CN116384155A - Home energy management optimization method considering user behavior uncertainty - Google Patents

Abstract

The invention relates to the technical field of intelligent energy conservation, and aims to provide a new household energy management model which can ensure the electricity economy and improve the comfort level of users. The household energy management optimization method is characterized in that uncertainty of user behavior is considered, the uncertainty of the user is quantified on the basis of a household energy management system model, the allowable start-stop time, the environmental temperature preference and the temporary hot water consumption of the user are represented by comfort level deviation coefficients, the uncertainty parameters of the user and the electricity cost are jointly used as objective functions to solve, meanwhile, the household energy management system model is operated in a rolling optimization mode in the day by adopting a model prediction control method, and an operation plan is adjusted in real time along with the operation state and prediction information of the system to realize household energy management optimization. The invention is mainly applied to intelligent energy-saving occasions.

Description

Home energy management optimization method considering user behavior uncertainty

technical field

The invention relates to the technical field of intelligent energy saving, in particular to a family energy management optimization method considering user behavior uncertainty.

Background technique

With the development of Internet technology and communication technology, the smart grid solves many difficulties in the current power grid, such as slow transmission of power information and insufficient system stability, by combining with the infrastructure of the power grid. Home intelligent micro-grids represented by home energy management systems are gradually developing to better meet people's requirements for power quality, comfort requirements, and environmental protection. The optimization of the home energy management system plays an important role in reducing user costs, satisfying user comfort, and promoting the on-site consumption of renewable energy, which is conducive to improving the smart power consumption level of home users. In the process of household smart electricity consumption, due to the large randomness of the user's behavior parameters, only the upper and lower limits of the probability distribution or the distribution interval can be known, and it is impossible to accurately solve it. If such uncertain factors are not considered, the estimated optimization scheme will deviate from the actual situation, and in severe cases, it will violate the actual constraints, and the optimal dispatching scheme cannot be obtained, which will affect the user's electricity economy and Comfort has a big impact.

Contents of the invention

In order to overcome the deficiencies of the existing technology and solve the impact of the user's subjective behavior on the operation of the equipment in the home energy management system, the present invention aims to propose a new model of home energy management to improve the comfort of the user while ensuring the economy of electricity consumption. For this reason, the technical solution adopted by the present invention is to consider the home energy management optimization method of user behavior uncertainty, quantify the user's uncertain behavior on the basis of the home energy management system model, and calculate the allowable start-stop time of the user's electrical equipment , environmental temperature preference, and temporary hot water consumption are represented by the comfort deviation coefficient, and the user uncertainty parameters and electricity cost are jointly used as the objective function to solve the problem. It operates in a rolling optimization mode within a day, and adjusts the operation plan in real time according to the system operating status and forecast information to achieve home energy management optimization.

Specific steps are as follows:

1) According to the power consumption characteristics of household appliances, household appliances are divided into distributed power generation equipment, energy storage equipment and household electricity load, and the operation model of household electrical equipment is established;

2) Analyze the user's subjective behavior, and model the user's behavior by expressing the allowable start and stop time of the user's electrical equipment, environmental temperature preference, and temporary hot water consumption with the comfort deviation coefficient;

3) Aiming at user behavior and household user cost, establish a household energy management model that considers user behavior uncertainty, select user net electricity cost and comfort = violation coefficient as the objective function, and solve the multi-objective optimization problem;

4) Using the model predictive control method, the rolling optimization solution is carried out on the established day-ahead household operation model, and the simulation optimization result is obtained.

The detailed steps are as follows:

1. Home energy management system model

In the home energy management system, home smart electrical equipment is the main control object. By classifying and modeling the electrical characteristics of different types of equipment, analyzing their impact on home electricity consumption, setting the dispatch cycle to 24h, the It is equal to N time periods, the duration of each time period is Δt, and each time period is recorded as i, then the operation model of various electrical equipment is as follows:

(1) Interruptible load: P _ICL,i is the operating power of the interruptible load in the i-th period; p _ICL,r is the rated power of the interruptible load, x _ICL,i is the operation of the interruptible load in the i-th period state, 0 is not running, 1 is running; [b _ICL , e _ICL ] is the allowable working period of the interruptible load, l _ICL is the total running time of the interruptible load, then the work constraint of the interruptible load is:

P _ICL,i = x _ICL,i p _ICL,r (1)

(2) Uninterruptible load: including the operational constraints of the interruptible load, and at the same time adding the supplementary constraints to ensure the uninterruptible characteristics are:

P _UICL,i is the operating power of the uninterruptible load in the i period; p _UICL,r is the rated power of the load. x _UICL,i is the running status of the load in the i-th period; [b _UICL ,e _UICL ] is the allowed running period, l _UICL is the total running time;

(3) Temperature-controlled loads: Temperature-controlled loads refer to those equipments that optimize the load operation by controlling the temperature. Let T _in,i and T _out,i represent the indoor and outdoor air temperatures in the i-th period respectively; C _ac is the indoor air temperature The thermal capacity of the air conditioner; p _ac,i is the operating power of the air conditioner in the i-th time period, and R is the thermal resistance of the house, then the operating model of the air conditioner and the upper and lower limits of the temperature constraints are:

T _in,min ≤T _in,i ≤T _in,max (6)

为注入到水箱中的冷水温度；V_cold,i为第i时段内需注入的冷水体积；V_total为水箱的总容量；C_st为千瓦时和焦耳的转换系数；p_st,i为第i时段电热水器的运行功率，则热水器的运行模型和温度上下限约束为：Let T _st,i represent the hot water temperature of the electric water heater in the i-th period;

is the temperature of cold water injected into the water tank; V _cold,i is the volume of cold water to be injected in the i-th period; V _total is the total capacity of the water tank; C _st is the conversion coefficient of kWh and Joule; p _st,i is the i-th period The operating power of the electric water heater, the operating model of the water heater and the upper and lower limit constraints of the temperature are:

T _st,min ≤T _st,i ≤T _st,max (8)

表示室外的温度(℃)，则光伏电池的功率特性为：(4) Distributed photovoltaic equipment: In the home energy management system, photovoltaic equipment converts light energy into electrical energy, and then converts it into power frequency current that can be used by household electrical equipment through converters and inverters, thereby providing For household electrical equipment or stored in energy storage equipment, let p _pv,i represent the photovoltaic output power, η _pv represent the photoelectric conversion efficiency of the photovoltaic system, S _pv represent the receiving area m ² of the photovoltaic panel, and I _pv,i represent Photovoltaic system solar radiation intensity (kW/m ² ),

Indicates the outdoor temperature (°C), then the power characteristics of the photovoltaic cell are:

分别代表家庭储能设备的充电和放电功率以及充电和放电的工作状态；η_ch为充电效率；η_dch为放电效率；/>

和/>

代表充电和放电的最大功率；ε代表自放电率，Q_r为蓄电池额定容量，则家庭储能设备的充电放电特性和运行约束如式(10)-式(15)所示：(5) Energy storage equipment: Let SOC be the state of charge of the energy storage equipment,

Represent the charging and discharging power of the household energy storage equipment and the working state of charging and discharging respectively; η _ch is the charging efficiency; η _dch is the discharging efficiency; />

and />

Represents the maximum power of charging and discharging; ε represents the self-discharge rate, Q _r is the rated capacity of the battery, then the charging and discharging characteristics and operating constraints of household energy storage equipment are shown in formula (10) - formula (15):

SOC _min ≤ SOC _i+1 ≤ SOC _max (11)

SOC ₉₆ ≤SOC _ini (15)

2. User Uncertain Behavior Analysis

(1) Uninterruptible Load Uncertainty Behavior

与/>

采用蒙特卡洛抽样的方法模拟N种不同场景，设b_UICL,j为场景j下的起始工作时段，e_UICL,j为场景j下的终止工作时段，x_UICL,i为不可中断负荷运行状态，p_UICL,i为不可中断负荷的用电功率，得到的不可中断负荷舒适度偏差C_UICL如下所示：In actual operation, due to user behavior, the allowable start time of the work will be delayed or the allowable cut-off time of the work will be advanced. Both of these situations will cause the original operation plan to not be completed normally. The change of time obeys the normal distribution, satisfying

with />

Use the Monte Carlo sampling method to simulate N different scenarios, let b _UICL,j be the initial working period under scenario j, e _UICL,j be the end working period under scenario j, and x _UICL,i be the uninterruptible load operation state, p _UICL,i is the electric power of the uninterruptible load, and the obtained uninterruptible load comfort deviation C _UICL is as follows:

(2) Uncertain behavior of interruptible load

Similar to non-interruptible loads, the running time of interruptible loads will also be affected by user behavior. The actual work start time b _ICL,j and the actual work deadline e _ICL,j of interruptible loads also satisfy the normal distribution. Monte Carlo simulation method, let b _ICL,j be the actual allowable running time of the interruptible load under scenario j, e _ICL,j be the actual cut-off running time of the interruptible load under scenario j, x _ICL,i be the interruptible load The operating state of p _ICL,i is the electric power of the interruptible load, and the obtained interruptible load comfort deviation C _ICL is as follows:

(3) Not sure about temperature preference

The user's preferred temperature occurrence time i _ac,b,j is uniformly distributed within a specified period of time, and the change time is 30 minutes, while the user's preferred temperature T _in,j,i obeys a normal distribution and remains constant within the occurrence time, using In the Montecat sampling method, let T _in,j,t be the user's preferred comfort temperature in the i-th period of scene j, and T _in,i be the indoor temperature in the i-th period, and the deviation of the user's temperature comfort degree obtained is:

(4) Uncertain hot water consumption

The user has temporary water consumption within [i _st,min ,i _st,max ], the start time i _st,b,j and the water consumption V _st,j follow a uniform distribution, and the water use time is 30 minutes, through the Montecat sampling method , Let Q _st,j,i be the heat consumed by the water heater in the i-th period under scene j, and T _ws,j,i be the hot water temperature in the i-th period under scene j. Q _i is the heat consumed by the water heater in the i period, and the user's hot water consumption comfort deviation C _st is as follows:

Q _st,j,i ＝c _water V _total (T _st,j,i+1 -T _st,j,i ) (20)

Q _i ＝c _water V _total (T _st,i+1 -T _st,i ) (22)

3. Home energy management optimization model considering user behavior uncertainty

Select the two objectives of the user’s net electricity cost and the comfort violation coefficient as the objective function of the optimization model, and use the weighted sum method to solve it. ω ₁ and ω ₂ are the weight coefficients of the user’s economy and comfort, ω ₁ +ω ₂ =1 ; By means of normalization, let C _cost be the daily electricity cost of the household energy management model considering user behavior, C _com be the sum of the user comfort violation indices in the model in one day, and C _cost,max and C _cost,min be iterations The maximum and minimum values of the electricity cost appearing in , C _com,max and C _{com,min are} the maximum and minimum values of the user comfort violation index appearing in the iteration,

p _grid,i represents the interactive power between the home energy management system and the grid in the i-th period, price _b,i represents the electricity purchase price of the user in the i-th period, price _s,i represents the electricity sales price of the user in the i-th period, and ρ is the penalty factor , ε _n is the occurrence probability of uncertain behavior n, and B is the set of uncertain behaviors. The objective function of the obtained home energy management system optimization model is:

In the household energy optimal dispatch model, in addition to the constraints of operating equipment, electric power balance constraints also need to be added. Let P _pv,i be the power generation power of distributed photovoltaic equipment in the i-th period, and P _δ,i be the power of the δ-type load in the i-th period. Electric power, including energy storage equipment and various electric loads. Then the electric power balance constraint is as follows

4. Real-time rolling optimization based on model predictive control

The model predictive control changes the running period t=1...N in the formula to t=i...N through the rolling optimization process. For example, at time t=i, the real operating data such as the water temperature of the water heater and the indoor temperature at time i are brought into the home energy management system together with the forecast data in the next optimization time domain T, and the optimal scheduling of the next optimization time domain T is obtained result. During the time period (i, i+1), use the optimized scheduling result to optimize the scheduling of the household electrical equipment, and repeat the above process until the end of the day when t=i+1.

Features and beneficial effects of the present invention are:

The invention adopts the method of multi-objective optimization, which can improve the user's comfort while ensuring lower electricity consumption cost, and reduce the impact of uncertain behavior on the user's electricity consumption. Compared with the general home energy management optimization based on user cost, the present invention can better meet the user's comfort requirements, and can self-adjust according to the user's preference for economy and comfort, and set different The preference factor is used to obtain the most suitable power consumption plan for the user.

Description of drawings:

Figure 1 Uninterruptible load uncertainty behavior.

Fig. 2 Interruptible load uncertainty behavior.

Figure 3. Time frame diagram of rolling optimization.

Figure 4 Electricity price curve for electricity sales.

Fig. 5 Photovoltaic power curve.

Figure 6 Outdoor temperature curve.

Figure 7 Hot water consumption curve.

Fig. 8 The running results under different ω ₁ .

Detailed ways

In order to solve the impact of the user's subjective behavior on the operation of the equipment in the home energy management system, the present invention introduces a random optimization method based on the home energy management system model, quantifies the user's uncertain behavior, and calculates the allowable start-stop time of the user's electrical equipment , environmental temperature preference, temporary hot water consumption, etc. are expressed by the comfort deviation coefficient, and the user uncertainty parameters and electricity cost are used as the objective function to solve the problem, which improves the comfort of the user while ensuring the economy of electricity consumption. At the same time, the model predictive control method is adopted to run the above model in a rolling optimization mode within a day, and adjust the operation plan in real time according to the system operation status and forecast information, which is more suitable for the actual operation optimization system.

1) According to the power consumption characteristics of household appliances, household appliances are divided into distributed power generation equipment, energy storage equipment and household electricity load, and an operation model of household electrical equipment is established.

2) Analyze the user's subjective behavior, and model the user's behavior by expressing the allowable start and stop time of the user's electrical equipment, environmental temperature preference, and temporary hot water consumption with the comfort deviation coefficient.

3) Aiming at user behavior and household user cost, establish a household energy management model that considers user behavior uncertainty, and select user net electricity cost and comfort level = violation coefficient as the objective function to solve the multi-objective optimization problem.

4) Using the model predictive control method, the rolling optimization solution is carried out on the established day-ahead household operation model, and the simulation optimization result is obtained.

The present invention will be described in detail below.

1. Home energy management system model

In the home energy management system, home smart electrical equipment is the main control object. By classifying and modeling the electrical characteristics of different types of equipment, it is more convenient to build an optimization model and analyze its impact on home electricity consumption. Set the scheduling cycle to 24h, divide it into N time periods, each time period is Δt, and record each time period as i, then the operation model of various electrical equipment is as follows:

(1) Interruptible load: For interruptible loads, the operating cycle is fixed, but the operating mode is more flexible, and can be scheduled arbitrarily within the operating period, and the suspension during operation will not cause a major impact on the equipment. Let P _ICL,i be the operating power of the interruptible load in the i-th period; p _ICL,r is the rated power of the interruptible load. x _ICL,i is the operating state of the interruptible load in the i-th period, 0 is not running, 1 is running; [b _ICL ,e _ICL ] is the allowable working period of the interruptible load, l _ICL is the total running duration. Then the interruptible load work constraint is:

P _ICL,i = x _ICL,i p _ICL,r (28)

(2) Uninterruptible loads: The operation model of uninterruptible loads is similar to that of interruptible loads, but due to the characteristics that they cannot be suspended and stopped during operation, in addition to the operational constraints of interruptible loads, it is necessary to increase the Supplementary constraints for the break feature. Let P _UICL,i be the operating power of the uninterruptible load in the i period; p _UICL,r is the rated power of the load. x _UICL,i is the running state of the load in the i-th period; [b _UICL ,e _UICL ] is the allowed running period, and l _UICL is the total running time. Then the supplementary constraints that guarantee its uninterruptible properties are:

(3) Temperature-controlled loads: Temperature-controlled loads refer to equipment that optimizes load operation by controlling temperature. In this paper, air conditioners and electric water heaters are selected as typical representatives. Let T _in,i and T _out,i represent the indoor and outdoor air temperature in the i-th period, respectively; C _ac is the heat capacity of the indoor air; p _ac,i is the operating power of the air conditioner in the i-th period. R is the thermal resistance of the house, then the operating model of the air conditioner and the upper and lower limits of the temperature constraints are:

T _in,min ≤T _in,i ≤T _in,max (33)

为注入到水箱中的冷水温度；V_cold,i为第i时段内需注入的冷水体积；V_total为水箱的总容量；C_st为千瓦时和焦耳的转换系数；p_st,i为第i时段电热水器的运行功率。则热水器的运行模型和温度上下限约束为：Similar to the operating model of the air conditioner, the operating state of the electric water heater is closely related to the temperature of the hot water. Let T _st,i represent the hot water temperature of the electric water heater in the i-th period;

is the temperature of cold water injected into the water tank; V _cold,i is the volume of cold water to be injected in the i-th period; V _total is the total capacity of the water tank; C _st is the conversion coefficient of kWh and Joule; p _st,i is the i-th period The operating power of the electric water heater. Then the operating model of the water heater and the upper and lower limit constraints of the temperature are:

T _st,min ≤T _st,i ≤T _st,max (35)

表示室外的温度(℃)，则光伏电池的功率特性为：(4) Distributed photovoltaic equipment: In the home energy management system, photovoltaic equipment converts light energy into electrical energy, and then converts it into power frequency current that can be used by household electrical equipment through converters and inverters, thereby providing For household electrical equipment or stored in energy storage equipment. Suppose p _pv,i represents the photovoltaic output power, η _pv represents the photoelectric conversion efficiency of the photovoltaic system, S _pv represents the receiving area m ² of the photovoltaic panel, and I _pv,i represents the solar radiation intensity of the photovoltaic system (kW/m ² ),

Indicates the outdoor temperature (°C), then the power characteristics of the photovoltaic cell are:

分别代表家庭储能设备的充电和放电功率以及充电和放电的工作状态；η_ch为充电效率；η_dch为放电效率；/>

和/>

代表充电和放电的最大功率；ε代表自放电率，Q_r为蓄电池额定容量，则家庭储能设备的充电放电特性和运行约束如式(10)-式(15)所示：(5) Energy storage equipment: Household energy storage devices represented by lead-acid batteries mainly use their charging and discharging functions to supply power to houses during high electricity price periods, and buy electricity from the grid during low electricity price periods, so as to reduce electricity costs. Purpose. Let SOC be the state of charge of the energy storage device,

Represent the charging and discharging power of the household energy storage equipment and the working state of charging and discharging respectively; η _ch is the charging efficiency; η _dch is the discharging efficiency; />

and />

Represents the maximum power of charging and discharging; ε represents the self-discharge rate, Q _r is the rated capacity of the battery, then the charging and discharging characteristics and operating constraints of household energy storage equipment are shown in formula (10) - formula (15):

SOC _min ≤ SOC _i+1 ≤ SOC _max (38)

SOC ₉₆ ≤SOC _ini (42)

2. User Uncertain Behavior Analysis

In the actual household electricity consumption process, there is a certain subjective uncertainty in the user's electricity consumption behavior, which leads to the fact that the actual operation plan does not conform to the optimal electricity consumption plan of the household energy management system, thereby reducing the robustness and reliability of the household energy management system. reliability. Therefore, it is necessary to quantitatively analyze the user's uncertain behavior, model the user's behavior, and reduce the impact on user comfort. The user's uncertain behavior is analyzed as follows:

(1) Uninterruptible Load Uncertainty Behavior

与/>

采用蒙特卡洛抽样的方法模拟N种不同场景，设b_UICL,j为场景j下的起始工作时段，e_UICL,j为场景j下的终止工作时段，x_UICL,i为不可中断负荷运行状态，p_UICL,i为不可中断负荷的用电功率，得到的不可中断负荷舒适度偏差C_UICL如下所示：In actual operation, due to user behavior, the start time of work permission may be delayed or the work permission deadline may be advanced, both of which will lead to the failure of the original operation plan to be completed normally. Figure 1 shows the uncertain behavior of uninterruptible loads. Therefore, it is assumed that the changes of the actual allowable working time and the actual cut-off working time of the uninterruptible load obey the normal distribution, satisfying

with />

Use the Monte Carlo sampling method to simulate N different scenarios, let b _UICL,j be the initial working period under scenario j, e _UICL,j be the end working period under scenario j, and x _UICL,i be the uninterruptible load operation state, p _UICL,i is the electric power of the uninterruptible load, and the obtained uninterruptible load comfort deviation C _UICL is as follows:

(2) Uncertain behavior of interruptible load

Similar to non-interruptible workloads, the running time of interruptible workloads is also affected by user behavior. Figure 2 shows the uncertain behavior of interruptible loads. Assuming that the actual work start time b _ICL,j of the interruptible load and the actual work cut-off time e _ICL,j also satisfy the normal distribution, using the method of Monte Carlo simulation, let b _ICL,j be the actual time of the interruptible load under scenario j Allowable running time, _eICL,j is the actual cut-off running time of the interruptible load in scenario j, xICL _,i is the running state of the interruptible load, pICL _,i is the power consumption of the interruptible load, and the obtained interruptible load The Comfort Deviation C _ICL looks like this:

(3) Not sure about temperature preference

The user's preference for temperature will also affect the user's comfort, which is mainly reflected in the change of the upper and lower limits of the user's preferred temperature. Assume that the user's preferred temperature occurrence time i _ac,b,j is uniformly distributed within a specified period of time, and the change time is 30 minutes, while the user's preferred temperature T _in,j,i obeys a normal distribution and remains constant within the occurrence time. Using the Montecat sampling method, let T _in,j,t be the user's preferred comfort temperature in the i-th period of scene j, and T _in,i be the indoor temperature in the i-th period, and the deviation of the user's temperature comfort degree obtained is:

(4) Uncertain hot water consumption

The operation of electric water heaters is accompanied by heat exchange and water consumption and replenishment. When the user uses a large amount of water outside the planned plan, a large amount of cold water will be suddenly injected into the water heater. However, due to the constant power of the electric water heater, the heat provided in a short period of time cannot quickly raise the temperature of the cold water to the lower limit of the temperature, resulting in a further drop in the temperature inside the water heater. Ultimately below the lower temperature limit, impacting user comfort. Therefore, assuming that the user has temporary water consumption within [i _st,min ,i _st,max ], the start time i _st,b,j and the water consumption V _st,j follow a uniform distribution, and the water use time is 30 minutes, through Montecut Sampling method, let Q _st,j,i be the heat consumed by the water heater in the i-th period under scene j, and T _ws,j,i be the hot water temperature in the i-th period under scene j. Q _i is the heat consumed by the water heater in the i period, and the user's hot water consumption comfort deviation C _st is as follows:

Q _st,j,i ＝c _water V _total (T _st,j,i+1 -T _st,j,i ) (47)

Q _i ＝c _water V _total (T _st,i+1 -T _st,i ) (49)

3. Home energy management optimization model considering user behavior uncertainty

Home energy management systems are generally aimed at better managing the planning of the operation of household electrical appliances. At the same time, in order to reduce the impact of user behavior on power consumption plan, it is necessary to minimize the user's comfort violation coefficient. Therefore, taking into account the user's economy and comfort, the two objectives of the user's net electricity cost and the comfort violation coefficient are selected as the objective function of the optimization model. In order to better solve the multi-objective optimization problem, the method of weighted sum is used to solve it. Let ω ₁ and ω ₂ be the weight coefficients of user economy and comfort, ω ₁ + ω ₂ = 1; due to the order of magnitude difference between economic goals and comfort goals, they cannot directly participate in the calculation. Through the normalization method, set C _cost is the one-day electricity cost of the household energy management model considering user behavior, C _com is the sum of user comfort violation indices in the model in one day, C _cost,max and C _cost,min are the maximum value of electricity cost that occurs in iterations and the minimum value, C _com,max and C _com,min are the maximum and minimum values of the user comfort violation index that appear in the iteration,

p _grid,i represents the interactive power between the home energy management system and the grid in the i-th period, price _b,i represents the electricity purchase price of the user in the i-th period, price _s,i represents the electricity sales price of the user in the i-th period, and ρ is the penalty factor , ε _n is the occurrence probability of uncertain behavior n, and B is the set of uncertain behaviors. The objective function of the obtained home energy management system optimization model is:

In the household energy optimal dispatch model, in addition to the constraints of operating equipment, electric power balance constraints also need to be added. Let P _pv,i be the power generation power of distributed photovoltaic equipment in the i-th period, and P _δ,i be the power of the δ-type load in the i-th period. Electric power, including energy storage equipment and various electric loads. Then the electric power balance constraint is as follows

4. Real-time rolling optimization based on model predictive control

The scheduling optimization model established above uses forecast data to formulate an operation plan within 24 hours at the beginning of a day's operation, so as to obtain a day's operation result. However, in actual operation, due to the error of forecast data, real-time parameters need to be considered, and the control strategy should be corrected according to the actual operating status of electrical equipment. However, the model predictive control changes the running period t=1...N in the formula to t=i...N through the rolling optimization process. For example, at time t=i, the real operating data such as the water temperature of the water heater and the indoor temperature at time i are brought into the home energy management system together with the forecast data in the next optimization time domain T, and the optimal scheduling of the next optimization time domain T is obtained result. In the (i, i+1) time period, the optimized scheduling results are used to optimize the scheduling of household electrical equipment. After t=i+1, repeat the above process until the end of the day. Therefore, only the running results of the first period of each optimization scheme will actually be run. Adjusting the operation plan in real time with the system operation status and forecast information can reduce the impact of uncertainty, which is more suitable for the actual operation optimization system, and the optimal home energy management operation plan can be obtained. Figure 3 is a frame diagram of the rolling optimization time.

The proposed model is verified with the relevant data of a user's home smart power system in spring. The scheduling duration is from 0:00 to 24:00, and the scheduling step is 15 minutes. There are 96 time periods in total. Each time period is represented by i∈{1,2,…,95,96}, and the household energy optimization scheduling model is established. The price of electricity sold in this paper is shown in Figure 4, and the price of electricity purchased is half of the price of electricity sold. The photovoltaic output power curve is shown in Figure 5. The operating parameters of various interruptible loads and non-interruptible loads are shown in Table 1, and the operating parameters of energy storage equipment are shown in Table 2.

Table 1 Interruptible and non-interruptible controllable load operating parameters

Table 2 Operating parameters of energy storage equipment

The rated power of the household air conditioner is 1.8kW, the heat capacity C _ac is 0.426 (kW h/°C), and the thermal resistance R of the house is 75.71kJ/°C. The rated power of the domestic electric water heater is 3.6kW, and the specific heat C _water of water is 4.26kJ/(kg·℃). The forecast curves of outdoor temperature and hot water consumption are shown in Fig. 6 and Fig. 7.

In the scene simulation of uncertain behavior, set the number of scenes N=100. The user's possible temporary water use period and temperature preference change time are set as 18:00-22:00, and both ω ₁ and ω ₂ are taken as 0.5. The probability distribution of the user's possible uncertain behavior is shown in Table 3, the occurrence probability of uncertain behavior ε _n =0.16667, the penalty factor ρ=100, ω ₁ =ω ₂ =0.5.

Table 3 Probability distribution parameters of uncertain behavior

In order to simulate the actual data measured in the actual operation, the forecast curve of the day before is adjusted, and the actual data is simulated by adding a normal distribution deviation to the forecast data. It is also assumed that the prediction error after 8 periods of forecast data is more obvious, which is regarded as the far forecast error, and the forecast error within 8 periods is regarded as the near forecast error. The set parameter prediction errors are shown in Table 4.

Using the scheduling method of the present invention, in order to verify the proposed optimization model considering user behavior uncertainty, three different optimization scenarios of the family are simulated. Scenario 1 considers only electricity cost in the objective function; Scenario 2 considers only user uncertainty behavior in the objective function; Scenario 3 considers both electricity cost and user uncertainty behavior in the objective function. The conclusions are shown in Table 4.

Table 4 Parameter prediction error

It can be seen from Table 4 that compared with Scenario 1, which only considers the cost of electricity consumption, Scenario 3 reduces the user's comfort deviation index from 7.05 to 0.72 by quantifying user behavior. In the optimization scheme considering the user’s uncertain behavior, in order to avoid the start-stop time of controllable loads exceeding the actual allowable start-stop time, the optimization scheme will try to shift the working time of such loads to the middle of the allowable operating period, although this Part of the electricity cost will be sacrificed, but it can effectively improve the user's comfort. In order to avoid the temperature being too low, the temperature control load will keep the hot water temperature and indoor temperature within a relatively high range during the temperature change time, so as to cope with the temporary hot water quantity and temperature preference changes at any time. Through the timely adjustment of the two types of loads, the impact of user uncertain behavior on user comfort can be well reduced. Analysis of Scenario 2 and Scenario 3 shows that when the comfort deviation index is completely used as the objective function, it cannot effectively reduce the impact of uncertain behavior on user comfort, on the contrary, it will significantly increase the user's electricity cost. The willingness to exchange higher electricity cost for lower comfort deviation index is relatively low. Therefore, the method of multi-objective optimization can improve the comfort of users while ensuring lower electricity costs, and reduce the impact of uncertain behavior on users' electricity consumption.

Considering that different users have different preferences for economy and comfort, different weight factors are set, and the household electricity cost coefficient and comfort deviation coefficient are simulated with a scale of 0.1. The conclusions obtained are shown in Figure 8. It can be seen that with the increase of the electricity cost coefficient, the comfort deviation index continues to increase, and the electricity cost gradually decreases. The operation results are mainly divided into three parts: when ω ₁ is in [0.1,0.3], the change of ω ₁ will affect the energy consumption The electricity cost has a great impact, but has little impact on the comfort deviation index. It is an applicable scheme for comfort-preferred users, whose requirements for comfort are far higher than those for economy. When ω ₁ is in [0.7,0.9], the change of ω ₁ will have a greater impact on the user comfort deviation index, but has little impact on the electricity cost. Get lower electricity costs. However, when ω ₁ is in [0.4,0.6], the change of ω ₁ has a great impact on the user comfort deviation index and electricity cost, and belongs to neutral preference users. Such users usually choose according to the experience after adopting the optimization scheme Appropriate coefficient. From the above results, it can be found that compared with the general household energy management optimization based on the user cost, the present invention can better meet the user's comfort requirements, and can automatically adjust according to the user's preference for economy and comfort. Adjust and set different preference factors to obtain the most suitable power consumption plan for users.

Table 5 Electricity cost and user comfort under different scenarios

Those skilled in the art can understand that the accompanying drawing is only a schematic diagram of a preferred embodiment, and the serial numbers of the above-mentioned embodiments of the present invention are for description only, and do not represent the advantages and disadvantages of the embodiments.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection of the present invention. within range.

Claims (3)

1. A home energy management optimization method considering user behavior uncertainty, characterized in that, on the basis of the home energy management system model, the user's uncertain behavior is quantified, and the user's electrical equipment is allowed to start and stop time, ambient temperature The preference and temporary hot water consumption are expressed by the comfort deviation coefficient, and the user uncertainty parameters and the electricity cost are jointly used as the objective function to solve. It operates in a rolling optimization mode, and adjusts the operation plan in real time according to the system operation status and forecast information to achieve home energy management optimization. 2. The family energy management optimization method considering user behavior uncertainty as claimed in claim 1, wherein the specific steps are as follows: 1) According to the power consumption characteristics of household appliances, household appliances are divided into distributed power generation equipment, energy storage equipment and household electricity load, and the operation model of household electrical equipment is established; 2) Analyze the user's subjective behavior, and model the user's behavior by expressing the allowable start and stop time of the user's electrical equipment, environmental temperature preference, and temporary hot water consumption with the comfort deviation coefficient; 3) Aiming at user behavior and household user cost, establish a household energy management model that considers user behavior uncertainty, select user net electricity cost and comfort = violation coefficient as the objective function, and solve the multi-objective optimization problem; 4) Using the model predictive control method, the rolling optimization solution is carried out on the established day-ahead household operation model, and the simulation optimization results are obtained. 3. The family energy management optimization method considering user behavior uncertainty as claimed in claim 1, wherein the detailed steps are as follows: 1. Home energy management system model In the home energy management system, home smart electrical equipment is the main control object. By classifying and modeling the electrical characteristics of different types of equipment, analyzing their impact on home electricity consumption, setting the dispatch cycle to 24h, the It is equal to N time periods, the duration of each time period is Δt, and each time period is recorded as i, then the operation model of various electrical equipment is as follows: (1) Interruptible load: P _ICL,i is the operating power of the interruptible load in the i-th period; p _ICL,r is the rated power of the interruptible load, x _ICL,i is the operation of the interruptible load in the i-th period state, 0 is not running, 1 is running; [b _ICL , e _ICL ] is the allowable working period of the interruptible load, l _ICL is the total running time of the interruptible load, then the work constraint of the interruptible load is: P _ICL,i = x _ICL,i p _ICL,r (1)

(2) Uninterruptible load: including the operational constraints of the interruptible load, and at the same time adding the supplementary constraints to ensure the uninterruptible characteristics are:

P _UICL,i is the operating power of the uninterruptible load in the i period; p _UICL,r is the rated power of the load. x _UICL,i is the running status of the load in the i-th period; [b _UICL ,e _UICL ] is the allowed running period, l _UICL is the total running time; (3) Temperature-controlled loads: Temperature-controlled loads refer to those equipments that optimize the load operation by controlling the temperature. Let T _in,i and T _out,i represent the indoor and outdoor air temperatures in the i-th period respectively; C _ac is the indoor air temperature The thermal capacity of the air conditioner; p _ac,i is the operating power of the air conditioner in the i-th time period, and R is the thermal resistance of the house, then the operating model of the air conditioner and the upper and lower limits of the temperature constraints are:

T _in,min ≤T _in,i ≤T _in,max (6) Let T _st,i represent the hot water temperature of the electric water heater in the i-th period;

is the temperature of cold water injected into the water tank; V _cold,i is the volume of cold water that needs to be injected in the i-th period; V _total is the total capacity of the water tank; C _st is the conversion coefficient between kWh and Joule; p _st,i is the i-th period The operating power of the electric water heater, the operating model of the water heater and the upper and lower limit constraints of the temperature are:

T _st,min ≤T _st,i ≤T _st,max (8) (4) Distributed photovoltaic equipment: In the home energy management system, photovoltaic equipment converts light energy into electrical energy, and then converts it into power frequency current that can be used by household electrical equipment through converters and inverters, thereby providing For household electrical equipment or stored in energy storage equipment, let p _pv,i represent the photovoltaic output power, η _pv represent the photoelectric conversion efficiency of the photovoltaic system, S _pv represent the receiving area m ² of the photovoltaic panel, and I _pv,i represent Photovoltaic system solar radiation intensity (kW/m ² ),

Indicates the outdoor temperature (°C), then the power characteristics of the photovoltaic cell are:

(5) Energy storage equipment: Let SOC be the state of charge of the energy storage equipment,

Represent the charging and discharging power of the household energy storage equipment and the working state of charging and discharging respectively; η _ch is the charging efficiency; η _dch is the discharging efficiency; />

and />

Represents the maximum power of charging and discharging; ε represents the self-discharge rate, Q _r is the rated capacity of the battery, then the charging and discharging characteristics and operating constraints of household energy storage equipment are shown in formula (10) - formula (15):

SOC _min ≤ SOC _i+1 ≤ SOC _max (11)

SOC ₉₆ ≤SOC _ini (15) 2. User Uncertain Behavior Analysis (1) Uninterruptible Load Uncertainty Behavior In actual operation, due to user behavior, the allowable start time of the work will be delayed or the allowable cut-off time of the work will be advanced. Both of these situations will cause the original operation plan to not be completed normally. The change of time obeys the normal distribution, satisfying

with />

Use the Monte Carlo sampling method to simulate N different scenarios, let b _UICL,j be the initial working period under scenario j, e _UICL,j be the end working period under scenario j, x _UICL,i be the uninterruptible load operation state, p _UICL , i is the electric power of the uninterruptible load, and the obtained uninterruptible load comfort deviation C _UICL is as follows:

(2) Uncertain behavior of interruptible load Similar to non-interruptible loads, the running time of interruptible loads will also be affected by user behavior. The actual work start time b _ICL,j and the actual work deadline e _ICL,j of interruptible loads also satisfy the normal distribution. Monte Carlo simulation method, let b _ICL,j be the actual allowable running time of the interruptible load under scenario j, e _ICL,j be the actual cut-off running time of the interruptible load under scenario j, x _ICL,i be the interruptible load The operating state of p _ICL,i is the electric power of the interruptible load, and the obtained interruptible load comfort deviation C _ICL is as follows:

(3) Not sure about temperature preference The user's preferred temperature occurrence time i _ac,b,j is uniformly distributed within a specified period of time, and the change time is 30 minutes, while the user's preferred temperature T _in,j,i obeys a normal distribution and remains constant within the occurrence time, using In the Montecat sampling method, let T _in,j,t be the user's preferred comfort temperature in the i-th period of scene j, and T _in,i be the indoor temperature in the i-th period, and the deviation of the user's temperature comfort degree obtained is:

(4) Uncertain hot water consumption The user has temporary water consumption within [i _st,min ,i _st,max ], the start time i _st,b,j and the water consumption V _st,j follow a uniform distribution, and the water use time is 30 minutes, through the Montecat sampling method , Let Q _st,j,i be the heat consumed by the water heater in the i-th period under scene j, and T _ws,j,i be the hot water temperature in the i-th period under scene j. Q _i is the heat consumed by the water heater in the i period, and the user's hot water consumption comfort deviation C _st is as follows:

Q _st,j,i ＝c _water V _total (T _st,j,i+1 -T _st,j,i ) (20)

Q _i ＝c _water ·V _total ·(T _st,i+1 -T _st,i ) (22) 3. Home energy management optimization model considering user behavior uncertainty Select the two objectives of the user’s net electricity cost and the comfort violation coefficient as the objective function of the optimization model, and use the weighted sum method to solve it. ω ₁ and ω ₂ are the weight coefficients of the user’s economy and comfort, ω ₁ +ω ₂ =1 ; By means of normalization, let C _cost be the daily electricity cost of the household energy management model considering user behavior, C _com be the sum of the user comfort violation indices in the model in one day, and C _cost,max and C _cost,min be iterations The maximum and minimum values of the electricity cost appearing in , C _com,max and C _{com,min are} the maximum and minimum values of the user comfort violation index appearing in the iteration, p _grid,i represents the interactive power between the home energy management system and the grid in the i-th period, price _b,i represents the electricity purchase price of the user in the i-th period, price _s,i represents the electricity sales price of the user in the i-th period, and ρ is the penalty factor , ε _n is the occurrence probability of uncertain behavior n, and B is the set of uncertain behaviors. The objective function of the obtained home energy management system optimization model is:

In the household energy optimal dispatch model, in addition to the constraints of operating equipment, electric power balance constraints also need to be added. Let P _pv,i be the power generation power of distributed photovoltaic equipment in the i-th period, and P _δ,i be the power of the δ-type load in the i-th period. Electric power, including energy storage equipment and various electric loads. Then the electric power balance constraint is as follows

4. Real-time rolling optimization based on model predictive control The model predictive control changes the running period t=1...N in the formula to t=i...N through the rolling optimization process. For example, at time t=i, the real operating data such as the water temperature of the water heater and the indoor temperature at time i are brought into the home energy management system together with the forecast data in the next optimization time domain T, and the optimal scheduling of the next optimization time domain T is obtained result. During the time period (i, i+1), use the optimized scheduling result to optimize the scheduling of the household electrical equipment, and repeat the above process until the end of the day when t=i+1.

Images