CN111898856B - Analysis method of physical-data fusion building based on extreme learning machine - Google Patents

Analysis method of physical-data fusion building based on extreme learning machine Download PDF

Info

Publication number
CN111898856B
CN111898856B CN202010571025.2A CN202010571025A CN111898856B CN 111898856 B CN111898856 B CN 111898856B CN 202010571025 A CN202010571025 A CN 202010571025A CN 111898856 B CN111898856 B CN 111898856B
Authority
CN
China
Prior art keywords
data
building
physical
model
temperature
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
Application number
CN202010571025.2A
Other languages
Chinese (zh)
Other versions
CN111898856A (en
Inventor
崔嘉
胡罗乐
杨俊友
孙峰
周小明
陈得丰
杨智斌
佟昊松
苑经纬
李桐
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenyang University of Technology
State Grid Liaoning Electric Power Co Ltd
Electric Power Research Institute of State Grid Liaoning Electric Power Co Ltd
Original Assignee
Shenyang University of Technology
State Grid Liaoning Electric Power Co Ltd
Electric Power Research Institute of State Grid Liaoning Electric Power Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenyang University of Technology, State Grid Liaoning Electric Power Co Ltd, Electric Power Research Institute of State Grid Liaoning Electric Power Co Ltd filed Critical Shenyang University of Technology
Publication of CN111898856A publication Critical patent/CN111898856A/en
Application granted granted Critical
Publication of CN111898856B publication Critical patent/CN111898856B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/067—Enterprise or organisation modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/211—Selection of the most significant subset of features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • 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
    • G06N3/045—Combinations of 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
    • 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
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Operations Research (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Evolutionary Biology (AREA)
  • Game Theory and Decision Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Quality & Reliability (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • Primary Health Care (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention relates to the field of power systems, in particular to an analysis method for building modeling based on physical-data fusion of an extreme learning machine. The method comprises the steps of data acquisition and pretreatment: building a building physical model based on an overall measurement and discrimination method by collecting and preprocessing building data and electrical data; training the building physical model, the collected and preprocessed user data, environment data and actual power consumption data by using an extreme learning machine to obtain a physical-data fusion model; the static parameters of the electricity consumption behavior to be analyzed are input into the physical-data fusion model through the physical model and the dynamic parameters, and an analysis result is obtained. The system comprises a data acquisition and preprocessing module and a building physical module; and a physical-data fusion module. The invention provides an analysis method for building modeling based on physical-data fusion of an extreme learning machine, which aims to solve the problems of low building model precision, slower response on a demand side and low load prediction precision at present.

Description

Analysis method of physical-data fusion building based on extreme learning machine
Technical Field
The invention relates to the field of power systems, in particular to an analysis method for building modeling based on physical-data fusion of an extreme learning machine.
Background
With the strong promotion of ubiquitous power internet of things construction, the arrangement of mass sensing terminals enables the analysis and control of the power consumption of digital houses to become a hotspot existing house energy efficiency analysis for research and application, and the house energy efficiency analysis is usually based on a power distribution network area or an aggregation model of a single building. The dynamic change of the external environment, uneven heat distribution in the building and the change state of the internal communication structure can cause parameter deviation of the building electricity model even when the activity of personnel, so that the accurate analysis of the house electricity consumption is affected.
Therefore, the construction of an accurate house electricity consumption model is important to achieving the aims of accurate analysis of electricity consumption, further energy conservation management and control and the like. At present, three main building energy consumption models are established, namely, a model is established by combining building energy consumption analysis software through modeling methods such as finite element and space vector method based on physical simulation modeling; secondly, based on a statistical or data driving model, an energy consumption model is established through a large amount of measured data and an intelligent algorithm, and thirdly, a modeling mode based on a physical-data fusion idea is adopted. In the three methods, the simulation method based on the physical model can effectively reflect the causal relation of building electric-thermal conversion, but due to inaccurate measurement data, the model precision is often poor due to high complexity of the physical model. The second and third methods can train and calibrate the electricity consumption model of the building by utilizing the historical data and weather data of the building, the calculation time is short and can be used for real-time operation decision, but the second method is very sensitive to sample size, and the demand of the training data set is far greater than that of the third method.
The existing analysis method based on the building electric energy consumption model can not completely meet the requirement of multi-element data, and an analysis method with high accuracy of an space model, faster response of a requirement side and high load prediction accuracy is urgently needed.
Disclosure of Invention
The invention aims to:
the invention aims to overcome the defects of the analysis method based on the traditional building modeling, and provides an analysis method of physical-data fusion building modeling based on an extreme learning machine. The method aims to solve the problems that an existing building model is low in precision, the response of a demand side is slow, and the load prediction precision is low.
The technical scheme is as follows:
the analysis method of the physical-data fusion building modeling based on the extreme learning machine comprises the following steps:
step 1, data acquisition and pretreatment:
step 2, building a building physical model based on an overall measurement and dialect method by collecting and preprocessing building data and electrical data;
training the building physical model, the collected and preprocessed user data, the collected and preprocessed environment data and the collected and preprocessed actual power consumption data by using an extreme learning machine to obtain a physical-data fusion model;
the static parameters of the electricity consumption behavior to be analyzed are input into the physical-data fusion model through the physical model and the dynamic parameters, and an analysis result is obtained.
The analysis system based on the physical-data fusion building modeling of the extreme learning machine comprises a data acquisition and preprocessing module and a building physical module; a physical-data fusion module; the acquisition and preprocessing module is used for acquiring and preprocessing data; building physical module, which is used to build building physical model based on the overall measurement and dialect method by collecting and preprocessing the building data and electric data; the physical-data fusion module is used for training the building physical model, the collected and preprocessed user data, the environment data and the actual measured data by using an extreme learning machine to obtain a physical-data fusion model; the static parameters of the electricity consumption behavior to be analyzed are input into the physical-data fusion model through the physical model and the dynamic parameters, and an analysis result is obtained.
The advantages and effects:
the invention has the following advantages and beneficial effects:
1. the building model is modeled and analyzed by adopting a physical-data fusion method, the traditional building model is modeled only from a single physical model, environmental factors and user behaviors are not considered, the building model is fused with the data model by adopting a machine learning method, the building physical model is fused with the data model by adopting the machine learning method, general main components are replaced by high-entropy causal data, the reliability of output results of different physical simplified model methods is rapidly calculated by adopting the machine learning method, and the method reduces the complexity of the built model on the basis of combining the physical model with the data model and improves the precision of the building model.
2. The building physical-data model and the actual load are trained by the extreme learning machine, and the method can mine knowledge from the history and experience data of the building load, so that direct or auxiliary decision making is provided for the analysis, control, planning and other services of the building load. Compared with other methods, the method has the following characteristics: (1) hidden layer nodes/neurons do not need iterative adjustment in the learning process; (2) not only belongs to a general single hidden layer feedforward network, but also belongs to a multi hidden layer feedforward network; (3) the same architecture can be used as a plurality of problems of feature learning, clustering, regression and classification; (4) the weight parameter solving mode can ensure the global optimum of the solving result. And (3) repeatedly training the ELM network by changing the activation function of the hidden layer node of the ELM algorithm of the extreme learning machine, and calculating the training accuracy under different activation functions. The method is used for fitting the built physical-data model with the actual load to the greatest extent through a neural algorithm, so that the accuracy of the physical-data model is greatly improved, and the load prediction accuracy of the building model is greatly improved.
In summary, the method for fusing the extreme learning machine and the physical-data is used for building modeling analysis for the first time, and can fully account for the thermal-electric conversion characteristics of a building and the dynamic changes of various environmental parameters and equipment electricity consumption behaviors, so that more accurate building electricity model construction, refined electricity consumption analysis and more accurate load prediction capability are realized.
Drawings
FIG. 1 is a diagram of a physical-data fusion algorithm framework;
FIG. 2 is a frame diagram of analysis of electrical behavior;
FIG. 3 is a schematic diagram of an extreme learning machine.
Detailed Description
The following description of the embodiments of the invention is presented in conjunction with the accompanying drawings to provide a better understanding of the invention to those skilled in the art.
The invention provides a physical-data fusion building modeling method based on an extreme learning machine, aiming at the problems of complex modeling structure, poor model precision and low load prediction precision caused by the problems of large sample demand, environmental data, user behavior and other dynamic factors in the modeling process in the traditional building modeling analysis process. The method is suitable for building models, improves the accuracy of the models on the basis of reducing the complexity of the models, and greatly increases the prediction accuracy of the load.
Based on the prior art, the invention provides a digital house electricity model construction method based on physical and data fusion modeling of an extreme learning machine by combining multiple data such as building, electricity, environment and user behavior, and can realize fine load electricity efficiency analysis. Compared with the traditional neural network model, the neural network model has the advantages of high learning speed, small training error and strong generalization capability. By combining data and a physical modeling method, the precision of a load model can be greatly improved, the response to the load on the demand side is improved, and the load prediction precision is greatly improved.
As shown in fig. 1 and 2, the analysis method of physical-data fusion building modeling based on the extreme learning machine is characterized in that: the method comprises the following steps:
step 1, data acquisition and pretreatment:
step 2, building a building physical model based on an overall measurement and dialect method by collecting and preprocessing building data and electrical data;
training the building physical model, the collected and preprocessed user data, the collected and preprocessed environment data and the collected and preprocessed actual power consumption data by using an extreme learning machine to obtain a physical-data fusion model;
the static parameters of the electricity consumption behavior to be analyzed are input into the physical-data fusion model through the physical model and the dynamic parameters, and an analysis result is obtained.
Building a building physical model by a general survey and debate method specifically comprises:
fusing a physical model and a data model through a machine learning method, describing the state of a building by using a physical simplified model reflecting a physical association relation, generating data with high entropy characteristics, and taking the data as the physical model and the data model to be input;
the expressions of the physical model and the data model are:
where k represents a time status tag, k+1 represents a future time, X k+1 A vector composed of system state characteristics predicted at the time of k+1; x is x k+1 ' is a state characteristic vector to be predicted which is preprocessed by a physical model at the moment k+1; f and h respectively reflect the mapping relation between the measured data and the data characteristics to be predicted in the physical model and the data model; u is a random error vector in the calculation of the data model; x is X k And Y k The vector of measurement data for the power system is different from the measurement data vector X k Processed by a physical model to measure a data vector Y k Processed by a data model.
The physical model in the physical-data fusion model provides high-entropy input characteristics for the data model, so that the data model can express the characteristics of the problem to be solved more accurately, and further, an accurate data model is built.
The building data model and the physical model are fused through a machine learning method, the conventional main components are replaced by causal data with high entropy, the reliability of output results of different physical simplified model methods is rapidly calculated through the machine learning method, the building physical-data fusion model is obtained, and then the building physical data-data model is started for calculation.
The physical simplified model comprises the following specific steps:
firstly, building physical models are built; the building physical model comprises a building physical model and a heat pump physical model
In the heating model of the intelligent building, the heat obtaining mode of the building from the outside mainly comprises the following three modes of heat conduction, heat convection and heat radiation. The heat storage capability and the heat conduction capability are conceptual bases for building an intelligent building RC network model. Walls, ceilings, floors and indoor air can be provided as heat storage elements in the model. They store heat as a function of their own mass and their specific heat capacity. Meanwhile, heat is not only used for storage, but also the intelligent building can exchange heat with the outside through the heat storage element in the building. The RC network model which is most commonly used at the present stage maintains the dynamic heating characteristic of the building while ensuring certain accuracy, so that the RC network model is widely applied.
TABLE 1 Unit analogy
As shown in table 1, by analogy to newton's law of cooling and ohm's law, heat is compared to the charge in the circuit, temperature difference is compared to voltage, heat transfer rate is compared to current, and thermal resistance is compared to resistance; meanwhile, in order to further study the transient thermal behavior of the intelligent building model, the heat capacity, namely the heat storage capacity of an object, is introduced; in the RC network model, wall, indoor and outdoor air are used as nodes in the circuit model.
The intelligent building model is as follows: assuming that the heating system is located in a technical room tr of the building; assuming that all parts of the heating system are linear with respect to heat loss from the technical chamber; these heat losses are expressed as heat pump losses hp, thermal resistances of the space heating storage tank sh and the domestic hot water storage tank dw; a ground gnd, which is assumed to be thermostatted, is discharged by transport and ventilation into the surrounding air amb; since it is assumed that the floor and all walls are constructed using heavy materials, the thermal resistance of the floor and the exterior wall is divided into several parts R on either side of the lumped heat capacity w1 、R w2 And R is f1 、R f2 ，R w1 、R w2 Is the equivalent thermal resistance of the outer wall, R f1 、R f2 Is the equivalent thermal resistance of the floor; modeling ventilation heat loss as thermal resistance R ve Solar energy and internal revenue are modeled as heat flow to all heat capacities;
The building model can be written as:
wherein C is in 、C wi 、C we 、C f1 、C tr Respectively representing indoor air, inner wall, outer wall, floor, technical room temperature equivalent capacitance, t represents time variable, Q in Indicating the heating capacity of the room air, Q wi Represents the heating capacity of the inner wall, Q we Represents the heating capacity of the outer wall, Q f1 Indicating the heating capacity of the floor, Q hp,loss Indicating heat loss from the heat pump;
T in 、T amb 、T wi 、T f1 、T sh 、T tr respectively representing indoor air, inflow indoor air, inner wall, floor, hot water pump, technical room temperature;
in the optimal control problem, heat loss of the heat pump to the machine room is ignored; the equation is discretized as:
wherein T is i And T i+1 The temperatures at the time i and the time i+1 are shown, respectively, and Δt represents the heating time
The heat flow to the different components is calculated as solar energy and internal gain:
the operating temperature of the building is linearized:
wherein Q represents the heating capacity of the heat pump, P represents the required power input, lt, mt and ht represent low temperature, medium temperature and high temperature respectively;
the heating system of the heat pump heating system is modeled as follows:
the heat pump provides heat for space heating or domestic hot water production:
indicating the heating capacity of the heat pump at low temperature at instant i,/and a method for controlling the heating capacity of the heat pump at low temperature>Indicating the heating capacity of the heat pump at low temperature at instant i, < > >The heating capacity of the water storage tank at low temperature at the moment i is shown, and the same is true; />Indicating the heating capacity of the heat pump at medium temperature at instant i,/and a method for controlling the heating capacity of the heat pump at medium temperature>Indicating the heating capacity of the heater at mid-temperature of the heat pump at time i,/>Indicating the heating capacity of the water storage tank at medium temperature at time i, < >>Indicating the heating capacity of the heat pump at high temperature at time i,indicating the heating capacity of the heater at high temperature of the heat pump at instant i,/and a method for controlling the heating capacity of the heater at high temperature of the heat pump at instant i>The heating capacity of the water storage tank at the high temperature at the moment i is shown;
the heat pump efficiency varies with ambient temperature, water temperature and compressor adjustment for the following reasons:
p hp representing the total power of the heat pump, p hp,ht Represents high temperature heating power, p hp,lt Indicating low temperature heating power, p hp,mt Indicating medium temperature heating power ζ l Indicating modulation degree ζ q Is a constraint coefficient;
the auxiliary heater efficiency is assumed to be 1:
auxiliary heater power for instant i, +.>Auxiliary addition for instant iThe heating capacity of the heater;
the spatial heat storage state is calculated by the following formula:
represents the heat quantity discharged at the moment i, T em,ret Indicating the compensated loss temperature;
there are other limitations:
ρ represents the density of the water,losing the volume of water in the heat; />Representing the volume of maximum supply water to the heat pump; epsilon no m Representing heat pump heating efficiency;
calculating the state of a domestic hot water storage tank:
C dw,lt Representing the equivalent capacitance of the water storage tank at low temperature, C dw,mt Representing the equivalent capacitance of the water storage tank at medium temperature, C dw,ht Representing the equivalent capacitance of the water storage tank at high temperature, C dw,vht Represents the equivalent capacitance of the water storage tank at the ultra-high temperature,indicating the heating capacity at ultra-high temperature at time i,/-, for example>The heating capacity of the auxiliary heater at the ultrahigh temperature at the moment i is shown;
other limitations:
indicates the temperature of the water storage tank at ultra-high temperature, < >>Indicates the temperature of the water storage tank at high temperature, < >>Indicates the temperature of the water storage tank at low temperature, < >>The temperature of the water storage tank is shown at medium temperature; t (T) mains Indicating the temperature of the thermostatic mixing valve after mixing the stored water with cold water from the water line; t (T) dw,dem Representing the required temperature of the water storage tank; c represents the specific heat capacity of water;
the multi-standard target is written as:
W EIE weight factor representing the influence of energy on environment, j is applicable to the power generation index of all power plants, p impact,j For the marginal effect of the power generation of the power plant,representing power generated by a power plant over a period of time at a power generation index;
defining different control targets for the control strategy of the heat pump building physical model;
from the perspective of an individual consumer, it is desirable for the consumer to minimize his own costs for providing thermal comfort and domestic hot water; this objective is achieved by minimizing the individual energy costs within a control range if thermal comfort is maintained; the cost of discomfort is added to the objective function from as a soft constraint:
By usingIndicating the electricity price at time steps i and +.>And->The power of the heat pump power supply and the auxiliary heater are respectively; quantifying the discomfort cost by multiplying the building operating temperature deviation outside the building comfort temperature boundary by the building discomfort cost p;
secondly, considering the viewpoint of environmental impact of energy use, minimizing the impact of the used energy on the environment; since the impact is different for each foundry, optimization attempts are made to run the most environmentally friendly factories as much as possible; the objective function is then written as:
the influence of power generation on different power plants is selected as the single-stage influence of roof photovoltaic power generation, offshore wind power generation, nuclear power generation, combined cycle natural gas power generation and open cycle gas turbine power generation; finally, considering the influence of power generation on the environment, not only the use of energy is important; for many power plants, a large amount of emissions is accompanied by the construction and decommissioning of the power plant;
representing a heat pump overload threshold,/->Indicating overload power +.>Representing the remaining load power;
since this is an approximation, the installed capacity cost is included in the objective function, here designated capacity price is pimpact; cap has little physical meaning;
If an overload condition occurs, the influence of the overload condition is far greater than the influence of power generation at the power just below an overload threshold; the three views of consumer cost, energy influence and productivity influence are integrated into a single multi-objective function; consumers of energy-dependent scale environmental impact power generation costs incorporating energy impact into the consumer cost function environmental impact weight factor energy introduction scale; when the factor is small, the controller ignores the influence of power generation; the factors are large, and the capacity influence point of view can be simply added to the cost function; to change the behavior of the controller, the overload threshold is also changed to minimize
l cap Representing the capacity limiting factor of the device,indicating the maximum residual load power.
Training by an extreme learning machine according to the building physical model, the collected and preprocessed user data, environment data and actual measured data, wherein the method specifically comprises the following steps:
using the optimal feature set as input of the extreme learning machine network; attaching a classification label to the data according to the actual power consumption data to serve as the output of the extreme learning machine network, so as to train the extreme learning machine network; meanwhile, selecting an activation function suitable for user electricity behavior analysis; after determining the activation function, training the extreme learning machine network for multiple times by changing the number of nodes of the hidden layer, and calculating the accuracy of training results under the nodes of different hidden layers; and meanwhile, the calculation complexity is reduced, and a physical-data fusion model is obtained.
According to a characteristic optimization strategy, a preferred characteristic set of the user power consumption data suitable for use is selected, the characteristic optimization of a resident user load curve is carried out, the preferred characteristic set comprises daily average load, valley power coefficient, average section power consumption percentage and peak time power consumption rate, and normalization processing is carried out on the characteristic set.
The fusion of the physical model and the data model in the step 3 is based on multiple training of the extreme learning machine to obtain the weight of the output layerThe method comprises the following steps:
s1 for N arbitrary different samples (x i ，t i ) When the hidden layer unit of the extreme learning machine isWhen the activation function is g (x), the mathematical model can be represented by the following formula:
wherein j=1, 2 …, N; w (w) i ＝[w i1 ，w i2 ,…,w in ]T is a weight vector connecting the input feature and the ith hidden layer unit; beta i ＝[β i1 ，β i2 ,…,β in ]T is a weight vector connecting the ith hidden layer unit and the output result; b i Bias for the i-th hidden layer cell: w (w) i ·x j Representing the inner product of the two;
s2, selecting a function g (-) meeting infinite microability as an activation function of a hidden layer node, and randomly setting a weight w between an input layer and the hidden layer and a threshold b of the hidden layer node;
s3, determining the number of hidden layer nodes, and further according to the following formula:
obtaining an hidden layer matrix; wherein: w (w) i Is hidden layer node O i A connection weight matrix with each node of the input layer; f (F) j Is the input feature of the jth sample, one sample is composed of n points; n is the output value corresponding to the j-th input sample, and the output of one sample consists of m points;
s4, calculating the Moor-Penrose generalized inverse of the output matrix H by adopting a singular value decomposition method, and then according to the following formula X:
calculating the weight of the output layerAnd finishing the network training of the extreme learning machine.
The invention adopts a characteristic optimization strategy to extract the optimal characteristic set of the load curve, and takes the optimal characteristic set as the input of the ELM network. Attaching a classification label to the data according to the real measurement data and taking the classification label as output of the ELM network so as to train the ELM network; the trained network is used for classifying the electricity utilization behaviors of the users. If the output result of the ELM network is the same as the original label of the set of data, the training is considered correct. And then respectively calculating the accuracy of classification of the training set and the testing set, and comparing the influence of different parameters on the performance of the training result by changing the input parameters of the algorithm. Changing the activation function of hidden layer nodes of the ELM algorithm, repeatedly training the ELM network, and calculating the training accuracy under different activation functions. Selecting an activation function suitable for user electricity behavior analysis by taking the accuracy as an evaluation index; after the activation function is determined, changing the number of hidden layer nodes, repeating the step of training the ELM network, and calculating the accuracy of training results under different hidden layer nodes. On the basis of ensuring accuracy, the calculation complexity is reduced, and the number of nodes suitable for model analysis hidden layer is selected.
The analysis system of building modeling is fused to physics-data based on extreme learning machine, its characterized in that: the system comprises a data acquisition and preprocessing module and a building physical module; a physical-data fusion module;
the acquisition and preprocessing module is used for acquiring and preprocessing data;
building physical module, which is used to build building physical model based on the overall measurement and dialect method by collecting and preprocessing the building data and electric data;
the physical-data fusion module is used for training the building physical model, the collected and preprocessed user data, the environment data and the actual measured data by using an extreme learning machine to obtain a physical-data fusion model;
the static parameters of the electricity consumption behavior to be analyzed are input into the physical-data fusion model through the physical model and the dynamic parameters, and an analysis result is obtained.
The acquisition and preprocessing module specifically comprises: the collected data comprise building construction data, electrical data, user data, environment data and actual measurement numbers; preprocessing the collected building construction data, electrical data, user data and environment data, cleaning abnormal data, and repairing error data based on a Neville algorithm of Lagrange interpolation.
The building physical module specifically comprises:
fusing a physical model and a data model through a machine learning method, describing the state of a building by using a physical simplified model reflecting a physical association relation, generating data with high entropy characteristics, and taking the data as the physical model and the data model to be input;
the expressions of the physical model and the data model are:
where k represents a time status tag, k+1 represents a future time, X k+1 A vector composed of system state characteristics predicted at the time of k+1; x is x k+1 ' is a state characteristic vector to be predicted which is preprocessed by a physical model at the moment k+1; f and h respectively reflect the mapping relation between the measured data and the data characteristics to be predicted in the physical model and the data model;u is a random error vector in the calculation of the data model; x is X k And Y k The vector of measurement data for the power system is different from the measurement data vector X k Processed by a physical model to measure a data vector Y k Processing the data through a data model;
the physical simplification model module specifically comprises:
firstly, building physical models are built; the building physical model comprises a building physical model and a heat pump physical model
By analogy to Newton's law of cooling and ohm's law, heat is analogized to charge in the circuit, temperature difference is analogized to voltage, heat transfer rate is analogized to current, and thermal resistance is analogized to resistance; meanwhile, in order to further study the transient thermal behavior of the intelligent building model, the heat capacity, namely the heat storage capacity of an object, is introduced; in the RC network model, wall, indoor and outdoor air are used as nodes in the circuit model,
The intelligent building model is as follows: assuming that the heating system is located in a technical room tr of the building; assuming that all parts of the heating system are linear with respect to heat loss from the technical chamber; these heat losses are expressed as heat pump losses hp, thermal resistances of the space heating storage tank sh and the domestic hot water storage tank dw; a ground gnd, which is assumed to be constant temperature, is discharged by means of transport and ventilation into the surrounding air amb, irrespective of temperature variations; since it is assumed that the floor and all walls are constructed using heavy materials, the thermal resistance of the floor and the exterior wall is divided into several parts R on either side of the lumped heat capacity w1 、R w2 And R is f1 、R f2 ，R w1 、R w2 Is the equivalent thermal resistance of the outer wall, R f1 、R f2 Is the equivalent thermal resistance of the floor; modeling ventilation heat loss as thermal resistance R ve Solar energy and internal revenue are modeled as heat flow to all heat capacities;
the building model can be written as:
/>
wherein C is in 、C wi 、C we 、C f1 、C tr Respectively representing indoor air, inner wall, outer wall, floor, technical room temperature equivalent capacitance, t represents time variable, Q in Indicating the heating capacity of the room air, Q wi Represents the heating capacity of the inner wall, Q we Represents the heating capacity of the outer wall, Q f1 Indicating the heating capacity of the floor, Q hp,loss Indicating heat loss from the heat pump;
T in 、T amb 、T wi 、T f1 、T sh 、T tr respectively representing indoor air, inflow indoor air, inner wall, floor, hot water pump, technical room temperature;
In the optimal control problem, heat loss of the heat pump to the machine room is ignored; the equation is discretized as:
wherein T is i And T i+1 The temperatures at the time i and the time i+1 are shown, respectively, and Δt represents the heating time
The operating temperature of the building is linearized:
wherein Q represents the heating capacity of the heat pump, P represents the required power input, lt, mt and ht represent low temperature, medium temperature and high temperature respectively;
modeling the heating system of the heat pump heating system is as follows
The heat pump provides heat for space heating or domestic hot water production:
indicating the heating capacity of the heat pump at low temperature at instant i,/and a method for controlling the heating capacity of the heat pump at low temperature>Indicating the heating capacity of the heat pump at low temperature at instant i, < >>The heating capacity of the water storage tank at low temperature at the moment i is shown, and the same is true; />Indicating the heating capacity of the heat pump at medium temperature at instant i,/and a method for controlling the heating capacity of the heat pump at medium temperature>Indicating the heating capacity of the heater at mid-temperature of the heat pump at time i,/>Indicating the heating capacity of the water storage tank at medium temperature at time i, < >>Indicating the heating capacity of the heat pump at high temperature at time i,indicating the heating capacity of the heater at high temperature of the heat pump at instant i,/and a method for controlling the heating capacity of the heater at high temperature of the heat pump at instant i>The heating capacity of the water storage tank at the high temperature at the moment i is shown;
the heat pump efficiency varies with ambient temperature, water temperature and compressor adjustment for the following reasons:
p hp Representing the total power of the heat pump, p hp,ht Represents high temperature heating power, p hp,lt Indicating low temperature heating power, p hp,mt Indicating medium temperature heating power ζ l Indicating modulation degree ζ q Is a constraint coefficient;
the auxiliary heater efficiency is assumed to be 1:
auxiliary heater power for instant i, +.>The heating capacity of the auxiliary heater at the moment i;
the spatial heat storage state is calculated by the following formula:
represents the heat quantity discharged at the moment i, T em,ret Indicating the compensated loss temperature;
there are other limitations:
ρ represents the density of the water,losing the volume of water in the heat; />Representing the volume of maximum supply water to the heat pump; epsilon no m Representing heat pump heating efficiency;
calculating the state of a domestic hot water storage tank:
/>
C dw,lt representing the equivalent capacitance of the water storage tank at low temperature, C dw,mt Representing the equivalent capacitance of the water storage tank at medium temperature, C dw,ht Representing the equivalent capacitance of the water storage tank at high temperature, C dw,vht Represents the equivalent capacitance of the water storage tank at the ultra-high temperature,indicating the heating capacity at ultra-high temperature at time i,/-, for example>The heating capacity of the auxiliary heater at the ultrahigh temperature at the moment i is shown;
other limitations:
indicates the temperature of the water storage tank at ultra-high temperature, < >>Indicates the temperature of the water storage tank at high temperature, < >>Indicates the temperature of the water storage tank at low temperature, < >>The temperature of the water storage tank is shown at medium temperature; t (T) mains Indicating the temperature of the thermostatic mixing valve after mixing the stored water with cold water from the water line; t (T) dw,dem Representing the required temperature of the water storage tank; c represents the specific heat capacity of water;
the multi-standard target is written as:
W EIE weight factor representing the influence of energy on environment, j is applicable to the power generation index of all power plants, p impact,j For the marginal effect of the power generation of the power plant,representing power generated by a power plant over a period of time at a power generation index;
defining different control targets for the control strategy of the heat pump building physical model;
from the perspective of an individual consumer, it is desirable for the consumer to minimize his own costs for providing thermal comfort and domestic hot water; this objective is achieved by minimizing the individual energy costs within a control range if thermal comfort is maintained; the cost of discomfort is added to the objective function from as a soft constraint:
by usingIndicating the electricity price at time steps i and +.>And->The power of the heat pump power supply and the auxiliary heater are respectively; quantifying the discomfort cost by multiplying the building operating temperature deviation outside the building comfort temperature boundary by the building discomfort cost p;
secondly, considering the viewpoint of environmental impact of energy use, minimizing the impact of the used energy on the environment; since the impact is different for each foundry, optimization attempts are made to run the most environmentally friendly factories as much as possible; the objective function is then written as:
The influence of power generation on different power plants is selected as the single-stage influence of roof photovoltaic power generation, offshore wind power generation, nuclear power generation, combined cycle natural gas power generation and open cycle gas turbine power generation; finally, considering the influence of power generation on the environment, not only the use of energy is important; for many power plants, a large amount of emissions is accompanied by the construction and decommissioning of the power plant;
representing a heat pump overload threshold,/->Indicating overload power +.>Representing the remaining load power;
since this is an approximation, the installed capacity cost is included in the objective function, here designated capacity price is pimpact; cap has little physical meaning;
if an overload condition occurs, the influence of the overload condition is far greater than the influence of power generation at the power just below an overload threshold; the three views of consumer cost, energy influence and productivity influence are integrated into a single multi-objective function; consumers of energy-dependent scale environmental impact power generation costs incorporating energy impact into the consumer cost function environmental impact weight factor energy introduction scale; when the factor is small, the controller ignores the influence of power generation; the factors are large, and the capacity influence point of view can be simply added to the cost function; to change the behavior of the controller, the overload threshold is also changed to minimize
l cap Representing the capacity limiting factor of the device,indicating the maximum residual load power.
The physical-data fusion module is specifically:
physical modelThe fusion of the data model is based on the multiple training of the extreme learning machine to obtain the weight of the output layerThe method comprises the following steps:
s1 for N arbitrary different samples (x i ，t i ) When the hidden layer unit of the extreme learning machine isWhen the activation function is g (x), the mathematical model can be represented by the following formula:
/>
wherein j=1, 2 …, N; w (w) i ＝[w i1 ，w i2 ,…,w in ]T is a weight vector connecting the input feature and the ith hidden layer unit; beta i ＝[β i1 ，β i2 ,…,β in ]T is a weight vector connecting the ith hidden layer unit and the output result; b i Bias for the i-th hidden layer cell: w (w) i ·x j Representing the inner product of the two;
s2, selecting a function g (-) meeting infinite microability as an activation function of a hidden layer node, and randomly setting a weight w between an input layer and the hidden layer and a threshold b of the hidden layer node;
s3, determining the number of hidden layer nodes, and further according to the following formula:
obtaining an hidden layer matrix; wherein: w (w) i Is hidden layer node O i A connection weight matrix with each node of the input layer; f (F) j Is the input feature of the jth sample, one sample is composed of n points; n is the output value corresponding to the j-th input sample, and the output of one sample consists of m points;
S4, calculating the Moor-Penrose generalized inverse of the output matrix H by adopting a singular value decomposition method, and then according to the following formula X:
calculating the weight of the output layerAnd finishing the network training of the extreme learning machine.
As shown in fig. 1 and 2, the electricity load sample data collected by the building is first divided into building data, electric data, environment data, and behavior data. By establishing a physical simple model and guiding and correcting the physical model by real electricity data, the physical analysis method can provide information with high entropy characteristics for the data analysis method, and is beneficial to improving the analysis efficiency of the data model, and the principle is as follows: the input features comprise building models to be predicted, so that the search space can be reduced and the calculation complexity can be reduced when the data model parameters are solved through an optimization process; at the same time, the method is helpful for establishing a better data model, namely: the input characteristics of high entropy make the object of establishing the data model more definite, have more pertinence when model parameter optimization solves, avoid sinking into the local optimum, thus improve the rationality of the data model. The data driving method can solve the problem of rule loss caused by model simplification and the like in the physical analysis method. And obtaining a final physical-data fusion model. And then inputting the electricity behavior data to be predicted and analyzed into an electricity analysis prediction model based on physical-data fusion to obtain a final physical-data fusion model.
As shown in fig. 3, the feature optimization strategy is used to extract the optimal feature set of the load curve, and the optimal feature set is used as the input of the ELM network. Attaching a classification label to the data according to the real measurement data and taking the classification label as output of the ELM network so as to train the ELM network; the trained network is used for classifying the electricity utilization behaviors of the users. If the output result of the ELM network is the same as the original label of the set of data, the training is considered correct. And then respectively calculating the accuracy of classification of the training set and the testing set, and comparing the influence of different parameters on the performance of the training result by changing the input parameters of the algorithm. Changing the activation function of hidden layer nodes of the ELM algorithm, repeatedly training the ELM network, and calculating the training accuracy under different activation functions. Selecting an activation function suitable for user electricity behavior analysis by taking the accuracy as an evaluation index; after the activation function is determined, changing the number of hidden layer nodes, repeating the step of training the ELM network, and calculating the accuracy of training results under different hidden layer nodes. On the basis of ensuring accuracy, the calculation complexity is reduced, and the number of nodes suitable for model analysis hidden layer is selected.

Claims (8)

1. The analysis method of the physical-data fusion building modeling based on the extreme learning machine is characterized by comprising the following steps of: the method comprises the following steps:
Step 1, data acquisition and pretreatment:
step 2, building a building physical model based on an overall measurement and dialect method by collecting and preprocessing building data and electrical data;
training the building physical model, the collected and preprocessed user data, the collected and preprocessed environment data and the collected and preprocessed actual power consumption data by using an extreme learning machine to obtain a physical-data fusion model;
static parameters of electricity consumption behaviors to be analyzed are input into a physical-data fusion model through a physical model and dynamic parameters, so that an analysis result is obtained;
building a building physical model by a general survey and debate method specifically comprises:
fusing a physical model and a data model through a machine learning method, describing the state of a building by using a physical simplified model reflecting a physical association relation, generating data with high entropy characteristics, and taking the data as the physical model and the data model to be input;
the expressions of the physical model and the data model are:
wherein k represents timeStatus tag, k+1 denotes future time, X k+1 A vector composed of system state characteristics predicted at the time of k+1; x is x k+1 ' is a state characteristic vector to be predicted which is preprocessed by a physical model at the moment k+1; f and h respectively reflect the mapping relation between the measured data and the data characteristics to be predicted in the physical model and the data model; u is a random error vector in the calculation of the data model; x is X k And Y k The vector of measurement data for the power system is different from the measurement data vector X k Processed by a physical model to measure a data vector Y k Processing the data through a data model;
the physical simplified model comprises the following specific steps:
firstly, building physical models are built; the building physical model comprises a building physical model and a heat pump physical model;
by analogy to Newton's law of cooling and ohm's law, heat is analogized to charge in the circuit, temperature difference is analogized to voltage, heat transfer rate is analogized to current, and thermal resistance is analogized to resistance; meanwhile, in order to further study the transient thermal behavior of the intelligent building model, the heat capacity, namely the heat storage capacity of an object, is introduced; in the RC network model, wall, indoor and outdoor air are used as nodes in the circuit model;
the intelligent building model is as follows: assuming that the heating system is located in a technical room tr of the building; assuming that all parts of the heating system are linear with respect to heat loss from the technical chamber; these heat losses are expressed as heat pump losses hp, thermal resistances of the space heating storage tank sh and the domestic hot water storage tank dw; a discharge of air amb into the surroundings by transport and ventilation and to the ground, which is assumed to be at constant temperature; since it is assumed that the floor and all walls are constructed using heavy materials, the thermal resistance of the floor and the exterior wall is divided into several parts R on either side of the lumped heat capacity w1 、R w2 And R is f1 、R f2 ，R w1 、R w2 Is the equivalent thermal resistance of the outer wall, R f1 、R f2 Is the equivalent thermal resistance of the floor; modeling ventilation heat loss as thermal resistance R ve Solar energy and internal revenue are modeled as heat flow to all heat capacities;
the building model is written as:
wherein C is in 、C wi 、C we 、C f1 、C tr Respectively representing indoor air, inner wall, outer wall, floor, technical room temperature equivalent capacitance, t represents time variable, Q in Indicating the heating capacity of the room air, Q wi Represents the heating capacity of the inner wall, Q we Represents the heating capacity of the outer wall, Q f1 Indicating the heating capacity of the floor, Q hp,loss Indicating heat loss from the heat pump;
T in 、T amb 、T wi 、T f1 、T sh 、T tr respectively representing indoor air, inflow indoor air, inner wall, floor, hot water pump, technical room temperature;
in the optimal control problem, heat loss of the heat pump to the machine room is ignored; the equation is discretized as:
wherein T is i And T i+1 The temperatures at the time i and the time i+1 are respectively represented, and Δt represents the heating time;
the operating temperature of the building is linearized:
wherein Q represents the heating capacity of the heat pump, P represents the required power input, lt, mt and ht represent low temperature, medium temperature and high temperature respectively;
modeling a heating system of a heat pump heating system as follows;
the heat pump provides heat for space heating or domestic hot water production:
indicating the heating capacity of the heat pump at low temperature at instant i,/and a method for controlling the heating capacity of the heat pump at low temperature >Indicating the heating capacity of the heat pump at low temperature at instant i, < >>The heating capacity of the water storage tank at low temperature at the moment i is shown, and the same is true; />Indicating the heating capacity of the heat pump at medium temperature at instant i,/and a method for controlling the heating capacity of the heat pump at medium temperature>Indicating the heating capacity of the heater at mid-temperature of the heat pump at time i,/>Indicating the heating capacity of the water storage tank at medium temperature at time i, < >>Indicating the heating capacity of the heat pump at high temperature at instant i,/and a method for controlling the heating capacity of the heat pump at high temperature>Indicating the heating capacity of the heater at high temperature of the heat pump at instant i,/and a method for controlling the heating capacity of the heater at high temperature of the heat pump at instant i>The heating capacity of the water storage tank at the high temperature at the moment i is shown;
the heat pump efficiency varies with ambient temperature, water temperature and compressor adjustment for the following reasons:
p hp representing the total power of the heat pump, p hp,ht Represents high temperature heating power, p hp,lt Indicating low temperature heating power, p hp,mt Indicating medium temperature heating power ζ l Indicating modulation degree ζ q Is a constraint coefficient;
the auxiliary heater efficiency is assumed to be 1:
auxiliary heater power for instant i, +.>The heating capacity of the auxiliary heater at the moment i;
the spatial heat storage state is calculated by the following formula:
represents the heat quantity discharged at the moment i, T em,ret Indicating the compensated loss temperature;
there are other limitations:
ρ represents the density of the water,losing the volume of water in the heat; />Representing the volume of maximum supply water to the heat pump; epsilon nom Representing heat pump heating efficiency;
calculating the state of a domestic hot water storage tank:
C dw,lt representing the equivalent capacitance of the water storage tank at low temperature, C dw,mt Representing the equivalent capacitance of the water storage tank at medium temperature, C dw,ht Representing the equivalent capacitance of the water storage tank at high temperature, C dw,vht Represents the equivalent capacitance of the water storage tank at the ultra-high temperature,indicating the heating capacity at ultra-high temperature at time i,/-, for example>The heating capacity of the auxiliary heater at the ultrahigh temperature at the moment i is shown;
other limitations:
indicates the temperature of the water storage tank at ultra-high temperature, < >>Indicates the temperature of the water storage tank at high temperature, < >>Indicates the temperature of the water storage tank at low temperature, < >>The temperature of the water storage tank is shown at medium temperature; t (T) mains Indicating the temperature of the thermostatic mixing valve after mixing the stored water with cold water from the water line; t (T) dw,dem Representing the required temperature of the water storage tank; c represents the specific heat capacity of water;
the multi-standard target is written as:
W EIE weight factor representing the influence of energy on environment, j is applicable to the power generation index of all power plants, p impact,j For the marginal effect of the power generation of the power plant,representing power generated by a power plant over a period of time at a power generation index;
defining different control targets for the control strategy of the heat pump building physical model;
from the perspective of an individual consumer, it is desirable for the consumer to minimize his own costs for providing thermal comfort and domestic hot water; this objective is achieved by minimizing the individual energy costs within a control range if thermal comfort is maintained; the cost of discomfort is added to the objective function from as a soft constraint:
By usingIndicating the electricity price at time steps i and +.>And->The power of the heat pump power supply and the auxiliary heater are respectively; quantifying the discomfort cost by multiplying the building operating temperature deviation outside the building comfort temperature boundary by the building discomfort cost p;
secondly, considering the viewpoint of environmental impact of energy use, minimizing the impact of the used energy on the environment; since the impact is different for each foundry, optimization attempts are made to run the most environmentally friendly factories as much as possible; the objective function is then written as:
the influence of power generation on different power plants is selected as the single-stage influence of roof photovoltaic power generation, offshore wind power generation, nuclear power generation, combined cycle natural gas power generation and open cycle gas turbine power generation; finally, considering the influence of power generation on the environment, not only the use of energy is important; for many power plants, a large amount of emissions is accompanied by the construction and decommissioning of the power plant;
representing a heat pump overload threshold,/->Indicating overload power +.>Representing the remaining load power;
since this is an approximation, the installed capacity cost is included in the objective function, here designated capacity price is pimpact; cap has little physical meaning;
If an overload condition occurs, the influence of the overload condition is far greater than the influence of power generation at the power just below an overload threshold; the three views of consumer cost, energy influence and productivity influence are integrated into a single multi-objective function; consumers of energy-dependent scale environmental impact power generation costs incorporating energy impact into the consumer cost function environmental impact weight factor energy introduction scale; when the factor is small, the controller ignores the influence of power generation; the factors are large, and the capacity influence point of view is simply added to the cost function; in order to change the behavior of the controller, the overload threshold is also changed in order to minimize the amount;
l cap representing the capacity limiting factor of the device,indicating the maximum residual load power.
2. The method for analyzing physical-data fusion building modeling based on extreme learning machine according to claim 1, wherein the method comprises the following steps: training by an extreme learning machine according to the building physical model, the collected and preprocessed user data, environment data and actual measured data, wherein the method specifically comprises the following steps:
using the optimal feature set as input of the extreme learning machine network; attaching a classification label to the data according to the actual power consumption data to serve as the output of the extreme learning machine network, so as to train the extreme learning machine network; meanwhile, selecting an activation function suitable for user electricity behavior analysis; after determining the activation function, training the extreme learning machine network for multiple times by changing the number of nodes of the hidden layer, and calculating the accuracy of training results under the nodes of different hidden layers; and meanwhile, the calculation complexity is reduced, and a physical-data fusion model is obtained.
3. The method for analyzing physical-data fusion building modeling based on extreme learning machine according to claim 1, wherein the method comprises the following steps: according to a characteristic optimization strategy, a preferred characteristic set of the user power consumption data suitable for use is selected, the characteristic optimization of a resident user load curve is carried out, the preferred characteristic set comprises daily average load, valley power coefficient, average section power consumption percentage and peak time power consumption rate, and normalization processing is carried out on the characteristic set.
4. The method for analyzing physical-data fusion building modeling based on extreme learning machine according to claim 1, wherein the method comprises the following steps: the fusion of the physical model and the data model in the step 3 is based on multiple training of the extreme learning machine to obtain the weight of the output layerThe method comprises the following steps:
s1 for N arbitrary different samples (x i ，t i ) When the hidden layer unit of the extreme learning machine isWhen the activation function is g (x), the mathematical model can be represented by the following formula:
wherein j=1, 2 …, N; w (w) i ＝[w i1 ，w i2 ,…,w in ]T is a weight vector connecting the input feature and the ith hidden layer unit; beta i ＝[β i1 ，β i2 ,…,β in ]T is a weight vector connecting the ith hidden layer unit and the output result; b i Bias for the i-th hidden layer cell: w (w) i ·x j Representing the inner product of the two;
s2, selecting a function g (-) meeting infinite microability as an activation function of a hidden layer node, and randomly setting a weight w between an input layer and the hidden layer and a threshold b of the hidden layer node;
S3, determining the number of hidden layer nodes, and further according to the following formula:
obtaining an hidden layer matrix; wherein: w (w) i Is hidden layer node O i A connection weight matrix with each node of the input layer; f (F) j Is the input feature of the jth sample, one sample is composed of n points; n is the output value corresponding to the j-th input sample, and the output of one sample consists of m points;
s4, calculating the Moor-Penrose generalized inverse of the output matrix H by adopting a singular value decomposition method, and then according to the following formula X:
calculating the weight of the output layerAnd finishing the network training of the extreme learning machine.
5. The analysis system of building modeling is fused to physics-data based on extreme learning machine, its characterized in that: the system comprises a data acquisition and preprocessing module and a building physical module; a physical-data fusion module;
the acquisition and preprocessing module is used for acquiring and preprocessing data;
building physical module, which is used to build building physical model based on the overall measurement and dialect method by collecting and preprocessing the building data and electric data;
the physical-data fusion module is used for training the building physical model, the collected and preprocessed user data, the environment data and the actual measured data by using an extreme learning machine to obtain a physical-data fusion model;
The static parameters of the electricity consumption behavior to be analyzed are input into the physical-data fusion model through the physical model and the dynamic parameters, and an analysis result is obtained.
6. The extreme learning machine-based physical-data fusion building modeling analysis system of claim 5, wherein: the acquisition and preprocessing module specifically comprises: the collected data comprise building construction data, electrical data, user data, environment data and actual measurement numbers; preprocessing the collected building construction data, electrical data, user data and environment data, cleaning abnormal data, and repairing error data based on a Neville algorithm of Lagrange interpolation.
7. The extreme learning machine-based physical-data fusion building modeling analysis system of claim 5, wherein: the building physical module specifically comprises:
fusing a physical model and a data model through a machine learning method, describing the state of a building by using a physical simplified model reflecting a physical association relation, generating data with high entropy characteristics, and taking the data as the physical model and the data model to be input;
the expressions of the physical model and the data model are:
where k represents a time status tag, k+1 represents a future time, X k+1 A vector composed of system state characteristics predicted at the time of k+1; x is x k+1 ' is a state characteristic vector to be predicted which is preprocessed by a physical model at the moment k+1; f and h respectively reflect the mapping relation between the measured data and the data characteristics to be predicted in the physical model and the data model; u is a random error vector in the calculation of the data model; x is X k And Y k The vector of measurement data for the power system is different from the measurement data vector X k Processed by physical modelMeasuring the data vector Y k Processing the data through a data model;
the physical simplification model module specifically comprises:
firstly, building physical models are built; the building physical model comprises a building physical model and a heat pump physical model
By analogy to Newton's law of cooling and ohm's law, heat is analogized to charge in the circuit, temperature difference is analogized to voltage, heat transfer rate is analogized to current, and thermal resistance is analogized to resistance; meanwhile, in order to further study the transient thermal behavior of the intelligent building model, the heat capacity, namely the heat storage capacity of an object, is introduced; in the RC network model, wall, indoor and outdoor air are used as nodes in the circuit model,
the intelligent building model is as follows: assuming that the heating system is located in a technical room tr of the building; assuming that all parts of the heating system are linear with respect to heat loss from the technical chamber; these heat losses are expressed as heat pump losses hp, thermal resistances of the space heating storage tank sh and the domestic hot water storage tank dw; a ground gnd, which is assumed to be thermostatted, is discharged by transport and ventilation into the surrounding air amb; since it is assumed that the floor and all walls are constructed using heavy materials, the thermal resistance of the floor and the exterior wall is divided into several parts R on either side of the lumped heat capacity w1 、R w2 And R is f1 、R f2 ，R w1 、R w2 Is the equivalent thermal resistance of the outer wall, R f1 、R f2 Is the equivalent thermal resistance of the floor; modeling ventilation heat loss as thermal resistance R ve Solar energy and internal revenue are modeled as heat flow to all heat capacities;
the building model is written as:
wherein C is in 、C wi 、C we 、C f1 、C tr Respectively representing indoor air, inner wall, outer wall, floor, technical room temperature equivalent capacitance, t represents time variable, Q in Indicating the heating capacity of the room air, Q wi Represents the heating capacity of the inner wall, Q we Represents the heating capacity of the outer wall, Q f1 Indicating the heating capacity of the floor, Q hp,loss Indicating heat loss from the heat pump;
T in 、T amb 、T wi 、T f1 、T sh 、T tr respectively representing indoor air, inflow indoor air, inner wall, floor, hot water pump, technical room temperature;
in the optimal control problem, heat loss of the heat pump to the machine room is ignored; the equation is discretized as:
wherein T is i And T i+1 The temperatures at the time i and the time i+1 are shown, respectively, and Δt represents the heating time
The operating temperature of the building is linearized:
wherein Q represents the heating capacity of the heat pump, P represents the required power input, lt, mt and ht represent low temperature, medium temperature and high temperature respectively;
modeling the heating system of the heat pump heating system is as follows
The heat pump provides heat for space heating or domestic hot water production:
indicating the heating capacity of the heat pump at low temperature at instant i,/and a method for controlling the heating capacity of the heat pump at low temperature >Indicating the heating capacity of the heat pump at low temperature at instant i, < >>The heating capacity of the water storage tank at low temperature at the moment i is shown, and the same is true; />Indicating the heating capacity of the heat pump at medium temperature at instant i,/and a method for controlling the heating capacity of the heat pump at medium temperature>Indicating the heating capacity of the heater at mid-temperature of the heat pump at time i,/>Indicating the heating capacity of the water storage tank at medium temperature at time i, < >>Indicating the heating capacity of the heat pump at high temperature at instant i,/and a method for controlling the heating capacity of the heat pump at high temperature>Indicating the heating capacity of the heater at high temperature of the heat pump at instant i,/and a method for controlling the heating capacity of the heater at high temperature of the heat pump at instant i>The heating capacity of the water storage tank at the high temperature at the moment i is shown;
the heat pump efficiency varies with ambient temperature, water temperature and compressor adjustment for the following reasons:
p hp representing the total power of the heat pump, p hp,ht Represents high temperature heating power, p hp,lt Indicating low temperature heating power, p hp,mt Indicating medium temperature heatingThermal power ζ l Indicating modulation degree ζ q Is a constraint coefficient;
the auxiliary heater efficiency is assumed to be 1:
auxiliary heater power for instant i, +.>The heating capacity of the auxiliary heater at the moment i;
the spatial heat storage state is calculated by the following formula:
represents the heat quantity discharged at the moment i, T em,ret Indicating the compensated loss temperature;
there are other limitations:
ρ represents the density of the water,losing the volume of water in the heat; />Representing the volume of maximum supply water to the heat pump; epsilon nom Representing heat pump heating efficiency;
calculating the state of a domestic hot water storage tank:
C dw,lt representing the equivalent capacitance of the water storage tank at low temperature, C dw,mt Representing the equivalent capacitance of the water storage tank at medium temperature, C dw,ht Representing the equivalent capacitance of the water storage tank at high temperature, C dw,vht Represents the equivalent capacitance of the water storage tank at the ultra-high temperature,indicating the heating capacity at ultra-high temperature at time i,/-, for example>The heating capacity of the auxiliary heater at the ultrahigh temperature at the moment i is shown;
other limitations:
indicates the temperature of the water storage tank at ultra-high temperature, < >>Indicates the temperature of the water storage tank at high temperature, < >>Indicates the temperature of the water storage tank at low temperature, < >>The temperature of the water storage tank is shown at medium temperature; t (T) mains Indicating the temperature of the thermostatic mixing valve after mixing the stored water with cold water from the water line; t (T) dw,dem Representing the required temperature of the water storage tank; c represents the specific heat capacity of water;
the multi-standard target is written as:
W EIE weight factor representing the influence of energy on environment, j is applicable to the power generation index of all power plants, p impact,j For the marginal effect of the power generation of the power plant,representing power generated by a power plant over a period of time at a power generation index;
defining different control targets for the control strategy of the heat pump building physical model;
from the perspective of an individual consumer, it is desirable for the consumer to minimize his own costs for providing thermal comfort and domestic hot water; this objective is achieved by minimizing the individual energy costs within a control range if thermal comfort is maintained; the cost of discomfort is added to the objective function from as a soft constraint:
By usingIndicating the electricity price at time steps i and +.>And->The power of the heat pump power supply and the auxiliary heater are respectively; quantifying the discomfort cost by multiplying the building operating temperature deviation outside the building comfort temperature boundary by the building discomfort cost p;
secondly, considering the viewpoint of environmental impact of energy use, minimizing the impact of the used energy on the environment; since the impact is different for each foundry, optimization attempts are made to run the most environmentally friendly factories as much as possible; the objective function is then written as:
the influence of power generation on different power plants is selected as the single-stage influence of roof photovoltaic power generation, offshore wind power generation, nuclear power generation, combined cycle natural gas power generation and open cycle gas turbine power generation; finally, considering the influence of power generation on the environment, not only the use of energy is important; for many power plants, a large amount of emissions is accompanied by the construction and decommissioning of the power plant;
representing a heat pump overload threshold,/->Indicating overload power +.>Representing the remaining load power;
since this is an approximation, the installed capacity cost is included in the objective function, here designated capacity price is pimpact; cap has little physical meaning;
If an overload condition occurs, the influence of the overload condition is far greater than the influence of power generation at the power just below an overload threshold; the three views of consumer cost, energy influence and productivity influence are integrated into a single multi-objective function; consumers of energy-dependent scale environmental impact power generation costs incorporating energy impact into the consumer cost function environmental impact weight factor energy introduction scale; when the factor is small, the controller ignores the influence of power generation; the factors are large, and the capacity influence point of view is simply added to the cost function; to change the behavior of the controller, the overload threshold is also changed to minimize
l cap Representing the capacity limiting factor of the device,indicating the maximum residual load power.
8. The extreme learning machine-based physical-data fusion building modeling analysis system of claim 5, wherein: the physical-data fusion module is specifically:
the fusion of the physical model and the data model is based on the multiple training of the extreme learning machine to obtain the weight of the output layerThe method comprises the following steps:
s1 for N arbitrary different samples (x i ，t i ) When the hidden layer unit of the extreme learning machine isWhen the activation function is g (x), the mathematical model can be represented by the following formula:
wherein j=1, 2 …, N; w (w) i ＝[w i1 ，w i2 ,…,w in ]T is a weight vector connecting the input feature and the ith hidden layer unit; beta i ＝[β i1 ，β i2 ,…,β in ]T is a weight vector connecting the ith hidden layer unit and the output result; b i Bias for the i-th hidden layer cell: w (w) i ·x j Representing the inner product of the two;
s2, selecting a function g (-) meeting infinite microability as an activation function of a hidden layer node, and randomly setting a weight w between an input layer and the hidden layer and a threshold b of the hidden layer node;
s3, determining the number of hidden layer nodes, and further according to the following formula:
obtaining an hidden layer matrix; wherein: w (w) i Is hidden layer node O i A connection weight matrix with each node of the input layer; f (F) j Is the input feature of the jth sample, one sample is composed of n points; n is the output value corresponding to the j-th input sample, and the output of one sample consists of m points;
s4, calculating the Moor-Penrose generalized inverse of the output matrix H by adopting a singular value decomposition method, and then according to the following formula X:
calculating the weight of the output layerAnd finishing the network training of the extreme learning machine. />
CN202010571025.2A 2020-04-02 2020-06-22 Analysis method of physical-data fusion building based on extreme learning machine Active CN111898856B (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN2020102534371 2020-04-02
CN202010253437 2020-04-02

Publications (2)

Publication Number Publication Date
CN111898856A CN111898856A (en) 2020-11-06
CN111898856B true CN111898856B (en) 2024-03-22

Family

ID=73206865

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010571025.2A Active CN111898856B (en) 2020-04-02 2020-06-22 Analysis method of physical-data fusion building based on extreme learning machine

Country Status (1)

Country Link
CN (1) CN111898856B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112669446B (en) * 2020-12-24 2024-04-19 联通(浙江)产业互联网有限公司 Building scene modeling method and device
CN113065190B (en) * 2021-04-15 2022-05-27 天津大学 Uncertainty-based residential building heat supply amount calculation method
CN114237057B (en) * 2021-12-20 2023-09-19 东南大学 Dynamic modeling method, system, equipment and medium for intelligent building electricity analysis

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015172560A1 (en) * 2014-05-16 2015-11-19 华南理工大学 Central air conditioner cooling load prediction method based on bp neural network
CN105160437A (en) * 2015-09-25 2015-12-16 国网浙江省电力公司 Load model prediction method based on extreme learning machine
CN107423839A (en) * 2017-04-17 2017-12-01 湘潭大学 A kind of method of the intelligent building microgrid load prediction based on deep learning
CN108321793A (en) * 2018-01-17 2018-07-24 东北电力大学 The active distribution network of integrated intelligent building flexible load models and Optimization Scheduling
CN109270841A (en) * 2018-10-11 2019-01-25 东北电力大学 A kind of energy flexibility regulation method of the intelligent building based on Model Predictive Control
FR3077119A1 (en) * 2018-01-23 2019-07-26 Saint-Gobain Isover METHOD AND DEVICE FOR PREDICTIVE DETERMINATION OF A THERMAL COMFORT INDICATOR OF A LOCAL

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015172560A1 (en) * 2014-05-16 2015-11-19 华南理工大学 Central air conditioner cooling load prediction method based on bp neural network
CN105160437A (en) * 2015-09-25 2015-12-16 国网浙江省电力公司 Load model prediction method based on extreme learning machine
CN107423839A (en) * 2017-04-17 2017-12-01 湘潭大学 A kind of method of the intelligent building microgrid load prediction based on deep learning
CN108321793A (en) * 2018-01-17 2018-07-24 东北电力大学 The active distribution network of integrated intelligent building flexible load models and Optimization Scheduling
FR3077119A1 (en) * 2018-01-23 2019-07-26 Saint-Gobain Isover METHOD AND DEVICE FOR PREDICTIVE DETERMINATION OF A THERMAL COMFORT INDICATOR OF A LOCAL
CN109270841A (en) * 2018-10-11 2019-01-25 东北电力大学 A kind of energy flexibility regulation method of the intelligent building based on Model Predictive Control

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
"基于极限学习机的居民用电行为分类分析方法";陆俊 等;《电力***自动化》;第43卷(第2期);第97-104页 *
"基于物理-数据融合的数字化楼宇用电模型构建方法";汤奕 等;《供用电》(第10期);第16-21、29页 *

Also Published As

Publication number Publication date
CN111898856A (en) 2020-11-06

Similar Documents

Publication Publication Date Title
CN111898856B (en) Analysis method of physical-data fusion building based on extreme learning machine
Li et al. Assessment of linear emulators in lightweight Bayesian calibration of dynamic building energy models for parameter estimation and performance prediction
Noye et al. A review of advanced ground source heat pump control: Artificial intelligence for autonomous and adaptive control
Darivianakis et al. The power of diversity: Data-driven robust predictive control for energy-efficient buildings and districts
CN113112077B (en) HVAC control system based on multi-step prediction deep reinforcement learning algorithm
Starke et al. Multi-objective optimization of a solar-assisted heat pump for swimming pool heating using genetic algorithm
KR102194002B1 (en) energy operation management system by using optimized physical learning model and machine learning methods
KR101301123B1 (en) Prediction method for cooling and heating load
Hosamo et al. Digital Twin of HVAC system (HVACDT) for multiobjective optimization of energy consumption and thermal comfort based on BIM framework with ANN-MOGA
Lu et al. Data augmentation strategy for short-term heating load prediction model of residential building
CN111754037A (en) Long-term load hybrid prediction method for regional terminal integrated energy supply system
CN114239991A (en) Building heat supply load prediction method, device and equipment based on data driving
Wang et al. Multi-criteria comprehensive study on predictive algorithm of heating energy consumption of district heating station based on timeseries processing
Abtahi et al. Control-oriented thermal network models for predictive load management in Canadian houses with on-Site solar electricity generation: application to a research house
Sun et al. Research on control strategy integrated with characteristics of user's energy-saving behavior of district heating system
Sretenović et al. Support vector machine for the prediction of heating energy use
Wang Typical building thermal and thermal load forecasting based on wavelet neural network
Han et al. Physical-data fusion modeling method for energy consumption analysis of smart building
Alden et al. Artificial intelligence method for the forecast and separation of total and hvac loads with application to energy management of smart and nze homes
CN114060902A (en) Heating prediction model establishing method, heating system and control method thereof
Zhang et al. Regulation capability evaluation of individual electric heating load based on radial basis function neural network
Li et al. Gossen’s First Law in the Modeling for Demand Side Management: A First Heat Pump Case Study
Zhang et al. A day-ahead operation regulation method for solar water heating based on model predictive control
Ariwoola et al. An integrated hybrid thermal dynamics model and energy aware optimization framework for grid-interactive residential building management
Fan et al. Ultrashort-term heat load prediction for heat exchange stations based on dual-source data feature fusion

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