Chicken feed production line intelligent ingredient regulation and control system based on Internet of Things edge calculation

Abstract

The invention discloses a chicken feed production line intelligent ingredient regulation and control system based on Internet of Things edge computing, which relates to the technical field of feed production management and comprises an intelligent sensing terminal module, an edge computing center module, a cloud management and control platform module, a network communication module, an intelligent algorithm control module and an execution control module. The intelligent sensing terminal module is used for collecting ingredient production data of a chicken feed production line in real time through a sensor. Through a double-strategy cooperation mechanism of the intelligent algorithm control module, raw material conveying lag is compensated in advance through model prediction control, disturbance of humidity, vibration and the like is corrected in real time through self-adaptive fuzzy control, batching errors are effectively reduced, uneven nutrition of chicken flocks caused by formula errors is avoided, and a high-precision execution mechanism in the execution control module is matched, so that the control accuracy is improved. Through frequency conversion control dynamic matching feeding rotation speed and near infrared spectrum uniformity monitoring, the nutrition consistency is improved, and the breeding growth income is remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of feed production management, in particular to a chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing. BACKGROUND

[0002] Chicken feed is a nutrient substance required for chicken growth, egg production and health maintenance, to ensure normal physiological function and efficient production, with the rapid development of large-scale breeding and smart agriculture, chicken feed production presents the upgrading demand of high precision, high flexibility and high synergy.

[0003] At present, the formula adjustment error rate of the traditional chicken feed production line is high, which easily leads to uneven growth of chicken flock, high feed-meat ratio, and sudden condition response lag of raw material moisture fluctuation, caking and the like, and cannot meet the real-time demand of dynamic ingredient, resulting in unstable batch quality. SUMMARY

[0004] The purpose of the present application is to provide a chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing, to solve the problem of high formula adjustment error rate of the traditional chicken feed production line, which easily leads to uneven growth of chicken flock, high feed-meat ratio, and sudden condition response lag of raw material moisture fluctuation, caking and the like, and cannot meet the real-time demand of dynamic ingredient, resulting in unstable batch quality.

[0005] To achieve the above purpose, the present application provides the following technical scheme: a chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing, comprising an intelligent sensing terminal module, an edge computing hub module, a cloud control platform module, a network communication module, an intelligent algorithm control module and an execution control module.

[0006] The intelligent perception terminal module is used to collect the batching production data of the chicken feed production line in real time through sensors, and the production data includes raw material identification information, weight monitoring information, environmental perception information and device state information; the edge computing hub module (2) is used to provide real-time data processing and local control decision processing for the data collected by the intelligent perception terminal module (1), realize real-time control, device diagnosis and digital twin simulation, process data locally and quickly, complete dynamic batching, device collaboration and fault emergency processing; the cloud management and control platform module is used to provide production visual monitoring, formula optimization, quality traceability and big data analysis, provide integrated interface support and internal and third-party platform data interconnection, provide remote device management and operation, realize unified management and decision optimization of multiple production lines; the network communication module is used to provide data communication channels between modules, control instruction execution priority transmission principle in downlink data, and sensor data is uploaded asynchronously using MQTT protocol; the intelligent algorithm control module is used to provide dynamic batching control algorithm optimization batching strategy to the edge computing hub module, and provide device health management algorithm to realize device fault prediction and anomaly detection, and the algorithm is deployed in the edge and the cloud; the execution control module is used to receive the control instruction of the edge computing hub module, and the actuator performs actions according to the control instruction.

[0007] Preferably, the intelligent perception terminal module includes a raw material state monitoring module and a device running state module.

[0008] The raw material monitoring module is used to obtain the weight and material level perception data of each raw material bin of the chicken feed, and simultaneously bind the raw material batch information based on RFID technology.

[0009] The device running state module is used to collect the state information of the feeding device through vibration sensors and temperature sensors, simultaneously monitor the real-time load of the motor using current sensors, and collect the environmental parameters of the chicken feed production workshop through temperature and humidity sensors.

[0010] Preferably, the edge computing hub module includes an edge computing node cluster module, a device synchronization control module, an edge intelligent service module and an edge-cloud collaborative data processing module.

[0011] The edge computing node cluster module realizes parallel processing of control algorithms and AI inference based on a heterogeneous computing platform, realizes concurrent access of sensor groups through an edge gateway, and simultaneously provides 72-hour local data caching;

[0012] The device synchronization control module is used to realize nanosecond-level time synchronization of sensors, controllers and actuators through a clock protocol, and control the device group to keep action synchronization.

[0013] The edge intelligent service module is used for real-time state diagnosis through equipment running state data, and simultaneously provides real-time mapping equipment state of a production line, and supports edge node simulation of different formula batching processes.

[0014] The edge cloud cooperative data processing module is used for data hierarchical processing between edge equipment and the cloud, including local storage processing, timing upload processing and asynchronous archiving processing.

[0015] Preferably, the cloud management and control platform module includes a production management and control module, a quality traceability module, a raw material procurement management module and an integrated interface module.

[0016] The production management and control module is used for providing production line running state visual display and chicken feed formula full life cycle management.

[0017] The quality traceability module realizes traceability query based on raw material batch information, mines the correlation between batching parameters and finished product quality through an association rule algorithm, and performs quality anomaly analysis.

[0018] The raw material procurement management module predicts raw material price fluctuations based on an LSTM neural network, assists procurement decision-making, and supports dynamic balance of feed inventory.

[0019] The integrated interface module provides a data intercommunication interface, synchronizes production plans, raw material consumption and cost data in real time, and simultaneously provides remote diagnosis assistance services.

[0020] Preferably, the network communication module includes a network layered transmission module and a secure communication protection module.

[0021] The network layered transmission module adopts a layered design of a field layer, a control layer and a management layer, the field layer connects sensors and actuators, realizes basic data acquisition and instruction issuance, the control layer is used for real-time communication between edge nodes, and the management layer completes data interaction between the edge and the cloud by means of a 5G industrial router and a wired VPN, and supports networking connection between different factories.

[0022] The secure communication protection module is used for constructing a security mechanism through dynamic device authentication and minimum permission principle verification of device access, and performing data encryption transmission.

[0023] Preferably, the intelligent algorithm control module includes a dynamic batching algorithm module and a device health management algorithm module.

[0024] The dynamic batching algorithm module includes a model predictive control module and an adaptive fuzzy control module, the model predictive control module utilizes an LSTM prediction model to rollingly optimize future 10-second feeding strategies based on real-time raw material fluidity data, and the adaptive fuzzy control module dynamically adjusts PID control parameters through fuzzy logic for nonlinear scenes of material caking and humidity mutation.

[0025] The device health management algorithm module includes vibration signal processing, anomaly detection and device remaining life prediction.

[0026] Preferably, the model predictive control module calculation formula is as follows:

[0027] x(k+1) = Ax(k) + Bu(k) + w(k);

[0028] y(k) = Cx(k) + v(k);

[0029] In the formula, x(k) is a device state vector at the current time (k), x(k+1) is a device state vector at the next time (k+1), u(k) is a device control input, including a feeder frequency and a valve opening, y(k) is a system output, A, B and C are a state transition matrix, a control matrix and an output matrix, w(k) and v(k) are process noise and measurement noise, respectively.

[0030] Rolling optimization with device future operation steps N p The minimum of inner prediction error and control input cost minimization is taken as an objective, actual output is close to a target value and control input change is gentle, and a rolling optimization output value J calculation formula is as follows:

[0031]

[0032] In the formula, y r (k+i) is a target batching amount at the i-th future step, Q is an output error weight matrix, R is a control input change weight matrix, N u is a control time domain, N u ≤N p , N p =100step, N u =50step, corresponding to 10-second rolling optimization;

[0033] The constraint condition is as follows:

[0034] u min ≤u(k+i)≤u max , i=0, 1,..., N u -1;

[0035] y min ≤y(k+i)≤y max , i=1,..., N p .

[0036] Preferably, the adaptive fuzzy control process includes the following steps:

[0037] A1, determine input variables based on real-time acquisition of parameters in the batching process: batching error, error rate, raw material humidity, and feeding equipment vibration value;

[0038] A2, fuzzy processing of input variables, converting precise quantities into fuzzy language variables;

[0039] A3, constructing a fuzzy rule base based on the process experience of chicken feed batching, and outputting control decisions through fuzzy reasoning algorithm to obtain a fuzzy output set;

[0040] A4, converting the fuzzy output obtained by reasoning into precise control instructions for the actuator;

[0041] A5, real-time monitoring of the final batching error, adjustment frequency, and actuator energy consumption through the edge computing hub module (2), and dynamically optimizing the fuzzy rule base.

[0042] Preferably, the execution control module includes a feeding control module and a mixing and conveying module;

[0043] The feeding control module includes main raw material feeding and auxiliary material addition, and the main raw material feeding is controlled by a combination of a servo screw feeder and a pneumatic butterfly valve, and the auxiliary material addition is realized by an electromagnetic metering pump and a vibrating feeder.

[0044] The mixing and conveying module controls the operation of the double-shaft paddle mixer through frequency control and mixing uniformity monitoring.

[0045] Compared with the prior art, the present application has the following advantages:

[0046] 1. In the present application, the double-strategy collaborative mechanism of the intelligent algorithm control module uses model predictive control to compensate for raw material transportation lag in advance, and self-adaptive fuzzy control to correct humidity, vibration and other disturbances in real time, effectively reducing the batching error and avoiding uneven nutrition of the chicken flock caused by formula error. Combined with the high-precision actuator in the execution control module, the feeding speed is dynamically matched through frequency control, and the near-infrared spectrum uniformity monitoring is monitored, which improves the consistency of nutrition and significantly improves the breeding growth benefit.

[0047] 2. In the present application, the intelligent sensing terminal module acquires raw material humidity, caking state and equipment operation data in real time, and the edge computing hub module realizes parallel processing of control algorithm and AI reasoning based on a heterogeneous computing platform, which shortens the response time to sudden conditions such as raw material moisture fluctuation and caking, and meets the real-time demand of dynamic batching.

[0048] 3. In the present application, the quality tracing module is based on the raw material batch information bound by RFID, and combines the association rule algorithm to mine the association relationship between batching parameters and finished product quality, which can quickly locate the source of quality abnormalities, trace and adjust in time, and reduce the quality fluctuation between batches.

[0049] 4、The present application, through the production control module realizes production line state visualization and formula full life cycle management, supports many formulas such as laying hens, broilers one-key switching, change production time is shortened, improve flexible production capacity, raw material procurement management module through LSTM neural network predicts raw material price fluctuation, combined with inventory dynamic balance strategy, make procurement cost reduce, at the same time, through the shelf life warning avoids raw material expiration waste. BRIEF DESCRIPTION OF DRAWINGS

[0050] Fig. 1 It is a system block diagram of the chicken feed production line intelligent ingredient regulation system based on the Internet of Things edge computing of the application;

[0051] Fig. 2 It is a system block diagram of the network communication module of the chicken feed production line intelligent ingredient regulation system based on the Internet of Things edge computing of the application;

[0052] Fig. 3 It is an execution control module control flow chart of the chicken feed production line intelligent ingredient regulation system based on the Internet of Things edge computing of the application.

[0053] In the figure: 1, intelligent sensing terminal module; 2, edge computing hub module; 3, cloud control platform module; 4, network communication module; 5, intelligent algorithm control module; 6, execution control module. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0055] Referring to Figs. 1-3 As shown in the figure: a chicken feed production line intelligent ingredient regulation system based on the Internet of Things edge computing, including intelligent sensing terminal module 1, edge computing hub module 2, cloud control platform module 3, network communication module 4, intelligent algorithm control module 5 and execution control module 6;

[0056] Intelligent sensing terminal module 1 is used for collecting the ingredient production data of chicken feed production line in real time through sensor, and the production data includes raw material identification information, weight monitoring information, environment sensing information and equipment state information; specifically, including raw material state monitoring module and equipment running state module;

[0057] The raw material monitoring module is used to acquire weight and material level sensing data of each raw material bin of the chicken feed, and bind raw material batch information based on RFID technology, so as to realize raw material whole life cycle traceability, solve the problems of traditional manual recording error and batch traceability difficulty, and provide accurate data support for quality abnormality analysis;

[0058] The equipment running state module is used to collect state information of the feeding equipment through vibration sensors and temperature sensors, monitor real-time load of the motor through current sensors, and collect environmental parameters of the chicken feed production workshop through temperature and humidity sensors, so as to capture real-time abnormal precursor data of the equipment, avoid production interruption caused by sudden equipment failure, and improve production line stability;

[0059] The edge computing hub module 2 is used to provide real-time data processing and local control decision processing for the data collected by the intelligent sensing terminal module 1, realize real-time control, equipment diagnosis and digital twin simulation of the equipment, and perform local rapid processing of the data to complete dynamic batching, equipment cooperation and fault emergency; specifically, the edge computing hub module 2 includes an edge computing node cluster module, an equipment synchronous control module, an edge intelligent service module and an edge-cloud collaborative data processing module;

[0060] The edge computing node cluster module realizes parallel processing of control algorithms and AI inference based on a heterogeneous computing platform, realizes concurrent access of a sensor group through an edge gateway, provides 72-hour local data caching at the same time, ensures that production can still be maintained when the network is disconnected, solves the delay and network disconnection risk caused by traditional dependence on the cloud, and improves the anti-interference ability of the system;

[0061] The equipment synchronous control module is used to realize nanosecond-level time synchronization of sensors, controllers and actuators through a clock protocol, control the equipment group to perform synchronous actions, maintain action cooperation, reduce multi-device cooperation errors by ensuring action cooperation of sensors, controllers and actuators, and improve batching accuracy;

[0062] The edge intelligent service module is used to perform real-time state diagnosis through equipment running state data, provide real-time mapping of equipment states of the production line, support the edge node to simulate the batching process of different formulations, simulate the effects of formulations through digital twin, verify the feasibility of new formulations without stopping production, shorten process debugging time, and support small-batch customized production;

[0063] The edge-cloud collaborative data processing module is used for hierarchical processing of data between edge devices and the cloud, including local storage processing, timing upload processing and asynchronous archiving processing, which reduces the bandwidth pressure of the cloud by processing data hierarchically, while ensuring that key data is not lost;

[0064] The hierarchical data processing includes the following three levels:

[0065] T1: including real-time batching error and equipment instantaneous load, local memory storage, retaining for 2 hours;

[0066] T2: Including batch ingredient records and device state daily reports, uploaded to the cloud every 5 minutes, and the edge node caches for 72 hours;

[0067] T3: Including historical production reports and model training samples, archived to the cloud asynchronously every morning, and the edge node does not store for a long time;

[0068] The cloud management platform module 3 is used to provide production visualization monitoring, formula optimization, quality traceability and big data analysis, provide integrated interface support and internal and third party platform data intercommunication, provide remote device management and operation and maintenance, realize unified management and decision optimization of multiple production lines; Specifically, it includes a production management module, a quality traceability module, a raw material procurement management module and an integrated interface module;

[0069] The production management module is used to provide visual display of the running state of the production line and full life cycle management of chicken feed formula;

[0070] The quality traceability module realizes traceability query based on raw material batch information, mines the correlation between ingredient parameters and finished product quality through association rule algorithm, performs quality anomaly analysis, and realizes rapid positioning of quality abnormal source;

[0071] The raw material procurement management module predicts raw material price fluctuations based on LSTM neural network, assists procurement decision, supports dynamic balance of feed inventory, reduces procurement cost, and at the same time avoids waste of expired raw materials;

[0072] The prediction of raw material price fluctuation adopts a rolling prediction mechanism: the latest 30-day data is used to update the model input every day to generate a 7-day price prediction sequence

[0073] By converting the price prediction into specific procurement actions, the inventory dynamic balance is realized:

[0074] Purchase timing judgment:

[0075] If the predicted price rises in the next 7 days (rise≥3%), and the current inventory is < safe threshold, trigger immediate purchase;

[0076] If the predicted price falls (fall≥3%), and the inventory is sufficient, delay the purchase.

[0077] Purchase quantity calculation: combined with the predicted price and the economic order quantity model, the formula is: purchase quantity = max (safe inventory - current inventory, EOQ 修正 );

[0078] Safe inventory = maximum daily demand × longest replenishment period × risk factor;

[0079] Maximum daily demand: take 1.2 times of the average daily demand in the past 30 days;

[0080] Longest replenishment cycle: average supplier delivery cycle + 2 days;

[0081] Risk coefficient: determined according to the price fluctuation range predicted by LSTM;

[0082]

[0083] In the formula: D is the monthly average demand, determined by the monthly production plan; S is the single purchase cost, including transportation fees, labor coordination fees, and other fixed costs; H is the monthly inventory holding cost; k is the price trend coefficient: if the predicted future price increase is ≥ 3%, k = 1.2; if the predicted price drop is ≥ 3%, k = 0.8; when the price is stable, k = 1.

[0084] Inventory linkage: the prediction results are synchronized to the inventory management module. When the shelf life of raw materials is approaching expiration and the predicted price is falling, the inventory is consumed preferentially to avoid waste.

[0085] The integrated interface module provides a data interconnection interface, synchronizes production plans, raw material consumption, and cost data in real time, and provides remote diagnosis assistance services, synchronizes data with ERP / MES systems in real time, realizes production plan and cost accounting automation, reduces manual input errors, and improves management efficiency;

[0086] The network communication module 4 is used to provide a data communication channel between modules. The control instructions in the downlink data are transmitted with priority, and the sensor data is uploaded asynchronously using the MQTT protocol. Specifically, it includes a network layered transmission module and a secure communication protection module.

[0087] The network layered transmission module adopts a layered design of field layer, control layer, and management layer. The field layer connects sensors and actuators to realize basic data acquisition and instruction issuance. The control layer is used for real-time communication between edge nodes. The management layer completes data interaction between the edge and the cloud with the help of a 5G industrial router and a wired VPN, supports networking connection between different factories, ensures priority transmission of control instructions, and asynchronous upload of sensor data without occupying critical bandwidth, ensuring the stability of real-time control.

[0088] The secure communication protection module is used to build a security mechanism by dynamically authenticating devices and verifying access with the principle of least privilege, performing data encryption transmission, preventing data tampering through dynamic authentication and encryption transmission, and intercepting attacks through industrial firewalls to ensure the safety of production data and devices.

[0089] The intelligent algorithm control module 5 is used to provide dynamic batching control algorithms to the edge computing hub module 2 to optimize batching strategies, and to provide device health management algorithms to realize device fault prediction and anomaly detection. The algorithms are deployed on the edge and in the cloud. Specifically, it includes a dynamic batching algorithm module and a device health management algorithm module.

[0090] The dynamic batching algorithm module includes a model predictive control module and an adaptive fuzzy control module. The model predictive control module uses an LSTM prediction model to rollingly optimize the feeding strategy for the next 10 seconds based on real-time data of raw material fluidity. The model predictive control calculation formula is as follows:

[0091] x(k+1) = Ax(k) + Bu(k) + w(k);

[0092] y(k) = Cx(k) + v(k);

[0093] In the formula, x(k) is a device state vector, the state vector dimension is 3, including raw material flow, feeder speed and valve opening, u(k) is a control input, including feeder frequency and valve opening, y(k) is a system output, A, B and C are state transition matrix, control matrix and output matrix, w(k) and v(k) are process noise and measurement noise respectively;

[0094] The rollingly optimized future operation step N p The minimum of the inner prediction error and the control input cost is taken as the target, so that the actual output approaches the target value and the control input changes gently. The rollingly optimized output value J calculation formula is as follows:

[0095]

[0096] In the formula, y r (k+i) is the target batching amount of the i-th future step, Q is an output error weight matrix, R is a control input change weight matrix, N u is a control time domain, N u ≤N p , N p =100step, N u =50step, corresponding to 10 seconds of rollingly optimization;

[0097] Constraint conditions:

[0098] u min ≤u(k+i)≤u max , i=0,1,…,N u -1;

[0099] y min ≤y(k+i)≤y max , i=1,…,N p ;

[0100] The adaptive fuzzy control dynamically adjusts the PID control parameters for the nonlinear scene of material caking and humidity mutation;

[0101] The adaptive fuzzy control process includes the following steps:

[0102] A1, determine input variables based on real-time acquisition of parameters in the batching process: batching error, error change rate, raw material humidity, and feeding equipment vibration value;

[0103] A2, fuzzy processing of input variables, converting precise quantities into fuzzy language variables;

[0104] The precise quantity is converted into a fuzzy language variable by combining the domain division and membership function, as follows:

[0105] Domain and fuzzy set division:

[0106] Batching error (E) and error change rate (EC): both domains are divided into 5 fuzzy sets, corresponding to the language values: negative large (NB), negative small (NS), zero (ZO), positive small (PS), and positive large (PB);

[0107] Raw material humidity (H): the domain is divided into 3 fuzzy sets: low (L, 8%-12%), medium (M, 12%-16%), and high (H, 16%-18%);

[0108] Vibration value (V): the domain is divided into 3 fuzzy sets: low (L, 0-0.05g), medium (M, 0.05-0.1g), and high (H, 0.1-0.5g).

[0109] Membership function selection: all variables use triangular membership functions, taking batching error (E) as an example, the function parameters are:

[0110] NB: vertex (-5, 1), base (-6, -4);

[0111] NS: vertex (-2, 1), base (-4, 0);

[0112] ZO: vertex (0, 1), base (-2, 2);

[0113] PS: vertex (2, 1), base (0, 4);

[0114] PB: vertex (5, 1), base (4, 6);

[0115] A3, construct a fuzzy rule base based on the process experience of chicken feed batching, and output control decisions through fuzzy reasoning algorithm to obtain fuzzy output set;

[0116] Mamdani reasoning method is used to calculate the rule activation strength by taking the minimum value, integrate multiple rule outputs, and finally obtain the fuzzy output set;

[0117] A4, convert the fuzzy output obtained by reasoning into precise control instructions, and apply them to the actuator;

[0118] The fuzzy output set is converted into precise control quantity by using the barycentric method, and the formula is as follows:

[0119]

[0120] Wherein, f is the precise control quantity, f i is the membership degree of the output fuzzy set, x i is the corresponding argument value;

[0121] A5, by edge computing hub module (2) real-time monitoring final batching error, adjustment times and actuator energy consumption, dynamic optimization fuzzy rule base;

[0122] Optimization logic:

[0123] When the error of two consecutive batches is out of limit, the rule weight correction is triggered: using gradient descent method, the weight of the activated rule is increased / decreased (adjustment amplitude ± 5%), and the effective rule is strengthened and the invalid rule is weakened;

[0124] When the raw material batch is replaced (such as the change of corn variety), the rule adaptation module is automatically called, and 10% of the core rules (such as humidity-rotation speed correlation rule) are pre-adjusted based on the characteristics of the new raw material (such as bulk density, fluidity);

[0125] At 2:00 every morning (non-production period), the rule base optimized on the same day is synchronized to the cloud backup, and compared with the historical optimal rule base, the version with performance improvement ≥3% is retained.

[0126] The device health management algorithm module includes vibration signal processing, abnormality detection and device remaining life prediction, so that the unplanned downtime is reduced and the maintenance cost is reduced;

[0127] Among them, the vibration signal processing extracts the bearing fault characteristic frequency through EMD empirical mode decomposition and envelope spectrum analysis;

[0128] The abnormality detection identifies new type of fault unsupervisedly through the combination of autoencoder and isolated forest algorithm;

[0129] The device remaining life prediction predicts the device remaining life cycle by inputting vibration, temperature and load data based on the RUL model of LSTM.

[0130] The execution control module 6 is used for receiving the control instruction of the edge computing hub module, and the actuator performs action according to the control instruction; specifically, it includes a feeding control module and a mixing conveying module;

[0131] The feeding control module includes main raw material feeding and auxiliary material adding, the main raw material feeding is controlled by the combination of servo screw feeder and pneumatic butterfly valve group, the auxiliary material adding is realized by electromagnetic metering pump and vibration feeder to solve the problem of uneven manual addition.

[0132] The mixed conveying module controls the operation of the double-shaft paddle mixer through frequency conversion control and mixed uniformity monitoring, improves the consistency of feed nutrition, and indirectly improves the breeding benefit.

[0133] The application is based on a closed-loop architecture of perception, edge processing, cloud optimization and execution feedback, and fuses Internet of Things and edge computing technology to realize intelligent ingredient regulation of the chicken feed production line, and the core principle is as follows:

[0134] Full-dimensional perception: through the intelligent perception terminal module, raw material information, equipment data and environmental parameters are collected in real time to provide a data basis for regulation.

[0135] Edge real-time decision-making: the edge computing hub module processes the collected data locally, realizes millisecond-level real-time control adjustment of the feeder speed and valve opening degree through a dynamic ingredient algorithm, simultaneously performs equipment diagnosis and digital twin simulation, and quickly responds to sudden conditions.

[0136] Cloud global optimization: the cloud management and control platform performs big data analysis on the non-real-time data uploaded by the edge, realizes formula full life cycle management, raw material procurement optimization, quality traceability and multi-production line cooperation, and provides global decision support.

[0137] Safe and efficient communication: the network communication module realizes layered transmission and security protection, guarantees the reliable transmission of data between sensors, edge nodes and the cloud, and ensures the priority execution of control instructions.

[0138] Precise execution feedback: the execution control module accurately acts according to the edge instructions to form a closed loop of perception, decision-making and execution, and finally realizes the improvement of ingredient accuracy, the reduction of cost and the optimization of production efficiency.

[0139] Although the application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements to part of the technical features, and any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the application shall be included in the protection scope of the application.

Claims

1. An Internet of Things edge computing-based intelligent ingredient regulation system for a chicken feed production line, characterized in that: Intelligent sensing terminal module (1), edge computing hub module (2), cloud management platform module (3), network communication module (4), intelligent algorithm control module (5) and execution control module (6) are included; The intelligent sensing terminal module (1) is used for collecting the ingredient production data of the chicken feed production line in real time through sensors, and the production data includes raw material identification information, weight monitoring information, environmental perception information and device state information; The edge computing hub module (2) is used for providing real-time data processing and local control decision processing for the data collected by the intelligent sensing terminal module (1), realizing real-time control of equipment, equipment diagnosis and digital twin simulation, locally and quickly processing data, completing dynamic ingredient, equipment cooperation and fault emergency processing; The cloud management platform module (3) is used for providing production visual monitoring, formula optimization, quality traceability and big data analysis, providing integrated interface support and internal and third-party platform data intercommunication, providing remote equipment management and operation, and realizing unified management and decision optimization of multiple production lines; The network communication module (4) is used for providing data communication channels between modules, control instruction execution priority transmission principle in downlink data, and sensor data asynchronous upload using MQTT protocol; The intelligent algorithm control module (5) is used for providing dynamic ingredient control algorithm to the edge computing hub module (2) to optimize ingredient strategy, and providing equipment health management algorithm to realize equipment fault prediction and anomaly detection, and the algorithm is deployed in edge and cloud; The execution control module (6) is used for receiving control instructions of the edge computing hub module, and the actuator performs actions according to the control instructions.

2. The chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing according to claim 1, characterized in that: The intelligent sensing terminal module (1) includes a raw material state monitoring module and a device running state module; The raw material monitoring module is used for obtaining the weight and material level perception data of each raw material bin of chicken feed, and simultaneously binding the raw material batch information based on RFID technology; The device running state module is used for collecting the state information of the feeding equipment through vibration sensors and temperature sensors, simultaneously monitoring the real-time load of the motor through current sensors, and collecting the environmental parameters of the chicken feed production workshop through temperature and humidity sensors.

3. The chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing according to claim 1, characterized in that: The edge computing hub module (2) includes an edge computing node cluster module, a device synchronous control module, an edge intelligent service module and an edge-cloud collaborative data processing module; The edge computing node cluster module realizes parallel processing of control algorithm and AI inference based on a heterogeneous computing platform, realizes concurrent access of sensor groups through an edge gateway, and simultaneously provides 72-hour local data cache; The device synchronous control module is used for realizing time synchronization among sensors, controllers and actuators through clock protocol, and controlling the device group to perform synchronous actions; The edge intelligent service module is used for performing real-time state diagnosis through device running state data, simultaneously providing real-time mapping device state of the production line, and supporting the edge node to simulate the ingredient process of different formulas; The edge-cloud collaborative data processing module is used for hierarchical processing of data between edge devices and the cloud, including local storage processing, timing upload processing and asynchronous archiving processing.

4. The chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing according to claim 1, characterized in that: The cloud management platform module (3) comprises a production management module, a quality traceability module, a raw material procurement management module and an integrated interface module; The production management module is used for providing visual display of production line running state and full life cycle management of chicken feed formula; The quality traceability module realizes traceability query based on raw material batch information, mines the correlation between formula parameters and finished product quality through an association rule algorithm, and performs quality anomaly analysis; The raw material procurement management module predicts raw material price fluctuations based on an LSTM neural network, assists in procurement decision-making, and supports dynamic balance of feed inventory; The integrated interface module provides a data intercommunication interface, synchronizes production plans, raw material consumption and cost data in real time, and simultaneously provides remote diagnosis assistance services.

5. The chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing according to claim 1, characterized in that: The network communication module (4) comprises a network layered transmission module and a secure communication protection module; The network layered transmission module adopts a layered design of a field layer, a control layer and a management layer, the field layer connects sensors and actuators to realize basic data acquisition and instruction issuance; the control layer is used for real-time communication between edge nodes; and the management layer completes data interaction between the edge and the cloud by means of a 5G industrial router and a wired VPN, and supports networking connection between different factories; The secure communication protection module is used for constructing a security mechanism by dynamic device authentication and minimum permission principle verification device access, and performing data encryption transmission.

6. The chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing according to claim 1, characterized in that: The intelligent algorithm control module (5) comprises a dynamic batching algorithm module and a device health management algorithm module; The dynamic batching algorithm module comprises a model predictive control module and an adaptive fuzzy control module; the model predictive control module uses an LSTM prediction model to rollingly optimize future 10-second feeding strategies based on real-time data of raw material fluidity; and the adaptive fuzzy control module dynamically adjusts PID control parameters for nonlinear scenes such as material caking and humidity mutation through fuzzy logic; The device health management algorithm module comprises vibration signal processing, anomaly detection and device remaining life prediction.

7. The chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing according to claim 6, characterized in that: The calculation formula of the model predictive control module is as follows: x(k+1)=Ax(k)+Bu(k)+w(k); y(k)=Cx(k)+v(k); In the formula, x(k) is a device state vector at a current time (k), x(k+1) is a device state vector at a next time (k+1), u(k) is a device control input, including a feeder frequency and a valve opening, y(k) is a system output, A, B and C are a state transition matrix, a control matrix and an output matrix, w(k) and v(k) are process noise and measurement noise respectively; Rolling optimization to future operation step N of the plant p The rolling optimization outputs a value J with the objective of minimizing the prediction error and the control input cost. The formula for computing J is as follows: where: y r (k+i) is the target batch quantity of the future i-th step, Q is the output error weight matrix, R is the control input change weight matrix, N u is the control time domain, N u ≤N p , N p = 100 steps, N u = 50 steps, corresponding to 10 seconds of rolling optimization; The constraint condition is as follows: u min ≤ u(k + i) ≤ u max , i = 0, 1,..., N u -1; y min ≤ y max , i = 1,..., N p .

8. The chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing according to claim 6, characterized in that: The adaptive fuzzy control process comprises the following steps: A1, determining input variables based on real-time acquisition of parameters in the batching process: batching error, error change rate, raw material humidity and feeding device vibration value; A2, fuzzy processing of the input variables, converting accurate quantities into fuzzy language variables; A3, constructing a fuzzy rule base based on process experience of chicken feed batching, outputting control decisions through a fuzzy reasoning algorithm to obtain a fuzzy output set; A4, converting the fuzzy output obtained by reasoning into accurate control instructions, and applying the control instructions to an actuator. A5, the final batching error, adjustment times and actuator energy consumption are monitored in real time by the edge computing hub module (2), and the fuzzy rule base is dynamically optimized.

9. The chicken feed production line intelligent ingredient regulation system based on Internet of Things edge computing according to claim 1, characterized in that: The execution control module (6) includes a feeding control module and a mixing and conveying module; The feeding control module includes main raw material feeding and auxiliary material addition, the main raw material feeding is jointly controlled by a servo screw feeder and a pneumatic butterfly valve group, and the auxiliary material addition is realized by an electromagnetic metering pump and a vibrating feeder. The mixing and conveying module controls the operation of the double-shaft paddle mixer through frequency conversion control and mixing uniformity monitoring.

Images