresponding fuzzy set through fuzzification technology; fuzzification is the process of converting an exact value into the membership corresponding to a fuzzy set, and fuzzification technologies include triangular membership function, trapezoidal membership function, etc.; for example, the value of the real-time difference corresponding to the fermentation liquid temperature is low, and it is inferred that the membership of a slight temperature deviation is 0.8, the membership of a moderate temperature deviation is 0.1, and the membership of a severe temperature deviation is 0.1;

[0150] Step a3: define fuzzy rules, which are defined based on expert knowledge or literature; for example, if the temperature has a slight deviation, the pH has a slight deviation, and the speed has a moderate deviation, then an adjustment instruction is inferred to be generated; if the temperature has a moderate deviation, the concentration has a serious deviation, and the speed has a serious deviation, then an early warning instruction is inferred to be generated;

[0151] Step a4: matching the fuzzified real-time difference with the fuzzy rule, performing fuzzy reasoning, and obtaining the fuzzy reasoning result, the fuzzy reasoning result is the membership of generating the adjustment instruction and the membership of generating the warning instruction; the fuzzy reasoning method is, for example, the Mamdani or Sugeno fuzzy reasoning method; the fuzzy reasoning result is, for example, the membership of generating the adjustment instruction is 0.2, and the membership of generating the warning instruction is 0.8;

[0152] Step a5: Defuzzify the fuzzy reasoning result to obtain the defuzzified result, which is the generated instruction; defuzzification is to convert the fuzzy reasoning result into a specific recognition result, and the defuzzification operation is: according to the fuzzy reasoning result, respectively obtain the membership of the generated adjustment instruction and the membership of the generated warning instruction and compare them, and take the instruction corresponding to the membership with a larger numerical value as the generated instruction.

[0153] It should be noted that the reason for using the fuzzy logic control method to determine the generation of early warning instructions or adjustment instructions is that the fuzzy logic control method can convert the precise real-time difference into the membership in the fuzzy set, thereby effectively handling uncertainty and ambiguity; enabling the system to respond to complex fermentation processes more flexibly, and by fuzzifying the deviation parameters, providing a more intuitive way to evaluate the state of the fermentation tank; in addition, the fuzzy logic control method combines fuzzy rules and reasoning methods, enabling the system to make intelligent decisions based on expert knowledge, and generate adjustment instructions or early warning instructions more reasonably, thereby improving the accuracy and response speed of fermentation process control, and improving the overall fermentation efficiency and product quality.

[0154] Deviation adjustment module, if an adjustment instruction is generated, the cloud computing platform calculates the adjustment amount corresponding to each deviation parameter and sends it to the corresponding device end; if an early warning instruction is generated, the cloud computing platform sends an early warning instruction to the corresponding device end.

[0155] Subtract the corresponding parameter in the optimal fermentation control parameter from each deviation parameter to obtain the adjustment amount corresponding to each deviation parameter; after receiving the adjustment amount corresponding to each deviation parameter, the device end controls the device corresponding to the deviation parameter in the adjustment device to adjust each deviation parameter, and when the value of each deviation parameter reaches the value of the corresponding parameter in the optimal fermentation control parameter, the control of the adjustment device is stopped; wherein the adjustment device includes a temperature control device, an acid and alkali adding device, a stirring device and an oxygen supply device; the temperature control device includes a heater and a cooler for increasing and decreasing the temperature of the fermentation liquid; the acid and alkali adding device automatically adds acid (such as hydrochloric acid) or alkali (such as sodium hydroxide) to the fermentation liquid in the fermentation tank through a pump or other conveying device to adjust the pH of the fermentation liquid; the stirring speed of the stirring device is adjusted by controlling the variable frequency drive in the stirring device; the oxygen supply device includes a gas distributor and an oxygen dissolver, and the dissolved oxygen concentration is adjusted by controlling the oxygen flow and oxygen concentration.

[0156] While the cloud computing platform sends warning instructions to the corresponding device end, it also sends the fuzzified real-time difference to the corresponding device end, so that relevant staff can clearly understand the actual deviation degree of the deviation parameter and accurately locate the source of the deviation problem, thereby improving the maintenance efficiency of relevant staff.

[0157] The interaction process between the cloud computing platform and the device side is as follows: Figure 2 As shown;

[0158] This embodiment collects and analyzes fermentation parameters in real time, takes into account multiple target requirements, dynamically calculates optimal fermentation control parameters, automatically identifies deviation parameters in the fermentation control parameters collected in real time, and uses fuzzy logic control methods to determine and generate early warning instructions or adjustment instructions, and performs corresponding early warning or deviation adjustment operations; intelligent monitoring and control of the fermentation process is achieved through a cloud computing platform, and the fermentation control parameters can be adaptively optimized according to the fermentation characteristic parameters of different stages; and deep learning technology is used in the optimization process, which can handle complex and highly nonlinear fermentation process control; the fermentation quality and fermentation yield can be effectively controlled, while saving labor costs and improving fermentation efficiency, thereby effectively improving the automation level and stability of the fermentation process in the fermenter.

[0159] Example 2

[0160] See also Figure 3 As shown, this embodiment further improves the design on the basis of embodiment 1. When the fermentation liquid temperature and stirring speed in the fermenter are adjusted, the dissolved oxygen concentration will also change accordingly. The reason is that increasing the stirring speed can increase the fluidity of the fermentation liquid, promote the contact between oxygen and the fermentation liquid, and enhance the solubility of oxygen, thereby increasing the dissolved oxygen concentration, while reducing the stirring speed will lead to uneven oxygen distribution, limit the effective transmission of dissolved oxygen, and thus reduce the dissolved oxygen concentration; when the fermentation liquid temperature increases, the solubility of oxygen in the fermentation liquid decreases, and at the same time accelerates the growth and metabolism of microorganisms, resulting in an increase in the demand for oxygen, further exacerbating the consumption of dissolved oxygen, and may also cause the viscosity of the fermentation liquid to change, affecting the contact efficiency between oxygen and the fermentation liquid, thereby causing a decrease in the dissolved oxygen concentration; when the fermentation liquid temperature decreases, the solubility of oxygen in the fermentation liquid increases, and slows down the growth and metabolism rate of microorganisms, thereby increasing the dissolved oxygen concentration; therefore, when controlling the dissolved oxygen concentration in the fermenter, it is necessary to simultaneously consider the influence of the changes in the fermentation liquid temperature and stirring speed on the dissolved oxygen concentration, so as to achieve accurate adjustment of the dissolved oxygen concentration; therefore, the cloud computing-based fermentation tank fermentation process control service system provided in this embodiment also includes a secondary analysis module.

[0161] The secondary analysis module is used for the cloud computing platform to recalculate the adjustment amount corresponding to the dissolved oxygen concentration if the deviation parameter is the fermentation liquid temperature or the stirring speed.

[0162] The fermentation liquid temperature marked as the deviation parameter is marked as the temperature deviation parameter, and the stirring speed marked as the deviation parameter is marked as the speed deviation parameter; the adjustment amount corresponding to the temperature deviation parameter is marked as the temperature adjustment amount, and the adjustment amount corresponding to the speed deviation parameter is marked as the speed adjustment amount; the influence values ​​of the temperature adjustment amount and the speed adjustment amount on the dissolved oxygen concentration are calculated respectively, and marked as the temperature influence value and the speed influence value respectively; the temperature influence value is added to the speed influence value to obtain the actual value of the dissolved oxygen concentration, the dissolved oxygen concentration in the optimal fermentation control parameter is subtracted from the actual value of the dissolved oxygen concentration, and the adjustment amount corresponding to the dissolved oxygen concentration is recalculated.

[0163] The calculation method of temperature impact value includes:

[0164] A temperature adjustment threshold is preset. The temperature adjustment threshold is determined by technicians in this field by gradually increasing the temperature of the fermentation liquid for multiple times during the fermentation process in the historical fermenter, and recording the change in the fermentation liquid temperature and the change in the dissolved oxygen concentration in sequence. The change in the fermentation liquid temperature when the change in the dissolved oxygen concentration is not linearly related to the change in the fermentation liquid temperature is taken as the maximum change, and the average of multiple maximum changes is taken as the temperature adjustment threshold.

[0165] comparing the temperature regulation amount to a temperature regulation threshold;

[0166] If the temperature adjustment amount is less than the temperature adjustment threshold, the temperature adjustment amount is substituted into the pre-built influence function to calculate the corresponding temperature influence value; the influence function is: ; In the formula, is the temperature influence value, is the impact factor, is the temperature adjustment amount, is the influence constant; the influence factor and the influence constant are determined by a technician in the field of fermentation in a historical fermentation tank, adjusting the temperature of the fermentation liquid for multiple times, and the temperature adjustment amount during each adjustment is less than the temperature adjustment threshold, and collecting and calculating the temperature influence value after each adjustment is completed; taking each temperature adjustment amount and the corresponding temperature influence value as a group of analysis sets, substituting each group of analysis sets into the influence function respectively, and jointly calculating the influence factor and the influence constant by substituting every two influence functions into the analysis sets, taking the mean value of all the calculated influence factors as the value of the influence factor in the preset influence function, and taking the mean value of all the calculated influence constants as the value of the influence constant in the preset influence function;

[0167] If the temperature adjustment amount is greater than or equal to the temperature adjustment threshold, the temperature adjustment amount, dissolved oxygen concentration and fermentation characteristic parameters are used as test data, and the test data are input into the trained change prediction model to predict the corresponding dissolved oxygen concentration change value. The specific training process of the change prediction model is consistent with the training process of the yield prediction model, and both are deep neural network models; the dissolved oxygen concentration change value is the value after the dissolved oxygen concentration changes under the influence of the temperature adjustment amount; the dissolved oxygen concentration change value is subtracted from the dissolved oxygen concentration in the real-time control parameter to obtain the temperature impact value.

[0168] The calculation method of speed influence value is the same as that of temperature influence value.

[0169] It should be noted that the purpose of dividing the temperature influence value and speed influence value into two calculation methods, linear and nonlinear, is to optimize the control strategy: using linear models in a small range can respond and adjust quickly, while using nonlinear models in a large range can more accurately capture complex system behaviors; this hierarchical control strategy can optimize the fermentation process under different conditions; improve prediction accuracy: linear models are suitable for simple situations and can quickly provide prediction results; nonlinear models can more accurately describe system behavior when making large adjustments, thereby improving the accuracy of the overall prediction; simplify calculation and analysis: the mathematical form of linear models is simple, which is easy to calculate and understand quickly, and is suitable for preliminary analysis and rapid estimation; while nonlinear models are complex, they can provide deeper analysis.

[0170] This embodiment takes into account the influence of fermentation liquid temperature and stirring speed on dissolved oxygen concentration, sets adjustment thresholds, identifies the adjustment range of fermentation liquid temperature and stirring speed, and calculates temperature influence values ​​and speed influence values ​​for fermentation liquid temperature and stirring speed in different adjustment ranges through a strategy combining linear and nonlinear models, so that the adjustment of dissolved oxygen concentration can simultaneously respond to changes in fermentation liquid temperature and stirring speed, realize fine adjustment of dissolved oxygen concentration, and realize rapid response adjustment when a small range changes and accurate numerical prediction when a large range changes, which not only improves the prediction accuracy, but also simplifies the calculation and analysis process; it is conducive to improving and optimizing the intelligent control effect of the entire fermentation process, thereby improving the fermentation efficiency and product quality in the fermentation process of the fermenter.

[0171] Example 3

[0172] See also Figure 4 As shown, the part not described in detail in this embodiment is described in Example 1 and Example 2, and a fermentation process control service method of a fermenter based on cloud computing is provided, and the method includes:

[0173] The fermentation parameters collected in real time by the receiving device during the fermentation process; the fermentation parameters include fermentation characteristic parameters and fermentation control parameters;

[0174] Analyze the fermentation characteristic parameters to obtain the optimal fermentation control parameters;

[0175] comparing the optimal fermentation control parameters with the fermentation control parameters collected in real time, and identifying deviation parameters in the fermentation control parameters collected in real time;

[0176] According to the deviation parameters, generate warning instructions or adjustment instructions;

[0177] If an adjustment instruction is generated, the adjustment amount corresponding to each deviation parameter is calculated and sent to the corresponding device end; if an early warning instruction is generated, the early warning instruction is sent to the corresponding device end.

[0178] Furthermore, the step of obtaining the optimal fermentation control parameters includes:

[0179] Step 1: Construct n parameter sets, set different numerical labels for each parameter set, and mark them as parameter labels, where n is an integer greater than 1;

[0180] Step 2: Preset the population size Z and the number threshold T;

[0181] Step 3: Initialize the population. The position of each individual in the initialization population is defined in a one-dimensional search space. The position of the individual corresponds to the parameter label one by one. The number of iterations t of the initialization population is 0.

[0182] Step 4: Determine the fitness function;

[0183] Step 5: Calculate the candidate position of each individual in the next iteration and mark it as the candidate position;

[0184] Step 6: Determine the behavior performed by each individual and update the candidate position of each individual;

[0185] Step 7: Calculate the dynamic boundary, generate the reverse solution corresponding to the current position of each individual and add it to the population, using the reverse solution as the reverse position;

[0186] Step 8: Calculate the fitness of each individual's current position, updated candidate position, and reverse position, and determine whether to replace the position;

[0187] Step 9: If the number of iterations t is less than the number threshold T, then set the number of iterations , and return to step 5; if the number of iterations t is greater than or equal to the number threshold T, proceed to step 10;

[0188] Step 10: Calculate the fitness corresponding to the current position of each individual, sort all fitnesses from large to small, obtain the current position of the individual corresponding to the fitness at the front, obtain the parameter set corresponding to the corresponding parameter label according to the obtained individual current position, and use it as the optimal fermentation control parameter;

[0189] The steps for obtaining the optimal fermentation control parameters at each device are the same.

[0190] Furthermore, the fermentation characteristic parameters include substrate parameters and microbial parameters; the fermentation control parameters include fermentation liquid temperature, fermentation liquid pH, stirring speed and dissolved oxygen concentration;

[0191] In the step 1, a control range is obtained, and the control range includes a fermentation liquid temperature range, a fermentation liquid pH range, a stirring speed range, and a dissolved oxygen concentration range; a value is randomly selected from each range in the control range to construct a parameter set, and a total of n parameter sets are constructed;

[0192] In step 3, the population is initialized , including Z individuals, is the Zth individual;

[0193] The parameter labels and the range of the one-dimensional search space are ; The expression of each individual's initial position is: ; In the formula, is the initial position of the ith individual, is the random coefficient of the ith individual, , ;

[0194] In step 4, the fitness function is expressed as: ; In the formula, For fitness, It is an indicator of fermentation products;

[0195] Methods for obtaining fermentation product indicators include:

[0196] The parameter set corresponding to the parameter label corresponding to the current position of the individual and the fermentation characteristic parameters are used as analysis data, and the analysis data are sequentially input into the trained prediction model to predict the corresponding fermentation yield, product quality and fermentation cycle; wherein the prediction model includes a yield prediction model, a quality prediction model and a cycle prediction model, the output of the yield prediction model is the fermentation yield, the output of the quality prediction model is the product quality, and the output of the cycle prediction model is the fermentation cycle;

[0197] The predicted fermentation yield, product quality and fermentation cycle are weighted and summed to calculate the fermentation product index.

[0198] Furthermore, in step 5, the method of exploring the candidate positions includes migrating to unknown positions, migrating to known positions, and migrating to other individual positions; wherein the unknown positions are positions that have not been explored by the individual; the known positions are positions that have been explored by the individual but currently do not have an individual; and the other individual positions are positions that have been explored by the individual and currently have an individual;

[0199] Preset judgment coefficient m, ; For each individual, two The random numbers between are and ; The calculation method of the candidate position includes:

[0200] like ,but ;

[0201] like ,but ; ; ; ; ;

[0202] like ,but ;

[0203] In the formula, is the candidate position of the ith individual, , , , Both A random number between is the current position of a random individual, is the behavior coefficient, is the position coefficient, is the influence coefficient, is the current position of the i-th individual, The candidate position of the i-th individual in the last iteration, for Random numbers within.

[0204] Furthermore, in step 6, the execution behavior includes following behavior and competing behavior;

[0205] The switching coefficient q is preset; the method of updating the candidate position of each individual includes:

[0206] Calculate the fitness of each individual's current position, and mark the current position of the individual with the largest fitness as the best position ;

[0207] The behavior coefficient Compare with the switching factor q;

[0208] like ,but ; ;

[0209] like , then calculate the success probability , success probability The expression is: ; Preset probability threshold , ; The probability of success With probability threshold For comparison, if ,but ;like ,but ; ; ;

[0210] In the formula, is the candidate position of the i-th individual after update, is the following coefficient, is the competition coefficient, e is the natural constant, for A random number between .

[0211] Furthermore, in step 7, the dynamic boundary is , obtain the parameter label corresponding to the current position of the individual with the smallest fitness in the population and the parameter label corresponding to the current position of the individual with the largest fitness, and compare them. is a parameter label with a small value, It is the parameter label with large value;

[0212] The methods for generating the inverse solution include:

[0213] ;

[0214] In the formula, is the reverse solution of the ith individual, for Random numbers in ;

[0215] If the generated inverse solution is greater than or less than , it is marked as a transcendental solution, a random number function is used to randomly generate a value within the dynamic boundary, and the randomly generated value is assigned to the transcendental solution;

[0216] In step 8, the fitness corresponding to the current position, updated candidate position and reverse position of each individual is sorted from large to small; if the fitness ranked first is the fitness corresponding to the current position, the current position of the corresponding individual is not replaced; if the fitness ranked first is not the fitness corresponding to the current position, the current position of the corresponding individual is replaced with the position corresponding to the fitness ranked first.

[0217] Furthermore, the method for identifying deviation parameters in the fermentation control parameters collected in real time includes:

[0218] Collect x groups of historical control parameters, where the historical control parameters are fermentation control parameters collected at historical moments; obtain the optimal fermentation control parameters corresponding to the x groups of historical control parameters at corresponding moments, and mark them as historical optimal parameters, where the historical optimal parameters correspond to the historical control parameters one by one; subtract the corresponding parameters in the corresponding historical control parameters from each parameter in each group of historical optimal parameters and take the absolute value, obtain the parameter difference of each parameter in the fermentation control parameters, and mark them as historical difference; calculate the difference mean of each parameter according to the historical difference of each parameter in the fermentation control parameters, and then calculate the difference standard deviation of each parameter;

[0219] The fermentation control parameters collected in real time are marked as real-time control parameters; each parameter in the real-time control parameters is subtracted from the corresponding parameter in the optimal fermentation control parameters and the absolute value is taken to obtain the parameter difference of each parameter in the fermentation control parameters, and the difference is marked as the real-time difference; the parameters whose real-time difference is greater than or equal to the corresponding difference standard deviation are marked as deviation parameters, and the parameters whose real-time difference is less than the corresponding difference standard deviation are not marked.

[0220] Furthermore, the step of determining whether to generate a warning instruction or an adjustment instruction includes:

[0221] Step a1: Establish fuzzy sets and divide each deviation parameter into multiple fuzzy sets;

[0222] Step a2: convert the real-time difference corresponding to each deviation parameter into the membership degree of each corresponding fuzzy set through fuzzification technology;

[0223] Step a3: define fuzzy rules;

[0224] Step a4: matching the fuzzified real-time difference with the fuzzy rule, performing fuzzy reasoning, and obtaining the fuzzy reasoning result, the fuzzy reasoning result is the membership degree of generating the adjustment instruction and the membership degree of generating the warning instruction;

[0225] Step a5: Defuzzify the fuzzy reasoning result to obtain the defuzzified result, which is the generated instruction; the defuzzification operation is: according to the fuzzy reasoning result, respectively obtain the membership of the generated adjustment instruction and the membership of the generated warning instruction and compare them, and take the instruction corresponding to the membership with a larger value as the generated instruction.

[0226] Further, the corresponding parameter in the optimal fermentation control parameter is subtracted from each deviation parameter to obtain the adjustment amount corresponding to each deviation parameter; after the device end receives the adjustment amount corresponding to each deviation parameter, the device corresponding to the deviation parameter in the adjustment device is controlled to adjust each deviation parameter, and when the value of each deviation parameter reaches the value of the corresponding parameter in the optimal fermentation control parameter, the control of the adjustment device is stopped; wherein the adjustment device includes a temperature control device, an acid and alkali adding device, a stirring device and an oxygen supply device;

[0227] While sending the warning instruction to the corresponding device end, the fuzzified real-time difference is also sent to the corresponding device end.

[0228] Furthermore, the method further comprises: if the deviation parameter is the fermentation liquid temperature or the stirring speed, recalculating the adjustment amount corresponding to the dissolved oxygen concentration;

[0229] The fermentation liquid temperature marked as the deviation parameter is marked as the temperature deviation parameter, and the stirring speed marked as the deviation parameter is marked as the speed deviation parameter; the adjustment amount corresponding to the temperature deviation parameter is marked as the temperature adjustment amount, and the adjustment amount corresponding to the speed deviation parameter is marked as the speed adjustment amount; the influence values ​​of the temperature adjustment amount and the speed adjustment amount on the dissolved oxygen concentration are calculated respectively, and marked as the temperature influence value and the speed influence value respectively; the temperature influence value is added to the speed influence value to obtain the actual value of the dissolved oxygen concentration, and the dissolved oxygen concentration in the optimal fermentation control parameter is subtracted from the actual value of the dissolved oxygen concentration to recalculate the adjustment amount corresponding to the dissolved oxygen concentration;

[0230] The calculation method of temperature impact value includes:

[0231] Preset a temperature adjustment threshold, and compare the temperature adjustment amount with the temperature adjustment threshold;

[0232] If the temperature adjustment amount is less than the temperature adjustment threshold, the temperature adjustment amount is substituted into the pre-built influence function to calculate the corresponding temperature influence value;

[0233] If the temperature adjustment amount is greater than or equal to the temperature adjustment threshold, the temperature adjustment amount, dissolved oxygen concentration and fermentation characteristic parameters are used as test data, and the test data are input into the trained change prediction model to predict the corresponding dissolved oxygen concentration change value. The training process of the change prediction model is consistent with the training process of the yield prediction model, and both are deep neural network models; the dissolved oxygen concentration change value is the value after the dissolved oxygen concentration changes under the influence of the temperature adjustment amount; the dissolved oxygen concentration change value is subtracted from the dissolved oxygen concentration in the real-time control parameter to obtain the temperature influence value;

[0234] The calculation method of speed influence value is the same as that of temperature influence value.

[0235] Example 4

[0236] See also Figure 5 As shown, the present application also provides an electronic device, including: a processor 11 ( Figure 5 Only one processor is shown in the figure), a memory 12, and a computer program 13 stored in the memory 12 and executable on the processor 11. When the processor 11 executes the computer program 13, the steps in the above-mentioned cloud computing-based fermentation tank fermentation process control service method are implemented. Alternatively, when the processor 11 executes the computer program 13, the functions of each module / unit in the above-mentioned cloud computing-based fermentation tank fermentation process control service system are implemented.

[0237] Exemplarily, the computer program 13 may be divided into one or more modules / units, which are stored in the memory 12 and executed by the processor 11 to complete the present application. The one or more modules / units may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program 13 in the terminal device.

[0238] Those skilled in the art will understand that Figure 5 These are merely examples of electronic devices and do not constitute a limitation of the electronic device. The electronic device may include more or fewer components than those shown in the figure, or a combination of certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0239] The processor 11 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0240] The memory 12 may be an internal storage unit of the terminal device, such as a hard disk or memory of the terminal device. The memory 12 may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card (FlashCard), etc. equipped on the terminal device. Further, the memory 12 may also include both an internal storage unit of the terminal device and an external storage device. The memory 12 is used to store the computer program and other programs and data required by the terminal device. The memory 12 may also be used to temporarily store data that has been output or is to be output.

[0241] Example 5

[0242] One embodiment of the present application discloses a computer-readable storage medium. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are executed by a processor, a fermentation tank fermentation process management and control service method based on cloud computing can be executed. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0243] In addition, according to the embodiments of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, for example: a fermentation tank fermentation process control service method based on cloud computing. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0244] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art 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.

[0245] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A fermentation process control service method for a fermenter based on cloud computing, characterized in that: Applied in cloud computing platforms, including: The fermentation parameters collected in real time by the receiving device during the fermentation process; the fermentation parameters include fermentation characteristic parameters and fermentation control parameters; Analyze the fermentation characteristic parameters to obtain the optimal fermentation control parameters; The step of obtaining the optimal fermentation control parameters comprises: Step 1: Construct n parameter sets, set different numerical labels for each parameter set, and mark them as parameter labels, where n is an integer greater than 1; Step 2: Preset the population size Z and the number threshold T; Step 3: Initialize the population. The position of each individual in the initialization population is defined in a one-dimensional search space. The position of the individual corresponds to the parameter label one by one. The number of iterations t of the initialization population is 0. Step 4: Determine the fitness function; Step 5: Calculate the candidate position of each individual in the next iteration and mark it as the candidate position; Step 6: Determine the behavior performed by each individual and update the candidate position of each individual; Step 7: Calculate the dynamic boundary, generate the reverse solution corresponding to the current position of each individual and add it to the population, using the reverse solution as the reverse position; Step 8: Calculate the fitness of each individual's current position, updated candidate position, and reverse position, and determine whether to replace the position; Step 9: If the number of iterations t is less than the number threshold T, then set the number of iterations , and return to step 5; if the number of iterations t is greater than or equal to the number threshold T, proceed to step 10; Step 10: Calculate the fitness corresponding to the current position of each individual, sort all fitnesses from large to small, obtain the current position of the individual corresponding to the fitness at the front, obtain the parameter set corresponding to the corresponding parameter label according to the obtained individual current position, and use it as the optimal fermentation control parameter; The steps for obtaining the optimal fermentation control parameters at each device are consistent; In step 5, the method of exploring the candidate positions includes migrating to unknown positions, migrating to known positions, and migrating to other individual positions; wherein the unknown positions are positions that have not been explored by the individual; the known positions are positions that have been explored by the individual but where no individual currently exists; and the other individual positions are positions that have been explored by the individual and where the individual currently exists; Preset judgment coefficient m, ; For each individual, two The random numbers between are and ; The calculation method of the candidate position includes: like ,but ; like ,but ; ; ; ; ; like ,but ; In the formula, is the candidate position of the ith individual, , , , Both A random number between is the current position of a random individual, is the behavior coefficient, is the position coefficient, is the influence coefficient, is the current position of the i-th individual, The candidate position of the i-th individual in the last iteration, for Random numbers within comparing the optimal fermentation control parameters with the fermentation control parameters collected in real time, and identifying deviation parameters in the fermentation control parameters collected in real time; According to the deviation parameters, generate warning instructions or adjustment instructions; If an adjustment instruction is generated, the adjustment amount corresponding to each deviation parameter is calculated and sent to the corresponding device end; if an early warning instruction is generated, the early warning instruction is sent to the corresponding device end.

2. The fermentation process control service method of a fermenter based on cloud computing according to claim 1, characterized in that: The fermentation characteristic parameters include substrate parameters and microbial parameters; the fermentation control parameters include fermentation liquid temperature, fermentation liquid pH, stirring speed and dissolved oxygen concentration; In the step 1, a control range is obtained, and the control range includes a fermentation liquid temperature range, a fermentation liquid pH range, a stirring speed range, and a dissolved oxygen concentration range; a value is randomly selected from each range in the control range to construct a parameter set, and a total of n parameter sets are constructed; In step 3, the population is initialized , including Z individuals, is the Zth individual; The parameter labels and the range of the one-dimensional search space are ; The expression of each individual's initial position is: ; In the formula, is the initial position of the ith individual, is the random coefficient of the ith individual, , ; In step 4, the expression of the fitness function is: ; In the formula, For fitness, It is an indicator of fermentation products; Methods for obtaining fermentation product indicators include: The parameter set corresponding to the parameter label corresponding to the current position of the individual and the fermentation characteristic parameters are used as analysis data, and the analysis data are sequentially input into the trained prediction model to predict the corresponding fermentation yield, product quality and fermentation cycle; wherein the prediction model includes a yield prediction model, a quality prediction model and a cycle prediction model, the output of the yield prediction model is the fermentation yield, the output of the quality prediction model is the product quality, and the output of the cycle prediction model is the fermentation cycle; The predicted fermentation yield, product quality and fermentation cycle are weighted and summed to calculate the fermentation product index.

3. The fermentation process control service method of a fermenter based on cloud computing according to claim 2, characterized in that: In step 6, the execution behavior includes following behavior and competing behavior; Preset switching coefficient q; The method of updating the candidate position of each individual includes: Calculate the fitness of each individual's current position, and mark the current position of the individual with the largest fitness as the best position ; The behavior coefficient Compare with the switching factor q; like ,but ; ; like , then calculate the success probability , success probability The expression is: ; Preset probability threshold , ; The probability of success With probability threshold For comparison, if ,but ;like ,but ; ; ; In the formula, is the candidate position of the i-th individual after update, is the following coefficient, is the competition coefficient, e is the natural constant, for A random number between .

4. The fermentation process control service method of a fermenter based on cloud computing according to claim 3, characterized in that: In step 7, the dynamic boundary is , obtain the parameter label corresponding to the current position of the individual with the smallest fitness in the population and the parameter label corresponding to the current position of the individual with the largest fitness, and compare them. is a parameter label with a small value, It is the parameter label with large value; The methods for generating the inverse solution include: ； In the formula, is the reverse solution of the ith individual, for Random numbers in ; If the generated inverse solution is greater than or less than , it is marked as a transcendental solution, a random number function is used to randomly generate a value within the dynamic boundary, and the randomly generated value is assigned to the transcendental solution; In step 8, the fitness corresponding to the current position, the updated candidate position and the reverse position of each individual is sorted from large to small; If the fitness ranked first is the fitness corresponding to the current position, the current position of the corresponding individual will not be replaced; if the fitness ranked first is not the fitness corresponding to the current position, the current position of the corresponding individual will be replaced with the position corresponding to the fitness ranked first.

5. The fermentation process control service method of a fermenter based on cloud computing according to claim 4, characterized in that: The method for identifying deviation parameters in fermentation control parameters collected in real time comprises: Collect x groups of historical control parameters, where the historical control parameters are fermentation control parameters collected at historical moments; obtain the optimal fermentation control parameters corresponding to the x groups of historical control parameters at corresponding moments, and mark them as historical optimal parameters, where the historical optimal parameters correspond to the historical control parameters one by one; subtract the corresponding parameters in the corresponding historical control parameters from each parameter in each group of historical optimal parameters and take the absolute value, obtain the parameter difference of each parameter in the fermentation control parameters, and mark them as historical difference; calculate the difference mean of each parameter according to the historical difference of each parameter in the fermentation control parameters, and then calculate the difference standard deviation of each parameter; The fermentation control parameters collected in real time are marked as real-time control parameters; each parameter in the real-time control parameters is subtracted from the corresponding parameter in the optimal fermentation control parameters and the absolute value is taken to obtain the parameter difference of each parameter in the fermentation control parameters, and the difference is marked as the real-time difference; the parameters whose real-time difference is greater than or equal to the corresponding difference standard deviation are marked as deviation parameters, and the parameters whose real-time difference is less than the corresponding difference standard deviation are not marked.

6. The fermentation process control service method of a fermenter based on cloud computing according to claim 5, characterized in that: The step of determining and generating a warning instruction or an adjustment instruction comprises: Step a1: Establish fuzzy sets and divide each deviation parameter into multiple fuzzy sets; Step a2: convert the real-time difference corresponding to each deviation parameter into the membership degree of each corresponding fuzzy set through fuzzification technology; Step a3: define fuzzy rules; Step a4: matching the fuzzified real-time difference with the fuzzy rule, performing fuzzy reasoning, and obtaining the fuzzy reasoning result, the fuzzy reasoning result is the membership degree of generating the adjustment instruction and the membership degree of generating the warning instruction; Step a5: Defuzzify the fuzzy reasoning result to obtain the defuzzified result, which is the generated instruction; the defuzzification operation is: according to the fuzzy reasoning result, respectively obtain the membership of the generated adjustment instruction and the membership of the generated warning instruction and compare them, and take the instruction corresponding to the membership with a larger value as the generated instruction.

7. The fermentation process control service method of a fermenter based on cloud computing according to claim 6, characterized in that: Subtract the corresponding parameter in the optimal fermentation control parameter from each deviation parameter to obtain the adjustment amount corresponding to each deviation parameter; After receiving the adjustment amount corresponding to each deviation parameter, the device end controls the device corresponding to the deviation parameter in the adjustment device, and adjusts each deviation parameter. When the value of each deviation parameter reaches the value of the corresponding parameter in the optimal fermentation control parameter, the control of the adjustment device is stopped; wherein the adjustment device includes temperature control equipment, acid and alkali adding equipment, stirring equipment and oxygen supply equipment; While sending the warning instruction to the corresponding device end, the fuzzified real-time difference is also sent to the corresponding device end.

8. The fermentation process control service method of a fermenter based on cloud computing according to claim 7, characterized in that: The method further comprises: if the deviation parameter is the fermentation liquid temperature or the stirring speed, recalculating the adjustment amount corresponding to the dissolved oxygen concentration; The fermentation liquid temperature marked as the deviation parameter is marked as the temperature deviation parameter, and the stirring speed marked as the deviation parameter is marked as the speed deviation parameter; the adjustment amount corresponding to the temperature deviation parameter is marked as the temperature adjustment amount, and the adjustment amount corresponding to the speed deviation parameter is marked as the speed adjustment amount; the influence values ​​of the temperature adjustment amount and the speed adjustment amount on the dissolved oxygen concentration are calculated respectively, and marked as the temperature influence value and the speed influence value respectively; the temperature influence value is added to the speed influence value to obtain the actual value of the dissolved oxygen concentration, and the dissolved oxygen concentration in the optimal fermentation control parameter is subtracted from the actual value of the dissolved oxygen concentration to recalculate the adjustment amount corresponding to the dissolved oxygen concentration; The calculation method of temperature impact value includes: Preset a temperature adjustment threshold, and compare the temperature adjustment amount with the temperature adjustment threshold; If the temperature adjustment amount is less than the temperature adjustment threshold, the temperature adjustment amount is substituted into the pre-built influence function to calculate the corresponding temperature influence value; If the temperature adjustment amount is greater than or equal to the temperature adjustment threshold, the temperature adjustment amount, dissolved oxygen concentration and fermentation characteristic parameters are used as test data, and the test data are input into the trained change prediction model to predict the corresponding dissolved oxygen concentration change value. The training process of the change prediction model is consistent with the training process of the yield prediction model, and both are deep neural network models; the dissolved oxygen concentration change value is the value after the dissolved oxygen concentration changes under the influence of the temperature adjustment amount; the dissolved oxygen concentration change value is subtracted from the dissolved oxygen concentration in the real-time control parameter to obtain the temperature influence value; The calculation method of speed influence value is the same as that of temperature influence value.

9. A fermentation tank fermentation process control service system based on cloud computing, implementing the fermentation tank fermentation process control service method based on cloud computing according to any one of claims 1 to 8, characterized in that: Applied in cloud computing platforms, including: The parameter acquisition module is used to receive the fermentation parameters collected in real time by the equipment end during the fermentation process; the fermentation parameters include fermentation characteristic parameters and fermentation control parameters; Parameter analysis module, used to analyze fermentation characteristic parameters and obtain optimal fermentation control parameters; A deviation identification module, used for comparing the optimal fermentation control parameters with the fermentation control parameters collected in real time, and identifying the deviation parameters in the fermentation control parameters collected in real time; Deviation judgment module, used to judge and generate warning instructions or adjustment instructions according to deviation parameters; The deviation adjustment module is used to calculate the adjustment amount corresponding to each deviation parameter and send it to the corresponding device end if an adjustment instruction is generated; if an early warning instruction is generated, the early warning instruction is sent to the corresponding device end.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the fermentation process management and control service method of the fermenter based on cloud computing according to any one of claims 1 to 8 is implemented.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed, the cloud computing-based fermentation tank fermentation process management and control service method according to any one of claims 1 to 8 is implemented.

Images