CN111340270B - Intelligent household energy optimization method based on information interaction - Google Patents

Intelligent household energy optimization method based on information interaction Download PDF

Info

Publication number
CN111340270B
CN111340270B CN202010084781.2A CN202010084781A CN111340270B CN 111340270 B CN111340270 B CN 111340270B CN 202010084781 A CN202010084781 A CN 202010084781A CN 111340270 B CN111340270 B CN 111340270B
Authority
CN
China
Prior art keywords
state
equipment
power
time
action
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
CN202010084781.2A
Other languages
Chinese (zh)
Other versions
CN111340270A (en
Inventor
傅质馨
李潇逸
朱俊澎
袁越
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.)
Hohai University HHU
Original Assignee
Hohai University HHU
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 Hohai University HHU filed Critical Hohai University HHU
Priority to CN202010084781.2A priority Critical patent/CN111340270B/en
Publication of CN111340270A publication Critical patent/CN111340270A/en
Application granted granted Critical
Publication of CN111340270B publication Critical patent/CN111340270B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • 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
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0283—Price estimation or determination
    • 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • 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
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S50/00—Market activities related to the operation of systems integrating technologies related to power network operation or related to communication or information technologies
    • Y04S50/14—Marketing, i.e. market research and analysis, surveying, promotions, advertising, buyer profiling, customer management or rewards

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Development Economics (AREA)
  • Human Resources & Organizations (AREA)
  • General Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Finance (AREA)
  • Game Theory and Decision Science (AREA)
  • Accounting & Taxation (AREA)
  • Tourism & Hospitality (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention discloses an intelligent optimization method for household energy based on information interaction, which comprises the following steps: (1) information acquisition: collecting photovoltaic power generation information, electricity price information and the use state of household electrical equipment in a household energy management system; (2) optimizing admission prejudgment: judging whether energy optimization is carried out or not according to the acquired information; (3) monitoring and determining a device state type set; (4) and (3) performing optimized control output on the equipment, which comprises the following steps: establishing a Markov decision process model based on the electricity price, the photovoltaic prediction deviation and the controllable equipment; solving to obtain an optimal control strategy under the premise of minimum electric quantity cost of a user; and converting the equipment set into equipment control information, and transmitting the equipment control information to corresponding equipment for regulation and control. According to the method, uncertain factors such as electricity price information and user random behaviors are considered, the application energy demand can be met on the power grid side, the power consumption cost of the user is reduced on the demand side, and the power consumption demand of the user is met.

Description

Intelligent household energy optimization method based on information interaction
Technical Field
The invention relates to electric quantity optimization scheduling, in particular to an intelligent household energy optimization method based on information interaction.
Background
With the gradual development of social economy and the rapid progress of various advanced technologies, people have new understanding on the use and sustainable development of energy. Meanwhile, the construction of the ubiquitous power internet of things requires the construction of an intelligent networked power energy system which is more efficient, safer, more reliable, more flexible and sustainable. The energy management system plays a crucial role both on the demand side and on the supply side as an integral part of a stable, efficient and economic operation of the power grid. The Home Energy Management System (HEMS) is embodied at the user side as an Energy Management technology, and can achieve the purpose of minimum power consumption or minimum Energy consumption of a user by combining various household devices, distributed Energy, real-time electricity price information and the like, and provide an optimization control strategy of various loads and distributed Energy which meets the user requirements. The household energy management system can well provide support for demand response and access of renewable energy sources while improving household power utilization efficiency, can further optimize a system scheduling strategy by combining load change and environmental change, and even makes a significant contribution to power grid peak shaving. The household energy management system is composed of various loads and distributed power supplies and has the characteristic of diversity facing to user requirements, so that a great amount of uncontrollable factors and uncertainty are necessarily existed in the household energy management system. When a large number of user groups are faced, due to different attention points of each user, optimization targets are different, such as the minimum electricity consumption cost or the optimal electricity consumption comfort level, and the like, so that an optimal household energy management control strategy is required to be provided in a targeted manner when different user requirements are considered. Uncertain factors faced by the home energy management system mainly include uncertainty of renewable energy source prediction, uncertainty of Real Time Price (RTP) in an interaction process with electric power market information, and various unpredictable user random behaviors generated by the management system serving as a control core in the interaction process with users, which all bring challenges to formulation of a home energy management control strategy.
The existing household energy management system generally aims at optimizing the operation of electric equipment, and the influence of uncertain factors such as photovoltaic prediction error, real-time electricity price and user random behavior on an optimization result is rarely considered in the optimization process.
Disclosure of Invention
The purpose of the invention is as follows: the invention aims to provide an intelligent household energy optimization method based on information interaction, which is used for solving the problem that the influence of uncertain factors such as photovoltaic prediction errors, real-time electricity prices, user random behaviors and the like on an optimization result is not considered in the conventional household energy optimization scheduling.
The technical scheme is as follows: the invention provides an intelligent household energy optimization method based on information interaction, which comprises the following steps:
(1) information acquisition: collecting photovoltaic power generation information, electricity price information and the use state of household electrical equipment in a household energy management system at intervals of a certain time step length;
(2) optimizing admission prejudgment: judging whether energy optimization is carried out or not according to the acquired information, and setting the following four judgment conditions:
a. the electricity demand being less than the photovoltaic output, i.e.
Figure BDA0002381655850000023
Wherein, PDTo demand power, PVTo generate a photovoltaic power for the photovoltaic panel,
Figure BDA0002381655850000024
setting a threshold value;
b. the electricity price in the electricity market fluctuates too much, i.e.
Figure BDA0002381655850000025
Wherein, pitIs the real-time electricity price information at the time of t, pit+1Is the electricity price information at the latter moment,
Figure BDA0002381655850000026
a set electricity price fluctuation threshold value;
c. the overall operating power of the household appliances is within the permitted range of the household energy load, i.e.
Ptotal<Plim t
Wherein, PtotalTotal power of the household appliances, Plim tIs the maximum limit value of the household energy load;
d. occurrence of random behavior or unplanned operation of the user, i.e.
Figure BDA0002381655850000021
Wherein, tdFor the start-up time of the apparatus, YsetA set of planned equipment operating times;
if the conditions are not met, the original control strategy is used for controlling the equipment, namely, the optimization cycle is not entered, and the equipment keeps the original running state;
if any one of the above conditions is satisfied, performing optimization control according to the following steps:
(3) monitoring and determining a device state type set: determining a set of device state types for a current time period based on the collected device usage states, including: an unplanned device set, a running device set, a controllable device set, and a deferrable device set;
(4) and (3) performing optimized control output on the equipment, which comprises the following steps:
establishing a Markov decision process model based on electricity price, photovoltaic prediction deviation and controllable equipment;
solving a Markov decision process model to obtain an optimal control strategy under the premise of minimum electric quantity cost of a user;
and converting the running equipment set, the controllable equipment set and the delay equipment set into equipment control information, and transmitting the equipment control information to corresponding equipment for regulation and control.
Further, the Markov decision process model is constructed by the following method:
(21) constructing a discrete set C of user's overall electricity coststotal，
Ctotal＝{C1,C2,C3,...Cn} (1)
Figure BDA0002381655850000022
Figure BDA0002381655850000031
Wherein, CnFor the charge at time n, cnIs the real-time electricity price information, deltac is the deviation of the electricity price information,
Figure BDA0002381655850000032
is the total power consumption of the device,
Figure BDA0002381655850000033
power to the controllable device; pPVIs photovoltaic power generation power; n is controllableNumber of devices, j being the controllable device count number, Δ PpvFor the photovoltaic prediction deviation, Δ t is the time step;
(22) variable C for electricity consumptionnRegarding as a form, forming a mapping of state values and electric quantity costs; the future power cost state is only related to the current power cost, and is independent of the previous power cost, and the sequence of power costs can be regarded as a markov chain:
Figure BDA0002381655850000034
wherein the content of the first and second substances,
Figure BDA0002381655850000035
is a system slave state S1Transition to State S2A state transition matrix of time;
(23) defining a state space S, an action space A, a transition probability matrix T and a reward function R:
state space S: under an optimization objective of minimizing power consumption and electricity consumption, a range of power consumption costs characterize the state space, and a probability density function is employed to partition the state boundaries:
Sk＝[Bk,Bk+1)＝{Cn|Bk≤Cn＜Bk+1} (5)
wherein, B represents a state boundary, and k represents a state serial number;
action set ASET: the action set is characterized by a series of action variables, and different variables represent the switch states of different equipment; the combination of devices in different operating states constitutes a set of controlled devices, the set of actions at time t being a subset of the set of controllable devices, thereby giving a set of actions:
ASET＝{Am,k＝1,2,3...mmax} (6)
Figure BDA0002381655850000036
wherein, the number ncCharacterizing different controllable devices, each action characterizing a switch combination of the different devices;
Figure BDA0002381655850000037
is a subset of state actions, m is a subset number; q is the maximum number of control devices in the subset;
transition probability matrix T: the probability of change between states is characterized; the state transition probability is approximated by using a Monte Carlo simulation method, and the user power consumption data of one quarter is used for simulation, so that the potential relation between the user behavior change and the power consumption cost can be effectively represented; when the action variable is considered, the probability representation is in the current state, and the probability of transferring the given action to the next state is adopted;
Figure BDA0002381655850000038
in the formula (I), the compound is shown in the specification,
Figure BDA0002381655850000039
in an action AmTransition probability of a lower transition from state i to state j,
Figure BDA00023816558500000310
is an action AmThe number of times when the lower level transits from state i to state j; at the same time, a definition of the state change under the action variable is given:
Figure BDA0002381655850000041
in the formula, C1∈Si，C2∈Sj，PkTo contain the operating power of the action equipment in the action set, deltakThe running state quantity of the equipment has influence on the whole energy consumption in the form of positive and negative values;
the reward function R: and adopting power, electricity price, energy consumption and equipment state time to characterize and establish the overall objective of the function:
R＝F(Pd,Md,Ct,Wt) (10)
wherein, Pd,Md,Ct,WtRespectively, the power, the running state, the electricity consumption set and the remaining waiting time of the device d;
the overall reward is characterized by a weighting of cost and satisfaction: the first part of the spending change is determined by the change of the state, the second part of the whole comfort degree is determined by the state of the equipment and the running time of the equipment given by the action, namely the return function is as follows:
Figure BDA0002381655850000042
wherein, alpha and beta are proportionality coefficients, MdIn order to be in the operating state of the device,
Figure BDA0002381655850000043
is the return value, W, from state i to state j in action a at time tsetFor a preset device latency, WtLatency remains for the device.
Further, the Markov decision process model can be solved by combining a value iteration method and a q learning method, and comprises the following steps:
a control policy is a mapping that characterizes a set of states to a set of actions:
An＝πn(Sn) (12)
the value function is the mapping from the state set to the whole return, so that the relation between the value function of the current state and the value function of the next state can be well represented by adopting a Bellman equation; equations (13), (14) give the bellman expectation formula and the bellman expectation formula considering the state action pair, respectively:
Figure BDA0002381655850000044
Figure BDA0002381655850000045
wherein S, a correspond to the state S and the action a, respectively;
by the principle of optimality:
v*＝max vπ(s) (15)
q*＝max qπ(s,a) (16)
giving the optimal bellman expectation equation:
Figure BDA0002381655850000051
Figure BDA0002381655850000052
based on the Bellman equation and the Bellman optimality equation, the Markov decision process model can be solved by adopting dynamic programming;
respectively adopting a value iteration formula (19)) and an iteration method (formula (20)) of q-learning to solve so as to obtain a final optimization strategy;
Figure BDA0002381655850000053
qk+1＝Es'[r+γmax qk(s',a')|s,a] (20)
the final optimization strategy is given by the optimal equations of equation (21) and equation (22), respectively:
Figure BDA0002381655850000054
Figure BDA0002381655850000055
furthermore, the planned external device set represents a device set which is started and stopped outside the allowable running time of the device due to the random behavior of the user, and is not controlled under the conventional condition, and the control is performed after the preset standard is exceeded or the permission of the user is obtained. Wherein, exceeding the preset standard means that the total load rate of the household equipment is less than the threshold value. The load rate may be expressed as a ratio of operating equipment power to total power.
Further, the running device set has a maximum waiting time of zero within the device allowed running time period, and is converted into device control information by the following formula:
Figure BDA0002381655850000056
the controllable device set is a connected device, and in the allowed operation time period of the device, there is a waiting time set, and the device state can be controlled and converted by the control information according to the following formula:
Figure BDA0002381655850000057
the set of deferrable devices are connected devices within a device allowed run time period and the run time is less than a maximum latency; the set of deferrable devices are priority control devices whose state transitions are determined based on their run time and maximum user latency as follows:
Figure BDA0002381655850000061
in the formulae (23), (24) and (25),
Figure BDA0002381655850000063
in the state of being connected to the network,
Figure BDA0002381655850000062
to allow latency, δ is the operating state of the device, WmaxIs the maximum allowable latency.
Has the advantages that: compared with the prior art, the method has the following advantages:
1. the method has good effect on energy optimization under the condition of uncertainty factors including real-time electricity price information and user random behaviors;
2. an optimization scheme is provided on the premise of meeting the daily power consumption requirements of users;
3. the high-efficiency distribution of energy is realized, and the electricity consumption cost of a user is greatly reduced.
Drawings
FIG. 1 is a diagram of a home energy management system architecture;
FIG. 2 is a flow chart of a method for intelligent optimization of household energy according to the present invention;
FIG. 3 is a diagram illustrating real-time electricity rate information for simulation analysis according to an embodiment of the present invention;
FIG. 4 is a photovoltaic power and baseline load curve of a simulation analysis of an embodiment of the present invention;
FIG. 5 is a functional value of a return function of a simulation analysis according to an embodiment of the present invention;
FIG. 6 is an optimal strategy curve for simulation analysis according to an embodiment of the present invention;
FIG. 7(a) is a diagram of a device scheduling policy before simulation analysis optimization according to an embodiment of the present invention;
fig. 7(b) is a diagram illustrating the device scheduling policy after simulation analysis and optimization according to the embodiment of the present invention.
Detailed Description
The invention is further described below with reference to the following figures and examples:
1 Home energy management System framework
1.1 System Whole framework
The intelligent optimization method for the household energy is established on the basis of a household energy management system, and as shown in figure 1, the household energy management system mainly comprises three parts: 1. the information sensing module can be realized by various sensors and is used for monitoring the running states, the electric energy conversion conditions and the real-time power price information of various electric equipment, photovoltaic systems and household appliances in real time; the household appliances may include daily household appliances such as a refrigerator, a water heater, an air conditioner, a dish washer, a washing machine, a clothes dryer, an electric rice cooker, a hot water kettle, and the like, and an electric vehicle. 2. The information storage and transmission module is used for storing sensing information and transmitting necessary information to the management center according to application requirements, and the information is transmitted to a sensor network consisting of wireless sensors so as to realize interconnection of equipment; 3. the system management center can enable the mobile phone end app and the PC end to be used for receiving and fusing information, making an intelligent optimization strategy and diagnosing and processing faults. The three aspects are mutually linked, namely, the management center carries out further fusion processing on the obtained perception information, an intelligent scheduling strategy is formulated according to the requirements of users, and alarm information is sent out if equipment faults are found in the process. The user is used as the main body of the system facing to the object, and can operate the management center according to the self requirement, and the main body is an autonomous control part. Meanwhile, the management center can control the equipment under the authority preset by the user, and the intelligent automatic control part is provided. The two control modes are combined, and intelligent power utilization is facilitated. The interaction of information flow and energy flow exists in the running process of the household energy management system. The flow of energy is realized by controlling the running states of daily electric equipment, a photovoltaic system and an electric automobile through a management center so as to interact with each other and a superior power grid. The information flow is that the equipment information is sensed in real time through the Internet of things technologies such as an intelligent electric meter and wireless sensing equipment and is transmitted to a management center, and meanwhile, the management center can obtain real-time electricity price information in time through communication equipment. Therefore, the household energy management system is a household internet of things system on a user side, various devices are basic units, wherein electric devices, a photovoltaic system and an electric automobile are also monitored objects, communication devices are bridges connecting the monitored objects and a management center, the management center is a core part, bears important responsibilities of intelligent optimized operation of the whole system, and is also an important link for interactive fusion of energy flow and information flow, and finally a specific application value is formed.
1.2 electric device model
(1) According to the characteristics of the electric equipment, multiple household equipment models are established and stored in a management control center of the system, and when the corresponding equipment needs to be controlled in the optimization process, the corresponding electric equipment models are optimized.
The domestic photovoltaic system is the photovoltaic board in this embodiment, and photovoltaic power is decided by illumination intensity and photovoltaic board area, and accessible following formula calculates and obtains:
Pv＝rηvAB (1)
wherein, PvIs the photovoltaic power generation power, r is the illumination intensity, etavFor conversion efficiency, a is the effective illumination area and B is the number of photovoltaic panels.
According to the characteristics of whether the electric equipment can participate in scheduling, the electric equipment is divided into rigid load, transferable load and interruptible load. As shown in fig. 1, the rigid load mainly includes a television, a computer, an electric lamp, and the like, and the use condition objectively reflects the use habit and the life style of the user. Since the start-stop of these devices directly affects the user's home experience, they are considered baseline load and are not scheduled without the user's permission. These devices can be directly activated when desired by the user by responding to settings; the model is as follows:
Figure BDA0002381655850000081
wherein, the first and the second end of the pipe are connected with each other,
Figure BDA0002381655850000082
for the power of device d at time t, σtThe switching value of the device at time t is 0, which is off, and 1, which is on. t is tstart、tendThe desired on and off times are set by the user. t is ttotalIs the nominal duration of the device d and λ is the maximum delay time of the device.
The flexible load refers to controllable equipment, and is characterized in that the running time of the equipment has high adjustability according to the characteristics of the equipment, and the influence on the use experience of a home user is small when the equipment is controlled within the allowed running time range. The flexible load is divided into the following two types: interruptible devices and transferable devices.
Interruptible devices include home humidifiers, clothes dryers, water pumps, etc., which can be freely turned off during their operation without significant impact on user life, modeled as follows:
Figure BDA0002381655850000083
the transferable equipment mainly comprises equipment such as a washing machine, an electric cooker, a dish washing machine and the like, the operation period of the transferable equipment cannot be interrupted, but the starting and the operation time of the transferable equipment can be flexibly set according to the requirements of users, and the model is as follows:
Figure BDA0002381655850000084
2 optimization model based on Markov decision process
2.1 establishment of Markov decision Process model
Although the real-time information perception can accurately acquire the use condition of the household equipment, certain deviation may be brought to the preparation of the optimal strategy due to the occurrence of various random conditions (such as electricity price change, photovoltaic power generation change, user random behaviors and the like), so that a model of a Markov decision process is established, and for most economic users, the main driving factor is the electricity consumption. The electricity cost of the user is expressed as follows:
Figure BDA0002381655850000091
Figure BDA0002381655850000092
wherein, CnFor the electricity cost at time n, cnFor real-time electricity price information, deltac is the deviation of electricity price information,
Figure BDA0002381655850000093
is the total power consumption of the device,
Figure BDA0002381655850000094
power to the controllable device; pPVIs photovoltaic power generation power; n is the number of controllable devices, j is the number of controllable device counts, Δ PpvFor the photovoltaic prediction deviation, Δ t is the time step; the behavior habit of the user is covered in PmustInternal; in formula (7)
Figure BDA0002381655850000095
Can intuitively reflect the energy consumption condition in the family, and simultaneously, the formula (6) and the specific electricity price cost CnAre directly linked. Based on the above overall electricity cost, the following markov decision process model is given. When the overall charge cost is viewed as a collection of discrete costs, i.e.
Ctotal＝{C1,C2,C3,...Cn} (8)
This results in a sequence of cost variables, a range of which can be considered as a state, forming a mapping of state values and energy costs. The future power cost state is only related to the current power cost, and is independent of the previous power cost, and the sequence of the cost at this time can be regarded as a markov chain:
Figure BDA0002381655850000096
wherein the content of the first and second substances,
Figure BDA0002381655850000097
is a system slave state S1Transition to State S2The state transition matrix of time.
Therefore, the method is further established as a Markov decision process model and mainly comprises the definitions of a state space (S), an action space (A), a transition probability matrix (T) and a reward function (R). The core of the Markov decision process problem is to provide an optimal strategy under the condition of maximizing a reward function, and the specific purpose of the Markov decision process problem is to provide an optimal equipment control strategy on the premise of minimum electric quantity cost of a user by mapping.
State space (S): under an optimization objective of minimizing power consumption costs and electricity consumption, a range of power consumption costs characterize the state space, and a probability density function is employed to partition the state boundaries.
Sk＝[Bk,Bk+1)＝{Cn|Bk≤Cn＜Bk+1} (10)
Action set (a): the action set is characterized by a series of action variables, with different variables representing the switch states of different devices. The operation state of the equipment is given by the information control stage, the combination of different operation states forms a control equipment set, and the action set at the time t is a subset of the controllable set, so that the action set is given.
A＝{Am,k＝1,2,3...mmax} (11)
Figure BDA0002381655850000098
The numerical serial numbers characterize different controllable devices, and each action characterizes a switch combination of a different device.
Transition probability matrix (T): the probability of change between states is characterized. The invention approximates the state transition probability by a Monte Carlo simulation method. The simulation is carried out by using the user electricity consumption data in a quarter, so that the potential relation between the user behavior change and the electricity consumption can be effectively represented. When the action variables are considered, the probability characterization adopts the probability of the given action transferring to the next state under the current state.
Figure BDA0002381655850000101
In the formula (I), the compound is shown in the specification,
Figure BDA0002381655850000102
in an action AmTransition probability of a lower transition from state i to state j,
Figure BDA0002381655850000103
is an action AmThe number of times when the state is shifted from state i to state j. At the same time, a definition of the state change under the action variable is given:
Figure BDA0002381655850000104
in the formula, C1∈Si，C2∈Sj，PkTo contain the operating power of the action equipment in the action set, deltakThe operating state variable of the device has an influence on the overall energy consumption in the form of positive and negative values.
Reward function (R): the definition of the return function is an important part in the Markov decision process, and directly influences the selection of the whole optimal strategy set. The return function provided by the invention adopts power, electricity price and energy consumption and equipment state time to represent the whole target:
R＝F(Pd,Md,Ct,Wt) (15)
wherein, Pd,Md,Ct,WtRespectively, power, operating state, power cost set, and remaining latency of device d.
The overall reward is characterized by a weighting of cost and satisfaction. The first part cost change is determined by the change of the state, and the second part overall comfort degree is determined by the state of the equipment given by the action and the running time of the equipment.
Figure BDA0002381655850000105
Wherein, alpha and beta are proportionality coefficients, MdIn order to be in the operating state of the device,
Figure BDA0002381655850000106
the value is reported back when the state i transits from the action a to the state j in the time t. The return function is divided into three parts: a) S1/S2 represents the state proportion of the energy consumption variation trend; b) Md-Pd characterizes the degree of power reduction of the user equipment; c) W/WsetAdequate delay control, which characterizes the device latency limitation, is expressed in terms of time delay ratios.
2.2 model solution
The Markov decision process model can be solved by a value iteration and q learning method. The aim of solving in the invention is to give an optimal control strategy under the condition of maximizing the reward function, namely, giving the optimal control strategy of the equipment under the condition of meeting the condition of least electricity consumption cost of a user. The concrete steps of model solution are as follows:
a control policy is a mapping that characterizes a set of states to a set of actions.
An＝πn(Sn) (17)
The value function is the mapping from the state set to the overall return, so that the relation between the value function of the current state and the value function of the next state can be well represented by adopting the Bellman equation. Equations (18), (19) give the bellman expectation formula and the bellman expectation formula taking into account the state action pairs, respectively:
Figure BDA0002381655850000111
Figure BDA0002381655850000112
by the principle of optimality:
v*＝max vπ(s) (20)
q*＝max qπ(s,a) (21)
giving the optimal bellman expectation equation:
Figure BDA0002381655850000113
Figure BDA0002381655850000114
where S, a correspond to the state S and the action a, respectively.
Based on the Bellman equation and the Bellman optimality equation, the Markov decision process model can be solved by adopting dynamic programming. In the invention, a value iteration formula (24)) and a q-learning iteration method (formula (25)) are respectively adopted to solve, so that a final optimization strategy is obtained.
Figure BDA0002381655850000115
qk+1＝Es'[r+γmax qk(s',a')|s,a] (25)
Final optimization strategy pi*The optimal equations of the equations (26) and (27) are given:
Figure BDA0002381655850000116
Figure BDA0002381655850000117
2.3 household energy management system optimization method
The intelligent optimization method for the household energy comprises four stages of information acquisition, optimized admission prejudgment, equipment state type set detection and determination and equipment optimized control output. The traditional household energy management system generally focuses on optimization of electric equipment operation, and the invention focuses more on the influence of uncertain factors such as photovoltaic prediction error, real-time electricity price and user random behavior on an optimization result while optimizing the electric equipment operation. In the three stages of the optimization process, judgment is carried out according to the real-time electricity price information and the photovoltaic state information at each time step, and corresponding calculation results are given. The method specifically comprises the following steps:
(1) information acquisition: collecting photovoltaic power generation information, electricity price information and the use state of household electrical equipment in a household energy management system at intervals of a certain time step length; in the embodiment, the transmission time interval of the sensor and the real-time electricity price information interval are 5min, so that a time step of 5min is given.
(2) Optimizing admission prejudgment: in order to improve the effective operation times and efficiency of the system, the control mechanism of the equipment is judged before entering the control center, and the equipment is allowed to enter when the judgment condition is met, otherwise, the equipment is operated according to the normal condition. Therefore, the operation efficiency of the household energy management system is effectively improved by utilizing the event trigger mechanism. Judging whether energy optimization is carried out or not according to the acquired information, and setting the following four judgment conditions:
a. when the required energy consumption is less than the photovoltaic output, the surplus electricity is fed back to the power grid to earn profits.
Figure BDA0002381655850000121
Wherein, PDTo demand photovoltaic power, PVTo generate a photovoltaic power for the photovoltaic panel,
Figure BDA0002381655850000122
to set a threshold.
It should be noted that, in order to improve the renewable energy utilization rate and the degree of consumption, the present invention adopts a strategy of preferentially utilizing photovoltaic output to cope with load power consumption. When the photovoltaic is insufficient, electricity is purchased from the power grid to ensure normal operation of the load.
b. And according to the power price constraint, when the power price fluctuation of the power market is overlarge, giving a scheduling strategy again.
Figure BDA0002381655850000123
Wherein, pitIs the real-time electricity price information at the time of t, pit+1Is the electricity price information at the later time,
Figure BDA0002381655850000124
to set the threshold.
c. The overall operating power of the device needs to be within the allowable range of the home load.
Ptotal＜Plimt (30)
Wherein, PtotalFor the power of the whole household appliance, PlimtIs the home energy limit.
d. When the user acts randomly or operates outside the plan, the equipment needs to be rescheduled.
Figure BDA0002381655850000125
Wherein, tdFor the start-up time of the apparatus, YsetA set of run times for the in-plan equipment.
If the conditions are not met, the original control strategy is used for controlling the equipment, namely, the optimization cycle is not entered, and the equipment keeps the original running state;
if any one of the above conditions is satisfied, performing optimization control according to the following steps:
(3) detecting and determining a device state type set: determining a set of device state types for a current time period based on the collected device usage states, including: an unplanned equipment set, a running equipment set, a controllable equipment set and a deferrable equipment set;
(4) and performing optimized control output on the equipment, wherein the optimized control output comprises the following steps:
solving the established Markov decision process model to obtain an optimal control strategy on the premise of minimum electric quantity cost of the user;
and converting the running equipment set, the controllable equipment set and the delay equipment set into equipment control information, and transmitting the equipment control information to corresponding equipment for regulation and control.
Because the state judgment of various devices exists, the following four device state sets are defined, namely a planned device set, an operating device set, a controllable device set and a delay device set. The unplanned equipment set represents equipment starting and stopping outside the allowable running time of the equipment, is random power utilization behavior of a user, is not controlled under the conventional condition, and is controlled after exceeding a preset standard or obtaining user permission. Wherein, exceeding the preset standard means that the total load rate of the household equipment is less than the threshold value. The load rate may be expressed as a ratio of operating equipment power to total power. The method of the invention adjusts the equipment scheduling strategy in the optimization control stage within each time step range, thereby well combining the real-time electricity price to carry out the energy utilization distribution of the equipment.
Running a device set: the maximum waiting time is zero during the allowed operation period of the device. And giving a control strategy of the current stage to the running equipment set, and forcibly converting the running state of the connected equipment.
Figure BDA0002381655850000131
Controllable device set: the device is connected, and in the allowed operation period of the device, the waiting time is set, and the state of the device can be controlled by the control information and can be converted.
Figure BDA0002381655850000132
Set of deferrable devices: within the device allowed run time period, the device is connected and the run time is less than the maximum wait time. Such a set of devices is a priority control device. The transition of the state is determined according to the running time and the maximum user waiting time.
Figure BDA0002381655850000133
Wherein the content of the first and second substances,
Figure BDA0002381655850000134
in the state of being connected to the network,
Figure BDA0002381655850000135
to allow for latency, δ is the operating state of the device, WmaxIs the maximum allowable latency. The overall optimization scheduling control flow is shown in figure 2.
3 simulation results analysis
3.1 example parameter settings
In order to verify the effectiveness of the algorithm, the invention designs a comparison case for verification. The device simulation information is as follows, table 1, fig. 3 is real-time electricity rate information, and fig. 4 is power situation of photovoltaic and baseline load.
TABLE 1 operating conditions of the apparatus
Figure BDA0002381655850000141
3.2 simulation results analysis
The scheduling policy and scheduling result of the home energy management system are analyzed according to fig. 5 to 7:
fig. 5 shows the overall reported value, and it can be seen that the overall state transition tends to transition to the less expensive state, which meets the target feature trend.
Fig. 6 shows the overall scheduling strategy, and it can be seen from the figure that the control mechanism is frequently triggered to adjust the scheduling strategy during the load peak period in the evening (18:00-21: 00). In the daytime, particularly in the midday time (11:00-13:00), when the photovoltaic electric quantity is sufficient, the scheduling strategy is adjusted less because the mode of preferentially consuming the photovoltaic output to deal with the load power utilization is adopted, so that the triggering times of the control mechanism are reduced, and the efficiency of optimizing the scheduling is improved. The electric automobile is charged when no load needs to be controlled and the price of electricity is low at night (22:00-4: 00).
Figure 7 shows a comparison of the optimization strategies before and after scheduling. As can be seen from the figure, in the morning (6:00-8:00), due to the fact that the electricity price floats upwards, the starting time of the electric kettle and the humidifier is pushed backwards after the dispatching strategy is adopted, and electricity utilization cost is reduced. In the noon (11:00-13:00), although the rigid load energy consumption is increased, the photovoltaic power is sufficient, so that the load power consumption is not too high, and normal power consumption can be ensured even when random behaviors of users occur. In the evening period (18:00-21:00), the electricity price fluctuation is large due to the electricity utilization peak, the load operation period is properly shifted backwards after the dispatching strategy is adopted, and the electricity utilization cost is reduced to the maximum extent under the condition of meeting the electricity utilization requirement. At night (22:00-4:00), the electric automobile is used as a controllable load, no trip plan is provided, and the running time of the electric automobile is moved to midnight after a scheduling strategy is adopted, so that the electricity utilization cost is further reduced. On the whole, the optimization scheme can provide a reasonable control scheme while meeting the power consumption requirement.
In order to reflect the effectiveness of the scheduling strategy more intuitively, 5 family energy management scene results are respectively given in table 2, and are respectively the electricity consumption of residents not accessing the photovoltaic panel, the optimized scheduling electricity consumption of residents not accessing the photovoltaic panel, the electricity consumption of residents accessing the photovoltaic equipment, the electricity consumption of residents enjoying photovoltaic internet surfing subsidies and the optimized scheduling electricity consumption of residents accessing the photovoltaic equipment. As can be seen from the table, the user cost is high when the scheduling is not carried out, and the user cost can be effectively reduced after the optimization. After the household photovoltaic panel is adopted, the optimization strategy can remarkably reduce the user cost while improving the photovoltaic consumption, and the optimization rate of the system is improved.
TABLE 2 comparison of optimization results for each scene
Figure BDA0002381655850000151
In order to verify the effectiveness of the different solving algorithms provided by the present invention, the optimal strategy results obtained by two iterative methods, namely value iteration and q-learning, are provided in table 3. As can be seen from table 3, the optimization results under the two iterative methods have a small difference, but each of the two iterative methods has the following characteristics: the value iteration method can provide a stable control scheme which is biased to fit the power utilization habits of the user, and the requirements of experience-type users are better met. The q-learning method can provide more variable strategy sets and more exploration action spaces, thereby greatly reducing the energy cost of users and better meeting the requirements of economical users.
TABLE 3 comparison of optimization results for different optimal strategies
Figure BDA0002381655850000152
In order to verify whether the decision result of the system can still meet the user requirement when various uncertain factors exist, optimization effect comparison under the conditions of prediction error and random user behavior is given, as shown in table 4. Photovoltaic prediction error (PV), real-time electricity price prediction error (RTP), user Random Behavior (RBE) and the influence (ALL) caused by comprehensively considering the uncertain factors are respectively analyzed in the table. The prediction data in the invention are all given by the traditional prediction method, and the errors are all within an acceptable range. The final result shows that, aiming at the influence of the photovoltaic prediction error, in the daytime, under the normal electricity utilization habit of the user, the electricity utilization load is less, the output of the photovoltaic system is concentrated, so that more photovoltaic output can be fed back to the power grid in the daytime, and under the condition that the electricity price of the photovoltaic internet is not changed, the influence of the photovoltaic prediction error on the whole scheduling strategy result is less. Real-time electricity price prediction errors and random behaviors of users can greatly influence the overall electricity consumption cost, and the scheduling strategy provided by the invention can effectively reduce the influence caused by the uncertain factor change. For a single household user, although the overall error before and after optimization is not very large, the electricity utilization benefit brought by the optimization for a long time is considerable, and the user can obtain more comfortable electricity utilization experience.
TABLE 4 comparison of optimized results in the presence of multiple uncertain factors
Figure BDA0002381655850000161

Claims (5)

1. An intelligent household energy optimization method based on information interaction is characterized by comprising the following steps:
(1) information acquisition: acquiring photovoltaic power generation information, electricity price information and the use state of household electrical equipment in a household energy management system at intervals of a certain time step length;
(2) optimizing admission prejudgment: judging whether energy optimization is carried out or not according to the acquired information, and setting the following four judgment conditions:
a. the electricity demand being less than the photovoltaic output, i.e.
Figure FDA0003651320130000011
Wherein, PDTo demand power, PVFor the purpose of generating photovoltaic power for the photovoltaic panel,
Figure FDA0003651320130000012
setting a threshold value;
b. the electricity price in the electricity market fluctuates too much, i.e.
Figure FDA0003651320130000013
Wherein, pitIs the real-time electricity price information at the time of t, pit+1Is the electricity price information at the later time,
Figure FDA0003651320130000014
a set electricity price fluctuation threshold value;
c. the overall operating power of the household appliance is within the permitted range of the household energy load, i.e.
Ptotal＜Plimt，
Wherein, PtotalTotal power of the household appliances, PlimtMaximum limit value for household energy load;
d. occurrence of random behavior or unplanned operation of the user, i.e.
Figure FDA0003651320130000015
Wherein, tdFor the start-up time of the apparatus, YsetA set of planned equipment operating times;
if the conditions are not met, the original control strategy is used for controlling the equipment, namely, the optimization cycle is not entered, and the equipment keeps the original running state;
if any one of the above conditions is satisfied, performing optimization control according to the following steps:
(3) monitoring and determining a device state type set: determining a current time period device state type set based on the collected device usage states, including: an unplanned device set, a running device set, a controllable device set, and a deferrable device set;
(4) and (3) performing optimized control output on the equipment, which comprises the following steps:
establishing a Markov decision process model based on the electricity price, the photovoltaic prediction deviation and the controllable equipment;
solving a Markov decision process model to obtain an optimal control strategy under the premise of minimum electric quantity cost of a user;
and converting the running equipment set, the controllable equipment set and the delay equipment set into equipment control information, and transmitting the equipment control information to corresponding equipment for regulation and control.
2. The method of claim 1, wherein the Markov decision process model is constructed by:
(21) constructing a discrete set C of user's overall electricity coststotal，
Ctotal＝{C1，C2，C3，...Cn} (1)
Figure FDA0003651320130000021
Figure FDA0003651320130000022
Wherein, CnFor the electricity cost at time n, cnIs the real-time electricity price information, deltac is the deviation of the electricity price information,
Figure FDA0003651320130000023
is the total power consumption of the device,
Figure FDA0003651320130000024
power to the controllable device; pPVIs photovoltaic power generation power; n is the number of controllable devices, j is the number of controllable device counts, Δ PpvFor the photovoltaic prediction deviation, Δ t is the time step;
(22) variable C for electricity consumptionnRegarding as a form, forming a mapping of state values and electric quantity costs; the future power cost state is only related to the current power cost, and is independent of the previous power cost, and the sequence of power costs can be regarded as a markov chain:
Figure FDA0003651320130000025
wherein, the first and the second end of the pipe are connected with each other,
Figure FDA0003651320130000026
is a system slave state S1Transition to State S2A state transition matrix of time;
(23) defining a state space S, an action space A, a transition probability matrix T and a reward function R:
state space S: under an optimization objective of minimizing power consumption and electricity consumption, a range of power consumption costs characterize the state space, and a probability density function is employed to partition the state boundaries:
Sk＝(Bk，Bk+1)＝{Cn|Bk≤Cn＜Bk+1} (5)
wherein, B represents a state boundary, and k represents a state serial number;
action set ASET: the action set is characterized by a series of action variables, and different variables represent the switch states of different equipment; the combination of devices in different operating states constitutes a set of controlled devices, the set of actions at time t being a subset of the set of controllable devices, thereby giving a set of actions:
ASET＝{Am,k＝1,2,3...mmax} (6)
Figure FDA0003651320130000027
wherein, the number ncCharacterizing different controllable devices, each action characterizing a switch combination of the different devices;
Figure FDA0003651320130000028
is a subset of state actions, m is a subset number; q is the maximum number of control devices in the subset;
transition probability matrix T: the probability of change between states is characterized; approximating the state transition probability by using a Monte Carlo simulation method, and simulating by using user power consumption data of one quarter to effectively represent the potential relation between user behavior change and power consumption cost; when the action variable is considered, the probability representation is in the current state, and the probability of the given action transferred to the next state is adopted;
Figure FDA0003651320130000029
in the formula (I), the compound is shown in the specification,
Figure FDA0003651320130000031
in an action AmTransition probability of a lower transition from state i to state j,
Figure FDA0003651320130000032
is an action AmThe number of times when the lower level transits from state i to state j; meanwhile, a definition of state change under action variables is given:
Figure FDA0003651320130000033
in the formula, C1∈Si，C2∈Sj，PkTo contain the operating power of the action equipment in the action set, deltakIs a deviceThe operating state quantity of (a) has an influence on the overall energy consumption in the form of positive and negative values;
the reward function R: and adopting power, electricity price, energy consumption and equipment state time to characterize and establish the overall objective of the function:
R＝F(Pd,Md,Ct,Wt) (10)
wherein, Pd，Md，Ct，WtRespectively, the power, the running state, the electricity consumption set and the remaining waiting time of the device d;
the overall reward is characterized by a weighting of cost and satisfaction: the first part of the spending change is determined by the change of the state, the second part of the whole comfort degree is determined by the state of the equipment and the running time of the equipment given by the action, namely the return function is as follows:
Figure FDA0003651320130000034
wherein, alpha and beta are proportionality coefficients, MdIn order to be in the operating state of the device,
Figure FDA0003651320130000035
is the return value when the state i transits from the action a to the state j in the time t, WsetFor a preset device latency, WtThe latency remains for the device.
3. The method of claim 2, wherein the Markov decision process model can be solved by a combination of value iteration and q learning methods, comprising the steps of:
a control policy is a mapping that characterizes a set of states to a set of actions:
An＝πn(Sn) (12)
the value function is the mapping from the state set to the overall return, and therefore a Bellman equation is adopted for representing the relationship between the value function of the current state and the value function of the next state; equations (13), (14) give the bellman expectation formula and the bellman expectation formula considering the state action pair, respectively:
Figure FDA0003651320130000036
Figure FDA0003651320130000037
wherein S, a correspond to the state S and the action a, respectively;
by the principle of optimality:
v*＝max vπ(s) (15)
q*＝max qπ(s,a) (16)
giving the optimal bellman expectation equation:
Figure FDA0003651320130000041
Figure FDA0003651320130000042
based on the Bellman equation and the Bellman optimality equation, the Markov decision process model can be solved by adopting dynamic programming;
solving by respectively adopting a value iteration formula (19) and an iteration method formula (20) of q-learning so as to obtain a final optimization strategy;
Figure FDA0003651320130000043
qk+1＝Es'[r+γmax qk(s',a')|s,a] (20)
the final optimization strategy is given by the optimal equations of equation (21) and equation (22), respectively:
Figure FDA0003651320130000044
Figure FDA0003651320130000045
4. the method of claim 3, wherein the planned out-of-plan device set characterizes a device set that is started and stopped outside a device's allowable run time due to random behavior of the user, and is not controlled conventionally, and is controlled when a preset criterion is exceeded or user permission is obtained.
5. The method of claim 3, wherein the set of operating devices has a maximum latency of zero within a device allowed operating time period, and wherein the maximum latency is converted to device control information by:
Figure FDA0003651320130000046
the controllable device set is a connected device, and in the allowed operation time period of the device, there is a waiting time set, and the device state can be controlled and converted by the control information according to the following formula:
Figure FDA0003651320130000051
the set of deferrable devices are connected devices within a device allowed run time period and run time is less than a maximum latency; the set of deferrable devices are priority control devices whose state transitions are determined based on their run time and maximum user latency as follows:
Figure FDA0003651320130000052
in the formulae (23), (24) and (25),
Figure FDA0003651320130000053
in the state of being connected to the network,
Figure FDA0003651320130000054
to allow latency, δ is the operating state of the device, WmaxIs the maximum allowable latency.
CN202010084781.2A 2020-02-10 2020-02-10 Intelligent household energy optimization method based on information interaction Active CN111340270B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010084781.2A CN111340270B (en) 2020-02-10 2020-02-10 Intelligent household energy optimization method based on information interaction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010084781.2A CN111340270B (en) 2020-02-10 2020-02-10 Intelligent household energy optimization method based on information interaction

Publications (2)

Publication Number Publication Date
CN111340270A CN111340270A (en) 2020-06-26
CN111340270B true CN111340270B (en) 2022-07-12

Family

ID=71182155

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010084781.2A Active CN111340270B (en) 2020-02-10 2020-02-10 Intelligent household energy optimization method based on information interaction

Country Status (1)

Country Link
CN (1) CN111340270B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113434995A (en) * 2020-10-13 2021-09-24 天津大学 Intelligent electricity utilization management method for commercial users considering uncertainty of travel of electric automobile
CN112734188B (en) * 2020-12-30 2023-11-24 杭州电子科技大学 Household medical care dispatching optimization method based on two-stage mixed element heuristic algorithm
DE102022203741A1 (en) 2022-04-13 2023-10-19 Robert Bosch Gesellschaft mit beschränkter Haftung Method for determining a target value of a charging power for charging a battery of a vehicle using a charging device
DE102022203739A1 (en) 2022-04-13 2023-10-19 Robert Bosch Gesellschaft mit beschränkter Haftung Method for determining a charging power for charging a battery of a vehicle using a charging device
CN115994627B (en) * 2023-03-22 2023-07-14 湖南大学 Residential building flexible load day-ahead optimal scheduling method, device, equipment and medium

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104778504A (en) * 2015-03-18 2015-07-15 南京邮电大学 Electricity utilization arrangement optimization method for intelligent household electrical appliances
CN105719030A (en) * 2016-03-29 2016-06-29 武汉大学 Method for electric vehicle load prediction based on efficiency maximization principle
FR3042043A1 (en) * 2015-08-04 2017-04-07 Bioaxial Suivi Par Gabriel Sirat METHOD AND DEVICE FOR OPTICAL MEASUREMENT
CN109599856A (en) * 2018-11-12 2019-04-09 国网天津市电力公司电力科学研究院 Electric car management of charging and discharging optimization method and device in a kind of more building of microgrid
CN109709910A (en) * 2018-11-30 2019-05-03 中国科学院广州能源研究所 A kind of home energy source Optimized Operation management system and method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104778504A (en) * 2015-03-18 2015-07-15 南京邮电大学 Electricity utilization arrangement optimization method for intelligent household electrical appliances
FR3042043A1 (en) * 2015-08-04 2017-04-07 Bioaxial Suivi Par Gabriel Sirat METHOD AND DEVICE FOR OPTICAL MEASUREMENT
CN105719030A (en) * 2016-03-29 2016-06-29 武汉大学 Method for electric vehicle load prediction based on efficiency maximization principle
CN109599856A (en) * 2018-11-12 2019-04-09 国网天津市电力公司电力科学研究院 Electric car management of charging and discharging optimization method and device in a kind of more building of microgrid
CN109709910A (en) * 2018-11-30 2019-05-03 中国科学院广州能源研究所 A kind of home energy source Optimized Operation management system and method

Also Published As

Publication number Publication date
CN111340270A (en) 2020-06-26

Similar Documents

Publication Publication Date Title
CN111340270B (en) Intelligent household energy optimization method based on information interaction
Sedhom et al. IoT-based optimal demand side management and control scheme for smart microgrid
US10911257B2 (en) Context-aware smart home energy manager
US9310792B2 (en) Scheduling and modeling the operation of controllable and non-controllable electronic devices
Ramchurn et al. Agent-based control for decentralised demand side management in the smart grid
Du et al. Appliance commitment for household load scheduling
EP1263108A1 (en) Community energy comsumption management
CN114418249B (en) Operation control method and device for light storage flexible system
CN110209135B (en) Family energy optimization scheduling method based on micro cogeneration multi-time scale
GB2448896A (en) Energy management system
CN111598478B (en) Comprehensive energy demand response quantity calculation method
GR20190100088A (en) Method for improving the energy management of a nearly zero energy building
CN111027747A (en) Household energy control method considering user comfort risk preference
CN113328432A (en) Family energy management optimization scheduling method and system
Mohsenzadeh et al. Applying fuzzy techniques to model customer comfort in a smart home control system
AU2023263448A1 (en) Controlling an energy exchange
CN111679573A (en) Household energy consumption optimization method and device for residents
CN106685313B (en) Power generation control method and device for photovoltaic power station and photovoltaic power station
Yang et al. A novel dynamic load-priority-based scheduling strategy for home energy management system
CN110059895A (en) Electric appliance demand response Potentials method based on user behavior
CN112688312B (en) Comprehensive household energy management method considering multiple uncertain factors
CN106411671A (en) Peak valley power utilization oriented intelligent household electrical appliance control method and gateway
CN111585272B (en) Family demand side response method based on community cluster and centralized energy storage
CN206370695U (en) A kind of photovoltaic plant
CN117353454A (en) Power grid user side energy consumption management system and control method thereof

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