CN116307076A - Industrial park energy efficiency management and control method based on Internet of things - Google Patents
Industrial park energy efficiency management and control method based on Internet of things Download PDFInfo
- Publication number
- CN116307076A CN116307076A CN202310051194.7A CN202310051194A CN116307076A CN 116307076 A CN116307076 A CN 116307076A CN 202310051194 A CN202310051194 A CN 202310051194A CN 116307076 A CN116307076 A CN 116307076A
- Authority
- CN
- China
- Prior art keywords
- energy consumption
- data
- park
- internet
- things
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000005265 energy consumption Methods 0.000 claims abstract description 121
- 239000002245 particle Substances 0.000 claims abstract description 41
- 239000013598 vector Substances 0.000 claims abstract description 27
- 238000001514 detection method Methods 0.000 claims abstract description 26
- 238000005457 optimization Methods 0.000 claims abstract description 16
- 238000007619 statistical method Methods 0.000 claims abstract description 16
- 238000007781 pre-processing Methods 0.000 claims abstract description 9
- 230000008569 process Effects 0.000 claims abstract description 6
- 238000007726 management method Methods 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 16
- 230000002159 abnormal effect Effects 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 11
- 238000004458 analytical method Methods 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 7
- 238000004378 air conditioning Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 6
- 230000013016 learning Effects 0.000 claims description 5
- 238000004891 communication Methods 0.000 claims description 4
- 238000010219 correlation analysis Methods 0.000 claims description 4
- 238000013500 data storage Methods 0.000 claims description 4
- 230000005611 electricity Effects 0.000 claims description 4
- 238000001556 precipitation Methods 0.000 claims description 4
- 230000001932 seasonal effect Effects 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 3
- 230000015572 biosynthetic process Effects 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 claims description 3
- BTCSSZJGUNDROE-UHFFFAOYSA-N gamma-aminobutyric acid Chemical compound NCCCC(O)=O BTCSSZJGUNDROE-UHFFFAOYSA-N 0.000 claims description 3
- 238000012417 linear regression Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 230000009326 social learning Effects 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000005286 illumination Methods 0.000 claims description 2
- 230000009286 beneficial effect Effects 0.000 abstract description 4
- 238000012544 monitoring process Methods 0.000 abstract description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000012550 audit Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000013439 planning Methods 0.000 description 1
- 229920006395 saturated elastomer Polymers 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 230000001550 time effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Images
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
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16Y—INFORMATION AND COMMUNICATION TECHNOLOGY SPECIALLY ADAPTED FOR THE INTERNET OF THINGS [IoT]
- G16Y10/00—Economic sectors
- G16Y10/35—Utilities, e.g. electricity, gas or water
-
- 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
- Y02P80/00—Climate change mitigation technologies for sector-wide applications
- Y02P80/10—Efficient use of energy, e.g. using compressed air or pressurized fluid as energy carrier
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Tourism & Hospitality (AREA)
- Software Systems (AREA)
- Marketing (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Development Economics (AREA)
- Water Supply & Treatment (AREA)
- Entrepreneurship & Innovation (AREA)
- Public Health (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Medical Informatics (AREA)
- Operations Research (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Game Theory and Decision Science (AREA)
- Mathematical Physics (AREA)
- Accounting & Taxation (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an industrial park energy efficiency control method based on the Internet of things, and relates to the fields of the Internet of things, energy efficiency control and the like. The method comprises the processes of data acquisition, statistical analysis, energy consumption prediction, energy consumption early warning and the like, wherein various sensors and detection instruments are used for acquiring energy consumption data of an industrial park, and the energy consumption data are uploaded to an edge computing gateway for statistical analysis; parameter optimization is carried out by using an improved particle swarm intelligent optimization algorithm, a support vector regression energy consumption prediction model based on particle swarm optimization is established, real-time prediction of the total energy consumption of the park is realized, and over-value early warning is carried out based on an energy consumption prediction result so as to reduce the energy consumption of the park. The intelligent park energy consumption prediction method based on the Internet of things realizes the application of the Internet of things in intelligent park energy efficiency management and control, performs data preprocessing in combination with the edge computing gateway, considers and builds time and season characteristics, builds a multivariable energy consumption prediction model, is beneficial to park energy consumption monitoring and optimization, and improves energy utilization efficiency.
Description
Technical Field
The invention belongs to the technical field of energy efficiency control, and particularly relates to an industrial park energy efficiency control method based on the Internet of things.
Background
With the development of the internet of things technology, management modes of various parks are gradually intelligent and systematic. The energy consumption problem is always one of core problems which need to be focused on in the production operation of the park, the prediction and optimization of the energy consumption is an important support for tasks such as overall planning, energy audit and the like of the park, and the method has important significance in improving the energy efficiency level and optimizing the management and control process. The intelligent park is complex in structure, changeable in scene and easy to influence by the restriction of various factors in energy consumption, and the application of the Internet of things technology provides an effective method for energy efficiency management and control of the intelligent park.
The traditional energy consumption prediction is generally based on a mechanism model and is built on physical parameters and thermodynamic equilibrium equations, so that the calculated amount is large, the model complexity is high, and the application is complicated. The prediction method based on the data driving is used for completing the prediction of the future energy consumption by means of past historical data, wherein the data driving method based on the machine learning algorithm is easier to obtain more accurate prediction. The support vector regression model (SVR) is a new model of a machine learning algorithm, namely a Support Vector Machine (SVM) method, and the SVR can map data to a high-dimensional feature space, convert an optimization problem into a convex quadratic programming problem and realize modeling of nonlinear data.
The industrial park energy efficiency management and control method based on the Internet of things is researched, and the park energy consumption is predicted by using a support vector regression model.
Disclosure of Invention
In order to solve the problem of energy efficiency control of the intelligent park in the prior art, the invention aims to provide an industrial park energy efficiency control method based on the Internet of things, and an SVR energy consumption prediction model based on improved PSO optimization is established, so that the method has the characteristics of high prediction speed, high precision and long prediction time effect, and can support and improve the energy efficiency control of the industrial park.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
an industrial park energy efficiency management and control method based on the Internet of things comprises the following steps:
step 1, collecting energy consumption data and environment data of an industrial park, uploading the energy consumption data and the environment data to an edge computing gateway and processing abnormal values;
step 2, a total energy consumption model is established, statistical analysis is carried out on the collected data, and sample data are generated; the statistical analysis comprises data preprocessing, data storage, correlation analysis and data normalization;
step 3, establishing a support vector regression-improved particle swarm optimization (PSO-SVR) energy consumption prediction model, and carrying out model training by using the sample data to carry out energy consumption prediction;
and 4, carrying out energy consumption early warning by combining the current period electricity consumption according to the energy consumption predicted value.
In one embodiment, step 1, the detection sensor, the detection instrument and the data acquisition module for data acquisition are connected in a distributed manner through the node of the internet of things; the detection sensor comprises a temperature sensor and a humidity sensor, and the detection instrument comprises a smart meter; the temperature sensor and the humidity sensor adopt an NB-IoT communication protocol for data transmission; the intelligent ammeter collects electric parameters of power consumption equipment, including voltage, current, power, harmonic wave and power factor.
In one embodiment, the step 1, the initial data sequence received at the edge computing gateway is expressed as:
{x 1 (1),…,x 1 (k),x 2 (1),…,x 2 (k),…,x i (1),…,x i (k),time}
wherein x is i (k) Kth data representing i factors, i factors referring to the same type of detection sensor or detection instrumentThe collected data, k is any positive integer except 0, refers to the number of the detection sensors or the detection meters, the time represents the receiving time of the data, and the abnormal value processing is carried out on the received initial data sequence: abnormal data is detected by using a quarter bit distance (IQR) abnormal value identification method, and the abnormal value is modified to be a normal value by adopting a method of replacing the median of adjacent data.
In one embodiment, the step 2, the total energy consumption model is expressed as follows:
wherein P is total For the total energy consumption of the park, M is the number of air conditioning units and P ac,i The energy consumption of the ith air conditioning unit equipment is N, the number of industrial energy consumption equipment is P equip,j For the energy consumption of the j-th industrial energy consumption equipment, Q is the number of auxiliary equipment including illumination and elevators, P other,k Energy consumption for the kth auxiliary device;
setting a park energy consumption index PPUE as an index for evaluating park energy efficiency, wherein:
P equip indicating the total energy consumption of the industrial energy consumption equipment, the PPUE value should be greater than 1, and the closer to 1, the higher the energy use efficiency.
In one embodiment, the statistical analysis is performed as follows:
s1: preprocessing data at an edge computing gateway, and computing indoor average temperature and humidity and total energy consumption data;
s2: acquiring data information including the flow of people, the weather of the place, the wind speed, the air pressure and the precipitation of an industrial park at a back-end server; the pearson correlation coefficient r is adopted to reflect the correlation among different types of data, and a plurality of characteristics with the highest correlation are selected as partial input characteristics of sample data, namely the average air temperature, the average humidity, the indoor average temperature and the people flow;
s3: performing feature construction and complementation, wherein the construction time refers to the feature: adding weekends, holidays and seasonal features;
s4: normalization of the data is necessary, and the variable range is unified, and the formula is as follows:
wherein x is min Is the minimum value of the sample data; x is x max Is the maximum value of the sample data; x is x * Is normalized value;
final formation of final sample data sequences y and x i As a training input to the energy consumption prediction model, it is expressed as:
wherein y (k) represents the total energy consumption value of the kth data, x i (k) Kth data representing factor i.
In one embodiment, the statistical analysis further includes trend analysis, flow direction statistics, monthly energy consumption analysis, energy consumption to cycle ratio analysis.
In one embodiment, the step 3, using a support vector regression method, performs linear regression in the space by mapping the training samples to the Gao Weina kernel-induced feature space, by the following formula:
wherein the output is a fitting function f (x), x i ,x j A spatial feature vector is input for the sample low-dimensional data,α i is Lagrangian multiplied bySon, b is displacement, < >>Is a radial basis with a wide convergence domain, which acts as a kernel function, σ being a kernel parameter;
the penalty parameter C and the kernel parameter sigma in the support vector regression are optimized by adopting an improved Particle Swarm Optimization (PSO) algorithm, and the process is as follows:
s1: constructing a support vector regression model and setting parameter information of a particle swarm algorithm;
s2: initializing a punishment parameter C and a nuclear parameter sigma of a support vector regression model, initializing the position and the speed of particles in a particle swarm algorithm, and then calculating the fitting goodness R2 of the support vector regression model of a training set as fitness;
s3: continuously iterating, updating the position and the speed of the particles, calculating the corresponding fitness, and recording a penalty parameter C and a core parameter sigma corresponding to the global optimal fitness; let each particle be composed of 2-dimensional parameter vectors (C, sigma), the position of the ith particle in the 2-dimensional solution space be x i =(x i1 ,x i2 ) Speed v i =(v i1 ,v i2 ) Recording the individual extremum of the current moment as p i =(p i1 ,p i2 ) Global extremum is noted g i =(g i1 ,g i2 ) The method comprises the steps of carrying out a first treatment on the surface of the In each iteration, the particle tracks the states of the individual extremum, the global extremum and the previous moment to adjust the position and the speed at the current moment, and the speed and the position formula are as follows:
v id (t+1)=ωv id (t)+c 1 rand()(p id -x id (t))+c 2 rand()(g id -x id (t))
x id (t+1)=x id (t)+v id (t+1)
wherein rand () is [0,1 ]]The random number in between, d is the dimension of the solution space, c 1 Is the individual learning factor of each particle, c 2 Is a social learning factor for each particle, ω is a weight factor; p is p id G, for the optimal position obtained by single search id Obtained for global searchAn optimal position to arrive; x is x id (t) is the position of the ith particle at the t-th time, v id (t) is the speed of the ith particle at the t-th time.
S4: bringing the optimal punishment parameter C and the nuclear parameter sigma obtained in the step S3 into a support vector regression model to obtain an optimal energy consumption prediction model, predicting in a test set, and using R 2 The performance of the model is checked by fitting coefficients and MSE indexes;
s5: and carrying out energy consumption real-time prediction according to the obtained energy consumption prediction model.
In one embodiment, the step S3, iteratively updates ω, as follows:
wherein omega max Is the maximum weight factor omega min As the minimum weight factor, N max The total iteration times are found when the PSO algorithm is optimized.
In one embodiment, the step 4 sets a threshold of total energy consumption in the campus, and if the predicted value of energy consumption exceeds the threshold or is 20% higher than the historical contemporaneous data, an early warning is sent; setting a park PPUE threshold according to the experience value and the park actual condition, and carrying out early warning if the actual PPUE exceeds the value.
In one embodiment, step 4 sends the early warning information to the associated mobile terminal to remind the professional manager to make relevant plans in time, and corresponding energy optimization scheduling measures are formulated to realize energy supply and demand balance in the park.
Compared with the prior art, the invention has the beneficial effects that:
the energy consumption and environmental data of the park are collected through the Internet of things equipment, a total energy consumption model is built, the data are subjected to preliminary processing by using an edge computing gateway, and the calculated amount of a back-end server is reduced.
And (3) comprehensively considering weather factors and time characteristics of environmental data, carrying out characteristic engineering, constructing an influence factor data analysis system, realizing influence factor characteristic screening, and finally selecting important influence factors with high correlation with energy consumption, thereby being beneficial to improving the prediction precision of a subsequent prediction model.
The improved PSO algorithm is utilized to optimize SVR model parameters, so that the model fitting degree is improved, the mean square error is reduced, the applicability of the model is enhanced, the training efficiency is improved, and the prediction precision is ensured.
Drawings
FIG. 1 is a flow chart of the overall method of the present invention.
FIG. 2 is a flowchart of an improved PSO-SVR energy consumption prediction algorithm of the present invention.
Detailed Description
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings and examples. The embodiments described below by referring to the drawings are illustrative and intended to explain the present invention and should not be construed as limiting the invention.
Referring to fig. 1, in some embodiments, the present invention provides a park energy efficiency management and control method based on the internet of things, where the function modules include a data acquisition module, a statistical analysis module, an energy consumption prediction module, and an energy consumption early warning module, and the overall method includes:
1. and (5) collecting equipment data. And (3) collecting various energy consumption data and environment data of the industrial park by using a data collecting module, and uploading the energy consumption data and the environment data to an edge computing gateway for abnormal value processing.
Specifically, in the invention, the park environment and the equipment are provided with corresponding detection sensors and detection meters, and the detection sensors and the detection meters are connected with the data acquisition module in a distributed mode through the nodes of the Internet of things. The detection sensor mainly comprises environment sensing sensors such as a temperature sensor, a humidity sensor and the like. The detection instrument mainly comprises an intelligent ammeter which is used for collecting electric parameters such as voltage, current, power, harmonic wave, power factor and the like of various power consumption devices. The present invention employs NB-IoT communication protocols for data transmission.
The initial data sequence received at the edge computing gateway may be expressed as:
{x 1 (1),…,x 1 (k),x 2 (1),…,x 2 (k),…,x i (1),…,x i (k),time}
wherein x is i (k) The kth data of the i factor is represented, the i factor refers to data collected by the same type of detection sensor or detection instrument, k is any positive integer except 0, the k is the number of the type of sensor or detection instrument, and time is the receiving time of the data, namely the generation time of the data.
In practical application, the ammeter or the sensor occasionally has the fault conditions of sensitivity abnormality, meter fault, communication interference, protocol conversion and the like, so that data acquisition errors are caused, the accuracy of subsequent model training is affected, and the result is poor. In this step, abnormal data is detected using a quarter bit distance (IQR) abnormal value recognition method, and the abnormal value is modified to a normal value by using a method of replacing the median of adjacent data.
2. And carrying out statistical analysis based on a statistical analysis module. Establishing a total energy consumption model, carrying out statistical analysis on the acquired data, and generating sample data; wherein the statistical analysis includes data preprocessing, data storage, correlation analysis, and data normalization.
The total energy consumption model established by the invention is as follows:
wherein P is total For the total energy consumption of the park, M is the number of air conditioning units and P ac,i The energy consumption of the ith air conditioning unit equipment is N, the number of industrial energy consumption equipment is P equip,j For the energy consumption of the jth industrial energy consumption device, Q is the number of other auxiliary devices including lighting and elevators which are connected to the power distribution system, P other,k Is the energy consumption of the kth auxiliary device.
Setting a park energy consumption index PPUE as an index for evaluating park energy efficiency, wherein the formula is as follows:
P equip representing industrial energy-consuming equipmentThe total energy consumption, PPUE, should be greater than 1, and closer to 1 indicates higher energy use efficiency.
Specifically, the edge computing gateway performs preliminary preprocessing of data, and the back-end server performs operations such as data storage, correlation analysis, data normalization and the like, and can perform trend analysis, flow statistics, monthly energy consumption analysis and energy consumption comparison ring ratio analysis on the energy consumption data of the park. The specific steps of the statistical analysis are as follows:
1) And preprocessing the initial data sequence at the edge computing gateway, and computing the data such as indoor average temperature, indoor average humidity, total energy consumption and the like.
2) And the back-end server is connected with the network to acquire the data of the traffic, the local weather, the wind speed, the air pressure, the precipitation and the like of the industrial park. The invention reflects the relativity between different types of data (namely between each variable and the total energy consumption) by adopting the pearson relativity coefficient r, and the formula is shown as follows when the pearson relativity coefficient is used for a sample:
wherein: n is the number of samples, X i ,Y i Is the i sample value corresponding to the variable X, Y,is the average number of X samples and the average number of Y samples.
Specifically, in this step, 7 variables including the calculated indoor average temperature, indoor average humidity, outdoor air temperature of the campus, wind speed, air pressure, precipitation amount, and campus people flow are used as variables X i With total energy consumption value as variable Y i Their respective pearson correlation coefficients are calculated. Finally, 4 characteristics of the indoor average temperature, the indoor average humidity, the outdoor average air temperature and the people flow with highest correlation are selected as partial input characteristics of sample data, and other variables with lower correlation are removed.
3) Performing feature construction and complementation, wherein the construction time refers to the feature: adding weekends, holidays and seasonal features. Specifically, in this step, according to the time value in the initial data sequence of the data acquisition module, a time feature is constructed, for example, by adding a weekend feature through 0,1, the feature of the weekend is 1, and the feature of the non-weekend is 0; adding holiday feature 1 which meets the holiday date feature, otherwise, assigning 0; the seasonal features are assigned with values of 0,1, 2 and 3 in sequence of spring, summer, autumn and winter.
4) In the original data, the numerical value difference among variables is larger, and the prediction performance of a model is affected, so that normalization processing is necessary for the data, the data is prevented from falling into a saturated region, the original characteristics of the data are also maintained, and in the invention, the maximum and minimum value normalization is carried out by adopting the following formula, and the variable range is unified.
Wherein x is min Is the minimum value of the sample data; x is x max Is the maximum value of the sample data; x is x * Is normalized value.
Final formation of final sample data sequences y and x i As a training input to the energy consumption prediction model, it is expressed as:
wherein y (k) represents the total energy consumption value of the kth data, x i (k) Kth data representing factor i.
Therefore, the invention takes the historical energy consumption data as a training sample, and obtains sample data after the steps are processed.
3. And the energy consumption prediction module predicts the energy consumption. And (3) using an SVR support vector regression method, optimizing parameters in the SVR by adopting an improved PSO algorithm, establishing a support vector regression-improved particle swarm optimization (PSO-SVR) energy consumption prediction model, and carrying out model training by using the sample data to carry out energy consumption prediction.
The SVR support vector regression method is to perform linear regression in the Gao Weina kernel-induced feature space by mapping training samples to the space. The initial SVR model is realized by the following formula:
wherein the output is a fitting function f (x), x i ,x j A spatial feature vector is input for the sample low-dimensional data,α i is Lagrangian multiplier, b is displacement,>is a Radial Basis Function (RBF) with a wide convergence domain, which acts as a kernel function, σ being a kernel parameter.
Radial Basis Function (RBF) is used as a kernel function to map samples to a high-dimensional space, so that nonlinear problems can be converted into linear solutions conveniently, and dimension disasters can be avoided. The PSO-SVR method is suitable for learning small sample data, is not easy to be subjected to over fitting and has strong generalization capability.
When a Support Vector Regression (SVR) model is established, a penalty parameter C and a kernel parameter sigma have very important influence on the prediction performance of the model; the penalty parameter C is used for weighing the weight of the loss, and the kernel parameter sigma influences the radial action range of the kernel function to determine the range and the distribution characteristic of the training sample data. The present invention employs an improved PSO algorithm to find the optimal combination of parameters C, σ.
Specifically, in this step, the process of optimizing the SVR model parameters using the modified Particle Swarm Optimization (PSO) algorithm is as follows, and the algorithm steps are shown in fig. 2:
1) And constructing an SVR model and setting parameter information of a particle swarm algorithm. For example, the number of iterations is set to 20, the population size to 20, and the learning factor to 2.
2) Initializing punishment parameters (C) and kernel parameters (sigma) of the SVR model, initializing positions and speeds of particles in a particle swarm algorithm, and then calculating the fitness (R2) of the SVR model of the training set as fitness;
3) Continuously iterating, updating the position and the speed of the particles, calculating the corresponding fitness, and recording C and sigma corresponding to the global optimal fitness; let each particle be composed of 2-dimensional parameter vectors (C, sigma), the position of the ith particle in the 2-dimensional solution space be x i =(x i1 ,x i2 ) Speed v i =(v i1 ,v i2 ) Recording the individual extremum of the current moment as p i =(p i1 ,p i2 ) Global extremum is noted g i =(g i1 ,g i2 ). In each iteration, the particle tracks the states of the individual extremum, the global extremum and the previous moment to adjust the position and the speed at the current moment, and the speed and the position formula are as follows:
v id (t+1)=ωv id (t)+c 1 rand()(p id -x id (t))+c 2 rand()(g id -x id (t))
x id (t+1)=x id (t)+v id (t+1)
wherein rand () is [0,1 ]]The random number in between, d is the dimension of the solution space, c 1 Is the individual learning factor of each particle, c 2 Is a social learning factor for each particle, ω is a weight factor; p is p id G, for the optimal position obtained by single search id The optimal position obtained by global searching is obtained; x is x id (t) is the position of the ith particle at the t-th time, v id (t) is the speed of the ith particle at the t-th time.
Meanwhile, in order to avoid the phenomenon that a basic PSO algorithm is easy to oscillate near a global optimal solution, the weight omega is iteratively updated according to the following formula:
wherein omega max Is the maximum weight factor omega min As the minimum weight factor, N max The total iteration times are found when the PSO algorithm is optimized.
4) Substituting the optimal punishment parameter C and the nuclear parameter sigma obtained in the step 3) into an SVR model to obtain an optimal SVR prediction model, predicting in a test set, and using R 2 The fitting coefficients and the mean square error MSE indicator verify the model performance.
5) And carrying out energy consumption real-time prediction according to the obtained energy consumption prediction model, and providing data support for subsequent energy consumption early warning.
4. And the energy consumption early warning module carries out energy consumption early warning. After the energy consumption prediction result is obtained, the energy consumption early warning is carried out by combining the electricity consumption of the current period, so that the energy efficiency level is improved. For example, one month may be one period.
Specifically, setting a threshold value of total energy consumption of the park, and if the predicted value of the energy consumption exceeds the threshold value or is 20% higher than the historical contemporaneous data, sending out early warning; setting a park PPUE threshold according to the experience value and the park actual condition, and carrying out early warning if the actual PPUE exceeds the value. Therefore, analysis and early warning based on the campus history contemporaneous energy consumption data are realized. Meanwhile, the early warning information can be sent to the associated mobile terminal so as to remind professional managers of timely making relevant plans and making corresponding energy optimization scheduling measures, and energy supply and demand balance of the park is achieved.
Further, in one embodiment, energy efficiency management and control can be performed by comprehensively utilizing strategies such as energy consumption tracing, peak clipping and valley filling. The energy consumption tracing counts the energy consumption data of each unit and each department of the park, and the energy consumption data of each department can be shown by a histogram, a graph or a pie chart according to the time dimension such as year, month and day; the peak load shedding strategy mainly utilizes the electricity price difference of peak and valley periods, counts the interruptible load, pauses the interruptible load in the peak period, and transfers to the valley period for reducing the cost, and continuously improves part of main energy consumption departments and production lines to encourage the unnecessary load to be reduced in the peak period.
In conclusion, the intelligent energy consumption prediction method based on the intelligent park realizes the application of the Internet of things in energy efficiency management and control of the intelligent park, combines an edge computing gateway to perform data preprocessing, establishes a multi-variable energy consumption prediction model on the basis of considering time and season characteristics, is beneficial to monitoring and optimizing energy consumption of the park, and improves energy utilization efficiency.
The present invention is not limited to the above embodiments, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the present invention, and these modifications and substitutions are intended to be included in the scope of the present invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.
Claims (10)
1. The industrial park energy efficiency management and control method based on the Internet of things is characterized by comprising the following steps of:
step 1, collecting energy consumption data and environment data of an industrial park, uploading the energy consumption data and the environment data to an edge computing gateway and processing abnormal values;
step 2, a total energy consumption model is established, statistical analysis is carried out on the collected data, and sample data are generated; the statistical analysis comprises data preprocessing, data storage, correlation analysis and data normalization;
step 3, establishing a support vector regression-improved particle swarm optimization (PSO-SVR) energy consumption prediction model, and carrying out model training by using the sample data to carry out energy consumption prediction;
and 4, carrying out energy consumption early warning by combining the current period electricity consumption according to the energy consumption predicted value.
2. The industrial park energy efficiency control method based on the Internet of things according to claim 1, wherein in the step 1, a detection sensor and a detection instrument for data acquisition are connected with a data acquisition module in a distributed mode through nodes of the Internet of things; the detection sensor comprises a temperature sensor and a humidity sensor, and the detection instrument comprises a smart meter; the temperature sensor and the humidity sensor adopt an NB-IoT communication protocol for data transmission; the intelligent ammeter collects electric parameters of power consumption equipment, including voltage, current, power, harmonic wave and power factor.
3. The industrial park energy efficiency management and control method based on the internet of things according to claim 1, wherein the initial data sequence received at the edge computing gateway is expressed as:
{x 1 (1),…,x 1 (k),x 2 (1),…,x 2 (k),…,x i (1),…,x i (k),time}
wherein x is i (k) The kth data of the i factor is represented, the i factor refers to data collected by the same type of detection sensor or detection instrument, k is any positive integer except 0, the k is the number of the type of detection sensor or detection instrument, time represents the receiving time of the data, and the abnormal value processing is carried out on the received initial data sequence: abnormal data is detected by using a quarter bit distance (IQR) abnormal value identification method, and the abnormal value is modified to be a normal value by adopting a method of replacing the median of adjacent data.
4. The industrial park energy efficiency management and control method based on the internet of things according to claim 1, wherein the step 2, the total energy consumption model is expressed as follows:
wherein P is total For the total energy consumption of the park, M is the number of air conditioning units and P ac,i The energy consumption of the ith air conditioning unit equipment is N, the number of industrial energy consumption equipment is P equip,j For the energy consumption of the j-th industrial energy consumption equipment, Q is the number of auxiliary equipment including illumination and elevators, P other,k Energy consumption for the kth auxiliary device;
setting a park energy consumption index PPUE as an index for evaluating park energy efficiency, wherein:
P equip indicating the total energy consumption of the industrial energy consumption equipment, the PPUE value should be greater than 1, and the closer to 1, the higher the energy use efficiency.
5. The industrial park energy efficiency management and control method based on the internet of things according to claim 4, wherein the statistical analysis process is as follows:
s1: preprocessing data at an edge computing gateway, and computing indoor average temperature and humidity and total energy consumption data;
s2: acquiring data information including the flow of people, the weather of the place, the wind speed, the air pressure and the precipitation of an industrial park at a back-end server; the pearson correlation coefficient r is adopted to reflect the correlation among different types of data, and a plurality of characteristics with the highest correlation are selected as partial input characteristics of sample data, namely outdoor average air temperature, indoor average humidity, indoor average temperature and people flow;
s3: performing feature construction and complementation, wherein the construction time refers to the feature: adding weekends, holidays and seasonal features;
s4: normalization of the data is necessary, and the variable range is unified, and the formula is as follows:
wherein x is min Is the minimum value of the sample data; x is x max Is the maximum value of the sample data; x is x * Is normalized value;
final formation of final sample data sequences y and x i As a training input to the energy consumption prediction model, it is expressed as:
wherein y (k) represents the total energy consumption value of the kth data, x i (k) Kth data representing factor i.
6. The internet of things-based industrial park energy efficiency management method according to claim 4, wherein the statistical analysis further comprises trend analysis, flow direction statistics, monthly energy consumption analysis, and energy consumption duty cycle analysis.
7. The method for controlling energy efficiency of industrial park based on internet of things according to claim 4, wherein the step 3 uses a support vector regression method to perform linear regression in a Gao Weina kernel-induced feature space by mapping training samples to the space, and is implemented by the following formula:
wherein the output is a fitting function f (x), x i ,x j A spatial feature vector is input for the sample low-dimensional data,α i is Lagrangian multiplier, b is displacement,>is a radial basis with a wide convergence domain, which acts as a kernel function, σ being a kernel parameter;
the penalty parameter C and the kernel parameter sigma in the support vector regression are optimized by adopting an improved Particle Swarm Optimization (PSO) algorithm, and the process is as follows:
s1: constructing a support vector regression model and setting parameter information of a particle swarm algorithm;
s2: initializing a punishment parameter C and a nuclear parameter sigma of a support vector regression model, initializing the position and the speed of particles in a particle swarm algorithm, and then calculating the fitting goodness R2 of the support vector regression model of a training set as fitness;
s3: continuously iterating, updating the position and the speed of the particles, calculating the corresponding fitness, and recording a penalty parameter C and a core parameter sigma corresponding to the global optimal fitness; let each particle be composed of 2-dimensional parameter vectors (C, sigma), the position of the ith particle in the 2-dimensional solution space be x i =(x i1 ,x i2 ) Speed v i =(v i1 ,v i2 ) Recording the individual extremum of the current moment as p i =(p i1 ,p i2 ) Global extremum is noted g i =(g i1 ,g i2 ) The method comprises the steps of carrying out a first treatment on the surface of the In each iteration, the particle tracks the states of the individual extremum, the global extremum and the previous moment to adjust the position and the speed at the current moment, and the speed and the position formula are as follows:
v id (t+1)=ωv id (t)+c 1 rand()(p id -x id (t))+c 2 rand()(g id -x id (t))
x id (t+1)=x id (t)+v id (t+1)
wherein rand () is [0,1 ]]The random number in between, d is the dimension of the solution space, c 1 Is the individual learning factor of each particle, c 2 Is a social learning factor for each particle, ω is a weight factor; p is p id G, for the optimal position obtained by single search id The optimal position obtained by global searching is obtained; x is x id (t) is the position of the ith particle at the t-th time, v id (t) is the speed of the ith particle at the t-th time.
S4: bringing the optimal punishment parameter C and the nuclear parameter sigma obtained in the step S3 into a support vector regression model to obtain an optimal energy consumption prediction model, predicting in a test set, and using R 2 The performance of the model is checked by fitting coefficients and MSE indexes;
s5: and carrying out energy consumption real-time prediction according to the obtained energy consumption prediction model.
8. The industrial park energy efficiency management and control method based on the internet of things according to claim 4, wherein the step S3 of iteratively updating ω comprises the following formula:
wherein omega max Is the maximum weight factor omega min Is the least weight factorSon, N max The total iteration times are found when the PSO algorithm is optimized.
9. The method for controlling energy efficiency of industrial park based on the internet of things according to claim 1, wherein in the step 4, a threshold value of total energy consumption of the park is set, and if the predicted value of energy consumption exceeds the threshold value or is 20% higher than the historical contemporaneous data, an early warning is sent; setting a park PPUE threshold according to the experience value and the park actual condition, and carrying out early warning if the actual PPUE exceeds the value.
10. The industrial park energy efficiency management and control method based on the internet of things according to claim 1 or 9, wherein in the step 4, early warning information is sent to the associated mobile terminal to remind professional managers to make relevant plans in time, and corresponding energy optimization scheduling measures are formulated to achieve park energy supply and demand balance.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310051194.7A CN116307076A (en) | 2023-02-02 | 2023-02-02 | Industrial park energy efficiency management and control method based on Internet of things |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310051194.7A CN116307076A (en) | 2023-02-02 | 2023-02-02 | Industrial park energy efficiency management and control method based on Internet of things |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116307076A true CN116307076A (en) | 2023-06-23 |
Family
ID=86834922
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310051194.7A Pending CN116307076A (en) | 2023-02-02 | 2023-02-02 | Industrial park energy efficiency management and control method based on Internet of things |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116307076A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116596279A (en) * | 2023-07-14 | 2023-08-15 | 中通服建设有限公司 | Intelligent park energy consumption scheduling system |
CN116843361A (en) * | 2023-06-30 | 2023-10-03 | 国网信通亿力科技有限责任公司 | Power grid audit informatization management method and system |
CN117896417A (en) * | 2024-01-17 | 2024-04-16 | 西安交通大学城市学院 | Embedded industry thing networking controller |
CN118041159A (en) * | 2024-04-11 | 2024-05-14 | 深圳市创马优精密电子有限公司 | Intelligent feedback-based motor drive control board energy consumption optimization method and system |
-
2023
- 2023-02-02 CN CN202310051194.7A patent/CN116307076A/en active Pending
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116843361A (en) * | 2023-06-30 | 2023-10-03 | 国网信通亿力科技有限责任公司 | Power grid audit informatization management method and system |
CN116843361B (en) * | 2023-06-30 | 2024-02-06 | 国网信通亿力科技有限责任公司 | Power grid audit informatization management method and system |
CN116596279A (en) * | 2023-07-14 | 2023-08-15 | 中通服建设有限公司 | Intelligent park energy consumption scheduling system |
CN116596279B (en) * | 2023-07-14 | 2024-04-26 | 中通服建设有限公司 | Intelligent park energy consumption scheduling system |
CN117896417A (en) * | 2024-01-17 | 2024-04-16 | 西安交通大学城市学院 | Embedded industry thing networking controller |
CN118041159A (en) * | 2024-04-11 | 2024-05-14 | 深圳市创马优精密电子有限公司 | Intelligent feedback-based motor drive control board energy consumption optimization method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Li et al. | When weather matters: IoT-based electrical load forecasting for smart grid | |
Amasyali et al. | A review of data-driven building energy consumption prediction studies | |
CN116307076A (en) | Industrial park energy efficiency management and control method based on Internet of things | |
CN108388962B (en) | Wind power prediction system and method | |
Moradzadeh et al. | Heating and cooling loads forecasting for residential buildings based on hybrid machine learning applications: A comprehensive review and comparative analysis | |
CN115358347B (en) | Method for predicting remaining life of intelligent electric meter under different subsystems | |
CN114372360A (en) | Method, terminal and storage medium for power load prediction | |
CN112418495A (en) | Building energy consumption prediction method based on longicorn stigma optimization algorithm and neural network | |
CN117669960A (en) | New energy power prediction method based on multivariable meteorological factors | |
Chen et al. | MultiCycleNet: multiple cycles self-boosted neural network for short-term electric household load forecasting | |
CN115481918A (en) | Active sensing and predictive analysis system for unit state based on source network load storage | |
Tao et al. | Reserve evaluation and energy management of micro-grids in joint electricity markets based on non-intrusive load monitoring | |
CN108346009A (en) | A kind of power generation configuration method and device based on user model self study | |
CN109858668B (en) | Coordination prediction method for power load region in thunder and lightning climate | |
CN117621898B (en) | Intelligent parking lot charging pile charging control method and system considering power grid electricity price | |
Kumar et al. | Forecasting indoor temperature for smart buildings with ARIMA, SARIMAX, and LSTM: A fusion approach | |
Ibrahim et al. | LSTM neural network model for ultra-short-term distribution zone substation peak demand prediction | |
CN113111419A (en) | Method and system for establishing and predicting air-conditioning load prediction model in office building | |
CN115907228A (en) | Short-term power load prediction analysis method based on PSO-LSSVM | |
Nichiforov et al. | Efficient Load Forecasting Model Assessment for Embedded Building Energy Management Systems | |
Mughal et al. | Techno-economic assessment of photovoltaics by predicting daily global solar radiations using hybrid ANN-PSO model | |
CN113449909A (en) | Cold load prediction method of cold accumulation air conditioner based on improved long-term and short-term memory network | |
Cho et al. | Application of Parallel ANN-PSO to Hourly Solar PV Estimation | |
Jain et al. | Peak load forecasting using machine learning algorithms | |
CN108876078B (en) | Method for calculating energy consumption system performance improvement strategy and energy consumption system monitoring device |
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 |