CN113837428B - Sensor optimization layout and temperature and humidity prediction algorithm for raw smoke maintenance - Google Patents
Sensor optimization layout and temperature and humidity prediction algorithm for raw smoke maintenance Download PDFInfo
- Publication number
- CN113837428B CN113837428B CN202110726689.6A CN202110726689A CN113837428B CN 113837428 B CN113837428 B CN 113837428B CN 202110726689 A CN202110726689 A CN 202110726689A CN 113837428 B CN113837428 B CN 113837428B
- Authority
- CN
- China
- Prior art keywords
- temperature
- sensor
- humidity
- stack
- sensors
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- Human Resources & Organizations (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Strategic Management (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Molecular Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Biology (AREA)
- Development Economics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Game Theory and Decision Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a sensor optimization layout and a temperature and humidity prediction algorithm for raw tobacco curing, which belong to the field of tobacco, and comprise the following steps: step 1, analyzing sensor data, including temperature and humidity data and sensor position data; step 2, analyzing the arrangement positions of the sensors, including analyzing the data of abnormal points, analyzing the positions of the sensors which can be arranged in the smoke stack, and analyzing the number of the sensors; and 3, planning the arrangement positions of the sensors according to the temperature and humidity information, wherein the method comprises the steps of determining the number of the sensors required by the stack A and determining the layout scheme of the stack A. According to the invention, the problems of the number, unit price, installation cost, disassembly and assembly cost, coverage rate and the like of the temperature and humidity sensors are comprehensively considered, the cost for arranging the temperature and humidity sensors is minimized, the objective function for sensing the coverage rate is maximized, and the multi-raw tobacco stack temperature and humidity sensing layout optimization mathematical model is established.
Description
Technical Field
The invention belongs to the field of tobacco, and particularly relates to an optimized layout of a sensor for raw tobacco curing and a temperature and humidity prediction algorithm.
Background
Under the current industrial structure of the Chinese tobacco industry, the main problem of logistics center construction is the construction of a storage system, tobacco companies are arranged according to the division of administrative regions, each administrative region can have a corresponding storage goods yard for stacking raw cigarettes, and environmental factors such as sunlight, temperature, humidity and the like between different goods yards are also different. The original tobacco warehouse is pushed to be converted and upgraded from 'storage' to 'curing' according to the tobacco group job title, each goods yard is required to carry out fine maintenance on an original tobacco stack, and the scientific management and control on the microenvironment of the original tobacco stack are achieved by arranging temperature/humidity sensors on the original tobacco stack. The monitoring function of the tobacco stack microenvironment is realized through temperature/humidity, and the more the number of temperature/humidity sensor arrangements are, the better the accuracy of monitoring data in the tobacco stack is. However, because the space volume of a single smoke stack is larger, and the number of smoke stacks in a goods yard is also larger, if the temperature/humidity sensor is not reasonably planned, the labor resource and the purchase cost of the sensor required by the whole process are greatly increased. Therefore, arranging a limited number of sensors in a monitoring system to obtain as much micro-environmental information as possible is a complex sensor layout optimization problem, which is also a core fundamental problem of tobacco company logistics center sensor layout optimization operations.
Disclosure of Invention
According to the invention, the problems of the number, unit price, installation cost, disassembly and assembly cost, coverage rate and the like of the temperature and humidity sensors are comprehensively considered, the cost for arranging the temperature and humidity sensors is minimized, the objective function for sensing the coverage rate is maximized, and the multi-raw tobacco stack temperature and humidity sensing layout optimization mathematical model is established.
In order to achieve the above purpose, the present invention is implemented by the following technical scheme: the sensor optimization layout and the temperature and humidity prediction algorithm comprise the following steps: step 1, analyzing sensor data, including temperature and humidity data and sensor position data; step 2, analyzing the arrangement positions of the sensors, including analyzing the data of abnormal points, analyzing the positions of the sensors which can be arranged in the smoke stack, and analyzing the number of the sensors; and 3, planning the arrangement positions of the sensors according to the temperature and humidity information, wherein the method comprises the steps of determining the number of the sensors required by the stack A and determining the layout scheme of the stack A.
Preferably, the number of sensors and the layout positions in the step 1, the step 2 and the step 3 are determined to ensure the purposes of minimum total layout cost and maximum coverage; cost of raw tobacco stack a in the process of arranging the sensor:
wherein:
M sensor_A =numel(X A_i )
X A_i =(x A_i ,y A_i ,z A_i )
total cost of arranging the sensors:
coverage of temperature/humidity sensor in the a-th raw tobacco stack:
setting: the sensing radius of the temperature/humidity sensor in the raw tobacco stack is R damp_A
The total coverage is:
wherein:
objective function:
Ω abnor_A ;
a is a first raw tobacco stack;
k layer of original tobacco stack;
P A the cost required for arranging the temperature/humidity sensor for the A-th smoke stack;
M sensor_A the number of sensors arranged in the A-th raw tobacco stack;
P sensor unit price of temperature/humidity sensor;
P installation the cost required for installing a temperature/humidity sensor;
P smokestack_A disassembly of the pack required during installation of the temperature/humidity sensor on the A-th packFixed cost of unloading and re-stacking;
X A_i position set X of temperature/humidity sensors in A-th raw tobacco stack A_i =(x A_i ,y A_i ,z A_i );
P tatal The total cost required in the process of arranging the temperature/humidity sensor;
R damp_A the sensing radius of the temperature/humidity sensor of the A-th raw tobacco stack;
V perception_A the temperature and humidity sensor senses the volume in the smoke stack A in space;
V butt_A the volume of the A-th raw tobacco stack;
B cover_A space coverage of a temperature/humidity sensor in the A-th raw tobacco stack;
B total total coverage of the temperature/humidity sensor;
ω k representing the weight of the k-th layer;
α A_1 representing a path from an ith delivery point to a jth receiving point;
α A_2 representing a path from a j-th receiving point to a j' th receiving point;
Ω A representing a set of possible placement positions of the temperature/humidity sensor within the pack a space;
Ω abnor_A representing a temperature anomaly position set of the temperature/humidity sensor in the space of the smoke stack A;
Ω obstacle_A indicating that the temperature/humidity sensor cannot be arranged in the pack a space.
Preferably, the cost of the raw tobacco stack A in the process of arranging the sensors is as follows:
constraint conditions:
the number of temperature/humidity sensors cannot exceed the upper and lower limits of the number of sensors arranged for each raw tobacco stack:
0≤M sensor_A ≤80M sensor_A ∈Z
the abnormal temperature and humidity points in the raw tobacco stack A are included in the perception range of the temperature/humidity sensor:
Ω abnor_A ={(x abnor_A_i ,y abnor_A_i ,z abnor_A_i )|(x abnor_A_i ,y abnor_A_i ,z abnor_A_i )
a smoke stack temperature anomaly point }
Spatial range limitations of the sensor in the stack:
Ω A ={(x A_i ,y A_i ,z A_i )|x A_min ≤x A_i ≤x A_max ,y A_min ≤y A_i ≤y A_max ,z A_min ≤z A_i ≤z A_max }
X A_i =(x A_i ,y A_i ,z A_i )∈Ω(i=1,2,···,M;A=1,2,···,N)
temperature/humidity sensor position limitations cannot be placed in the pack space:
Ω obstacle_A ={(x obstacle_A ,y obstacle_A ,z obstacle_A )|(x obstacle_A ,y obstacle_A ,z obstacle_A )
a position where the A-th pack cannot be provided with a sensor }
A is a first raw tobacco stack;
k layer of original tobacco stack;
P A the cost required for arranging the temperature/humidity sensor for the A-th smoke stack;
M sensor_A the number of sensors arranged in the A-th raw tobacco stack;
P sensor unit price of temperature/humidity sensor;
P installation the cost required for installing a temperature/humidity sensor;
P smokestack_A the fixed cost for disassembling and re-stacking the original tobacco stacks required in the process of installing the temperature/humidity sensor on the A-th original tobacco stack;
X A_i position set X of temperature/humidity sensors in A-th raw tobacco stack A_i =(x A_i ,y A_i ,z A_i );
P tatal The total cost required in the process of arranging the temperature/humidity sensor;
R damp_A the sensing radius of the temperature/humidity sensor of the A-th raw tobacco stack;
V perception_A the temperature and humidity sensor senses the volume in the smoke stack A in space;
V butt_A the volume of the A-th raw tobacco stack;
B cover_A space coverage of a temperature/humidity sensor in the A-th raw tobacco stack;
B total total coverage of the temperature/humidity sensor;
ω k representing the weight of the k-th layer;
α A_1 representing a path from an ith delivery point to a jth receiving point;
α A_2 representing a path from a j-th receiving point to a j' th receiving point;
Ω A representing a set of possible placement positions of the temperature/humidity sensor within the pack a space;
Ω abnor_A representing a temperature anomaly position set of the temperature/humidity sensor in the space of the smoke stack A;
Ω obstacle_A indicating that the temperature/humidity sensor cannot be arranged in the pack a space.
Preferably, the step 1, the sensor data analysis, including temperature and humidity data and sensor position data; the method comprises the steps of inputting historical data collected by a wireless temperature and humidity sensor as initial data to a clustering algorithm to be clustered through introducing a K-means clustering algorithm, then conducting wireless temperature and humidity sensor layout optimization on a clustering result in a tobacco stack space through improving a particle swarm algorithm, and finally inputting and learning the historical data collected in the earlier stage through adopting an LSTM neural network and predicting the temperature and humidity in the tobacco stack.
Preferably, the K-means clustering algorithm is to randomly select K sample points as initial clustering centers, calculate the distance from each data to the selected sample centers, distribute each sample point to each corresponding cluster according to the principle of the closest distance, find a new clustering center by calculating the mean value of each cluster, and iterate until convergence conditions are met;
the distance between any two samples in the computation space is:
the new cluster center after iteration is completed is as follows:
6. the raw smoke maintenance sensor optimization layout and temperature and humidity prediction algorithm according to claim 4, wherein the raw smoke maintenance sensor optimization layout and temperature and humidity prediction algorithm is characterized in that: the particle swarm algorithm (Particle Swarm Optimization, PSO) updates the evolution formula as follows:
v ij (t+1)=w·v ij (t)+c 1 r 1 (t)[p ij (t)-x ij (t)]+c 2 r 2 (t)[p gi (t)-x ij (t)]
x ij (t+1)=x ij (t)+v ji (t+1)
however, as the PSO algorithm has lower convergence efficiency in the later solving process, in order to expand the search space and improve the search efficiency, a self-adaptive strategy improvement algorithm of dynamic acceleration constant speed and linear decreasing inertia weight is introduced, and the improvement formula is as follows:
preferably, to solve the drawbacks of the conventional LSTM neural network and to make the algorithm more suitable for temperature and humidity prediction in an original smoke stack, a time-based back propagation algorithm (back propagation trough time, BPTT) and an adaptive momentum estimation algorithm (adaptive moment estimation, adam) improve the LSTM network model training process.
The invention has the beneficial effects that:
according to the invention, the problems of the number, unit price, installation cost, disassembly and assembly cost, coverage rate and the like of the temperature and humidity sensors are comprehensively considered, the cost for arranging the temperature and humidity sensors is minimized, the objective function for sensing the coverage rate is maximized, and the multi-raw tobacco stack temperature and humidity sensing layout optimization mathematical model is established.
Drawings
FIG. 1 is a schematic diagram of a temperature and humidity sensor layout optimization system;
FIG. 2, algorithm flow chart;
FIG. 3 is a flow chart of an improved particle swarm algorithm;
FIG. 4 is a flowchart of an improved LSTM neural network algorithm;
FIG. 5, raw tobacco stack model;
FIG. 6 is a schematic diagram of the layout position of the first layer of sensors of the raw tobacco stack;
Detailed Description
In order to facilitate the understanding and implementation of the present invention by those skilled in the art, the technical solution of the present invention will be further described with reference to the accompanying drawings and specific embodiments.
The internal texture of the grass stack is uniform
The temperature/humidity sensor layout optimization problem is to solve the optimal solution of the construction model, and the purposes of minimum total layout cost and maximum coverage are achieved by optimizing the distribution of the number of sensors and the layout scheme. According to the relevant investigation, the total layout cost of the sensors is related to the number of sensors, while the coverage is related to the number of sensor arrangements and perceived radius.
Cost of raw tobacco stack a in the process of arranging the sensor:
wherein:
M sensor_A =numel(X A_i )
X A_i =(x A_i ,y A_i ,z A_i )
total cost of arranging the sensors:
coverage of temperature/humidity sensor in the a-th raw tobacco stack:
setting: the sensing radius of the temperature/humidity sensor in the raw tobacco stack is R damp_A
The total coverage is:
wherein:
objective function:
Ω abnor_A
constraint conditions:
the number of temperature/humidity sensors cannot exceed the upper and lower limits of the number of sensors arranged for each raw tobacco stack:
0≤M sensor_A ≤80 M sensor_A ∈Z
the abnormal temperature and humidity points in the raw tobacco stack A are included in the perception range of the temperature/humidity sensor:
Ω abnor_A ={(x abnor_A_i ,y abnor_A_i ,z abnor_A_i )|(x abnor_A_i ,y abnor_A_i ,z abnor_A_i )
a smoke stack temperature anomaly point }
Spatial range limitations of the sensor in the stack:
Ω A ={(x A_i ,y A_i ,z A_i )|x A_min ≤x A_i ≤x A_max ,y A_min ≤y A_i ≤y A_max ,z A_min ≤z A_i ≤z A_max }
X A_i =(x A_i ,y A_i ,z A_i )∈Ω(i=1,2,···,M;A=1,2,···,N)
temperature/humidity sensor position limitations cannot be placed in the pack space:
Ω obstacle_A ={(x obstacle_A ,y obstacle_A ,z obstacle_A )|(x obstacle_A ,y obstacle_A ,z obstacle_A )
a position where the A-th pack cannot be provided with a sensor }
In the design process of an original tobacco curing temperature and humidity sensor layout scheme algorithm, the historical data collected by a wireless temperature and humidity sensor is used as initial data to be input into a clustering algorithm for clustering by introducing a K-means clustering algorithm, then the wireless temperature and humidity sensor layout optimization is carried out on the tobacco stack space by means of a clustering result by improving a particle swarm algorithm, and finally the historical data collected in the earlier stage is input and learned by adopting an LSTM neural network and the temperature and humidity in the tobacco stack are predicted.
K-means clustering algorithm;
the K-means clustering algorithm is to randomly select K sample points as initial clustering centers, calculate the distance from each data to the selected sample centers, distribute each sample point to each corresponding cluster according to the principle of the closest distance, find a new clustering center by calculating the mean value of each cluster, and iterate until convergence conditions are met.
The distance between any two samples in the computation space is:
the new cluster center after iteration is completed is as follows:
particle swarm algorithm:
particle swarm optimization (Particle Swarm Optimization, PSO) is a method proposed by James Kennedy and Russel Eberhart et al to reference the activity laws of a bird swarm, which simulates the process of birds searching for food. By randomly initializing a solution for the particles, the particles start searching on an initial track according to the initial solution, if the searching result is not good, the original searching track is deviated to start searching for a better solution, and if other particles find a better solution in the searching process, the other particles follow searching, and the updated evolution formula is as follows:
v ij (t+1)=w·v ij (t)+c 1 r 1 (t)[p ij (t)-x ij (t)]+c 2 r 2 (t)[p gi (t)-x ij (t)]
x ij (t+1)=x ij (t)+v ji (t+1)
however, as the PSO algorithm has lower convergence efficiency in the later solving process, in order to expand the search space and improve the search efficiency, a self-adaptive strategy improvement algorithm of dynamic acceleration constant speed and linear decreasing inertia weight is introduced, and the improvement formula is as follows:
the specific implementation concept is as shown in fig. 3:
traditional LSTM neural network:
the LSTM neural network is a special RNNs and can well solve the problem of long-time dependence. The main characterization is the relationship between the current output of a sequence and the previous data. The LSTM algorithm flow mainly comprises three parts: first, it is decided to select to discard information from the old state of the cell, and the information is recorded as S-layer input through forget gate, S-layer output [0,1]Numbers of intervals, wherein 1 is fully reserved, 0 is fully forgotten, f t For the reserved amount of old state information, the current input is x t The input at the previous time is h t-1 Delta is S layer function, W f The forget gate is marked as a weight vector of an S layer, b f The amount of information reserved for the S-layer threshold is:
f i =δ(W f ·[h t-1 ,x t ]+b f )
then select and determine new information stored in the cell state, update the cell state
And (3) determining and updating the abnormal operation rate through the S layer:
i t =δ(W i ·[h t-1 ,x i ]+b i )
creating a new candidate value vector in an addable state by performing one operation through the tanh layer:
updating the cell state:
a decision unit state output section that decides a unit state to be output through an S layer of the "output gate":
O t =δ(W 0 ·[h t-1 ,x t ]+b 0 )
h t =O t tanh(C t )
LSTM neural network principle
An LSTM network is a special RNN network, which mainly comprises the following three characteristics:
1) The recurrent neural network is capable of producing an output at each time node, and the connections between hidden units are recurrent;
2) The cyclic neural network can generate one output at each time node, and the output on the time node is only in cyclic connection with the hidden unit of the next time node;
3) The recurrent neural network comprises hidden units with recurrent connections and is capable of processing sequence data and outputting a single prediction.
RNNs, however, have difficulty in dealing with long dependency sequence problems.
In order to solve the defects of the traditional LSTM neural network and make the algorithm more suitable for the temperature and humidity prediction in the original tobacco stack, a time-based back propagation algorithm (back propagation trough time, BPTT) and an adaptive momentum estimation algorithm (adaptive moment estimation, adam) improve the training process of the LSTM network model. The specific implementation concept is as follows in fig. 4:
in the original smoke maintenance sensor layout problem, an LSTM neural network is introduced, collected historical temperature and humidity data is used as a sample to be input into the LSTM neural network for learning, and an LSTM algorithm is improved based on a back propagation algorithm (back propagation trough time, BPTT) and an adaptive momentum estimation algorithm (adaptive moment estimation, adam). And calculating and transmitting a parameter gradient error by the BPTT, and updating the network weight by an Adam optimization algorithm.
Partial simulation and comparison of sensor layout optimization model:
in a certain simulation experiment, simulation analysis is carried out on one layer of a certain raw tobacco stack, wherein the overall dimension of the raw tobacco stack, the external temperature and humidity environment, the arrangement position and other information of a temperature and humidity sensor, the parameters of the temperature and humidity sensor and the like are all known. The size and shape of a certain stack are shown in fig. 5, the arrangement positions of the sensors are shown in fig. 6, and the temperature and humidity data of a certain moment of a sensor node are shown in table 1:
TABLE 1 temperature and humidity node information for first layer of some raw tobacco stack
The results of the partial simulation are shown below, wherein tables 2, 3, 4 and 5 are the position and temperature and humidity data of the temperature and humidity clustering centers with 7 and 8 clustering centers respectively.
Table 2 temperature clustering results when 2K =8
Table 3 temperature clustering results when 3K =8
Table 4 humidity clustering results at 4K =7
Table 5k=8 humidity clustering results
The foregoing has shown and described the basic principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the above-described embodiments, and that the above-described embodiments and descriptions are only preferred embodiments of the present invention, and are not intended to limit the invention, and that various changes and modifications may be made therein without departing from the spirit and scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (5)
1. The sensor optimization layout and temperature and humidity prediction algorithm for raw smoke maintenance is characterized in that: the sensor optimization layout and the temperature and humidity prediction algorithm comprise the following steps: step 1, analyzing sensor data, including temperature and humidity data and sensor position data; step 2, analyzing the arrangement positions of the sensors, including analyzing the data of abnormal points, analyzing the positions of the sensors which can be arranged in the smoke stack, and analyzing the number of the sensors; step 3, planning the arrangement positions of the sensors according to the temperature and humidity information, wherein the step comprises the steps of determining the number of the sensors required by the stack A and determining the layout scheme of the stack A;
the number and layout positions of the sensors are determined in the step 1, the step 2 and the step 3, so that the purposes of minimum total layout cost and maximum coverage are ensured; cost of raw tobacco stack a in the process of arranging the sensor:
P A =M sensor_A (P sensor +P installation )+P smokestack_A
wherein:
M sensor_A =numel(X A_i )
X A_i =(x A_i ,y A_i ,z A_i )
total cost of arranging the sensors:
coverage of temperature/humidity sensor in the a-th raw tobacco stack:
setting: the sensing radius of the temperature/humidity sensor in the raw tobacco stack is R damp_A
The total coverage is:
wherein:
objective function:
Ω abnor_A ;
the cost of the original smoke stack A in the process of arranging the sensors is as follows, and the constraint condition of the coverage rate of the temperature/humidity sensor in the A-th original smoke stack is as follows:
constraint conditions:
the number of temperature/humidity sensors cannot exceed the upper and lower limits of the number of sensors arranged for each raw tobacco stack:
0≤M sensor_A ≤80 M sensor_A ∈Z
the abnormal temperature and humidity points in the raw tobacco stack A are included in the perception range of the temperature/humidity sensor:
Ω abnor_A ={(x abnor_A_i ,y abnor_A_i ,z abnor_A_i )|(x abnor_A_i ,y abnor_A_i ,z abnor_A_i )
a smoke stack temperature anomaly point }
Spatial range limitations of the sensor in the stack:
Ω A ={(x A_i ,y A_i ,z A_i )|x A_min ≤x A_i ≤x A_max ,y A_min ≤y A_i ≤y A_max ,z A_min ≤z A_i ≤z A_max }
X A_i =(x A_i ,y A_i ,z A_i )∈Ω(i=1,2,···,M;A=1,2,···,N)
temperature/humidity sensor position limitations cannot be placed in the pack space:
Ω obstacle_A ={(x obstacle_A ,y obstacle_A ,z obstacle_A )|(x obstacle_A ,y obstacle_A ,z obstacle_A )
a position where the A-th pack cannot be provided with a sensor }
A is a first raw tobacco stack;
k layer of original tobacco stack;
P A the cost required for arranging the temperature/humidity sensor for the A-th smoke stack;
M sensor_A the number of sensors arranged in the A-th raw tobacco stack;
P sensor unit price of temperature/humidity sensor;
P installation the cost required for installing a temperature/humidity sensor;
P smokestack_A disassembly of the pack required during installation of the temperature/humidity sensor on the A-th pack
Fixed cost of unloading and re-stacking;
X A_i position set X of temperature/humidity sensors in A-th raw tobacco stack A_i =(x A_i ,y A_i ,z A_i );
P tatal The total cost required in the process of arranging the temperature/humidity sensor;
R damp_A the sensing radius of the temperature/humidity sensor of the A-th raw tobacco stack;
V perception_A the temperature and humidity sensor senses the volume in the smoke stack A in space;
V butt_A the volume of the A-th raw tobacco stack;
B cover_A space coverage of a temperature/humidity sensor in the A-th raw tobacco stack;
B total total coverage of the temperature/humidity sensor;
ω k representing the weight of the k-th layer;
α A_1 representing a path from an ith delivery point to a jth receiving point;
α A_2 representing a path from a j-th receiving point to a j' th receiving point;
Ω A representing a set of possible placement positions of the temperature/humidity sensor within the pack a space;
Ω abnor_A representing a temperature anomaly position set of the temperature/humidity sensor in the space of the smoke stack A;
Ω obstacle_A indicating that the temperature/humidity sensor cannot be arranged in the pack a space.
2. The raw smoke maintenance sensor optimization layout and temperature and humidity prediction algorithm according to claim 1, wherein the raw smoke maintenance sensor optimization layout and temperature and humidity prediction algorithm is characterized in that: the sensor data analysis comprises temperature and humidity data and sensor position data; the method comprises the steps of inputting historical data collected by a wireless temperature and humidity sensor as initial data to a clustering algorithm to be clustered through introducing a K-means clustering algorithm, then conducting wireless temperature and humidity sensor layout optimization on a clustering result in a tobacco stack space through improving a particle swarm algorithm, and finally inputting and learning the historical data collected in the earlier stage through adopting an LSTM neural network and predicting the temperature and humidity in the tobacco stack.
3. The raw smoke maintenance sensor optimization layout and temperature and humidity prediction algorithm according to claim 2, wherein the raw smoke maintenance sensor optimization layout and temperature and humidity prediction algorithm is characterized in that: the K-means clustering algorithm is characterized in that K sample points are randomly selected to serve as initial clustering centers, distances from each data to the selected sample centers are calculated, each sample point is distributed to each corresponding cluster according to the principle of closest distance, a new clustering center is found by calculating the mean value of each cluster, and iteration is carried out until convergence conditions are met;
the distance between any two samples in the computation space is:
the new cluster center after iteration is completed is as follows:
4. the raw smoke maintenance sensor optimization layout and temperature and humidity prediction algorithm according to claim 2, wherein the raw smoke maintenance sensor optimization layout and temperature and humidity prediction algorithm is characterized in that: the particle swarm algorithm (Particle Swarm Optimization, PSO) updates the evolution formula as follows:
v ij (t+1)=w·v ij (t)+c 1 r 1 (t)[p ij (t)-x ij (t)]+c 2 r 2 (t)[p gi (t)-x ij (t)]
x ij (t+1)=x ij (t)+v ji (t+1)
however, as the PSO algorithm has lower convergence efficiency in the later solving process, in order to expand the search space and improve the search efficiency, a self-adaptive strategy improvement algorithm of dynamic acceleration constant speed and linear decreasing inertia weight is introduced, and the improvement formula is as follows:
5. the raw smoke maintenance sensor optimization layout and temperature and humidity prediction algorithm according to claim 2, wherein the raw smoke maintenance sensor optimization layout and temperature and humidity prediction algorithm is characterized in that: in order to solve the defects of the traditional LSTM neural network and enable the algorithm to be more suitable for temperature and humidity prediction in an original tobacco stack, the LSTM network model training process is improved by a time-based back propagation algorithm and an adaptive momentum estimation algorithm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110726689.6A CN113837428B (en) | 2021-06-29 | 2021-06-29 | Sensor optimization layout and temperature and humidity prediction algorithm for raw smoke maintenance |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110726689.6A CN113837428B (en) | 2021-06-29 | 2021-06-29 | Sensor optimization layout and temperature and humidity prediction algorithm for raw smoke maintenance |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113837428A CN113837428A (en) | 2021-12-24 |
CN113837428B true CN113837428B (en) | 2023-08-25 |
Family
ID=78962726
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110726689.6A Active CN113837428B (en) | 2021-06-29 | 2021-06-29 | Sensor optimization layout and temperature and humidity prediction algorithm for raw smoke maintenance |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113837428B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114980131A (en) * | 2022-03-24 | 2022-08-30 | 红云红河烟草(集团)有限责任公司 | Raw cigarette tray wireless sensor layout optimization method based on improved particle swarm optimization |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR102129589B1 (en) * | 2019-02-12 | 2020-07-08 | 서울대학교산학협력단 | Method for sensor network development and system |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20130062410A (en) * | 2011-11-10 | 2013-06-13 | 한국전자통신연구원 | Apparatus for automatically arranging building energy control sensors and method thereof |
CA3078987C (en) * | 2017-10-11 | 2023-06-13 | Oneevent Technologies, Inc. | Fire detection system |
US11729190B2 (en) * | 2019-10-29 | 2023-08-15 | General Electric Company | Virtual sensor supervised learning for cyber-attack neutralization |
-
2021
- 2021-06-29 CN CN202110726689.6A patent/CN113837428B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR102129589B1 (en) * | 2019-02-12 | 2020-07-08 | 서울대학교산학협력단 | Method for sensor network development and system |
Non-Patent Citations (1)
Title |
---|
基于改进粒子群算法和特征点集的无线传感器网络覆盖问题研究;丁旭;吴晓蓓;黄成;;电子学报(第04期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN113837428A (en) | 2021-12-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Huang et al. | A static bike repositioning model in a hub-and-spoke network framework | |
CN104318329A (en) | Power load forecasting method of cuckoo search algorithm improved support vector machine | |
CN114678080B (en) | Converter end point phosphorus content prediction model, construction method and phosphorus content prediction method | |
CN102200759A (en) | Nonlinear kernelled adaptive prediction method | |
CN113704956A (en) | Urban road online microscopic simulation method and system based on digital twin technology | |
Zadeh et al. | An efficient metamodel-based multi-objective multidisciplinary design optimization framework | |
CN112101684A (en) | Plug-in hybrid electric vehicle real-time energy management method and system | |
CN113837428B (en) | Sensor optimization layout and temperature and humidity prediction algorithm for raw smoke maintenance | |
CN113449919B (en) | Power consumption prediction method and system based on feature and trend perception | |
CN112884236B (en) | Short-term load prediction method and system based on VDM decomposition and LSTM improvement | |
WO2024077969A1 (en) | Lstm-svr subway station temperature prediction method based on characteristic of multiple periods | |
CN114167898A (en) | Global path planning method and system for data collection of unmanned aerial vehicle | |
CN111898867A (en) | Airplane final assembly production line productivity prediction method based on deep neural network | |
CN110807490A (en) | Intelligent prediction method for construction cost of power transmission line based on single-base tower | |
CN111815026A (en) | Multi-energy system load prediction method based on feature clustering | |
CN111311001B (en) | Bi-LSTM network short-term load prediction method based on DBSCAN algorithm and feature selection | |
CN115271237A (en) | Industrial data quality prediction method based on improved PSO-GA and SVM | |
CN116187835A (en) | Data-driven-based method and system for estimating theoretical line loss interval of transformer area | |
CN112036598A (en) | Charging pile use information prediction method based on multi-information coupling | |
CN111882363A (en) | Sales prediction method, system and terminal | |
Zhou et al. | A new Co/Co $ _2 $ prediction model based on labeled and unlabeled process data for sintering process | |
CN111882114A (en) | Short-term traffic flow prediction model construction method and prediction method | |
CN112257348B (en) | Method for predicting long-term degradation trend of lithium battery | |
Tian et al. | Dynamic operation optimization based on improved dynamic multi-objective dragonfly algorithm in continuous annealing process. | |
CN117454765A (en) | Copper smelting furnace spray gun service life prediction method based on IPSO-BP neural network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |