Research on Decentralized Centralized Control Model for Fast Load Response

With the introduction of a high proportion of new energy, the security and stability of the power system face enormous challenges. Against the backdrop of insufficient traditional flexible regulation resources, it is urgent to tap into the adjustable potential of load side rapid response resources. This article conducts a modeling of the dynamic response process of fast response load, and proposes a decentralized centralized control optimization model for fast response load in response to the demand of the power grid. Through simulation analysis of different examples, the optimization effect of this method on the load curve of the power system is obtained, achieving the suppression of new energy output fluctuations and the maintenance of real-time supply and demand balance in the power system.


Introduction
Against the backdrop of increasing energy scarcity and environmental degradation, the development and application of new energy have become a necessary path.New energy generation has characteristics such as intermittency and volatility [1] .If directly integrated into the power grid, it will pose a threat to the safe and stable operation of the power system, requiring a large amount of reserve capacity.Therefore, large-scale new energy generation integration and consumption are relatively difficult.In the context of insufficient traditional flexible regulation resources, it is urgent to tap the adjustable potential of load side rapid response resources [2] .
The rapid development of information and communication technology provides a new solution to the above problems, and participates in the generalized demand response by regulating the rapid response resources [3] .Therefore, this article first analyzes and considers the establishment of a model for flexible load participation in frequency regulation, and proposes a decentralized centralized aggregation framework [4] .Finally, a control strategy based on moving average is proposed to cope with the aggregation control of fast response loads [5] .

Air conditioning-room model
In power systems with a high proportion of renewable energy generation, traditional generation side resources are unable to cope with rapidly fluctuating regulatory demands.Demand side flexible load resources based on IoT networks have the same regulatory ability as generation side resources, and have faster response speed and higher response accuracy.In order to better grasp and utilize flexible load resources, this chapter conducted modeling work on the dynamic response process of rapid response loads [6] . The In the formula, T air is the indoor temperature; C air is the heat capacity of air; T mass is the thermal and mass temperature of the air conditioner Degree; G mass is the thermal conductivity between air conditioning and air; G env is the thermal conductivity between the internal air and the outside world; Q air is the indoor air heat flux; Tout is the outdoor temperature; C mass is the heat capacity of the air conditioning heat mass; Q mass is the heat, mass, and flux of the air conditioner.
The air conditioning operates periodically within the set temperature range, and the control of the air conditioning load includes two types: direct switch control and temperature control.Considering that temperature control has a relatively small impact on user comfort, this article adopts the set temperature of the air conditioning load to adjust the operating power of the air conditioning load.Changing the set temperature of the air conditioning load will change the time the power load is in on and off mode, thereby affecting its average power [7][8] .

Electric vehicle load
The connection between electric vehicles (EVs) and the power grid is achieved through charging, and the charging process of EVs is essentially the energy storage process of lithium batteries.At present, EV charging methods are mainly divided into charging mode and battery replacement mode, using dedicated sockets to obtain electricity from the power grid.The charging mode can be divided into regular charging and fast charging.Regular charging requires a longer time; Fast charging utilizes high current charging, which not only brings significant impact to the power grid but also reduces battery life.The battery replacement mode is to leave the empty battery at the charging station and use a small current for long-term charging.
Therefore, the charging process for establishing a single electric vehicle is shown in the figure 1.

Fig1 Charging curve of a single electric vehicle
The mathematical expression for the charging power P(t) during period t during the charging process is as follows: In the formula: P r is the rated charging power, determined by the charging mode; t 1 and t 2 determine the degree to which the charging power of a lithium battery decreases when it approaches full charge.
When regulating electric vehicles, for users, the shorter the waiting time for charging, the better, and the lower the charging cost.Therefore, a desire degree model is introduced to describe the enthusiasm of users to participate in supply-demand friendly interactions.The user desire degree C c is defined as the weighted sum of the percentage of user charging cost savings and the queuing time factor: In the formula, a and b are weighting coefficients, a+b=1.In fact, different users have different expectations for cost and time savings, and the weighting coefficients should also be different.For simplicity, here, let a=0.5 and b=0.5.
saves users a percentage of costs.Assuming that the user chooses to charge at time k, and the charging price at that time is p i , and the required amount of electricity for the user to charge is Q ch , the percentage of charging costs saved by the user after using the real-time electricity price is as follows: If an electric vehicle reaches the charging station and the charging device is idle, it can be charged immediately; If the charging device is not idle, electric vehicles need to queue up.
The average queue length L q for electric vehicles and the average queue time T q for electric vehicle users are: In the formula, l n is the probability of n electric vehicles charging; N s is the number of available charging devices in the charging station; n is the number of electric vehicles being charged.

Energy storage equipment
At present, energy storage equipment can be divided into mechanical energy storage, electromagnetic energy storage, chemical energy storage, and phase change energy storage according to the form of electrical energy storage.Considering the energy characteristics of China, the lithium-ion battery in chemical energy storage has the characteristics of large energy density, small self-discharge, small size, environmental friendliness, etc [9][10] .
Therefore, this article selects high-capacity batteries as an example to study the model of energy storage equipment.
The State of Charge (SOC) of a battery reflects the ratio of its remaining capacity to its rated capacity, as shown in the following equation: In the formula, SOC(t) is the state of charge of the lithium battery at time t; C net (t) is the current remaining power; C bat is the rated capacity.The charging and discharging process at different times will affect its state of charge value, and its dynamic process is shown in the following equation: In the formula, SOC (t+1) is the state of charge of the lithium battery at time t+1, P bat-t is the charging and discharging power (in kW), charging is positive, discharging is negative, is the charging and discharging duration (in hours), and C bat is the rated capacity of the battery (in kWh).In order to extend the service life of lithium batteries, it is necessary to limit their state of charge to a certain range, and the constraint conditions are as follows: In the formula, SOC min is the minimum state of charge value, and when the battery discharges to SOC min , it is not allowed to continue discharging; SOC max is the maximum state of charge value, and when the battery is charged to SOC max , it is not allowed to continue charging.

  
In the formula: P i is the active power of the load; Q i is the reactive power of the load; S n is the set total power; U a is the actual voltage; U n is the rated voltage; Z is the proportion of constant impedance load; Z ɵ The angle of the constant impedance load; I represents the proportion of constant current load; I ɵ The angle of constant current load; P is the proportion of constant power load; P ɵ The angle of a constant power load.

Fast response load resource dispersion centralized control strategy
The previous article demonstrated through modeling and case analysis that flexible loads have the potential for regulation, but the individual capacity of fast response loads is usually small, and only after certain means of aggregation can significant regulation potential be realized.This article proposes a decentralized centralized control strategy for fast response loads to meet different needs of the power grid, achieving goals such as frequency stability control of the power system, increasing new energy penetration rate, and reducing user purchasing costs.
In the formula: P tie-line (t) is the power of the interconnection line between the microgrid and the power grid; line tie P  avg is the average interconnection line power between the microgrid and the power grid; M avg is the average electricity price; M rt (t) is the real-time electricity price; is the total power of load transfer; is predict output for photovoltaic systems; is predict power for load; is predict power for the battery; is the maximum power of the battery; is the state of charge of the battery; is the reference value of battery charge state; is the power of the i-th flexible load; N DR is the number of flexible loads.
The above formula can calculate the power of the interconnection line based on real-time electricity prices, which is negatively correlated with real-time electricity prices.That is, the higher the electricity price, the less electricity the microgrid purchases from the main grid.Calculate the transferable load for the next period of time through load pre measurement, photovoltaic prediction, and battery power calculation.In addition, due to the relationship between battery power and battery state of charge, it is necessary to meet the schedulability constraint.Finally, the flexible load power is aggregated to obtain a fast response total.
During operation, there may be a deviation between the load power and the predicted value before operation.To maintain a constant power of the connecting line during operation, it is necessary to adjust the power of the energy storage equipment in real time.The specific aggregation control strategy is shown in the figure 3.
Fig3 Aggregation control strategy for energy storage equipment The battery charging and discharging power is calculated based on the tie line power and demand response load between the microgrid and the power grid determined before operation, as well as the real-time load power: In the formula: is the real-time total power of the battery; P PV (t) is the real-time photovoltaic power; P load (t) is the real-time load power; P DR (t) is the load transfer power; P bat, j (t) is the real-time power of the jth battery; N bat is the total number of batteries.

example
In order to analyze the impact of flexible load generalized participation demand response on regional power grid, this section takes a small micro grid as an example to conduct simulation analysis on the calculation examples in the following table 1, including whether there is photovoltaic power generation, whether there is energy storage device and whether there is flexible load demand response.
Table 1 Simulation Example photovoltaics Energy storage Demand response example 1 The figure 4 shows the basic load of the entire microgrid, excluding air conditioning and flexible resources.The peak load on the residential side is usually concentrated around 11:00 am, and there are also obvious peaks at 8:00 am and 17:00 noon.From the above simulation results, it can be seen that the air conditioning load usually increases rapidly from 6:00, and the growth rate of the air conditioning load slows down after 10:00.Due to the fact that the temperature at this time also begins to maintain a relatively stable state, the air conditioning load begins to enter a relatively stable period.
In the original microgrid, the peak load appeared around 18:00.Due to the significant increase in air conditioning load during the day, the microgrid's load during the day was also higher than the basic load.The power changes of microgrids are basically consistent with real-time electricity prices, so microgrids will also bear relatively high electricity bills.
The introduction of photovoltaics can significantly reduce the daytime electricity purchase of microgrids.In weather conditions with good weather conditions and high photovoltaic power generation output, during the period from 9:00 to 16:00, the output of photovoltaics is basically greater than the load, but there is no energy storage, resulting in the phenomenon of solar abandonment.Due to the lack of photovoltaic output at night, during the peak electricity price period from 16:00 to 20:00, the load of the microgrid is still high, requiring higher electricity bills to be borne.
After the addition of energy storage, energy storage can first eliminate the phenomenon of light abandonment, releasing a large amount of electricity during high electricity prices at night, effectively reducing the high load during peak periods.
After the demand response is added, the obvious advantages are: the load during the 17:00-20:00 high electricity price period is further reduced by transferring to the daytime, the load during the 2:00-6:00 low electricity price period is significantly increased, and the effect of electricity cost saving brought by load transfer is obvious; When the photovoltaic output fluctuates, the energy storage battery combined with demand response is better than only the energy storage battery to restrain the load fluctuation.
From the simulation results of the four examples shown in the above figure, it can be seen that during the peak electricity price period, the power of the interconnection line in Example 1 also reached a peak, with a peak value of about 298kW, requiring high electricity purchase costs; Example 2 relies on photovoltaic output.Due to the photovoltaic output exceeding the load during the period from 9:00 to 13:00, the power of the connecting line decreased to 0 and some electrical energy was wasted; Example 3 uses energy storage equipment to store additional photovoltaic output, while also storing a certain amount of electricity during the low electricity price period in the early morning, effectively reducing the power of the interconnection line during the high electricity price period.It decreases to 124kW during the 17:00-19:00 period and 130kW during the 19:00-21:00 period; Example 4 relies on the transfer of flexible loads and further reduces the power of the connecting line during peak hours on the basis of Example 3. The power is reduced to 65 kW and 85 kW during the periods of 17:00 to 19:00 and 19:00 to 21:00, respectively.

Conclusion
This article establishes an uncertainty output model for renewable energy generation, including wind power and photovoltaic power, and analyzes the power grid regulation demand considering the uncertainty of new energy generation.To address its uncertainty, a response dynamic process model for multiple flexible loads is established based on GridLAB-D, covering generalized fast response load resources such as air conditioning, electric vehicles, and energy storage, It has achieved the suppression of fluctuations in new energy output and the maintenance of real-time supply-demand balance in the power system.

Fig4Fig 6
Fig4 basic load of the entire microgrid The results obtained through numerical simulation are shown in the following figure 5 and figure 6.
Equivalent Thermal Parameters (ETP) model is a commonly used residential air conditioning room model.The equivalent thermal parameter model takes into account various factors such as weather conditions, residential thermal parameters, solar radiation, and air conditioning set temperature, and can be expressed as: