METHOD FOR REGULATING PRIMARY FREQUENCY OF POWER GRID BASED ON AIR CONDITIONING LOAD CLUSTER IN LARGE BUILDING

Abstract

The present disclosure relates to a method for primary frequency regulation of an electric network based on large building air conditioning loads cluster. The method includes the use of a two layer control structure with a central coordinating layer and a local control layer. Each local controller performs a thermal model parameter identification and a local air conditioning autonomous control, and uploads local information to the central controller at the end of each communication interval tgap, the central controller broadcasts coordinating information to each local controller. Based on the coordinating information sent from the central controller, each local controller determines whether a power deviation is beyond an action dead zone at the beginning of each action period tact, if beyond, then perform a frequency regulation control action, else, perform no action and estimate operation states of all the air conditionings at the beginning of the next action period.

Description

This application is based on and claims priority to Chinese Patent Application No. 201610390839.X, filed on Jun. 6, 2016, the entire contents of which are incorporated herein by reference.

FIELD

The present disclosure generally relates to a technical field of operation and control of a power system, and more particularly, to a method for regulating primary frequency of a power grid based on an air conditioning load cluster in a large building.

BACKGROUND

Primary frequency regulation capacity is one of main representations of capabilities of a power system to realize a balance between power generation and loads and to respond to an accident or a disturbance. Traditional primary frequency regulation is generally provided by hydropower generating units or large-scale thermal power generating units having a quickly regulating capability with a response time of seconds. With the connection of large-scale new energy power generation, moment of inertia of the power system is significantly reduced, and randomness and uncontrollability of the power generated in new energy generation may increase frequency regulation burden of the power system. Meanwhile, connection of high voltage direct current transmission replaces a local power source, which further reduces the primary frequency regulation capacity of the system. Therefore, how to develop potentials of the load side to involve in the primary frequency regulation of the system becomes a pressing issue.

Based on load side response mechanism, a viable idea is to organize and manage numerous controllable thermal loads with low monomer power to participate in the primary frequency regulation of power system. Via collecting and processing information and certain control means, cluster loads with thermal energy storage effect, such as air conditionings, may be able to participate in ancillary services, while ensuring comfort of an end user is not significantly affected. This is because that the controllable thermal loads are energy type loads, users care about the total thermal energy released by a power consumption equipment to a thermal environment during a period of time rather than the power at each moment. While an error signal of the primary frequency regulation is an impulse type signal with integration within a period of time approaching to zero, thus will not cause significant change in the final energy output. Meanwhile, the controllable thermal loads have occupied more and more proportions of the total loads and have great potentials. In America, the controllable thermal loads in buildings account for more than 35% of the total power consumption loads of the whole power grid. Air conditioning loads also grow fast in China, where the air conditioning loads may account for more than 20% of the maximum loads of the power grid in summer. Therefore an air conditioning cluster may have a great potential of being a reserve frequency regulation means.

In addition, autonomous temperature dead zone control set for ensuring comfort of users of an air conditioning may cause total power of the air conditioning cluster to change against a requirement of a power-frequency response at some time point, that is a so-called rebound effect. The rebound effect may greatly limit thermal energy storage loads to participate in a load side response and to provide ancillary services to the system.

SUMMARY

The present disclosure aims to solve at least one of the problems existing in the related art to at least some extent.

Embodiments of the present disclosure provide a method for regulating primary frequency of a power grid based on an air conditioning load cluster in a large building. According embodiments of the present disclosure, a method for regulating primary frequency of a power grid based on an air conditioning load cluster in a large building is provided, in which, a two-layer control structure including a central coordinating layer and a local control layer is used in the air conditioning load cluster, the central coordinating layer includes a central controller, the local control layer includes N local controllers, N air conditionings, and temperature sensors and frequency sensors provided in rooms the air conditionings located in.

The method includes the following steps:

1) performing, by each local controller, a thermal model parameter identification and an air conditioning autonomous control to obtain local information corresponding to each of the air conditionings, and uploading the local information to the central controller at an end of each communication interval t_gap, and broadcasting, by the central controller, coordinating information to each local controller;

2) when a communication between the central controller and each of the local controllers in step 1) is finished, based on the coordinating information sent from the central controller, determining, by each local controller, whether a power deviation in the air conditionings is beyond an action dead zone at a beginning of each action period t_act, if yes, a frequency regulation control action is performed, else, no action is performed and operation states of all the air conditionings at a beginning of a next action period are estimated. if a current time reaches to a beginning of a next communication interval, step 1) is executed, else, step 2) is repeated.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram showing a two-layer control structure used in embodiments of the present disclosure.

FIG. 2 is a flow chart showing a method regulating primary frequency of a power grid based on an air conditioning load cluster in a large building according to embodiments of the present disclosure; and

FIG. 3 is a flow chart showing a method regulating primary frequency of a power grid based on an air conditioning load cluster in a large building according to embodiments of the present disclosure.

DETAILED DESCRIPTION

A method for regulating primary frequency of a power grid based on an air conditioning load cluster in a large building provided in the present disclosure will be described in combination with embodiments and with reference to drawings as follows.

The present embodiment is shown as FIG. 1, a two-layer, i.e. a central coordinating layer and a local control layer, control structure is used in an air conditioning loads cluster. The central coordinating layer includes a central controller. The local control layer includes N local controllers. N air conditionings and temperature sensors and frequency sensors provided in rooms the air conditionings located in.

In some embodiments, the central controller and the local controllers may communicate in both-way at every communication interval. The local controllers acquire data from the temperature sensors at each temperature sampling period. A communication among the central controller and each of the local controllers is in a way of wireless communication. A communication between each of the local controllers and each of the air conditionings, each of the temperature sensors or the frequency sensors may be in a way of wireless communication or wire communication. The local controllers regulate and control the local air conditionings once during each action period according to local information and coordinating information transmitted from the central controller.

As shown in FIG. 2 and FIG. 3, a method for regulating primary frequency of a power grid based on an air conditioning load cluster in a large building provided in the present disclosure includes following acts.

In block 1, each local controller performs a thermal model parameter identification and an air conditioning autonomous control to obtain local information corresponding to each of the air conditionings, and uploads the local information to the central controller at an end of each communication interval t_gap, the central controller broadcasts coordinating information to each local controller.

In an embodiment of the present disclosure, the communication interval t_gap may be a period between 15 seconds to 1 minute.

In block 2, when a communication between the central controller and each of the local controllers in block 1 is finished, based on the coordinating information sent from the central controller, each local controller begins to determine whether a power deviation in the air conditionings is beyond an action dead zone at a beginning of each action period t_act, if yes, a frequency regulation control action is performed, else, no action is performed and operation states of all the air conditionings at a beginning of a next action period are estimated.

If a current time reaches to a beginning of a next communication interval, block 1 is executed, else, block 2 is repeated.

In an embodiment of the present disclosure, the action period t_act may be 1 second or other preset times.

In an embodiment of the present disclosure, the next action period is an action period t_act interval from current moment, i.e. there is one action period t_act between the beginning of the next action period and the current moment.

In some embodiments, block 1 includes following sub-acts.

In block 11, each local controller i, i=1.2 . . . N, performs the room thermal model parameter identification according to air temperature data recorded at each temperature acquisition period to obtain thermal model parameters corresponding to each room. N is a number of the local controllers.

In some embodiments, a precision degree of the thermal model parameters is determined according to a hardware storage capability of the local controller and an error requirement between a thermal model identification curve and an actual temperature curve.

In some embodiments, for each air conditioning room i, i=1.2 . . . N, three precision degrees of thermal model may be determined. The three precision degrees of thermal model include a zero-order thermal model, a first-order thermal model, and a second-order thermal model represented by equations (1)-(3) respectively.

ΔT_i=α_iΔt_i  (1)

ΔT_i=α_ie^γiΔti−α_i  (2)

ΔT_i=α_i1e^γi1Δti+α_i2e^γi2Δti−α_i1−α_i2  (3)

where, numbers of parameters to be identified in the three precision degrees of thermal models are 1, 2, and 4 respectively, i.e. α_i in equation (1) is a thermal model parameter to be identified in the zero-order thermal model, α_i,γ_i in equation (2) are thermal model parameters to be identified in the first-order thermal model, α_i1,γ_i1,α_i2,γ_i2 in equation (3) are thermal model parameters to be identified in the second-order thermal model, ΔT_i (an initial value of switching temperature for short) is a difference between a current temperature Ta_i and an indoor temperature Ta_i^tog when an on-off state of the air conditioning is last switched, and Δt_i (an initial value of switching time for short) is a difference between a current time and a time t_i^tog when an on-off state of the air conditioning is last switched, and,

ΔT_i=Ta_i−Ta_i^tog  (4)

Δt_i=t_i−t₋^tog  (5)

The initial value of switching temperature Ta_i^tog and the initial value of switching time t_i^tog are taken as parameters of the thermal model as well as α_i in equation (1), α_i,γ_i in equation (2), or α_i1,γ_i1,α_i2,γ_i2 in equation (3).

In block 12, the local controller i identifies parameters of the thermal model corresponding to room i according to air temperature data recorded at each temperature acquisition period t_temp (a period between 1 to 4 seconds) in a communication interval t_gap to obtain thermal model parameters corresponding to each room.

In some embodiments, k is denoted as a number of times that the temperature is recorded, and the on-off state state_i (where state_i=1 corresponding to state ON, state_i=0 corresponding to state OFF) of the air conditioning i is recorded at each time a temperature is recorded. When a zero-order (linear) thermal model is selected, a corresponding parameter identification model is

$\min \left\{\sum _k{\left[\left({\mathrm{Ta}}_i^k-{\mathrm{Ta}}_i^{\mathrm{tog}}\right)-\alpha \left(t_i^k-t_i^{\mathrm{tog}}\right)\right]}^2\right\},$

when an on-off state of the air conditioning i is state_i=1, state parameter α_i^ON corresponding to the ON state is identified according to currently recorded k sets of switching temperature and switching time, when the on-off state of the air conditioning i is state_i=0, state parameter α_i^OFF corresponding to the OFF state is identified according to currently recorded k sets of switching temperature and switching time. When a first-order model is selected, a corresponding parameter identification model is

$\min \left\{\sum _k{\left[\left({\mathrm{Ta}}_i^k-{\mathrm{Ta}}_i^{\mathrm{tog}}\right)-\left(\alpha e^{\gamma \left(t_i^k-t_i^{\mathrm{tog}}\right)}-\alpha \right)\right]}^2\right\},$

similarly, two sets of parameters α_i^ON,γ_i^ON and α_i^OFF,γ_i^OFF are identified in different states ON and OFF respectively. When a two-order model is selected, a corresponding parameter identification model is

$\min \left\{\sum _k{\left[\left({\mathrm{Ta}}_i^k-{\mathrm{Ta}}_i^{\mathrm{tog}}\right)-\left({\alpha }_1e^{{\gamma }_1\left(t_i^k-t_i^{\mathrm{tog}}\right)}+{\alpha }_2e^{{\gamma }_2\left(t_i^k-t_i^{\mathrm{tog}}\right)}-{\alpha }_1-{\alpha }_2\right)\right]}^2\right\},$

similarly, two sets of parameters α_i1^ON,γ_i1^ON,α_i2^ON,γ_i2^ON and α_i1^OFF,γ_i1^OFF,α_i2^OFF,γ_i2^OFF are identified in different states ON and OFF respectively.

In some embodiments of the present disclosure, a same thermal model is selected to use for all the local controllers. In an embodiment, a first-order model is selected to use for all the local controllers, the identified parameters of the i th room are α_i^ON,γ_i^ON and α_i^OFF,γ_i^OFF via common algorithms.

In block 13, each local controller performs the air conditioning autonomous control, according to following equations.

$\begin{array}{cc}{\mathrm{state}}_i=\{\begin{array}{cc}1,& {\mathrm{Ta}}_i\ge {\stackrel{\_}{T}}_i\&{\mathrm{state}}_i=0\\ 0,& {\mathrm{Ta}}_i\le {\underset{\_}{T}}_i\&{\mathrm{state}}_i=1\end{array}& \left(6\right)\\ {\stackrel{\_}{T}}_i={\mathrm{Ts}}_i+{\Delta }_i,{\underset{\_}{T}}_i={\mathrm{Ts}}_i-{\Delta }_i& \left(7\right)\end{array}$

In the above equations, i=1.2 . . . N, Ta_i is an air temperature in the i th room, Δ_i is a temperature control dead zone corresponding to the i th air conditioning, T_i corresponds to an upper bound of the temperature control dead zone Δ_i, T_i corresponds to a lower bound of the temperature control dead zone Δ_i, Ts_i is the required temperature set by the user, state_i is the on-off state of the i th conditioning (state_i=1 corresponds to state ON, state_i=0 correspond to state OFF).

Denoting an i th local controller, an i th air conditioning and an i th room with mark i, i=1.2 . . . N. An air temperature in the i th room is Ta_i. An on-off state of the i th air conditioning is state_i (where state_i=1 corresponding to state ON, state_i=0 corresponding to state OFF). In some embodiments, it is assumed that the i th air conditioning is a constant power air conditioning with an operation power P_i and with on-off state controlled only.

Each temperature sensor acquires an indoor air temperature of a corresponding room in real-time. Each local controller acquires the temperature data from a corresponding temperature sensor every temperature acquisition period t_temp (for example, a period between 1 to 4 seconds).

A required temperature Ts_i of each air conditioning i is set directly by the user, and each air conditioning i has a temperature control dead zone Δ_i, which is a factory setting attribute, and in an embodiment, Δ_i is assumed to be 1° C. Equation (6) shows that if state_i=0, i.e. the air conditioning is in an OFF state, when the room air temperature Ta_i rises to the upper bound T_i of the temperature control dead zone of the i th local air conditioning, the i th local controller turns on the i th air conditioning autonomously; if state_i=1, i.e. the air conditioning is in an ON state, when the room air temperature Ta_i drops to the lower bound T_i of the temperature control dead zone of the i th local air conditioning, the i th local controller turns off the i th air conditioning autonomously.

In block 14, at an end moment (i.e. a communication moment) of the communication interval t_gap between the local controller and the central controller, each local controller uploads the local information to the central controller.

The local information includes the indoor air temperature acquired most recently Ta_i of the room, the on-off state state_i of the air conditioning, the operation power P_i, the required temperature Ts_i, the temperature control dead zone Δ_i, and the thermal model parameters α_i^ON,γ_i^ON, α_i^OFF,γ_i^OFF, Ta_itog, and t_i^tog.

In block 15, the central controller collects all the local information from the local controllers and broadcasts all collected information to each local controller as the coordinating information, and the central controller obtains a reference power P0_i of each air conditioning after the thermal model parameters corresponding to each local controller are collected, a sum of reference powers of all the air conditionings is obtained as a reference power P0 of all the air conditionings, and the reference power of all the air conditionings P0 is broadcasted to each local controller.

The coordinating information includes the indoor air temperature Ta_i, the on-off state state_i, the operation power P_i, the required temperature Ts_i, the temperature control dead zone Δ_i, and the thermal model parameters α_i^ON,γ_i^ON, α_i^OFF,γ_i^OFF, Ta_i^tog, and t_i^tog.

In some embodiments, the reference power P0_i corresponds to an average power of the i th air conditioning during an on-off period T_i (referring to a time period during which the i th air conditioning switches its on-off state in one cycle according to a local autonomous control logic) in a communication interval t_gap.

Block 15 may include following acts.

The central controller calculates a first time t_i⁽¹⁾, a second time t₁⁽¹⁾, a third time t_i⁽¹⁾, and a forth time t_i⁽¹⁾ by solving the following equations respectively according to the upper T_i, the lower temperature bound T_i, the thermal model parameters α_i^ON,γ_i^ON, α_i^OFF,γ_i^OFF, Ta_i^tog, and t_i^tog, the required temperature Ts_i, and the temperature control dead zone Δ_i.

T _i −Ta_i^tog=α_i^ONe^{γiON(ti(1)-titog}₎−α_i^ON

T _i −Ta_i^tog=α_i^ONe^{γiON(ti(2)-titog}₎−α_i^ON

T _i −Ta_i^tog=α_i^OFFe^{γiOFF(ti(3)-titog}₎−α_i^OFF

T _i −Ta_i^tog=α_i^OFFe^{γiOFF(ti(4)-titog}₎−α_i^OFF

In above equations, t_i(1) is a moment when the indoor temperature is equal to the upper bound temperature T_i and the air conditioning is in an “ON” state; t_i⁽²⁾ is a moment when the indoor temperature is equal to the lower bound temperature T_i and the air conditioning is in an “ON” state; t_i⁽³⁾ is a moment when the indoor temperature is equal to the upper bound temperature T_i and the air conditioning is in an “OFF” state; t_i⁽⁴⁾ is a moment when the indoor temperature is equal to the lower bound temperature T_i and the air conditioning is in an “OFF” state.

A total time period Ton_i when the i^th air conditioning is in an “ON” state in an on-off period T_i, and a total time period Toff_i when the i^th air conditioning is in an “OFF” state in an on-off period T_i are obtained according to following equations.

Ton _i =t _i ⁽¹⁾ −t _i ⁽²⁾  (8)

Toff _i =t _i ⁽⁴⁾ −t _i ⁽³⁾  (9)

The reference power P0_i of each air conditioning is calculated as:

$\begin{array}{cc}P0_i=\frac{{\mathrm{Ton}}_i}{T_i}P_i=\frac{{\mathrm{Ton}}_i}{{\mathrm{Ton}}_i+{\mathrm{Toff}}_i}P_i& \left(10\right)\end{array}$

In which, P0_i is an reference power of the i^th air conditioning, Ton_i is the total time period when the i^th air conditioning is in an “ON” state in an on-off period T_i, Toff_i is the total time period when the i^th air conditioning is in an “OFF” state in an on-off period T_i, P_i is an operation power of the i^th air conditioning.

The reference power of all the air conditionings P0 is obtained by summing all the reference powers P0_i of the air conditionings according to following equation.

$\begin{array}{cc}P0=\sum _iP0_i& \left(11\right)\end{array}$

The central controller broadcasts the reference power P0 of all the air conditionings to each local controller.

In some embodiments, a control objective of primary frequency regulation response of the air conditioning cluster is set as making a difference ΔP between a real-time total power P(t) of all the air conditionings and the reference power P0 of all the air conditionings to be directly proportional to a real-time frequency deviation Δf, satisfying following equation.

ΔP=P(t)−P0=K(f(t)−f₀)=KΔf  (12)

In which, f₀ is a reference frequency, being 50 Hz for Chinese mainland, f(t) is a real-time frequency obtained by the frequency sensor, K is a power-frequency response coefficient and set to a same value for all the local controllers. K may be determined according to a ratio of a total power of the air conditioning cluster to a maximum frequency fluctuation in history. The greater K is, the more the air conditioning involves in the frequency regulation, and the smaller K is, the less the air conditioning involves in the frequency regulation.

Block 2) includes the following actions.

In block 21, a frequency of the power grid is acquired by a frequency sensor every action period t_act, and each local controller calculates a power deviation δ of all the air conditionings according to the acquired frequency of the power grid at the beginning of each action period t_act.

In some embodiments, each local controller calculates the real-time total power P(t) of all the air conditionings according to the received coordinating information (air temperature Ta_i, on-off state state_i and power Pi of the air conditionings in all the rooms) broadcasted by the central controller via following equation:

$\begin{array}{cc}P\left(t\right)=\sum _iP_i*{\mathrm{state}}_i& \left(13\right)\end{array}$

where. i=1.2 . . . N, P_i is the operation power of the i_th air conditioning, state_i is an on-off state of the i^th air conditioning.

Then, the power deviation δ of all the air conditionings is calculated according to following equation.

δ=P(t)−P0−KΔf,

where, P(t) is the real-time total power of all the air conditionings, P0 is the reference power of all the air conditionings, K is a power-frequency response coefficient set for all the local controllers, Δf is a real-time frequency deviation.

In block 22, each local controller determines whether the power deviation δ is in the action dead zone ξ, when the power deviation δ is in the action dead zone ξ, the air conditioning does not participate in the frequency regulation control, when the power deviation δ is not in the action dead zone ξ, the air conditioning participates in the frequency regulation control action in the present action period.

In some embodiments, ξ may be set according to accuracy requirement, for example, in an embodiment, ξ is 1 KW.

In some embodiments, if|ξ|≧ξ, the power deviation δ is determined to be in the action dead zone ξ.

In some embodiments, block 22 includes following actions.

In block 221, a temperature priority Tpri_i of each local controller is obtained according to following equation.

$\begin{array}{cc}{\mathrm{Tpri}}_i=\{\begin{array}{cc}\left({\mathrm{Ts}}_i-{\mathrm{Ta}}_i\right)/{\Delta }_i,& {\mathrm{state}}_i=1\\ \left({\mathrm{Ta}}_i-{\mathrm{Ts}}_i\right)/{\Delta }_i,& {\mathrm{state}}_i=0\end{array}& \left(14\right)\end{array}$

In which, Tpri_i is a temperature priority of i^th local controller, Ta_i is the indoor air temperature, Zs_i the required temperature corresponding to the i^th air conditioning set by a user, Δ_i s the temperature control dead zone, state_i is the on-off state of the i th air conditioning (ON corresponds 1, OFF corresponds to 0).

Equation (14) means that the air conditioning in a room where the air temperature is closer to a boundary of the temperature control dead zone corresponds a higher priority. When state_i=1, i.e. the air conditioning is in an “ON” state, the lower the air temperature Ta_i is, the higher the priority Tpri_i is, and the air conditioning will be turned off more preferentially in a local frequency regulation process; When state_i=0, i.e. the air conditioning is in an “OFF” state, the higher the air temperature Ta_i is, the higher the priority Tpri_i is, and the air conditioning will be turned on more preferentially in the local frequency regulation process.

In block 222, when δ>ξ, temperature priorities Tpri_i of air conditionings whose state_i=1 are selected, and an array qu_ON is generated accordingly with its rows arranged according to values of the temperature priorities Tpri_i in a descending order, the first column of the array is Tpri_i, the second column is P_i, the third column is i, and the number of rows in the array qu_ON is denoted as r, a minimum control set which can regulate the power deviation into the dead zone is selected according to r*=min{r|Σ_d=1^rqu_ON(d,2)≧δ−ξ}, a set I_ON of numbers of the air conditionings to be regulated in the present operation is extracted from the minimum regulation control set according to I_ON=qu_ON(j,3), j=1, 2, . . . , r*, and I_ON′={iεI_ON|Ta_i<Tgon_i} is calculated (in which, as a parameter represents a participating degree of the air conditioning in the frequency regulation, Tgon_i may be preset by users of the air conditionings and people who controls the frequency regulation system, for example, Tgon_i may be set as T_i+0.8Δ_i), if a number of an air conditioning controller i_localεI_ON′, an air conditioning corresponding to the an air conditioning controller i_local is controlled to participate in the present frequency regulation control, i.e. a state of the air conditioning corresponding to the an air conditioning controller i_local is switched (turn off the local air conditioning), else, no action is performed.

In block 223, when δ<−ξ, temperature priorities Tpri_i of air conditionings whose state_i0 are selected, and an array qu_OFF is generated accordingly with its rows arranged according to values of the temperature priorities Tpri_i in a descending order, the first column of the array is Tpri_i, the second column is P_i, the third column is i, and the number of rows in the array qu_OFF is denoted as r, a minimum control set which can regulate the power deviation into the dead zone is selected according to r*=min{r|Σ_d=1^rqu_OFF(d,2)≧−δ−ξ}, a set I_OFF of numbers of the air conditionings to be regulated in the present operation is extracted from the minimum regulation control set according to qu_OFF(j,3), j=1, 2, . . . , r*, and I_OFF′={iεI_OFF|Ta_i>Tgoff_i} is calculated (in which, as a parameter represents a participating degree of the air conditioning in the frequency regulation, Tgoff_i may be preset by users of the air conditionings and people who controls the frequency regulation system, for example, Tgoff_i may be set as T_i−0.4Δ_i), if a number an air conditioning controller i_localεI_OFF′, the air conditioning corresponding to the an air conditioning controller i_local is controlled to participate in the present frequency regulation control, i.e. a state of the local air conditioning corresponding to the an air conditioning controller i_local is switched (turn on the local air conditioning), else, no action is performed.

In block 23, after the frequency regulation in block 22 is finished, each local controller estimates the on-off states of all the air conditionings at a beginning of a next action period.

In some embodiments, the next action period is one action period t_act ahead from the present moment.

Because on-off states of some of the air conditionings have been changed in the present action period, each local controller estimates the on-off states of all the air conditionings at the beginning of the next action period after each frequency regulation in block 22 is finished.

The process (block 23) of estimating the on-off states of all the air conditionings includes following actions.

In block 231, the set I_ON′ or I_OFF′ in block 22) are obtained.

In block 232, it is set that i=1.

In block 233, it is determined whether iεI_ON to determine whether the i th air conditioning participates in the frequency regulation action, if iεI_ON, the present state of the i th air conditioning is “OFF”, i.e. state_i0, and the air temperature Ta_i^tog before the switch is flipped and the moment t_i^tog when the switch is flipped are recorded. if iεI_OFF, the present state of the i th air conditioning is “ON”, i.e. state_i=1, and the air temperature Ta_i^tog before the switch is shifted and the moment t_i^tog when the switch is shifted are recorded;

In block 234, let i=i+1, if i≦N block 233 is executed, else, block 24 is executed.

In block 24, each local controller estimates air temperatures in other rooms at the beginning of the next action period, and modifies on-off state state_i of the i th air conditioning at the beginning of the next action period t_act according to the coordinating parameters transmitted from the central controller and the on-off states of all the air conditionings estimated in block 23 via the autonomous control method.

Block 21 is executed when the next action period comes, or block 1 is executed when the next communication interval begins.

The autonomous control method can refer to equations (6) and (7) illustrated in block 13.

Block 24 may include following acts.

In block 241, a first-order thermal model is used for estimating the temperature in a present embodiment, for the i th air conditioning, a time variance relative to t_i^tog at moment t is Δt_i, and i=1.

In block 242, if the on-off state of the i th air conditioning stored locally is “ON”, i.e. state_i=1, the room air temperature stored locally is Ta_i(t)=α_i^ONe^γiONΔti−α_i^ON+Ta_i^tog; if the on-off state of the i th air conditioning stored locally is “OFF”, i.e. state_i=0, the room air temperature stored locally is Ta_i(t)=α_i^OFFe^γiOFFΔti−α_i^OFF+Ta_i^tog.

In block 243, if the room air temperature meets a condition Ta_i(t)≦T_i, the state of the i th air conditioning stored locally is state_i=0, and Ta_i^tog and t_i^tog are recorded, if the room air temperature meets a condition Ta_i(t)≧T_i, the state of the i th air conditioning stored locally is state_i=1, and Ta_i^tog and t_i^tog are recorded, else, the state of the i th air conditioning state_i remains unchanged.

In block 244, let i=i+1, if i≦N block 242 is executed, else, block 21 is executed when the next action period t_act comes, or, block 1 is executed if next communication moment comes.

The method for regulating primary frequency of a power grid based on an air conditioning load cluster in a large building according to embodiments of the present disclosure has following characteristics.

With the method according to embodiments of the present disclosure, by taking advantage of heat capacity of large buildings, the two layers control structure including the central coordinating layer and the local control layer is formed, via rapid control of the air conditioning cluster, it is now possible to involve the air conditioning cluster in primary frequency regulation with a linear power-frequency characteristic similar to that of an electric generator. Each air conditioning performs a primary frequency regulation response locally to increase the speed of entire response, and communicate with the central controller every a certain time interval to upload local information and obtain overall coordinating information so as to ensure accuracy of the entire power-frequency linear response. Meanwhile, a temperature monitor threshold guarantees comfort of users and useful life of the equipment will not be significantly affected.

With the method according to embodiments of the present disclosure, a two-layer control structure including a slow centralized coordination and a rapid distributed local control is provided. In the centralized coordinating layer (i.e. the central coordinating layer), information of each room, such as operation power and state of corresponding air conditioning, the temperature in the room, thermal model parameters of the room, etc., is collected according to coordinating control period, and then is broadcasted to each of the local controllers. The problems of long time delay in centralized control and lack of coordination in distributed control are solved, thus improving the control accuracy and solving the problem of rebound effect.

The controllers in the local control layer estimate operation states of all the air conditionings and temperatures in all rooms based on dynamic thermal models respectively, and sequence the air conditioning cluster accordingly. Whether a local frequency regulation will be triggered is determined by an order of the air conditioning in the sequence and a real-time frequency deviation. The method is a local algorithm, thus improving speed of response to the power-frequency deviation.

With the method according to embodiments of the present disclosure, loads of the air conditioning cluster is a linear response to a frequency deviation and the comfort of users is not effected, the contradiction of slow response in central control and lack of coordinating information in local decentralized control when an air conditioning cluster participating in primary frequency regulation is eliminated. The control accuracy is increased and the problem of rebound effect is solved.

Claims

1. A method for regulating primary frequency of a power grid based on an air conditioning load cluster in a large building, wherein, a two-layer control structure comprising a central coordinating layer and a local control layer is used in the air conditioning load cluster, the central coordinating layer comprises a central controller, the local control layer comprises N local controllers, N air conditionings, and temperature sensors and frequency sensors provided in rooms the air conditionings located in; andthe method comprises:1) performing, by each local controller, a thermal model parameter identification and an air conditioning autonomous control to obtain local information corresponding to each of the air conditionings, and uploading the local information to the central controller at an end of each communication interval tgap, and broadcasting, by the central controller, coordinating information to each local controller;2) when a communication between the central controller and each of the local controllers in step 1) is finished, based on the coordinating information sent from the central controller, determining, by each local controller, whether a power deviation in the air conditionings is beyond an action dead zone at a beginning of each action period tact, if yes, a frequency regulation control action is performed, else, no action is performed and operation states of all the air conditionings at a beginning of a next action period are estimated; if a current time reaches to a beginning of a next communication interval, step 1) is executed, else, step 2) is repeated.

2. The method according to claim 1, wherein, step 1) comprises: 1-1) performing, by each local controller i, i=1.2 . . . N, the room thermal model parameter identification according to air temperature data recorded at each temperature acquisition period to obtain thermal model parameters corresponding to each room;1-2) identifying, by the local controller i, parameters of the thermal model corresponding to room i according to air temperature data recorded at each temperature acquisition period ttemp in a communication interval tgap to obtain identified thermal model parameters of each room;1-3) performing, by each local controller, the air conditioning autonomous control according to following equations:

3. The method according to claim 2, wherein a precision degree of the thermal model parameters is determined according to a hardware storage capability of the local controller and an error requirement between a thermal model identification curve and an actual temperature curve.

4. The method according to claim 3, wherein the thermal model corresponding to ith room comprises a zero-order thermal model, a first-order thermal model, or a second-order thermal model, represented by equations (1)-(3) respectively: ΔTi=αiΔti  (1)ΔTi=αieγiΔti−αi  (2)ΔTi=αi1eγi1Δti+αi2eγi2Δti−αi1−αi2  (3)where, numbers of parameters to be identified in the three thermal models are 1, 2, and 4 respectively, αi in equation (1) is a thermal model parameter to be identified in the zero-order thermal model, αi,γi in equation (2) are thermal model parameters to be identified in the first-order thermal model, αi1,γi1,αi2,γi2 in equation (3) are thermal model parameters to be identified in the second-order thermal model, ΔTi is a difference between a current temperature Tai and an indoor temperature Taitog when an on-off state of the air conditioning is last switched, and Δti is a difference between a current time and a time titog when an on-off state of the air conditioning is last switched, where, ΔTi=Tai−Taitog  (4)Δti=ti−titog  (5)wherein, Taitog and titog are thermal model parameters.

5. The method according to claim 4, if the thermal model corresponding to ith room is the first-order thermal model, the thermal model parameters comprises αiON,γiON and αiOFF,γiOFF.

6. The method according to claim 2, wherein, the reference power P0i corresponds to an average power of the i th air conditioning during an on-off period Ti in a communication interval tgap, and step 1-5) comprises: obtaining, by the central controller, a first time ti(1), a second time ti(1), a third time ti(1), and a forth time ti(1) by solving the following equations respectively according to the upper Ti, the lower temperature bound Ti, the thermal model parameters αiON,γiON, αiOFF,γiOFF, Taitog, and titog, the required temperature Tsi, and the temperature control dead zone Δi: Ti−Taitog=αiONeγiON(ti(1)-titog)−αiON Ti−Taitog=αiONeγiON(ti(2)-titog)−αiON Ti−Taitog=αiOFFeγiOFF(ti(3)-titog)−αiOFF Ti−Taitog=αiOFFeγiOFF(ti(4)-titog)−αiOFF wherein, ti(1) is a moment when the indoor temperature is equal to the upper bound temperature Ti and the air conditioning is in an “ON” state; ti(2) is a moment when the indoor temperature is equal to the lower bound temperature Ti and the air conditioning is in an “ON” state; ti(3) is a moment when the indoor temperature is equal to the upper bound temperature Ti and the air conditioning is in an “OFF” state; ti(4) is a moment when the indoor temperature is equal to the lower bound temperature Ti and the air conditioning is in an “OFF” state;obtaining a total time period Toni when the ith air conditioning is in an “ON” state in an on-off period Ti, and a total time period Toffi when the ith air conditioning is in an “OFF” state in an on-off period Ti according to following equations: Toni=ti(1)−ti(2)  (8)Toffi=ti(4)−ti(3)  (9)obtaining the reference power P0i of each air conditioning as:

7. The method according to claim 1, wherein, step 2) comprises: 2-1) acquiring, by a frequency sensor, a frequency of the power grid every action period tact, and calculating, by each local controller, a power deviation δ of all the air conditionings according to the acquired frequency of the power grid at the beginning of each action period tact;2-2) determining, by each local controller, whether the power deviation δ is in the action dead zone ξ, when the power deviation δ is in the action dead zone ξ, the air conditioning does not participate in the frequency regulation control; when the power deviation δ is not in the action dead zone ξ, the air conditioning participates in the frequency regulation control action in the present action period;2-3) estimating, by each local controller, on-off states of all the air conditionings at a beginning of a next action period;2-4) estimating, by each local controller, air temperatures in other rooms at the beginning of the next action period, and modifying the on-off state statei of the i th air conditioning at the beginning of the next action period tact according to the coordinating parameters transmitted from the central controller and estimated on-off states of all the air conditionings via the air conditioning autonomous control in step 1-2), and executing step 2-1) when the next action period comes or executing step 1) when a next communication interval begins.

8. The method according to claim 7, wherein, step 2-1) comprises: calculating, by each local controller, a real-time total power P(t) of all the air conditionings according to the received coordinating information broadcasted by the central controller via following equation:

9. The method according to claim 7, wherein, step 2-2) comprises: 2-2-1) obtaining a temperature priority Tprii of each local controller according to following equation:

10. The method according to claim 1, wherein the central controller and the local controllers communicate in both-way at every communication interval, the local controllers acquire data from the temperature sensors at each temperature sampling period.

11. The method according to claim 1, wherein the local controllers regulate and control the air conditionings once during each action period according to local information and coordination information transmitted from the central controller.

12. The method according to claim 1, wherein each of the air conditionings is constant power air conditioning with an operation power and having two states; ON and OFF.

13. The method according to claim 1, wherein each temperature sensor acquires an indoor air temperature of a corresponding room in real-time, and the local controller acquires temperature data from the temperature sensors every the temperature acquisition period.