Optimized fermentation control method for winter warm type greenhouse organic garbage treatment and application

Abstract

The invention belongs to the technical field of fermentation control, and particularly relates to an optimized fermentation control method for organic garbage treatment of a winter warm type greenhouse and application. According to the method, aiming at the problems of incomplete monitoring, inaccurate pile turning time judgment, fixed control parameters and the like in traditional fermentation control, various temperature sensors, humidity sensors and the like are arranged on the upper layer, the middle layer and the lower layer of a fermentation tank, and the heights of the sensors are dynamically adjusted by utilizing the shrinkage rate of a pile body, so that fermentation parameters are comprehensively obtained. The composting state of the surface of the pile is evaluated through the YOLO algorithm, the pile turning time is judged in combination with a sensor data difference value, and the judgment accuracy is improved. A PID controller is improved, parameters of each link are optimized according to a historical error trend and an error change rate, and PID parameters are globally optimized by adopting an improved Hemma optimization algorithm to adapt to time-varying characteristics in a fermentation process. The organic garbage decomposition efficiency and quality are effectively improved, energy consumption is reduced, the method can be widely applied to agricultural organic garbage treatment, and a more efficient and accurate fermentation control scheme is provided for the field.

Description

Technical Field

[0001] The present invention belongs to the technical field of fermentation control, and particularly relates to an optimized fermentation control method and application for organic waste treatment in a winter-warm greenhouse. Background Technique

[0002] In the treatment of organic waste in a winter-warm greenhouse, it is necessary to control the fermentation of organic waste such as crop straws, vegetable stems and seedlings, and defective fruits and vegetables. In the process of fermentation, controlling the turning operation is crucial for the decomposition efficiency and quality. However, there are significant technical bottlenecks in traditional methods. Existing technologies mostly rely on a single sensor or empirical judgment to determine the turning time, and there are defects in the monitoring method: the fixed-position sensor does not consider the monitoring blind area caused by the volume shrinkage of the compost pile during fermentation, and only collects local parameters, lacking the differential analysis of multi-dimensional data in the upper, middle, and lower layers, making it difficult to comprehensively reflect the decomposition uniformity inside the compost pile; the turning trigger mechanism is single, either only based on internal environmental parameters or only based on the surface apparent state independently, without combining the surface decomposition degree with the gradient change of internal parameters, which is likely to cause premature or late turning, resulting in energy waste or insufficient decomposition. In addition, the parameters of the traditional PID controller are fixed and cannot adapt to the non-linear and time-varying characteristics such as the change of microbial activity and the fluctuation of greenhouse environment during the fermentation process. The adjustment of the turning speed lags behind, which is likely to cause problems such as uneven distribution of temperature and oxygen, and the parameter tuning depends on manual trial and error, lacking a global optimization mechanism, making it difficult to achieve optimal control in a complex environment. Summary of the Invention

[0003] In view of the technical problems existing in the above background technique, the present invention proposes an optimized fermentation control method and application for organic waste treatment in a winter-warm greenhouse.

[0004] To achieve the above object, the technical solution adopted by the present invention includes the following steps:

[0005] S1. Deploy a variety of sensors in the fermentation tank to monitor the parameters during the fermentation process. The sensors include a temperature sensor, a humidity sensor, an oxygen concentration sensor, and a pH value sensor;

[0006] S2. Periodically collect the surface images of the compost pile through a machine vision device, use an image recognition algorithm to evaluate the surface decomposition state of the compost pile, and combine the difference in sensor data between the upper and lower layers of the compost pile to judge the turning operation;

[0007] S3. Build an improved PID controller to control the turning speed. The improvement of the PID controller is

[0008] For the proportional link, dynamically weight the proportional coefficient K p (t) based on the historical error trend. The calculation formula is: where e(i) is the error at the i-th moment, max|e| is the maximum absolute value of the error within the window, and n represents the previous n moments;

[0009] For the integral link, the dynamic integral mode is switched according to the error change rate Δe(t), and the calculation method is: where δ is the transformation rate threshold, and E max is the maximum allowable error;

[0010] For the differential link, a linear prediction model is constructed based on the historical error sequence [e(t - n),..., e(t - 1)] where a and b are parameters to be determined;

[0011] The improved PID controller is finally

[0012] S4. The optimal parameters of the PID controller are obtained by using the improved hippopotamus optimization algorithm; the improved hippopotamus optimization algorithm realizes the optimization of the initial population by improving the reverse learning strategy;

[0013] S5. Multiple optimized control parameters are obtained to control the turning operation and optimize the fermentation process.

[0014] Preferably, in step S1, the positions of the sensors are divided into upper, middle, and lower layers; the upper layer is at 5 / 6 of the distance from the bottom of the pool, the middle layer is at 1 / 2 of the distance from the bottom of the pool, and the lower layer is at 1 / 6 of the distance from the bottom of the pool, and each layer includes the above four types of sensors, and the types and positions are evenly distributed.

[0015] Preferably, the height positions of the various sensors have dynamic height adjustment, and the calculation method of the dynamic height adjustment is where H is the original deployment height, H′ is the corrected deployment height, λ is the shrinkage rate of the pile body, t f represents the time of fermentation that has occurred, and T f represents the total estimated fermentation time.

[0016] Preferably, in step S2, the image recognition algorithm is used to evaluate the composting state on the surface of the pile through the YOLO algorithm, and the surface image data of the pile is input to obtain the composting state score M.

[0017] Preferably, the calculation method for judging the turning operation by combining the obtained composting state score M and the sensor error is: K·(1 - M) ≥ θ, where θ is the turning threshold, K is the comprehensive difference coefficient of the sensor data, and K = ω W ·ΔW + ω S ·ΔS + ω Y ·ΔY + ω P· |ΔP|, where ΔW, ΔS, ΔY, and ΔP are the upper and lower layer data differences of the temperature sensor, humidity sensor, oxygen concentration sensor, and pH value sensor respectively, and ω W , ω S , ω Y , ω P are the corresponding weight coefficients, and the turning pile operation is triggered when it is greater than or equal to the turning pile threshold.

[0018] Preferably, the implementation of the improved hippopotamus optimization algorithm is as follows: Initialize the population parameters, set the maximum number of iterations and the population size; during the iteration process, simulate the survival behavior of hippopotamuses to update the individual positions, calculate the fitness values of the PID controller parameters corresponding to each individual, and retain high-quality individuals through a competition mechanism. When the termination condition is met, output the optimal PID controller parameters.

[0019] Preferably, when initializing the population, an improved opposition-based learning strategy is adopted, a dynamic scaling factor is introduced, and according to the search space dimension and the current iteration stage, the parameter offset of the opposition solution generation is adjusted; for each individual parameter dimension, the search direction of the opposition-based learning is dynamically corrected based on the current population distribution entropy value.

[0020] Preferably, the process of optimizing the initialized population based on the improved opposition-based learning strategy includes: in the population initialization stage, for the randomly generated initial individuals, use the improved opposition-based learning strategy to generate an opposition individual set; construct a two-way evaluation mechanism, calculate the fitness of the initial individuals and the opposition individuals at the same time, and screen out the individuals with better fitness to form a new population; perform diversity detection on the new population. If it is lower than the preset threshold, supplement and generate individuals with spatial distribution differences through the improved opposition-based learning strategy.

[0021] Preferably, an optimized fermentation control method for organic waste treatment in a warm winter greenhouse described in any one of the above claims 1-8 is applied to agricultural organic waste treatment.

[0022] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Deploy multiple types of sensors on the upper, middle, and lower layers of the fermentation tank and dynamically adjust the height to solve the monitoring blind area problem; use the YOLO algorithm to evaluate the composting state on the surface of the pile body, and combine the sensor data difference to judge the turning pile operation to improve the judgment accuracy; build an improved PID controller, adopt the improved hippopotamus optimization algorithm to optimize the parameters, and dynamically adapt to the time-varying characteristics of the fermentation process. Description of the Drawings

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other accompanying drawings can be obtained based on these drawings without creative efforts.

[0024] Figure 1 It is a structural flowchart of an optimized fermentation control method for organic waste treatment in a warm winter greenhouse. Specific embodiments

[0025] In order to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described below with reference to the accompanying drawings and embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0026] Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the present invention is not limited to the limitations of the specific embodiments disclosed in the following specification.

[0027] Embodiment. During the process of organic waste treatment in a warm winter greenhouse, the accuracy of fermentation control directly affects the efficiency and quality of organic waste composting. Traditional methods have problems such as incomplete monitoring, single judgment of the turning pile timing, and fixed control parameters, resulting in energy consumption waste and incomplete composting. In order to achieve the effects of comprehensively monitoring the fermentation state, accurately judging the turning pile timing, dynamically optimizing control parameters to improve the fermentation efficiency and quality, and solving the deficiencies of traditional fermentation control technologies, the present invention proposes an optimized fermentation control method for organic waste treatment in a warm winter greenhouse. The specific process of the present invention is as Figure 1 shown.

[0028] First, a variety of sensors are deployed in the fermentation tank to monitor the parameters during the fermentation process. The sensors include a temperature sensor, a humidity sensor, an oxygen concentration sensor, and a pH value sensor. To solve the problems of monitoring blind spots of traditional single sensors and monitoring deviations caused by the shrinkage of the pile body, the present invention deploys temperature, humidity, oxygen concentration, and pH value sensors in the upper layer (5 / 6 from the bottom of the tank), middle layer (1 / 2 from the bottom of the tank), and lower layer (1 / 6 from the bottom of the tank) of the fermentation tank respectively. Four types of sensors are evenly distributed in each layer to obtain multi-dimensional data.

[0029] In order to eliminate the monitoring blind spots caused by the shrinkage of the compost pile during fermentation and ensure that the sensors continuously cover the effective monitoring area, and to solve the problem that the traditional fixed-position sensors cause data distortion due to the volume change of the compost pile and cannot accurately reflect the fermentation state, the present invention adopts the technical solution of a dynamic height adjustment mechanism. Specifically, on the basis of deploying sensors in the upper, middle, and lower layers of the fermentation tank, the shrinkage rate of the compost pile is introduced, and the height positions of various sensors have dynamic height adjustment. The calculation method of the dynamic height adjustment is where H is the original deployment height, H' is the corrected deployment height, λ is the shrinkage rate of the compost pile, and t f represents the elapsed fermentation time, and T f represents the total estimated fermentation time. The deployment height of the sensor is corrected in real time, so that the sensor height dynamically adapts to the shrinkage of the compost pile, ensuring that the sensor always remains in the core fermentation area in the middle of the compost pile, providing real-time and accurate data support for the judgment of the turning pile timing and the control of the fermentation process.

[0030] Next, considering the problem of the single traditional turning pile triggering mechanism and the easy deviation of the turning pile timing, the present invention periodically collects the surface images of the compost pile through a machine vision device, uses an image recognition algorithm to evaluate the compost state on the surface of the pile, and combines the difference in the data of the upper and lower layer sensors of the pile to judge the turning pile operation. Specifically, the evaluation of the compost state on the surface of the pile by using an image recognition algorithm is through the YOLO algorithm. By inputting the surface image data of the compost pile, a compost state score M is obtained. First, the YOLO algorithm needs to be trained. A large number of surface images of compost piles with different degrees of compost maturity are collected and divided into a training set, a validation set, and a test set. Different compost maturity level labels are marked for the compost maturity features such as color and texture in the images. The training set is input into the YOLO algorithm model, and the model continuously learns the features of the compost pile surface and the corresponding compost maturity level relationship through a convolutional neural network. During the training process, the parameters of the model are continuously adjusted to improve the recognition accuracy of different degrees of compost maturity. The validation set is used to evaluate the model performance to prevent overfitting. After training, the surface image data of the compost pile to be evaluated is input into the trained model. The model will identify the features of the compost pile in the image, judge its compost maturity according to the learned pattern, and output the corresponding compost state score M. The calculation method for judging the turning pile operation by combining the obtained compost state score M and the sensor error is: K·(1 - M) ≥ θ, where θ is the turning pile threshold, and K is the comprehensive difference coefficient of the sensor data, K = ω W ·ΔW + ω S ·ΔS + ω Y ·ΔY + ω P ·|ΔP|, where ΔW, ΔS, ΔY, and ΔP are the upper and lower layer data differences of the temperature sensor, humidity sensor, oxygen concentration sensor, and pH value sensor respectively, and ω W , ω S , ω Y , ωP is the corresponding weight coefficient, and the turning operation is triggered when it is greater than or equal to the turning threshold.

[0031] Next, considering that when the existing PID controls the turning speed, there will be a problem that the controller parameters are fixed and it is difficult to adapt to the time-varying characteristics of the fermentation process. The present invention first improves the PID controller. Specifically, the improvement of the PID controller is as follows: For the proportional link, the proportional coefficient K p (t) is dynamically weighted based on the historical error trend, and the calculation formula is: where e(i) is the error at the i-th moment, max|e| is the maximum absolute value of the error within the window, and n represents the previous n moments; for the integral link, the dynamic integral mode is switched according to the error change rate Δe(t), and the calculation method is: where δ is the conversion rate threshold, and E max is the maximum allowable error. For the derivative link, a linear prediction model is constructed based on the historical error sequence [e(t - n),..., e(t - 1)] where a and b are parameters to be determined; the improved PID controller is finally In the proportional link, the proportional coefficient of the traditional PID controller is fixed, making it difficult to adapt to the dynamic changes of the fermentation process. The present invention calculates the proportional coefficient by dynamically weighting based on the historical error, which can enhance the response to the error trend. For example, when the error shows an increasing trend during the fermentation process, the proportional coefficient will be adjusted accordingly, enabling the system to quickly respond to the error, improving the response speed of the system, and avoiding further expansion of the error. In terms of the integral link, the traditional integral link is prone to the problem of integral saturation. The present invention switches the integral mode according to the error change rate, and adopts a dynamic integral coefficient when the error change rate is large, which can effectively avoid integral saturation, enable the system to stably control even when the error is large, and prevent the system from overshooting. A linear prediction model is constructed for the derivative link to predict future errors. During the fermentation process, changes in microbial activity, environmental fluctuations, etc. will cause the error to change with a certain trend. By predicting future errors, the system can adjust the control parameters in advance, significantly improving the real-time performance and accuracy of the turning speed adjustment, and ensuring that parameters such as the temperature and oxygen concentration of the pile body are stable within a suitable range.

[0032] Meanwhile, on the basis of improving the PID controller, an improved hippopotamus optimization algorithm is adopted to obtain the optimal parameters of the PID controller. The population is initialized by improving the reverse learning strategy, a dynamic scaling factor and the population distribution entropy value are introduced to correct the search direction, and the PID parameters are globally optimized. Specifically, first, the integral of the squared error of the PID control performance index is defined as the objective function, the value ranges of the proportional coefficient, integral coefficient, and differential coefficient are set, and the algorithm parameters such as the population size and the maximum number of iterations are initialized. Then, the initial population is generated by improving the reverse learning strategy. Specifically, first, the initial range of the PID parameters is determined. For each parameter dimension of each initial individual, reverse individuals are generated according to the parameter boundary symmetry principle. The reverse individuals and the initial individuals are symmetrically distributed in the search space, so as to expand the coverage of the initial solutions and avoid local biases caused by random initialization. Then, a two-way evaluation mechanism is constructed. The fitness of the initial individuals and the reverse individuals is calculated simultaneously, and the top N individuals are selected according to the fitness to form a new population. This mechanism ensures that each individual in the initial population is a better solution in the current area, improving the overall quality of the initial population. Subsequently, the diversity of the new population is detected. By calculating the distribution entropy value between individuals, the uniformity of individuals in the parameter space is measured to evaluate the diversity. If the entropy value is lower than the preset threshold, it indicates that the individual distribution is too concentrated. Then, individuals with different spatial distributions are supplemented and generated by improving the reverse learning strategy: for the parameter dimensions where the population is aggregated, the perturbation amplitude of the reverse learning is dynamically adjusted, and new individuals are generated directionally in the sparse areas. After fitness screening, they are added to the population until the diversity reaches the standard. Then, it enters the iterative optimization stage. In each iteration, the population distribution entropy value is first calculated to evaluate the uniformity of individual distribution, and then the scaling factor is dynamically adjusted according to the iteration process to change the search step size (large step size for global exploration in the initial stage and small step size for local fine search in the later stage). Subsequently, the individual positions are updated by simulating the survival behavior of hippopotamuses. During the process, the population entropy value is introduced to correct the search direction. When the individual distribution is concentrated, it is guided to explore in the sparse areas to avoid local optima. After the update, it is checked whether the parameters exceed the preset range, and the out-of-bounds values are corrected at the boundaries. Then, the fitness of the new individuals is calculated, and the high-quality individuals are retained through the competition mechanism. The above iteration is repeated until the maximum number of iterations is reached. After termination, the parameters corresponding to the individual with the optimal fitness are output and applied to the PID controller.

[0033] The above are only the preferred embodiments of the present invention, and are not intended to limit the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention still belong to the protection scope of the technical solution of the present invention.

Claims

1. An optimized fermentation control method for organic waste treatment in a winter warm greenhouse, characterized in that, It includes the following steps: S1. Deploy a variety of sensors in the fermentation tank to monitor the parameters during the fermentation process. The sensors include a temperature sensor, a humidity sensor, an oxygen concentration sensor, and a pH value sensor. S2. Periodically collect the images of the heap surface through a machine vision device, use an image recognition algorithm to evaluate the composting state of the heap surface, and combine the difference in the sensor data of the upper and lower layers of the heap to make a judgment on the turning operation. S3. Build an improved PID controller to control the turning speed. The improvement of the PID controller is that for the proportional link, the proportional coefficient K p (t) is dynamically weighted based on the historical error trend, and the calculation formula is: where e(i) is the error at the i-th moment, max|e| is the maximum value of the absolute value of the error within the window, and n represents the previous n moments; For the integral link, the dynamic integral mode is switched according to the error change rate Δe(t), and the calculation method is as follows: where δ is the transformation rate threshold, and E max is the maximum allowable error; For the derivative link, a linear prediction model is constructed based on the historical error sequence [e(t - n),..., e(t - 1)] where a and b are parameters to be determined; The improved PID controller finally is S4. Use an improved hippopotamus optimization algorithm to obtain the optimal parameters of the PID controller. The improved hippopotamus optimization algorithm realizes the optimization of the initial population by improving the reverse learning strategy. S5. Obtain multiple optimized control parameters to control the turning operation and optimize the fermentation process.

2. The optimized fermentation control method for organic waste treatment in a winter-warm greenhouse according to claim 1, characterized in that, In step S1, the positions of the sensors are divided into upper, middle, and lower layers. The upper layer is at 5 / 6 of the pool bottom, the middle layer is at 1 / 2 of the pool bottom, and the lower layer is at 1 / 6 of the pool bottom. And each layer includes the above four sensors, and the types and positions are evenly distributed.

3. The optimized fermentation control method for organic waste treatment in a winter-warm greenhouse according to claim 2, wherein The height positions of the various types of sensors have dynamic height adjustment, and the calculation method of the dynamic height adjustment is where H is the original deployment height, H′ is the corrected deployment height, λ is the stack shrinkage rate, t f represents the time of fermentation that has occurred, and T f represents the total estimated fermentation time.

4. An optimized fermentation control method for organic waste treatment in a warm winter greenhouse according to claim 1, characterized in that, In step S2, using the image recognition algorithm to evaluate the composting state of the heap surface is through the YOLO algorithm. Input the image data of the heap surface to obtain the composting state score M.

5. An optimized fermentation control method for organic waste treatment in a winter-warm greenhouse according to claim 4, characterized in that The calculation method for judging the turning pile operation by combining the obtained compost maturity score M and the sensor error is: K·(1 - M)≥θ, where θ is the turning pile threshold, K is the comprehensive difference coefficient of sensor data, and K = ω W ·ΔW + ω S ·ΔS + ω Y ·ΔY + ω P ·|ΔP|, where ΔW, ΔS, ΔY, and ΔP are the upper and lower layer data differences of the temperature sensor, humidity sensor, oxygen concentration sensor, and pH value sensor respectively, and ω W , ω S , ω Y , ω P are the corresponding weight coefficients, and the turning pile operation is triggered when it is greater than or equal to the turning pile threshold.

6. An optimized fermentation control method for organic waste treatment in a winter-warm greenhouse according to claim 1, characterized in that, The implementation of the improved hippopotamus optimization algorithm is as follows: Initialize the population parameters, set the maximum number of iterations and the population size. During the iteration process, simulate the survival behavior of the hippopotamus to update the individual positions, calculate the fitness values of the PID controller parameters corresponding to each individual, and retain the high-quality individuals through a competition mechanism. When the termination condition is met, output the optimal PID controller parameters.

7. An optimized fermentation control method for organic waste treatment in a winter-warm greenhouse according to claim 6, characterized in that, When initializing the population, adopt an improved reverse learning strategy, introduce a dynamic scaling factor, and adjust the parameter offset of the reverse solution generation according to the search space dimension and the current iteration stage. For each individual parameter dimension, dynamically correct the search direction of the reverse learning based on the current population distribution entropy value.

8. An optimized fermentation control method for organic waste treatment in a winter-warm greenhouse according to claim 7, characterized in that, The process of optimizing the initial population based on the improved reverse learning strategy includes: in the population initialization stage, for the randomly generated initial individuals, use the improved reverse learning strategy to generate a set of reverse individuals; construct a two-way evaluation mechanism, calculate the fitness of the initial individuals and the reverse individuals at the same time, and select the individuals with better fitness to form a new population; perform diversity detection on the new population. If it is lower than the preset threshold, supplement and generate individuals with spatial distribution differences through the improved reverse learning strategy.

9. The optimized fermentation control method for the treatment of organic waste in a winter warm greenhouse according to any one of claims 1-8 above is applied to the treatment of agricultural organic waste.

Images