Intelligent monitoring system for compost parameters based on internet of things

Abstract

The application discloses a kind of compost parameter intelligent monitoring system based on Internet of Things, comprising: compost parameter acquisition module, for collecting compost environment parameter data;Edge communication module, for sending compost environment parameter data to remote cloud processing platform;Data preprocessing module, compost environment parameter data is preprocessed;Atlas construction module, based on parameter state vector, constructs compost parameter evolution atlas;Path planning module, for executing maximum incentive path search, optimal incentive guide path;Trend assessment module, for generating trend deviation risk index sequence;Early warning determination module, for setting risk early warning threshold, generating intelligent early warning prompt;User interaction module, for uploading relevant data to cloud platform database, and is shown in user end.This application fuses Internet of Things monitoring and atlas modeling method, realizes compost state intelligent evaluation and trend early warning, with the advantages of monitoring precision, response timely.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent agricultural environment monitoring technology, and particularly relates to a compost parameter intelligent monitoring system based on the Internet of Things. BACKGROUND

[0002] In the traditional agricultural waste treatment process, composting is a green and low-carbon resource utilization method, which is widely used in harmless treatment of organic matters such as livestock and poultry manure, straw, and garden waste. However, composting is essentially a complex microbial metabolic process, and its effect is affected by multiple environmental parameters such as temperature, humidity, oxygen concentration, ammonia concentration, and pH value. Moreover, the requirements for parameter conditions are significantly different at different stages. Therefore, how to realize accurate monitoring and dynamic adjustment of the composting process has become a key problem to improve composting quality, shorten the treatment period, and reduce operation and maintenance costs.

[0003] In the prior art, the monitoring method for composting parameters mostly uses fixed-point sensor collection and manual data recording, which not only discontinuously obtains information, but also has a lagging response, making it difficult to reflect the real dynamic evolution trend of the composting process. In addition, although some systems have initially introduced Internet of Things technology to realize automatic collection and remote uploading of composting parameters, they still mainly rely on static data display, lack deep modeling and trend prediction capabilities for changes in the composting process state, especially in key transition stages such as the high-temperature period, the cooling period, and the composting period. It is difficult to timely discover potential abnormal evolution risks.

[0004] Furthermore, the existing methods generally use simple threshold judgment or rule-driven stage recognition strategies in composting state modeling, which are difficult to handle the coupling relationship and dynamic evolution law among multiple parameters, leading to inaccurate judgment of the composting stage and non-targeted control intervention. At the same time, the current systems generally lack path-level analysis and intelligent evaluation mechanisms for parameter change trends, which cannot effectively identify whether the composting process evolves along the expected composting path, nor can they provide intelligent early warning prompts and visual feedback based on path deviation. Therefore, the existing technology still has significant deficiencies in multi-dimensional parameter fusion modeling, composting evolution path planning, and trend deviation risk identification, and cannot meet the actual needs of intelligent and accurate management of the composting process in current green agriculture. SUMMARY

[0005] One purpose of the present application is to provide an intelligent monitoring system for compost parameters based on the Internet of Things, which combines real-time monitoring technology and parameter evolution graph modeling method, and systematically constructs a state perception, trend evaluation and intelligent early warning mechanism for the whole composting process, which can dynamically track and model the evolution trend of key environmental parameters such as temperature, humidity, oxygen concentration, ammonia concentration and pH value in the composting process, realize the stage judgment and deviation risk prediction of the composting process based on the optimal incentive path, and has the advantages of accurate monitoring, timely response, high intelligent degree of auxiliary decision-making, and improves the scientificity and automation level of composting process management.

[0006] According to an embodiment of the present application, an intelligent monitoring system for compost parameters based on the Internet of Things comprises:

[0007] The compost parameter acquisition module is used to arrange a plurality of compost parameter acquisition nodes to collect compost environmental parameter data;

[0008] The edge communication module is used to send the compost environmental parameter data to a remote cloud processing platform;

[0009] The data preprocessing module is used to preprocess the compost environmental parameter data;

[0010] The graph construction module is used to construct a compost parameter evolution graph based on a parameter state vector;

[0011] The path planning module is used to execute maximum incentive path search in the parameter evolution graph after the composting stage is labeled, and identify the optimal incentive guide path between the starting node and the target composting node;

[0012] The trend evaluation module is used to calculate the state difference value of each node in the optimal incentive guide path and the target composting node, input a trend deviation risk scoring function, and generate a trend deviation risk index sequence;

[0013] The early warning judgment module is used to set a risk warning threshold, generate an intelligent early warning prompt, and output the corresponding parameter state vector, state difference value and trend deviation score;

[0014] The user interaction module is used to upload the compost parameter evolution graph, trend deviation risk index sequence and intelligent early warning information to the cloud platform database, and display the composting process state, trend evolution path and potential deviation risk information on the user side through a graphical interface.

[0015] Optionally, the modules are realized through the following methods:

[0016] S1, a plurality of compost parameter acquisition nodes are arranged in the composting area to collect temperature, humidity, oxygen concentration, ammonia concentration and pH data inside the compost pile in real time, and form compost environmental parameter data;

[0017] S2, transmit the composting environment parameter data to a remote cloud processing platform;

[0018] S3, pre-process the composting environment parameter data on the cloud processing platform to generate a parameter state vector;

[0019] S4, construct a composting parameter evolution graph based on the parameter state vector, taking the parameter state vector at each time as a node and the parameter change trend between adjacent time slices as a directed edge, forming a directed graph structure describing the dynamic evolution of the composting process;

[0020] S5, in the composting parameter evolution graph, set a target maturity node, and identify the optimal transition path from the current node to the target maturity node based on the maximum incentive path search method, forming an optimal incentive guide path;

[0021] S6, score the parameter evolution direction of each node in the optimal incentive guide path and the proximity of the target maturity state, generating a trend deviation risk index sequence;

[0022] When any value in the trend deviation risk index sequence exceeds a preset threshold, it is determined that the current composting process has a trend anomaly, and an intelligent early warning prompt is generated;

[0023] S7, upload the trend deviation risk index sequence, the composting parameter evolution graph and the current composting parameter state information to the cloud processing platform, and display the composting progress state, trend evolution path and potential deviation risk prompt through a graphical interface on the user side.

[0024] Optionally, the composting parameter collection node set integrates a temperature sensor, a humidity sensor, an oxygen concentration sensor, an ammonia concentration sensor and a pH sensor.

[0025] Optionally, the pre-processing includes outlier rejection, missing data completion, time synchronization and normalization.

[0026] Optionally, the S4 includes the following specific steps:

[0027] S41, at each sampling time point, based on the normalized parameter state vector of all composting parameter collection nodes, construct a composting parameter state matrix at the sampling time point;

[0028] S42, map the composting parameter state matrix corresponding to each time point to a composting state node, and define the node set of the composting parameter evolution graph;

[0029] S43, for any adjacent time points, calculate the state change value between the composting parameter state matrices using the Euclidean distance under the Frobenius norm;

[0030] S44, a directed edge is established between the compost state nodes of adjacent time points, and the edge weight of the directed edge is defined as the state change value, and a compost parameter evolution graph is constructed based on the node set and all directed edges:

[0031]

[0032] wherein G represents the compost parameter evolution graph, represents the node set of the compost parameter evolution graph, and ε represents the edge set of the compost parameter evolution graph, and represents a compost state node, represents a directed edge from the node to , and represents the edge weight of the directed edge.

[0033] S45, compost phase labels are introduced on the node set of the compost parameter evolution graph, and each state label in the compost phase label is defined according to the average temperature change rate in the node:

[0034]

[0035] wherein represents the average temperature change rate at time point t k , T i (t k ) represents the temperature value of node i at time point t k , T i (t k-1 ) represents the temperature value of node i at time point t k-1 , and N represents the total number of nodes.

[0036] The state label is divided into phases according to the average temperature change rate and the current average temperature value, including:

[0037] When the average temperature change rate is greater than the temperature rise rate threshold and the current average temperature value is less than the high temperature threshold, it indicates that the temperature continues to rise and has not yet entered the high temperature zone, and the node is labeled as the temperature rise period;

[0038] When the current average temperature value is greater than the high temperature threshold and the absolute value of the average temperature change rate is not greater than the temperature fluctuation threshold of the high temperature stable zone, it indicates that the temperature has reached the high temperature state and the temperature change tends to be stable, and the node is labeled as the high temperature period;

[0039] When the average temperature change rate is less than the negative temperature drop rate threshold and the current average temperature value is greater than the intermediate reference temperature, it indicates that the temperature appears a significant downward trend and is still in a higher temperature zone but is unstable, and the node is labeled as the temperature drop period.

[0040] When the absolute value of the average temperature change rate is less than the low change threshold and the current average temperature value is not greater than the intermediate reference temperature, it indicates that the temperature tends to be stable and the temperature has dropped to the normal temperature range, and the node is marked as the maturation period;

[0041] S46, redefine the compost parameter evolution graph after composting stage annotation, the compost parameter evolution graph after composting stage annotation includes a node set, an edge set and a state label;

[0042] The compost parameter evolution graph after composting stage annotation supports dynamic updating. When a new compost parameter state matrix corresponding to a new time point arrives, a new node is added in real time, a new directed edge is established according to the state change value, and the node set, the edge set and the state label are updated, so that the compost parameter evolution graph continues to evolve over time and is used for intelligent monitoring of the composting state.

[0043] Optionally, the S5 comprises the following specific steps:

[0044] S51, in the compost parameter evolution graph after composting stage annotation, determine the current node as a search starting node, and filter all nodes in the node set that satisfy the state label being the maturation period to form a target maturation node set;

[0045] S52, in the edge set of the compost parameter evolution graph, introduce a control incentive in addition to the original edge weight of each directed edge, the control incentive describes the net positive effect generated by the transition from a node to another node after performing a control action on the node:

[0046]

[0047] Wherein, Γ a,b represents the control incentive value generated by the transition from the node v a to the target maturation node v b , ΔΦ(v a , v b ) represents the improvement of the composting state on the target evaluation index function after performing the control action on the node v a , represents the comprehensive operation cost generated by the control action , v a represents the search starting node, v b represents the target maturation node, represents the control action;

[0048] The target evaluation index function describes the matching degree of the composting state at a certain time point, and is obtained by weighting the humidity stability index, the oxygen concentration rationality index, the ammonia concentration suppression index and the pH stability index. Before weighted summation, each index is converted into a dimensionless index by standard normalization.

[0049] The comprehensive operation cost is the total amount of resource consumption required for performing the control behavior on the node, which is obtained by weighted summation of a plurality of heterogeneous resource indicators, including energy consumption, water consumption and manual intervention, and each of the heterogeneous resource indicators is converted into a dimensionless indicator by standard normalization before weighted summation;

[0050] S53, construct a reinforcement compost parameter evolution graph, an edge weight of each edge of the reinforcement compost parameter evolution graph is driven by a state change value and a control incentive, and a reinforcement edge set is defined;

[0051] S54, based on the reinforcement compost parameter evolution graph, perform maximum incentive path search from the current node to the target mature node to generate an optimal incentive guide path:

[0052]

[0053] P * represents the optimal incentive guide path, and describes a path sequence with the largest cumulative control incentive value, represents a set of all feasible paths from the current node to the target mature node, (v a →v b ) represents any edge in the path, Γ a,b represents an incentive value corresponding to the edge (v a →v b ), argmax represents the variable value when the function value is maximum, w a,b represents the original edge weight of the edge (v a →v b ), β represents a cost suppression factor, which represents the tolerance degree of high cost path, P represents any path in the feasible path set, and ∈ represents “belongs to”.

[0054] Optionally, the S6 includes the following specific steps:

[0055] S61, based on the optimal incentive guide path, extract the compost parameter state matrix corresponding to each node in the path, and select the target mature node of the path;

[0056] S62, for any node in the optimal incentive guide path, calculate the state difference value between the current compost parameter state vector and the target compost parameter state vector by using the Euclidean distance under the Frobenius norm;

[0057] S63, input the state difference value of each node into a trend deviation risk score function to obtain the trend deviation risk score corresponding to each node on the path:

[0058]

[0059] wherein, ε(·) represents a trend deviation risk score function, Δ t represents a state difference value, δ represents a state difference tolerance threshold, α represents a risk score penalty coefficient, and e represents the base of natural logarithm;

[0060] S64, arranging the trend deviation risk scores of all nodes in path time sequence to form a trend deviation risk index sequence;

[0061] S65, setting a risk warning threshold, and if there is a trend deviation risk score greater than the risk warning threshold, it is determined that the current composting process has a trend anomaly deviating from the target composting state in the corresponding stage;

[0062] S66, in the case of determining that there is a trend anomaly, constructing intelligent warning prompt information including the trend deviation risk score, the state difference value and the corresponding composting parameter state vector, and uploading to a remote cloud processing platform through an Internet of Things communication module for remote viewing and auxiliary decision-making by a user.

[0063] The present application has the following advantages:

[0064] The composting parameter intelligent monitoring system based on the Internet of Things provided by the present application fully combines multi-parameter sensing, graph modeling and path trend evaluation technical means, can realize continuous, dynamic and intelligent monitoring of the whole composting process, through arranging multiple collection nodes in the composting processing area, real-time collection of key parameter information such as temperature, humidity, oxygen concentration, ammonia concentration and pH value, and efficient transmission of the parameter information to the cloud processing platform by using an edge communication module, realizing real-time acquisition and remote centralized management of data, and significantly improving the timeliness and reliability of composting monitoring.

[0065] In terms of data processing, the present application changes the isolated management mode of traditional composting parameters to a global modeling mode facing the evolution trend by constructing a parameter state vector and a composting parameter evolution graph, can accurately describe the dynamic change process of the composting state over time, further identifies the optimal path of the current composting state evolving to the target composting state by introducing a target composting node and a maximum incentive path search algorithm, and clearly determines the key evolution nodes and control timing in the process, providing a scientific basis for subsequent intelligent intervention.

[0066] In addition, the application innovatively introduces a trend deviation risk scoring mechanism to quantitatively evaluate the difference between each node state in the optimal evolution path and the target mature state, generate a trend deviation risk index sequence, and automatically issue an intelligent early warning prompt when the deviation degree exceeds a threshold, which effectively solves the problem that the existing technology cannot timely identify the deviation of the compost evolution, and enhances the predictability and decision support capability of the system. Combined with the graphical user interface and remote access function, the user can at any time master the compost state, trend evolution path and potential deviation risk, so as to realize efficient, intelligent and controllable compost whole-process management. BRIEF DESCRIPTION OF DRAWINGS

[0067] The accompanying drawings are included to provide a further understanding of the application, and constitute a part of the specification, which together with the embodiments of the application are used to explain the application, and do not constitute a limitation on the application. In the drawings:

[0068] Fig. 1 A schematic diagram of the overall structure of a compost parameter intelligent monitoring system based on the Internet of Things is proposed for the application;

[0069] Fig. 2 A flowchart of the maximum incentive path search algorithm of a compost parameter intelligent monitoring system based on the Internet of Things is proposed for the application. DETAILED DESCRIPTION

[0070] The application will now be described in further detail with reference to the drawings. These drawings are simplified schematic diagrams, and only illustrate the basic structure of the application in a schematic manner, and therefore only show the components related to the application.

[0071] REFERENCE Figs. 1-2 A compost parameter intelligent monitoring system based on the Internet of Things, comprising:

[0072] A compost parameter acquisition module for laying out a plurality of compost parameter acquisition nodes to collect compost environment parameter data;

[0073] An edge communication module for sending the compost environment parameter data to a remote cloud processing platform;

[0074] A data preprocessing module for preprocessing the compost environment parameter data;

[0075] A graph construction module for constructing a compost parameter evolution graph based on the parameter state vector;

[0076] A path planning module for performing maximum incentive path search in the parameter evolution graph labeled in the compost stage, and identifying the optimal incentive guide path between the starting node and the target mature node;

[0077] A trend evaluation module is configured to calculate state difference values of each node in the optimal incentive guide path and the target mature node, input a trend deviation risk scoring function, and generate a trend deviation risk index sequence;

[0078] An early warning determination module is configured to set a risk early warning threshold, generate an intelligent early warning prompt, and output corresponding parameter state vectors, state difference values, and trend deviation scores;

[0079] A user interaction module is configured to upload the compost parameter evolution graph, the trend deviation risk index sequence, and the intelligent early warning information to a cloud platform database, and display the composting process state, the trend evolution path, and the potential deviation risk information on the user side through a graphical interface.

[0080] The present application realizes the full-process automatic management of dynamic perception, graph modeling, path planning, and trend early warning of multiple parameters in the composting process by constructing a complete compost parameter intelligent monitoring system based on an Internet of Things architecture. The system modules are clear in function and highly efficient in cooperation, which not only improves the parameter monitoring accuracy and reaction speed in the composting process, but also reduces the frequency of human intervention and the error rate, enhances the controllability, visualization, and intelligence level of the composting process, and is suitable for large-scale agricultural waste resource processing scenarios, and has good popularization and application prospect.

[0081] In the present embodiment, the modules are realized through the following methods:

[0082] S1, multiple compost parameter collection nodes are arranged in the composting treatment area, and temperature, humidity, oxygen concentration, ammonia concentration, and pH data inside the compost pile are collected in real time to form compost environment parameter data;

[0083] S2, the compost environment parameter data is transmitted to a remote cloud processing platform;

[0084] S3, the compost environment parameter data is preprocessed on the cloud processing platform to generate parameter state vectors;

[0085] S4, a compost parameter evolution graph is constructed based on the parameter state vectors, and each parameter state vector at each moment is taken as a node, and the parameter change trend between adjacent time slices is taken as a directed edge to form a directed graph structure describing the dynamic evolution of the composting process;

[0086] S5, in the compost parameter evolution graph, a target mature node is set, and an optimal transition path from the current node to the target mature node is identified based on a maximum incentive path search method to form an optimal incentive guide path;

[0087] S6, the parameter evolution direction of each node in the optimal incentive guide path and the proximity degree to the target mature state are scored node by node to generate a trend deviation risk index sequence;

[0088] When the trend deviates from any value in the risk indicator sequence by more than a preset threshold, it is determined that there is a trend anomaly in the current composting process, and an intelligent early warning prompt is generated;

[0089] S7, upload the trend deviation risk indicator sequence, the composting parameter evolution graph and the current composting parameter state information to a cloud processing platform, and display the composting process state, the trend evolution path and the potential deviation risk prompt through a graphical interface on a user end.

[0090] The present application ensures the data closed loop from data collection to graphical display of the whole process by specifying the specific method interaction process between modules, improves the logical consistency and real-time linkage capability of the system, realizes the efficient transmission and execution path of the composting state from collection, processing, modeling, evaluation, early warning to user display by using standardized steps in series, makes the operation of each module more reliable and stable, and facilitates actual deployment, maintenance and expansion, and ensures that the composting process control is changed from "passive response" to "active guidance".

[0091] Optionally, the composting parameter collection node integrates temperature sensors, humidity sensors, oxygen concentration sensors, ammonia concentration sensors and pH sensors.

[0092] The present application adopts a multi-parameter integrated composting collection node integrating temperature, humidity, oxygen concentration, ammonia concentration and pH sensors, can comprehensively reflect the biochemical evolution characteristics of the composting environment, improves the accuracy and responsiveness of the composting state modeling through the linkage collection of multi-dimensional environmental parameters, has strong node adaptability and flexible layout, can stably operate in a high-temperature, high-humidity and corrosive composting environment, effectively guarantees the data collection quality, is the basic guarantee for efficient operation of the system, and at the same time provides sufficient raw data support for the subsequent intelligent analysis module.

[0093] In the embodiment, the preprocessing includes abnormality rejection, missing completion, time synchronization and normalization.

[0094] The present application effectively improves the time sequence integrity and analysis availability of the composting environmental parameter data by introducing the abnormality rejection, missing completion, time synchronization and normalization standardization data preprocessing process, the preprocessing module can automatically correct the mutation points, abnormal drifts and breakpoint missing in the collected data, provides data quality guarantee for subsequent state vector construction and graph modeling, improves the noise robustness and dynamic environment adaptability of the whole system, and lays a solid data foundation for realizing high-precision trend modeling and path judgment of the composting process.

[0095] In the embodiment, the S4 includes the following specific steps:

[0096] S41, constructing a compost parameter state matrix of the sampling time point based on the normalized parameter state vector of all compost parameter acquisition nodes at each sampling time point;

[0097] S42, mapping the compost parameter state matrix corresponding to each time point to a compost state node, and defining a node set of the compost parameter evolution graph;

[0098] S43, for any adjacent time points, using the Euclidean distance under the Frobenius norm to calculate the state change value between the compost parameter state matrices;

[0099] S44, establishing a directed edge between the compost state nodes of adjacent time points, and defining the edge weight of the directed edge as the state change value, and constructing the compost parameter evolution graph based on the node set and all directed edges:

[0100]

[0101] wherein G represents the compost parameter evolution graph, represents the node set of the compost parameter evolution graph, and ε represents the edge set of the compost parameter evolution graph, and represents the compost state node, represents the directed edge from the node to , and represents the edge weight of the directed edge;

[0102] S45, introducing a compost phase label on the node set of the compost parameter evolution graph, wherein each state label in the compost phase label is defined according to the average temperature change rate in the node:

[0103]

[0104] wherein represents the average temperature change rate at time point t k , T i (t k ) represents the temperature value of node i at time point t k , T i (t k-1 ) represents the temperature value of node i at time point t k-1 , and N represents the total number of nodes;

[0105] The state label is divided into phases according to the average temperature change rate and the current average temperature value, including:

[0106] when the average temperature change rate is greater than the temperature rise rate threshold and the current average temperature value is less than the high temperature threshold, indicating that the temperature continues to rise and has not yet entered the high temperature zone, the node is labeled as the temperature rise period;

[0107] When the current average temperature value is greater than the high temperature threshold and the absolute value of the average temperature change rate is not greater than the temperature fluctuation threshold of the high temperature stable zone, it indicates that the temperature has reached the high temperature state and the temperature change tends to be stable, and the node is marked as the high temperature period;

[0108] When the average temperature change rate is less than the negative temperature drop rate threshold and the current average temperature value is greater than the intermediate reference temperature, it indicates that the temperature has a significant downward trend and is still in a higher temperature zone but is unstable, and the node is marked as the temperature drop period;

[0109] When the absolute value of the average temperature change rate is less than the low change threshold and the current average temperature value is not greater than the intermediate reference temperature, it indicates that the temperature tends to be stable and the temperature has dropped to the normal temperature range, and the node is marked as the composting period;

[0110] S46, redefine the composting parameter evolution graph after the composting stage is marked, the composting parameter evolution graph after the composting stage is marked includes a node set, an edge set and a state label;

[0111] The composting parameter evolution graph after the composting stage is marked supports dynamic updating, when a new time point corresponding to the composting parameter state matrix arrives, a new node is added in real time, a new directed edge is established according to the state change value, and the node set, the edge set and the state label are updated, so that the composting parameter evolution graph continues to evolve over time and is used for intelligent monitoring of the composting state.

[0112] The composting parameter evolution graph is constructed with time as the main axis, the dynamic evolution law of the composting process is systematically described, the stage label division mechanism is introduced at the node layer, so that the time sequence, numerical change and trend relationship of each stage of composting can be clearly expressed by the graph model, the state difference quantification method based on the Frobenius norm and the Euclidean distance makes the edge weight in the graph have physical meaning, and provides structural support for subsequent path selection and trend deviation evaluation, the graph has the ability of continuous updating, and can reflect the parameter fluctuation and state change in the composting evolution process in real time.

[0113] In the embodiment, the S5 includes the following specific steps:

[0114] S51, in the composting parameter evolution graph after the composting stage is marked, determine a current node as a search starting node, and filter all nodes in the node set that satisfy the state label being the composting period to form a target composting node set;

[0115] S52, in the edge set of the composting parameter evolution graph, a control incentive is introduced in addition to the original edge weight of each directed edge, the control incentive describes the net positive effect generated by transferring from a node to another node after executing a control action on the node:

[0116]

[0117] wherein, Γ a,b denotes the control incentive value generated from the node v a to the target composting node v b , ΔΦ(v a , v b ) denotes the improvement of the composting state on the target evaluation index function after the node v a performs the control action, denotes the control action , v a denotes the search starting node, v b denotes the target composting node, denotes the control action;

[0118] the target evaluation index function describes the degree of matching of the composting state at a certain time point, and is obtained by weighting the humidity stability index, the oxygen concentration rationality index, the ammonia concentration inhibition index and the pH stability index, and each index is converted into a dimensionless index by a standard normalization method before weighted summation;

[0119] the comprehensive operation cost is the total amount of resource consumption required for performing the control action on the node, and the total amount of resource consumption is obtained by weighted summation of a plurality of heterogeneous resource indexes, the heterogeneous resource indexes include energy consumption, water consumption and manual intervention, and each heterogeneous resource index is converted into a dimensionless index by a standard normalization method before weighted summation;

[0120] S53, construct a reinforced composting parameter evolution graph, the edge weight of each edge of the reinforced composting parameter evolution graph is driven by the state change value and the control incentive, and a reinforced edge set is defined;

[0121] S54, based on the reinforced composting parameter evolution graph, perform maximum incentive path search from the current node to the target composting node to generate an optimal incentive guide path:

[0122]

[0123] wherein, P * denotes the optimal incentive guide path, describing the path sequence with the largest cumulative control incentive value, denotes a set of all feasible paths from the current node to the target composting node, (v a →v b ) denotes any edge in the path, Γ a,b denotes the incentive value corresponding to the edge (v a →v b ), and argmax denotes the variable value when the function value is maximum, wa,b denotes the original edge weight of edge (v a →v b ), i.e., the state change cost, β denotes the cost inhibition factor, which controls the tolerance degree of high cost path, P denotes any path in the feasible path set, and ∈ denotes "belongs to".

[0124] The present application introduces a control incentive and a comprehensive operation cost in a graph to construct a reinforced compost parameter evolution graph, realizes optimal incentive path planning from a current state to a mature state, and compared with a traditional state tracking method, adopts a maximum incentive path search strategy to quantize the benefits and costs of intervention behaviors, reduces resource consumption as much as possible while ensuring the achievement of the target, the mechanism provides an executable and adjustable path guiding mechanism for the system, significantly improves the scientificity, refinement and strategy of the compost process control, and provides support for subsequent intelligent intervention and prediction.

[0125] In the embodiment, the S6 includes the following specific steps:

[0126] S61, based on the optimal incentive guiding path, extracting the compost parameter state matrix corresponding to each node in the path, and selecting the target mature node of the path;

[0127] S62, for any node in the optimal incentive guiding path, using the Euclidean distance under the Frobenius norm to calculate the state difference value between the current compost parameter state vector and the target compost parameter state vector;

[0128] S63, inputting the state difference value of each node into a trend deviation risk score function to obtain the trend deviation risk score corresponding to each node on the path:

[0129]

[0130] Wherein, ε(·) represents the trend deviation risk score function, Δ t represents the state difference value, δ represents the state difference tolerance threshold, α represents the risk score penalty coefficient, and e represents the base number of natural logarithm;

[0131] S64, arranging the trend deviation risk scores of all nodes according to the path time sequence to form a trend deviation risk index sequence;

[0132] S65, setting a risk warning threshold, if there is a trend deviation risk score greater than the risk warning threshold, it is determined that the current composting process has a trend anomaly of deviating from the target mature state at the corresponding stage;

[0133] S66, in the case of determining that there is a trend anomaly, constructing an intelligent early warning prompt information including a trend deviation risk score, a state difference value and a corresponding compost parameter state vector, and uploading to a remote cloud processing platform through an Internet of Things communication module for remote viewing and auxiliary decision-making by a user.

[0134] The application constructs a trend rationality evaluation model based on an optimal incentive path, realizes node-by-node evaluation of the state deviation degree of each node in the path, and forms a trend deviation risk index sequence. The parameter state difference measurement and risk score function modeling method can accurately find potential trend abnormal behavior in the composting process and trigger an intelligent early warning mechanism to ensure that the composting process always progresses towards the expected maturity state. This mechanism improves the active monitoring and risk prevention and control capabilities of the system and is a key link for realizing the transition of composting state from "result judgment" to "process guidance".

[0135] Embodiment 1:

[0136] In a large-scale organic agricultural plantation, in order to realize the resource utilization of large-scale straw and livestock manure, a local agricultural cooperative constructed a centralized composting treatment site. Previously, the treatment site relied on manual inspection and experience-based management to monitor the composting process, which resulted in problems such as temperature and humidity control lag, insufficient oxygen supply, abnormal pH fluctuation, and the like, which often led to prolonged composting period, uneven composting degree, and unstable final compost quality, severely restricting the output efficiency of high-quality organic fertilizer.

[0137] In order to improve the above problems, the local cooperative introduced an intelligent composting parameter monitoring system based on the Internet of Things. The system was deployed in two main fermentation units of the composting treatment site. In the specific implementation process, first, multiple composting parameter collection nodes were arranged inside the composting pile according to the rules, each node integrated five sensors of temperature, humidity, oxygen concentration, ammonia concentration and pH, and real-time collected multi-dimensional parameters inside the pile. The collection nodes periodically uploaded data to the edge collection control unit through the LoRa communication network, and then transmitted to the remote cloud processing platform.

[0138] In the cloud platform, the system first cleans and preprocesses the original composting environment parameter data, eliminates abnormal values, fills in missing data, and completes time alignment and normalization processing. Then, the parameter state vector is constructed based on each parameter reported at each time point, and it is mapped as a graph node in the composting parameter evolution graph. The system uses the Euclidean distance between adjacent time point parameter vectors to construct the weight of the directed edge in the graph, thereby forming the dynamic graph structure of the application.

[0139] Further, the system labels each node in the graph based on a preset temperature change rate and absolute temperature threshold, such as the warming period, high temperature period, cooling period, and composting period, and on this basis, the highest path of control incentive value is calculated from the current node to the target composting node set by using a maximum incentive path search algorithm, and an optimal incentive guide path is constructed, wherein the edge weight of each hop in the path not only considers the state change, but also introduces the trade-off relationship between the composting degree improvement and the operation cost caused by the control behavior.

[0140] In order to monitor whether the composting process deviates from the optimal evolution trend, the system further constructs a trend rationality evaluation model to score the proximity of each node state in the optimal path to the target composting state, and form a trend deviation risk index sequence, and when the risk score exceeds a preset threshold, the system automatically triggers an early warning and pushes the deviation position, risk level, key parameter value and suggestion regulation information to the user terminal.

[0141] The user can view the graphical evolution graph of the composting parameters, the trend risk index sequence and the current composting state through the mobile terminal interface in the office, and the system supports historical backtracking, trend prediction and intervention suggestion linkage viewing, which improves the management efficiency and response speed of the operator. During the entire composting period, the system automatically completes the state tracking, abnormal early warning, path analysis and visual display, and assists the user in accurately identifying the core control nodes in different stages of composting processing, effectively avoiding problems such as overexposure, insufficient oxygen and ammonia volatilization abnormalities.

[0142] Through the operation and observation of the system, it can be seen that the composting period is significantly shortened compared with the traditional method, the composting degree uniformity is enhanced, the physical and chemical indicators of the finished compost are more stable, the worker intervention frequency is reduced, and the overall operation and maintenance cost is optimized, thereby improving the resource utilization efficiency and intelligent management level of agricultural waste.

[0143] Table 1 shows the performance comparison of the composting intelligent monitoring system and the traditional method.

[0144] Table 1 Comparison of composting intelligent monitoring system and traditional method

[0145]

[0146] As can be seen from Table 1, the intelligent monitoring system proposed in the present application is superior to the traditional manual method in terms of data acquisition frequency. The traditional method usually relies on personnel to record composting parameters at regular intervals, and the frequency is low and the periodicity is unstable, which is easy to miss the key change period. The system of the present application uses multiple sensing nodes to realize high-frequency automatic acquisition, ensuring data continuity and real-time, and providing a stable data basis for subsequent trend analysis.

[0147] In the data integrity dimension, the system integrates multiple parameters such as temperature, humidity, oxygen concentration, ammonia concentration and pH, and ensures the consistency and time integrity of data between different nodes through standardized and synchronous processing, which is more reliable than the data loss or delay often found in traditional records.

[0148] In terms of abnormality detection capability, the traditional method relies on manual judgment, which is highly subjective and lacks systematicness and early warning capability, and cannot quantitatively evaluate parameter fluctuations. The present application constructs a trend deviation risk scoring function, dynamically evaluates risks based on state difference changes in the evolution path, and realizes real-time identification and intelligent early warning of the deviation trend of maturity.

[0149] In terms of trend early warning response, the traditional method often reacts after the problem has become serious, often missing the best intervention opportunity. The present application sets a threshold based on the path deviation risk sequence, can identify potential risks at the first time, and automatically generates early warning prompts, improving the timeliness and reliability of the response.

[0150] In terms of accuracy of maturity state identification, the traditional method relies on experience for judgment, and the standard is not unified, so the judgment accuracy is greatly affected by individual experience. The present application integrates multi-dimensional indicators to construct a target evaluation index function, with clear standards and explicit analysis logic, and is more accurate and scientific in identifying different maturity stages.

[0151] In terms of frequency of human intervention, the traditional method requires workers to frequently inspect the pile and make adjustments, which is labor-intensive and inefficient. The present application reduces the frequency of human intervention through trend-guided paths and intelligent intervention suggestions, improving work efficiency.

[0152] In addition, the present application system supports cloud platform remote access function, and users can master the state of composting process and risk trend at any time through the graphical interface, which is of great significance for large distributed composting sites. In contrast, the traditional method does not have remote monitoring capability and the information feedback is lagging.

[0153] In terms of system scalability, the traditional architecture is usually complex to deploy and difficult to adjust flexibly, while the present application is based on modular design, each sensing node supports flexible deployment, and the overall system has good scalability and can be dynamically adjusted according to the application scale to adapt to various actual application requirements.

[0154] The above is only the preferred embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can make equivalent replacement or change within the technical scope disclosed by the present application according to the technical solution and inventive concept of the present application, which should be covered within the protection scope of the present application.

Claims

1. An intelligent monitoring system for composting parameters based on the Internet of Things, characterized in that, include: The composting parameter acquisition module is used to deploy multiple composting parameter acquisition nodes to collect composting environmental parameter data. The edge communication module is used to send composting environmental parameter data to a remote cloud processing platform; The data preprocessing module preprocesses the composting environmental parameter data; The graph construction module constructs a compost parameter evolution graph based on the parameter state vector; The path planning module is used to perform maximum incentive path search in the parameter evolution map after the composting stage is marked, and to identify the optimal incentive guidance path between the starting node and the target maturity node. The trend assessment module is used to calculate the state difference between each node in the optimal incentive guidance path and the target mature node. Input the trend deviation risk scoring function to generate a trend deviation risk index sequence. The early warning judgment module is used to set risk early warning thresholds, generate intelligent early warning prompts, and output the corresponding parameter state vector, state difference value, and trend deviation score. The user interaction module is used to upload composting parameter evolution maps, trend deviation risk indicator sequences, and intelligent early warning information to the cloud platform database, and to display the composting process status, trend evolution path, and potential deviation risk information on the user end through a graphical interface.

2. The intelligent monitoring system for composting parameters based on the Internet of Things according to claim 1, characterized in that, The modules are connected in the following way: S1. Multiple composting parameter acquisition nodes are set up in the composting treatment area to collect data on temperature, humidity, oxygen concentration, ammonia concentration and pH inside the compost pile in real time, forming composting environmental parameter data. S2. Transmit composting environmental parameter data to a remote cloud processing platform; S3. Preprocess the composting environmental parameter data on the cloud processing platform to generate parameter state vectors; S4. Construct a compost parameter evolution graph based on the parameter state vector. Use the parameter state vector at each time step as a node and the parameter change trend between adjacent time slices as directed edges to form a directed graph structure that describes the dynamic evolution of the composting process. S5. In the composting parameter evolution map, set the target maturation node, and identify the optimal transfer path from the current node to the target maturation node based on the maximum incentive path search method to form the optimal incentive guidance path. S6. Analyze the parameter evolution direction of each node in the optimal incentive guidance path and the degree of similarity to the target decay state, and generate a trend deviation risk index sequence. When the trend deviates from any value in the risk indicator sequence by more than a preset threshold, it is determined that there is an abnormal trend in the current composting process, and an intelligent early warning is generated. S7. Upload the trend deviation risk indicator sequence, composting parameter evolution map, and current composting parameter status information to the cloud processing platform, and display the composting process status, trend evolution path, and potential deviation risk warnings on the user terminal through a graphical interface.

3. The intelligent monitoring system for composting parameters based on the Internet of Things according to claim 2, characterized in that, The composting parameter acquisition node integrates a temperature sensor, a humidity sensor, an oxygen concentration sensor, an ammonia concentration sensor, and a pH sensor.

4. The intelligent monitoring system for composting parameters based on the Internet of Things according to claim 2, characterized in that, The preprocessing includes anomaly removal, missing data completion, time synchronization, and normalization.

5. The intelligent monitoring system for composting parameters based on the Internet of Things according to claim 2, characterized in that, S4 includes the following specific steps: S41. At each sampling time point, construct the compost parameter state matrix for the sampling time point based on the standardized parameter state vector of all compost parameter acquisition nodes. S42. Map the composting parameter state matrix corresponding to each time point to a composting state node, and define the node set of the composting parameter evolution graph. S43. For any adjacent time points, calculate the cost of state changes between the composting parameter state matrices using Euclidean distance under the Frobenius norm. S44. Establish directed edges for compost state nodes at adjacent time points. The weight of the directed edge is defined as the cost of state change. Construct a compost parameter evolution graph based on the node set and all directed edges. S45. Introduce composting stage labels on the node set of the composting parameter evolution map, wherein each state label in the composting stage label is defined according to the average temperature change rate in the node. The status label is divided into stages based on the average temperature change rate and the current average temperature value, including: When the average temperature change rate is greater than the temperature rise rate threshold and the current average temperature value is less than the high temperature threshold, it indicates that the temperature is continuously rising but has not yet entered the high temperature zone, and the node is marked as the temperature rise period. When the current average temperature value is greater than the high temperature threshold and the absolute value of the average temperature change rate is not greater than the temperature fluctuation threshold of the high temperature stable zone, it indicates that the temperature has reached a high temperature state and the temperature change tends to be stable. The node is marked as a high temperature period. When the average temperature change rate is less than the negative cooling rate threshold and the current average temperature value is greater than the intermediate reference temperature, it indicates that the temperature has a significant downward trend and is still in a relatively high temperature range but is no longer stable. The node is marked as the cooling period. When the absolute value of the average temperature change rate is less than the low change threshold and the current average temperature value is not greater than the intermediate reference temperature, it indicates that the temperature is stabilizing and has dropped to the normal temperature range. The node is marked as the decomposition period. S46. Redefine the composting parameter evolution graph after the composting stage is labeled, wherein the composting parameter evolution graph after the composting stage is labeled includes a node set, an edge set, and a state label; The composting parameter evolution graph labeled with the composting stage supports dynamic updates. When a new composting parameter state matrix corresponding to a new time point arrives, new nodes are added in real time, and new directed edges are established based on the cost of state changes. At the same time, the node set, edge set, and state label are updated to keep the composting parameter evolution graph continuously evolving over time and used for intelligent monitoring of composting status.

6. The intelligent monitoring system for composting parameters based on the Internet of Things according to claim 2, characterized in that, S5 includes the following specific steps: S51. In the composting parameter evolution map after the composting stage is marked, determine the current node as the search starting point node, and select all nodes that meet the state label as the maturity stage in the node set to form the target maturity node set. S52. In the edge set of the compost parameter evolution graph, each directed edge, in addition to its original edge weight, is introduced with a control incentive. The control incentive describes the net positive effect generated by moving from one node to another after executing a control action at a node: Among them, Γ a,b Indicates from node v a Transfer to target mature node v b The resulting control excitation value, ΔΦ(v) a ,v b ) indicates that at node v a The improvement in compost status on the target evaluation index function after implementing control actions. Indicates control behavior The resulting overall operational cost, v a Indicates the starting node of the search, v b Indicates the target maturation node, Indicates control behavior; The target evaluation index function describes the degree of matching of compost maturity at a certain point in time. It is obtained by weighting the humidity stability index, oxygen concentration rationality index, ammonia concentration inhibition index and pH stability index. Before the weighted summation, each index is converted into a dimensionless index through standard normalization. The comprehensive operational cost is the total resource consumption required to perform control actions on a node. The total resource consumption is obtained by weighted summation of multiple heterogeneous resource indicators, including energy consumption, water consumption, and human intervention. Before weighted summation, each heterogeneous resource indicator is converted into a dimensionless indicator through standard normalization. S53. Construct an enhanced composting parameter evolution graph, wherein the edge weight of each edge of the enhanced composting parameter evolution graph is driven by the state change cost and control incentive, and define an enhanced edge set. S54. Based on the enhanced composting parameter evolution map, perform a maximum incentive path search from the current node to the target maturity node to generate the optimal incentive guidance path: Among them, P * This represents the optimal incentive-guided path, describing the path sequence that maximizes the cumulative control incentive value. Represents the set of all feasible paths from the current node to the target mature node, (v a →v b ) represents any edge in the path, Γ a,b Represents the edge (v) a →v b The corresponding stimulus value, argmax represents the variable value when the function value reaches its maximum value, w a,b Represents the edge (v) a →v b The original edge weight of ) is the cost of state change, β represents the cost inhibition factor, which controls the tolerance for high-cost paths, P represents any path in the set of feasible paths, and ∈ means "belongs to".

7. The intelligent monitoring system for composting parameters based on the Internet of Things according to claim 2, characterized in that, S6 includes the following specific steps: S61. Based on the optimal incentive-guided path, extract the composting parameter state matrix corresponding to each node in the path, and select the target composting node of the path. S62. For any node in the optimal incentive guidance path, use the Euclidean distance under the Frobenius norm to calculate the state difference between the current composting parameter state vector and the target composting parameter state vector. S63. Input the state difference value of each node into the trend deviation risk scoring function to obtain the trend deviation risk score corresponding to each node on the path. S64. Arrange the trend deviation risk scores of all nodes in the path time sequence to form a trend deviation risk indicator sequence; S65. Set a risk warning threshold. If the trend deviation risk score is greater than the risk warning threshold, it is determined that the current composting process has an abnormal trend that deviates from the target composting state at the corresponding stage. S66. When an abnormal trend is determined to exist, construct an intelligent early warning information including a trend deviation risk score, a state difference value, and a corresponding composting parameter state vector, and upload it to a remote cloud processing platform through an IoT communication module for users to view remotely and assist in decision-making.

Images