WO2023160110A1 - 一种集群温控负荷系统调频方法、系统、电子设备及存储介质 - Google Patents

一种集群温控负荷系统调频方法、系统、电子设备及存储介质 Download PDF

Info

Publication number
WO2023160110A1
WO2023160110A1 PCT/CN2022/137165 CN2022137165W WO2023160110A1 WO 2023160110 A1 WO2023160110 A1 WO 2023160110A1 CN 2022137165 W CN2022137165 W CN 2022137165W WO 2023160110 A1 WO2023160110 A1 WO 2023160110A1
Authority
WO
WIPO (PCT)
Prior art keywords
load
state
cluster
temperature control
model
Prior art date
Application number
PCT/CN2022/137165
Other languages
English (en)
French (fr)
Inventor
潘玲玲
耿建
李峰
王勇
李亚平
周竞
刘建涛
刘俊
王礼文
徐鹏
郭晓蕊
毛文博
Original Assignee
中国电力科学研究院有限公司
国网北京市电力公司
国网江苏省电力有限公司
国家电网有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国电力科学研究院有限公司, 国网北京市电力公司, 国网江苏省电力有限公司, 国家电网有限公司 filed Critical 中国电力科学研究院有限公司
Publication of WO2023160110A1 publication Critical patent/WO2023160110A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H02GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02JCIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00Circuit arrangements for ac mains or ac distribution networks
    • H02J3/24Arrangements for preventing or reducing oscillations of power in networks
    • H02J3/241The oscillation concerning frequency
    • HELECTRICITY
    • H02GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02JCIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00Circuit arrangements for ac mains or ac distribution networks
    • H02J3/12Circuit arrangements for ac mains or ac distribution networks for adjusting voltage in ac networks by changing a characteristic of the network load
    • H02J3/14Circuit arrangements for ac mains or ac distribution networks for adjusting voltage in ac networks by changing a characteristic of the network load by switching loads on to, or off from, network, e.g. progressively balanced loading
    • H02J3/144Demand-response operation of the power transmission or distribution network
    • HELECTRICITY
    • H02GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02JCIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2203/00Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
    • H02J2203/10Power transmission or distribution systems management focussing at grid-level, e.g. load flow analysis, node profile computation, meshed network optimisation, active network management or spinning reserve management
    • HELECTRICITY
    • H02GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02JCIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2203/00Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
    • H02J2203/20Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04SSYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S20/00Management or operation of end-user stationary applications or the last stages of power distribution; Controlling, monitoring or operating thereof
    • Y04S20/20End-user application control systems
    • Y04S20/222Demand response systems, e.g. load shedding, peak shaving

Definitions

  • the invention belongs to the field of load frequency modulation control, and in particular relates to a frequency modulation method, system, electronic equipment and storage medium of a cluster temperature control load system.
  • Clustered household temperature-controlled loads mostly participate in short-term (second-level) demand response auxiliary projects, such as frequency regulation of distribution networks and balancing power demand.
  • second-level demand response auxiliary projects such as frequency regulation of distribution networks and balancing power demand.
  • the power demand of household temperature-controlled loads has great randomness, and its capacity is relatively small compared with industrial and commercial users; and large-scale industrial and commercial users have randomness in the power demand of temperature-controlled loads due to their own power consumption characteristics and operating procedures. Smaller and larger capacity. It is necessary to further explore the control technology of cluster temperature control load demand response.
  • the existing control models are less optimized in the selection of control objects.
  • the quality of the output performance depends on the time-varying characteristics of the given tracking signal, which cannot guarantee a superior response effect; on the other hand, it does not consider how to closely link the optimal control signal and the optimal load control object. for better control.
  • the purpose of the present invention is to invent a frequency modulation method, system, electronic equipment and storage medium of a cluster temperature control load system.
  • the overall performance is better.
  • the present invention adopts the following technical solutions:
  • a frequency regulation method for a cluster temperature control load system comprising:
  • the cluster temperature control load state space model is established, and the state transition probability is obtained based on the Markov chain to obtain the state transition matrix;
  • the cluster temperature control load optimization control model is established, and the total power change value of the cluster temperature control load is solved by rolling optimization to obtain the cluster temperature control load control scheme;
  • Controllable load selection is performed on the cluster temperature control load control scheme based on multi-scale priority ranking, and a controllable load selection result is output.
  • the establishment of a 2D state warehouse cluster temperature control load model based on the single air conditioner load model, and determining the schedulable capacity of the current cluster temperature control load specifically includes the following steps:
  • the current switch status of the cluster temperature control load it is divided into closed group and open group; respectively, according to the upper and lower limits of user comfort indoor air temperature and the upper and lower limits of the indoor substance temperature Divide the temperature interval into N i /2 indoor air temperature small intervals and N m /2 indoor material temperature small intervals to form N a *N m /4 closed group and open group state bins; and then form a 2D State bin transfer model;
  • the sum of the rated power of all air conditioners is the maximum adjustable power capacity, plus the currently turned on air conditioner load power The sum is used as the maximum power allowable value P max that can be scheduled and adjusted; take all the currently open bins in the 2D state bin, and the temperature ranges from low to high, and the sum of the rated power of all air-conditioning equipment except the last state bin that is closest to the upper limit of the allowable temperature is The maximum power capacity that can be lowered is obtained to obtain the minimum power allowable value P min that can be scheduled and adjusted by the cluster air conditioner group.
  • the second-order discretized differential equation of the equivalent thermal parameter model is:
  • ⁇ t represents the simulation step size
  • ⁇ a represents the indoor air temperature of the air-conditioning load
  • ⁇ m represents the indoor material temperature of the air-conditioning load
  • ⁇ s represents the ambient temperature of the air-conditioning load
  • R a and R m are the indoor air and indoor The equivalent thermal resistance of the substance
  • C a and C m are the equivalent heat capacities of the air temperature and the material temperature respectively; when the air conditioner load is turned on, Q a is the rated power of the air conditioner; when it is turned off, Q a is equal to 0.
  • the cluster temperature control load state space model is established, and the state warehouse transition probability is solved based on the Markov chain, thereby obtaining State transition matrix, including the following steps:
  • cluster temperature control load state space model is established, which is expressed by a time-varying discrete state space equation:
  • x(k) represents the system state vector at the kth moment, and the expression is as follows:
  • x(k) [x 1,off (k)x 2,off (k)...x N/2,off (k)x 1+N/2,on (k)x 2+N/2,on ( k)...x N,on (k)] T (3)
  • A(k) represents the system matrix at the k-th moment, and its element A ij (k) represents the transition probability of the air-conditioning load in state bin j being transferred to state bin i within the k-th time step;
  • u(k) represents the control signal at the kth moment
  • B(k) represents the input matrix at the kth moment
  • element B ij (k) represents the transition probability of the air-conditioning load in state bin j that needs to be switched to state bin i under the action of u(k), expressed as follows :
  • diag represents a diagonal matrix
  • diag sub represents a subdiagonal matrix
  • C(k) represents the output matrix at the kth moment, that is, the average power vector of the air-conditioning load of each state bin at the current moment k, expressed as follows:
  • m represents the total number of air-conditioning loads
  • Pa ave (k) represents the average power of the air-conditioning loads in each state cabin of the open group at the kth moment
  • Pagg (k) represents the observed value of the aggregated output power of the air-conditioning load group at the kth moment
  • the switch function vector of the state bin S i represents the switch state of the state bin i
  • w i represents the switch state of the i-th load
  • 0 is the off state
  • 1 is the on state
  • y(k) represents the output power of the cluster air-conditioning load model at the kth moment
  • x s represents the set value of the air conditioner temperature
  • x in,t represents the indoor temperature at time t
  • ⁇ x represents the allowable offset value of the air conditioner temperature set value
  • s AC,t represents the working state of the air conditioner at time t, the value A value of 0 means the air conditioner is turned off, and a value of 1 means it is turned on;
  • the state bins are numbered sequentially;
  • n i,j (k) represents the number of air-conditioning loads transferred from state bin i to state bin j in the k-th period ;
  • N represents the total number of state bins;
  • each state transition probability p i,j (k) can be obtained, and thus the state transition matrix P(k) can be obtained.
  • the calculation of the primary frequency regulation power change value of the cluster temperature control load according to the cluster temperature control load primary frequency regulation model specifically includes the following steps:
  • the cluster temperature control load primary frequency regulation system model includes a single-region frequency regulation system model and a cluster air conditioner group frequency regulation module;
  • the single-region frequency regulation system model is a closed-loop system with an integral regulation system, and the single-region
  • the power adjustment signal of the frequency modulation system model is converted into the steam turbine input power adjustment variable through the system secondary frequency modulation transfer function, the generator governor transfer function and the prime mover transfer function, and the input power adjustment variable and the load fluctuation variable participate in the system frequency modulation;
  • the group frequency modulation module includes the frequency modulation dead zone of the air conditioner group, the frequency modulation coefficient, the upper and lower limits of the dispatchable potential, and the response time delay of the air conditioner;
  • each air-conditioning device monitors the frequency deviation signal, it calculates the primary frequency regulation power change value ⁇ P AC of the cluster temperature-controlled load.
  • the calculation of the secondary frequency regulation power change value of the cluster temperature control load according to the secondary frequency regulation model of the cluster temperature control load specifically includes the following steps:
  • the control signal includes the system frequency deviation of the primary frequency modulation and the setting value of the system power deviation signal calculated by the regional AGC of the secondary frequency modulation, and the setting value of the regional AGC power deviation is determined by the setting ratio r;
  • r represents the setting magnification of the regional AGC power deviation on the air conditioning group, which is calculated by the following formula:
  • ⁇ P′ g (s) represents the secondary frequency modulation power deviation signal of the thermal power unit
  • ⁇ P′ AC (s) represents the secondary frequency modulation power deviation signal of the cluster air conditioner group
  • ⁇ P c (s) is calculated by the regional AGC Secondary frequency modulation power deviation signal
  • the total power change value ⁇ P AC' (s) of the cluster temperature control load is calculated according to formula (12).
  • the cluster temperature control load optimization control model is established according to the state transition matrix, and the total power change value of the cluster temperature control load is solved by rolling optimization to obtain the cluster temperature control load control scheme, which specifically includes the following steps:
  • the forecast duration is p
  • the forecast state at the k+p time is x(k+ ⁇
  • k), ⁇ 1,2,...,p; according to the cluster temperature control load state space model, Establish the state equation from the k+1th moment to the k+p moment:
  • a P internal block matrix A p [A(k+p-1)...A(k+1)A(k)] represents the predicted value of the state transition matrix of the system at the k+p time at the current k time, where The element A p (i,j) indicates that only the vector x(k
  • the optimal control objective function is:
  • W err represents the tracking error weight coefficient matrix between the model output and the actual target value
  • D(k) diag ⁇ C(k+1)C(k+2)...C(k+p) ⁇
  • R(k) [r(k+1)r(k+2)...r(k+p)] T
  • r(k+ ⁇ ) means to output the target trajectory value at k+ ⁇ moment
  • the optimal control model of the cluster air-conditioning load is transformed into a quadratic programming function:
  • the quadratic programming function is solved by rolling optimization.
  • the optimal control sequence composed of the number of air-conditioning load switches in the state warehouse in the control time domain p* ⁇ t is obtained.
  • the first component of the optimal sequence is issued.
  • k) wait for the arrival of the next scheduling cycle, repeat the above rolling optimization process, and obtain the cluster temperature control load control scheme.
  • the multi-scale prioritization is used to select the controllable load of the cluster temperature control load control scheme, and output the controllable load selection result, including the following steps:
  • NTD i, k is the normalized temperature distance of the i-th air-conditioning load at the current k moment; ⁇ represents the temperature dead zone, the difference between the upper and lower limits of user comfort temperature ⁇ high and ⁇ low ; ⁇ i,t Indicates the temperature of the i-th air-conditioning load at the current k moment, O k and C k respectively represent the on-group and off-group at the current k moment, and m is the total number of air-conditioning loads;
  • SIM i , (p,q) is the similarity index between the air-conditioning load power and the required adjustment power in the state bin (p,q)
  • P i is the rated power of the i-th air-conditioning load
  • P aim,(p, q) is the target power that the state bin (p, q) needs to respond to
  • N (p, q) represents the number of air-conditioning loads in the state bin (p, q);
  • NC i,k (C i,k ⁇ C k,min )/(C k,max ⁇ C k,max ) (20)
  • C i,k is the accumulative control times of air-conditioning load i at time k
  • C k,min and C k,max represent the minimum and maximum values of the controlled times of load k at the current moment
  • the multi-scale priority comprehensive index ⁇ open is obtained as follows:
  • K T , K S and K C are the corresponding weight coefficients
  • the air-conditioning loads in each 2D state bin corresponding to the open state group are sorted, and the controllable loads are selected according to the priority from high to low, and the controllable load selection results are obtained.
  • a cluster temperature control load system frequency modulation system including:
  • the acquisition unit is configured to acquire cluster temperature control load initialization parameters
  • the load model establishment unit is configured to establish a 2D state warehouse cluster temperature control load model according to the single air conditioner load model, and determine the dispatchable capacity of the current cluster temperature control load;
  • the space model building unit is configured to establish a cluster temperature-controlled load state space model based on the 2D state warehouse cluster temperature-controlled load model and the current cluster temperature-controlled load schedulable capacity, and solve the state-bin transition probability based on the Markov chain to obtain the state transition matrix;
  • the frequency modulation model calculation unit is configured to calculate the primary frequency modulation power change value of the cluster temperature control load according to the primary frequency modulation model of the cluster temperature control load, calculate the secondary frequency modulation power change value of the cluster temperature control load according to the secondary frequency modulation model of the cluster temperature control load, and then obtain the cluster Total power change value of temperature control load;
  • the control model optimization unit is configured to establish a cluster temperature control load optimization control model according to the state transition matrix, and the total power change value of the cluster temperature control load is solved by rolling optimization to obtain the cluster temperature control load control scheme;
  • the controllable load selection unit is configured to perform controllable load selection on the cluster temperature control load control scheme based on multi-scale priority ranking, and output a controllable load selection result.
  • An electronic device comprising a memory, a processor, and a computer program stored in the memory and operable on the processor, when the processor executes the computer program, the cluster temperature control load system frequency regulation method is implemented A step of.
  • a computer-readable storage medium wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for frequency regulation of the cluster temperature-controlled load system are implemented.
  • the invention provides a method for flexible frequency regulation of a cluster temperature-controlled load system, a cluster temperature-controlled load modeling method based on a time-varying Markov chain in a 2D state bin, and on this basis, the primary and secondary cluster temperature-controlled load power systems are established Frequency modulation model, and finally an innovative control method based on normalized temperature distance, power similarity and cumulative control times for multi-scale priority selection of loads is proposed.
  • the control process of cluster air conditioner groups participating in the primary and secondary frequency regulation of the power system under the load aggregator organization is given, and the effectiveness of cluster air conditioner groups participating in system frequency regulation is verified, and the frequency characteristics of the power grid can be improved when the load fluctuates. Enhance system operation security.
  • the method proposed by the invention can effectively improve the frequency characteristics of the power grid when the load fluctuates, and compared with the traditional method, the modeling accuracy of the temperature-controlled load is higher, and the control method has better control accuracy, response speed, and fairness of load participation in demand response. The overall performance is better.
  • FIG. 1 is a schematic flowchart of a method for frequency regulation of a cluster temperature-controlled load system provided by an embodiment of the present invention.
  • Fig. 2 is a thermal operation characteristic diagram of an exemplary unitary air conditioner provided by an embodiment of the present invention.
  • Fig. 3 is a schematic diagram of an exemplary 2D state bin transition model provided by an embodiment of the present invention.
  • Fig. 4 is a schematic diagram of an exemplary cluster temperature control load regulation power output variation curve with system frequency provided by an embodiment of the present invention.
  • FIG. 5 is a schematic diagram of an exemplary single-area primary frequency regulation simulation model based on cluster temperature control load provided by an embodiment of the present invention.
  • Fig. 6 is a schematic diagram of an exemplary single-region primary and secondary frequency regulation simulation model based on cluster temperature control load provided by an embodiment of the present invention.
  • FIG. 7 is a schematic flowchart of a cluster temperature control load control method provided by an embodiment of the present invention.
  • Fig. 8 is a schematic structural diagram of a frequency modulation system of a cluster temperature control load system provided by an embodiment of the present invention.
  • FIG. 9 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention.
  • Load frequency regulation When the frequency of the power system deviates from the rated value, the load can be controlled through the frequency deviation, so as to reduce the frequency fluctuation and maintain the safe operation of the power system.
  • Temperature-controlled load is an electric load controlled by a constant temperature heater, which has a certain thermal energy storage capacity. In an actual power system, the temperature-controlled load mostly refers to the air-conditioning load of residents or small commercial users.
  • Demand response that is, power demand response, which refers to the behavior of power suppliers to guide power users' power consumption patterns through power price signals or compensation incentives, so as to reduce or increase power loads and ensure stable operation of the power grid.
  • Model predictive control is a model-based closed-loop optimization control method, and its core idea is a rolling time-domain strategy. The basic idea is to predict the future dynamics of the system based on the current moment, solve the constraint programming problem to obtain the current optimal solution, and wait for the system to update the state, then roll forward the prediction time domain until the requirement is met.
  • Multi-scale prioritization In the present invention, it specifically refers to a load selection method based on three indicators of load normalized temperature distance, power similarity and cumulative control times.
  • the load selection is carried out through multi-scale prioritization, and the control is implemented.
  • the present invention provides a frequency regulation method for a cluster temperature control load system, including the following steps:
  • the air conditioner terminal automatically monitors the frequency deviation signal and converts it into a power signal; the load aggregator obtains the specific switch signal according to the control algorithm and sends it to the air conditioner group; the air conditioner group switches the equipment and responds to the action according to the switch signal;
  • Secondary frequency modulation accept the regional AGC command, and send the power deviation signal to the load aggregator; the load aggregator obtains the specific switching signal according to the control algorithm, and sends it to the air conditioning group; the air conditioning group switches the equipment according to the switching signal, and responds to the action;
  • the frequency regulation method of the cluster temperature-controlled load system of the present invention is embodied in three aspects: modeling of the cluster temperature-controlled load operating state, construction of the cluster temperature-controlled load frequency regulation model, and optimization of the control algorithm. Specifically:
  • the cluster temperature control load control method based on model predictive rolling optimization control adds a multi-scale priority sorting load selection process, and further simulates and analyzes the method's control accuracy, response speed, and fairness of load participation in demand response. The overall performance is better.
  • a method for flexible frequency regulation of a cluster temperature control load system comprising the following steps:
  • Step 1 Collect the initialization parameters of the cluster temperature-controlled loads.
  • the initialization parameters include the switch status, indoor temperature, and outdoor temperature of the current N temperature-controlled loads;
  • Step 2 Establish a 2D state warehouse cluster temperature control load model
  • Described step 2 comprises the following specific steps:
  • ⁇ t represents the simulation step size
  • ⁇ a represents the indoor air temperature of the air-conditioning load
  • ⁇ m represents the indoor material temperature of the air-conditioning load
  • ⁇ s represents the ambient temperature of the air-conditioning load
  • R a and R m are the indoor air and indoor The equivalent thermal resistance of the substance
  • C a and C m are the equivalent heat capacities of the air temperature and the material temperature respectively; when the air conditioner load is turned on, Q a is the rated power of the air conditioner; when it is turned off, Q a is equal to 0.
  • Figure 2 shows the thermal operating characteristics of a single air conditioner, where ⁇ +_ETP and ⁇ -_ETP represent the upper and lower limits of the air-conditioning load setting temperature respectively; ⁇ s_ETP represents the temperature setting value of the air-conditioning load; Turn-on time; ⁇ off is the turn-on time of the air conditioner.
  • Step 3 Solve the transition probability of the state bin based on the Markov chain; the step 3 includes the following specific steps:
  • x(k) represents the system state vector at the kth moment, and the expression is as follows:
  • x(k) [x 1,off (k)x 2,off (k)...x N/2,off (k)x 1+N/2,on (k)x 2+N/2,on ( k)...x N,on (k)] T (3)
  • x(k+1) represents the system state vector at the k+1th moment.
  • A(k) represents the system matrix at the kth moment, and its element A ij (k) represents the transition probability of the air-conditioning load in state bin j being transferred to state bin i in the kth time step.
  • u(k) represents the control signal at the kth moment, that is, the percentage of the air-conditioning load in each state bin at the current moment k that needs to be switched; when the signal is positive, it means the opening action, and when it is negative, it means the closing action.
  • B(k) represents the input matrix at the kth moment
  • element B ij (k) represents the transition probability of the air-conditioning load in state bin j that needs to be switched to state bin i under the action of u(k), expressed as follows :
  • diag represents a diagonal matrix
  • diag sub represents a subdiagonal matrix
  • C(k) represents the output matrix at the kth moment, that is, the average power vector of the air-conditioning load of each state bin at the current moment k, expressed as follows:
  • m represents the total number of air-conditioning loads
  • Pave (k) represents the average power of the air-conditioning loads of each state cabin of the open group at the kth moment
  • Pagg (k) represents the observed value of the aggregated output power of the air-conditioning load group at the kth moment.
  • S represents the switch function vector of each state bin
  • S i represents the switch state of state bin i
  • w i represents the switch state of the i-th load
  • 0 is the off state
  • 1 is the on state.
  • y(k) represents the output power of the cluster air conditioner load model at the kth moment.
  • x s represents the set value of the air conditioner temperature
  • x in,t represents the indoor temperature at time t
  • ⁇ x represents the allowable offset value of the air conditioner temperature set value
  • s AC,t represents the working state of the air conditioner at time t, the value A value of 0 means the air conditioner is off, and a value of 1 means it is on.
  • n i,j (k) represents the number of air-conditioning loads transferred from state bin i to state bin j in the k-th period ; N represents the total number of state bins.
  • step 31) From step 31), it can be seen that according to the different values of i and j, each state transition probability p i,j (k) can be obtained, thereby obtaining the state transition matrix P(k).
  • any column j of A(k) represents the state transition probability that the air-conditioning load of state bin i in the current k-th period is transferred to state bins 1 to N
  • any row i of P(k) represents the state transition probability of state bin i in the current k-th period
  • A(k) P T (k).
  • Step 4 Establish a cluster temperature control load primary frequency regulation model; the step 4 includes the following specific steps:
  • K g * represents the per unit value of the unified primary frequency modulation coefficient of the regional thermal power unit
  • ⁇ f * represents the per unit value of the system frequency deviation
  • K AC * represents the per unit value of the primary frequency modulation coefficient of the cluster air conditioner group
  • ⁇ K AC * represents The per unit value of the total power change of the cluster air conditioner group.
  • the single-region primary frequency regulation simulation model for cluster temperature-controlled loads is shown in Figure 5.
  • the primary frequency regulation system model of the cluster temperature control load is divided into two parts: the traditional single-area frequency regulation system model and the cluster air conditioner group frequency regulation module.
  • the traditional single-area frequency regulation system model is a closed-loop system with an integral regulation system, and the power regulation signal is converted into steam turbine input power regulation variables and load fluctuations through the system secondary frequency regulation transfer function, generator governor transfer function, and prime mover transfer function respectively. Variables participate in system tuning together.
  • the cluster air conditioner group frequency modulation module is composed of the frequency modulation dead zone of the air conditioner group, the frequency modulation coefficient, the upper and lower limits of the schedulable potential, and the air conditioner response time delay.
  • the transfer function of the secondary frequency modulation of the system is The transfer function of the generator governor is The prime mover transfer function is The load fluctuation variable is ⁇ P K (s), the frequency modulation coefficient is K AC , and the air conditioner response time delay is
  • K n is the static gain of the governor
  • T n is the time constant of the governor
  • R is the adjustment coefficient of the governor
  • K n is the static gain of the steam turbine
  • T T is the time constant of the steam turbine
  • K r is the reheat coefficient
  • T r is the reheat time constant
  • K P is the proportional adjustment coefficient of the second frequency modulation
  • K I is the integral adjustment coefficient of the second frequency modulation.
  • T AC is the temperature control load response time delay, which is between 0.1s and 0.5s.
  • Step 5 Establish a cluster temperature control load secondary frequency regulation model; the step 5 includes the following specific steps:
  • the control signals received by the cluster air conditioner group include the system frequency deviation ⁇ f (primary frequency modulation) and the setting value of the system power deviation signal calculated by the regional AGC (secondary frequency modulation), and the regional AGC power deviation
  • the setting value is determined by the setting ratio r.
  • the rest of the structure is the same as the primary frequency adjustment of the cluster air conditioner.
  • r represents the setting magnification of the regional AGC power deviation on the air conditioning group, which is calculated by the following formula:
  • ⁇ P′ g (s) represents the secondary frequency modulation power deviation signal of the thermal power unit
  • ⁇ P′ AC (s) represents the secondary frequency modulation power deviation signal of the cluster air conditioner group
  • ⁇ P c (s) is calculated by the regional AGC Secondary FM power deviation signal.
  • Step 6 MPC-based rolling optimization cluster temperature control load control; the step 6 includes the following specific steps:
  • the cluster temperature control load control method mainly includes steps:
  • Step 1 Establish a 2D state warehouse transfer model for cluster temperature control loads
  • Step 2 Establish the time-varying discrete state space equation of cluster temperature control load
  • Step 3 Obtain the k-moment control model based on the model prediction algorithm
  • Step 4 Multi-scale prioritization indicators for object selection
  • Step 5 Execute the model to predict the optimal control signal
  • step five it is judged whether k ⁇ k max holds true, if yes, continue to execute steps from step two for time k+1, and if no, end.
  • step 4 the primary frequency modulation or primary and secondary frequency modulation signals are obtained through step 4 or step 5, and the cluster temperature control load will be regulated through the following steps.
  • a P internal block matrix A p [A(k+p-1)...A(k+1)A(k)] represents the predicted value of the state transition matrix of the system at the k+p time at the current k time, where The element A p (i,j) indicates that only the vector x(k
  • the air-conditioning load of state bin j is transferred to the state Predicted value of transition probability for bin i. BP is completely similar and will not be repeated here.
  • W err represents the tracking error weight coefficient matrix of the model output and the actual target value, which is set as the unit matrix in the present invention
  • D(k) diag ⁇ C(k+1)C(k+2)...C(k+p ) ⁇
  • R(k) [r(k+1)r(k+2)...r(k+p)] T
  • r(k+ ⁇ ) means outputting the target trajectory value at k+ ⁇ time.
  • the value range of the control variable U(k) should be between -1 and 1; the value range of the state variable X(k) should be between 0 and 1. Therefore, by aiming at the minimum cluster output tracking error, the optimal control model of the cluster air-conditioning load can be transformed into a quadratic programming problem as shown below:
  • Rolling optimization solution Both 0 and 1 in the constraint condition of formula (17) are vector forms, and the quadratic programming model can be solved by calling the quadratic programming function provided by the MATLAB optimization toolbox. After the solution is obtained, the optimal control sequence composed of the number of air-conditioning load switches in the state bin in the control time domain p* ⁇ t is obtained, and only the first component u * (k
  • Step 7 Controllable load selection based on multi-scale prioritization; the step 7 includes the following specific steps:
  • NTD i,k is the normalized temperature distance of the i-th air-conditioning load at the current moment k, and is a dimensionless coefficient with a value range between 0 and 1.
  • represents the temperature dead zone, that is, the difference between the upper and lower limits of user comfort temperature ⁇ high and ⁇ low .
  • ⁇ i , t represents the temperature of the i-th air-conditioning load at the current time k
  • Ok and C k represent the on-group and off-group at the current k time, respectively
  • m is the total number of air-conditioning loads.
  • SIM i , (p,q) is the similarity index between the air-conditioning load power and the required adjustment power in the state bin (p,q)
  • P i is the rated power of the i-th air-conditioning load
  • P aim,(p, q) is the target power that the state bin (p, q) needs to respond to
  • N (p,q) represents the number of air-conditioning loads in the state bin (p,q).
  • NC i,k (C i,k ⁇ C k,min )/(C k,max ⁇ C k,max ) (20)
  • C i,k is the accumulative control times of air-conditioning load i at time k
  • C k,min and C k,max represent the minimum and maximum values of the controlled times of load k at the current moment.
  • K T , K S and K C are the corresponding weight coefficients respectively.
  • the air-conditioning loads in each 2D state bin corresponding to the open state group are sorted.
  • the smaller the value of ⁇ open the higher the priority of the air-conditioning load in the state bin.
  • the temperature control load is controlled to realize the frequency modulation function.
  • the present invention provides a cluster temperature control load system frequency modulation system, including:
  • the acquisition unit 10 is configured to acquire cluster temperature control load initialization parameters
  • the load model establishment unit 11 is configured to establish a 2D state warehouse cluster temperature control load model according to the single air conditioner load model, and determine the dispatchable capacity of the current cluster temperature control load;
  • the space model building unit 12 is configured to establish a cluster temperature-controlled load state space model based on the 2D state-warehouse cluster temperature-controlled load model and the current cluster temperature-controlled load schedulable capacity, and solve the state-bin transition probability based on the Markov chain, thereby obtaining the state transfer matrix;
  • the frequency regulation model calculation unit 13 is configured to calculate the primary frequency regulation power change value of the cluster temperature control load according to the primary frequency regulation model of the cluster temperature control load, calculate the secondary frequency regulation power change value of the cluster temperature control load according to the secondary frequency regulation model of the cluster temperature control load, and then obtain The total power change value of cluster temperature control load;
  • the control model optimization unit 14 is configured to establish a cluster temperature control load optimization control model according to the state transition matrix, and the total power change value of the cluster temperature control load is solved by rolling optimization to obtain a cluster temperature control load control scheme;
  • the controllable load selection unit 15 is configured to perform controllable load selection on the cluster temperature control load control scheme based on multi-scale priority ranking, and output a controllable load selection result.
  • the third object of the present invention is to provide an electronic device, including a memory 20, a processor 21, and a computer program stored in the memory 20 and operable on the processor 21, the When the processor executes the computer program, the steps of the frequency regulation method for the cluster temperature control load system are implemented.
  • the electronic device may further include a communication interface 22 for communicating with other external devices, for example, for data transmission, etc., which is not limited in this embodiment of the present invention.
  • the frequency regulation method of the cluster temperature control load system includes:
  • the cluster temperature control load state space model is established, and the state transition probability is obtained based on the Markov chain to obtain the state transition matrix;
  • the cluster temperature control load optimization control model is established, and the total power change value of the cluster temperature control load is solved by rolling optimization to obtain the cluster temperature control load control scheme;
  • Controllable load selection is performed on the cluster temperature control load control scheme based on multi-scale priority ranking, and a controllable load selection result is output.
  • the fourth object of the present invention is to provide a computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the processor executes the computer program, the steps of the frequency regulation method for the cluster temperature control load system are realized .
  • the frequency regulation method of the cluster temperature control load system includes:
  • the cluster temperature control load state space model is established, and the state transition probability is obtained based on the Markov chain to obtain the state transition matrix;
  • the cluster temperature control load optimization control model is established, and the total power change value of the cluster temperature control load is solved by rolling optimization to obtain the cluster temperature control load control scheme;
  • controllable load selection is performed on the cluster temperature control load control scheme, and the controllable load selection result is output.
  • the embodiments of the present invention may be provided as methods, systems, or computer program products. Accordingly, the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
  • computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions
  • the device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.
  • the invention provides a method for flexible frequency regulation of a cluster temperature-controlled load system, a cluster temperature-controlled load modeling method based on a time-varying Markov chain in a 2D state bin, and on this basis, the primary and secondary cluster temperature-controlled load power systems are established Frequency modulation model, and finally an innovative control method based on normalized temperature distance, power similarity and cumulative control times for multi-scale priority selection of loads is proposed.
  • the control process of cluster air conditioner groups participating in the primary and secondary frequency regulation of the power system under the load aggregator organization is given, and the effectiveness of cluster air conditioner groups participating in system frequency regulation is verified, and the frequency characteristics of the power grid can be improved when the load fluctuates. Enhance system operation security.
  • the method proposed by the invention can effectively improve the frequency characteristics of the power grid when the load fluctuates, and compared with the traditional method, the modeling accuracy of the temperature-controlled load is higher, and the control method has better control accuracy, response speed, and fairness of load participation in demand response. The overall performance is better.

Landscapes

  • Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Air Conditioning Control Device (AREA)

Abstract

提供一种集群温控负荷系统调频方法、系统、设备及介质,方法包括:获取集群温控负荷初始化参数;建立2D状态仓集群温控负荷模型;根据2D状态仓集群温控负荷模型,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵;根据集群温控负荷一次调频模型计算集群温控负荷一次调频功率变化值,根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值;采用滚动优化求解得到集群温控负荷控制方案;基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴选,并输出可控负荷遴选结果。该方法在控制精度、响应速度和负荷参与需求响应公平性等方面的综合性能较优。

Description

一种集群温控负荷系统调频方法、系统、电子设备及存储介质
相关申请的交叉引用
本申请基于申请号202210182642.2、申请日为2022年02月25日,申请名称为“一种集群温控负荷系统调频方法、系统、设备及介质”的中国专利申请提出,并要求该中国专利申请的优先权,该中国专利申请的全部内容在此以引入方式结合在本申请中。
技术领域
本发明属于负荷调频控制领域,尤其涉及一种集群温控负荷系统调频方法、系统、电子设备及存储介质。
背景技术
随着电力体制改革深化推进,电力需求侧管理作为电力改革的重要组成部分将获得实质性推动。随着我国智能电网建设和需求响应项目的逐步实施,集群温控负荷具有参与需求响应项目进行新能源消纳的巨大潜力,进行相关关键技术的研究具有重要的理论意义和工程价值。
目前,已有集群家用温控负荷削峰填谷和紧急负荷管理的需求响应项目。集群家用温控负荷多参与短时(秒级)的需求响应辅助项目,例如:配电网的频率调节以及平衡用电需求。然而,家用温控负荷的电力需求具有很大的随机性,且相对工商业用户来说其容量较小;而大型工商业用户由于自身的用电特性和运行流程,其温控负荷的电力需求随机性较小,且容量较大。需要进一步探索集群温控负荷需求响应的调控技术。
现有技术有考虑双质特性的二阶等效热参数模型,并用2D转移模型对集群温控负荷进行建模。在负荷控制方面,有采用模型预测控制,提前对温控负荷进行调整,实现滚动优化。然而现有的2D转移模型未能充分考虑负荷异构性且忽略环境温度时变特性,系统矩阵为常数矩阵,这将为负荷优化控制带来不利影响。
并且现有的控制模型在遴选控制对象的方面优化较少。一方面,输出性能的优劣依赖于给定跟踪信号的时变特性,无法保障优越的响应效果;另一方面,未考虑到如何将最优控制信号和最优负荷控制对象紧密联系在一起,以实现更好的控制效果。
发明内容
针对现有技术存在的问题,本发明目的为发明一种集群温控负荷系统 调频方法、系统、电子设备及存储介质,本发明的方法在控制精度、响应速度和负荷参与需求响应公平性等方面的综合性能较优。
为达到上述目的,本发明采用了以下技术方案:
一种集群温控负荷系统调频方法,包括:
获取集群温控负荷初始化参数;
根据单体空调负荷模型建立2D状态仓集群温控负荷模型,并确定当前集群温控负荷可调度容量;
根据2D状态仓集群温控负荷模型和当前集群温控负荷可调度容量,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵;
根据集群温控负荷一次调频模型计算集群温控负荷一次调频功率变化值,根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值,进而得到集群温控负荷总功率变化值;
根据状态转移矩阵建立集群温控负荷优化控制模型,集群温控负荷总功率变化值采用滚动优化求解得到集群温控负荷控制方案;
基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴选,并输出可控负荷遴选结果。
作为本发明的进一步改进,所述根据单体空调负荷模型建立2D状态仓集群温控负荷模型,并确定当前集群温控负荷可调度容量,具体包含以下步骤:
根据等效热参数模型的二阶离散化微分方程,建立单体空调负荷模型;
根据集群温控负荷当前开关状态将其分为关闭群和开启群;分别根据用户舒适度室内空气温度上下限值
Figure PCTCN2022137165-appb-000001
和室内物质温度的上下限值
Figure PCTCN2022137165-appb-000002
将温度区间等长度分为N i/2个室内空气温度小区间和N m/2个室内物质温度小区间,形成N a*N m/4个关闭群和开启群的状态仓;进而形成2D状态仓转移模型;
取2D状态仓中所有当期关闭仓,温度从低到高,除去第一个最接近容许温度下限状态仓所有的空调设备额定功率之和为最大可上调功率容量,加上当前已经开启空调负荷功率之和作为可调度调节的最大功率容许值P max;取2D状态仓中所有当前开启仓,温度从低到高,除去最后一个最接近容许温度上限的状态仓所有的空调设备额定功率之和为最大可下调功率容量,得到集群空调群组可调度调节的最小功率容许值P min
作为本发明的进一步改进,等效热参数模型的二阶离散化微分方程为:
Figure PCTCN2022137165-appb-000003
其中,Δt表示仿真步长,θ a表示空调负荷的室内气体温度;θ m表示空调负荷的室内物质温度;θ s表示空调负荷所处的环境温度;R a、R m分别为室内空气和室内物质的等值热阻;C a、C m分别为空气温度和物质温度的等值热容;当空调负荷开启时,Q a为空调额定功率;关闭时,Q a等于0。
作为本发明的进一步改进,所述根据2D状态仓集群温控负荷模型和当前集群温控负荷可调度容量,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵,包含以下步骤:
在2D状态仓集群温控负荷模型的基础上,建立集群温控负荷状态空间模型,用时变离散状态空间方程表示:
x(k+1)=A(k)x(k)+B(k)u(k)
y(k)=C(k)x(k)       (2)
其中,x(k)表示第k时刻系统状态向量,表达式如下:
x(k)=[x 1,off(k)x 2,off(k)…x N/2,off(k)x 1+N/2,on(k)x 2+N/2,on(k)…x N,on(k)] T  (3)
式中,N表示状态仓总数,N=N a*N m/2,其元素x i,off(k)表示第k时刻关闭群状态仓i内的空调负荷个数除以总负荷数比值,i=1,2,…,N/2;x j,on(k)表示第k时刻开启群状态仓j内的空调负荷个数除以总负荷数比值,j=N/2+1,N/2+2,…,N;x(k+1)表示第k+1时刻系统状态向量;
A(k)表示第k时刻的系统矩阵,其元素A ij(k)表示在第k个时间步长内,状态仓j的空调负荷转移到状态仓i的转移概率;
u(k)表示第k时刻的控制信号;
B(k)表示第k时刻的输入矩阵,其元素B ij(k)表示在u(k)作用下,需要进行开关动作的状态仓j的空调负荷转移到状态仓i的转移概率,表示如下:
Figure PCTCN2022137165-appb-000004
其中,diag表示对角矩阵,diag sub表示副对角矩阵;
C(k)表示第k时刻的输出矩阵,即当前时刻k各状态仓空调负荷的平均功率向量,表示如下:
C(k)=mP ave(k)*S      (5)
其中,
Figure PCTCN2022137165-appb-000005
m表示空调负荷总个数,P ave(k)表示第k时刻开启群各状态仓空调负荷的平均功率,P agg(k)表示第k时刻空调负荷群聚合输出功率的观测值;S表示各状态仓的开关函数向量,S i表示状态仓i的开关状态,w i表示第i个负荷的开关状态,0为关状态,1为开状态;
y(k)表示第k时刻集群空调负荷模型输出功率;
随机选取初始时刻的空间状态作为仿真初始状态,设定所有空调负荷初始温度在θ -_ETP和θ +_ETP间均匀分布;
基于式(7)对m个空调负荷进行模特卡洛随机模拟,得到每个空调负荷各个仿真时刻的时间-温度热运行曲线;
Figure PCTCN2022137165-appb-000006
式中:x s表示空调温度的设定值;x in,t表示t时刻的室内温度;Δx表示空调温度设定值允许的偏移值;s AC,t表示t时刻空调的工作状态,值为0表示空调关闭,值为1表示开启;
根据每个空调负荷在不同仿真时刻的2D温度,依次进行状态仓编号;
统计在相邻仿真时刻k到k+1间,状态仓i的空调负荷转移到状态仓j的个数,i,j=1,2,…,N;
计算一阶马尔可夫链第k时段状态仓i的空调负荷转移到状态仓j的状态转移概率:
Figure PCTCN2022137165-appb-000007
式中,n i,j(k)表示第k时段状态仓i的空调负荷转移到状态仓j的个数;n i(k)表示第k时段状态仓i中发生状态转移的总负荷数;N表示状态仓总数;
根据i和j的取值不同,可得各个状态转移概率p i,j(k),从而得到状态转移矩阵P(k)。
作为本发明的进一步改进,所述根据集群温控负荷一次调频模型计算 集群温控负荷一次调频功率变化值,具体包括以下步骤:
根据集群温控负荷的频率调节特性设置集群温控负荷一次调频系数;
建立集群温控负荷一次调频系统模型;所述集群温控负荷一次调频系统模型包括单区域调频系统模型、集群空调群组调频模块;单区域调频系统模型为具有积分调节系统的闭环系统,单区域调频系统模型的功率调整信号分别通过系统二次调频传递函数、发电机调速器传递函数及原动机传递函数转换为汽轮机输入功率调整变量,输入功率调整变量和负荷波动变量参与系统调频;集群空调群组调频模块包括空调群组的调频死区、调频系数、可调度潜力上下限外和空调响应时间延迟;
在所述集群温控负荷一次调频系统模型基础上,每台空调设备监测到频率偏差信号后,计算得到集群温控负荷一次调频功率变化值ΔP AC
作为本发明的进一步改进,所述根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值,具体包括以下步骤:
在集群温控负荷一次调频系统模型的基础上,建立集群温控负荷的二次调频仿真模型;集群温控负荷的二次调频仿真模型在集群空调一次调频模型的基础上,集群空调群组接收的控制信号包括一次调频系统频率偏差和二次调频的区域AGC计算出的系统功率偏差信号的整定值,区域AGC功率偏差整定值由整定倍率r决定;
其中,r表示区域AGC功率偏差在空调群组上的整定倍率,由下式进行计算:
Figure PCTCN2022137165-appb-000008
式中,ΔP′ g(s)表示为火电机组二次调频功率偏差信号,ΔP′ AC(s)表示为集群空调群组二次调频功率偏差信号,ΔP c(s)为区域AGC计算出的二次调频功率偏差信号;
在每台空调设备监测到的区域AGC二次调频功率偏差信号后,根据式(12)计算得到集群温控负荷总功率变化值ΔP AC′(s)。
作为本发明的进一步改进,所述根据状态转移矩阵建立集群温控负荷优化控制模型,集群温控负荷总功率变化值采用滚动优化求解得到集群温控负荷控制方案,具体包括以下步骤:
设预测时长为p,设当前k时刻条件下,第k+p时刻的预测状态为x(k+ζ|k),ζ=1,2,…,p;根据集群温控负荷状态空间模型,建立第k+1时刻到k+p时刻的状态方程:
X(k)=A P(k)x(k|k)+B P(k)U(k)    (13)
其中,
Figure PCTCN2022137165-appb-000009
Figure PCTCN2022137165-appb-000010
A P内分块矩阵A p=[A(k+p-1)…A(k+1)A(k)]表示当前k时刻下,系统第k+p时刻状态转移矩阵的预测值,其元素A p(i,j)表示仅已知当前时刻k各状态仓负荷个数向量x(k|k),系统在第k+p个时间步长内,状态仓j的空调负荷转移到状态仓i的转移概率预测值;
采用最小化跟踪误差作为目标函数,其优化控制目标函数为:
Figure PCTCN2022137165-appb-000011
其中,W err表示模型输出与实际目标值的跟踪误差权重系数矩阵,D(k)=diag{C(k+1)C(k+2)…C(k+p)},R(k)=[r(k+1)r(k+2)…r(k+p)] T,而r(k+ζ)表示在k+ζ时刻输出目标轨迹值;
通过以集群输出跟踪误差最小为目标,将集群空调负荷的优化控制模型转化得到二次规划函数:
Figure PCTCN2022137165-appb-000012
并对二次规划函数进行滚动优化求解,求解之后便得控制时域p*Δt内状态仓空调负荷开关个数构成的优化控制序列,在当前调度时刻仅下发该优化序列的第一个分量u *(k|k);等待下一个调度周期到来,重复上述滚动优化过程,得到集群温控负荷控制方案。
作为本发明的进一步改进,所述基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴选,并输出可控负荷遴选结果,包含以下步骤:
建立基于归一化温度距离的排序指标:
Figure PCTCN2022137165-appb-000013
其中,NTD i,k是第i个空调负荷在当前k时刻的归一化温度距离;δ表示温度死区,用户舒适度温度上、下限值θ high和θ low之差;θ i,t表示第i个空调负荷在当前k时刻的温度,O k和C k分别表示在当前k时刻的开启群和关闭群,m是空调负荷总个数;
建立基于功率相似度的排序指标:
Figure PCTCN2022137165-appb-000014
其中,SIM i, (p,q)为状态仓(p,q)中空调负荷功率与所需调整功率的相似度指数,P i为第i个空调负荷的额定功率,P aim,(p,q)为状态仓(p,q)需要响应的目标功率;N (p,q)表示状态仓(p,q)中的空调负荷个数;
建立基于累计控制次数的排序指标:
NC i,k=(C i,k-C k,min)/(C k,max-C k,max)     (20)
其中,C i,k为空调负荷i在k时刻的累计控制次数,C k,min和C k,max表示当前时刻k负荷已被控次数的最小值和最大值;
基于基于归一化温度距离的排序指标、建立基于功率相似度的排序指标及建立基于累计控制次数的排序指标,得到多尺度优先级综合指数Γ open如下式所示:
Figure PCTCN2022137165-appb-000015
其中,K T,K S和K C分别为相应的权重系数;
根据多尺度优先级综合指数Γ open的取值大小对开状态群对应的各2D状态仓内空调负荷进行排序,按优先级从高到低进行可控负荷遴选,得到可控负荷遴选结果。
一种集群温控负荷系统调频系统,包括:
获取单元,配置为获取集群温控负荷初始化参数;
负荷模型建立单元,配置为根据单体空调负荷模型建立2D状态仓集群温控负荷模型,并确定当前集群温控负荷可调度容量;
空间模型建立单元,配置为根据2D状态仓集群温控负荷模型和当前集群温控负荷可调度容量,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵;
调频模型计算单元,配置为根据集群温控负荷一次调频模型计算集群温控负荷一次调频功率变化值,根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值,进而得到集群温控负荷总功率变化值;
控制模型优化单元,配置为根据状态转移矩阵建立集群温控负荷优化控制模型,集群温控负荷总功率变化值采用滚动优化求解得到集群温控负荷控制方案;
可控负荷遴选单元,配置为基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴选,并输出可控负荷遴选结果。
一种电子设备,包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序,所述处理器执行所述计算机程序时实现 所述集群温控负荷系统调频方法的步骤。
一种计算机可读存储介质,所述计算机可读存储介质存储有计算机程序,所述计算机程序被处理器执行时实现所述集群温控负荷系统调频方法的步骤。
本发明的有益效果体现在:
本发明提供了一种集群温控负荷系统灵活调频的方法,基于2D状态仓时变马尔可夫链的集群温控负荷建模方法,在此基础上建立集群温控负荷电力系统一次和二次调频模型,最后创新性提出一种基于归一化温度距离、功率相似度和累计控制次数的多尺度优先级遴选负荷的优化控制方法。给出了集群空调群组在负荷聚合商组织下参与电力系统一次及二次调频的控制流程,验证了集群空调群组参与系统调频的有效性,并且在负荷波动时可改善电网的频率特性,增强系统运行安全性。本发明提出的方法可以在负荷波动时有效改善电网频率特性,且相比较于传统方法,温控负荷建模精度更高,其控制方法在控制精度、响应速度和负荷参与需求响应公平性等方面的综合性能较优。
附图说明
图1为本发明实施例提供的一种集群温控负荷系统调频方法的流程示意图。
图2为本发明实施例提供的一种示例性的单体空调的热运行特性图。
图3为本发明实施例提供的一种示例性的2D状态仓转移模型的示意图。
图4为本发明实施例提供的一种示例性的集群温控负荷调节功率输出随系统频率的变化曲线示意图。
图5为本发明实施例提供的一种示例性的基于集群温控负荷的单区域一次调频仿真模型的示意图。
图6为本发明实施例提供的一种示例性的基于集群温控负荷的单区域一次及二次调频仿真模型的示意图。
图7为本发明实施例提供的一种集群温控负荷控制方法的流程示意图。
图8为本发明实施例提供的一种集群温控负荷系统调频系统的结构示意图;
图9为本发明实施例提供的一种电子设备结构示意图。
具体实施方式
下面将参考附图并结合实施例来详细说明本发明。需要说明的是,在不冲突的情况下,本申请中的实施例及实施例中的特征可以相互组合。
以下详细说明均是示例性的说明,旨在对本发明提供进一步的详细说明。除非另有指明,本发明所采用的所有技术术语与本申请所属领域的一 般技术人员的通常理解的含义相同。本发明所使用的术语仅是为了描述具体实施方式,而并非意图限制根据本发明的示例性实施方式。
相关术语解释
负荷调频:当电力系统频率偏离额定值时,可以通过频率偏差量对负荷进行控制,从而达到减小频率波动,维持电力系统运行安全的作用。
温控负荷:温控负荷是一种以恒温加热器控制的电力负荷,具有一定的热能存储能力。在实际电力系统中,温控负荷多指居民或小型商业用户的空调负荷。
需求响应:即电力需求响应,是指电力供应商通过电价信号或补偿激励等手段引导电力用户用电模式的行为,从而达到减少或增加电力负荷,保障电网稳定运行。
模型预测控制:模型预测控制是一种基于模型的闭环优化控制方法,其核心思想是滚动的时域策略。基本思想是基于当前时刻预测系统未来动态,求解约束规划问题得到当前最优解,等系统更新状态后,预测时域向前滚动,直到达到要求。
多尺度优先级排序:在本发明在特指一种基于负荷归一化温度距离、功率相似度和累计控制次数三个指标的负荷遴选方法。当控制命令下达时,通过多尺度优先级排序进行负荷选择,实施控制。
如图1所示,为本发明提供一种集群温控负荷系统调频方法,包括以下步骤:
S1,获取集群温控负荷初始化参数;
S2,根据单体空调负荷模型建立2D状态仓集群温控负荷模型,并确定当前集群温控负荷可调度容量;
S3,根据2D状态仓集群温控负荷模型和当前集群温控负荷可调度容量,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵;
S4,根据集群温控负荷一次调频模型计算集群温控负荷一次调频功率变化值,根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值,进而得到集群温控负荷总功率变化值;
S5,根据状态转移矩阵建立集群温控负荷优化控制模型,集群温控负荷总功率变化值采用滚动优化求解得到集群温控负荷控制方案;
S6,基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴选,并输出可控负荷遴选结果。
参见图1,上述步骤相关的具体步骤如下:
读入N台空调空调相关参数、室温T t0 k,in和外温度T t0 out
读入初始时刻空调群组可调度容量[ΔP t0 min,ΔP t0 max];
确定t时刻各状态仓空调运行状态及室温T t0 k,in,确定t时刻空调群组可调度容量[ΔP t min,ΔP t max];
一次调频:空调终端自动监测频率偏差信号并将其转换为功率信号;负荷聚合商根据控制算法得到具体开关信号,并发送至空调群组;空调群组根据开关信号,切换设备,响应动作;
二次调频:接受区域AGC指令,将功率偏差信号发送至负荷聚合商;负荷聚合商根据控制算法得到具体开关信号,并发送至空调群组;空调群组根据开关信号,切换设备,响应动作;
t=t+1,更新T t out数据;
判断k<k max是否成立,若为是,则继续确定t+1时刻各状态仓空调运行状态及室温T t0 k,in,确定t+1时刻空调群组可调度容量,并执行调频步骤,若为否,则结束。
本发明集群温控负荷系统调频方法,体现在集群温控负荷运行状态建模、集群温控负荷调频模型构建以及优化控制算法三个方面,具体而言:
1)充分考虑集群温控负荷异构性和多样性基础上提出基于2D状态仓转移时变马尔可夫链的集群温控负荷建模方法,证明了相比传统时不变马尔可夫链建模方法,该方法具有更高的建模精度。
2)集群温控负荷一次和二次调频的建模方法,充分利用了负荷侧需求响应潜力,有一定的可行性和优越性。
3)基于模型预测滚动优化控制的集群温控负荷控制方法,增加了多尺度优先级排序的负荷遴选过程,并进一步仿真分析该方法在控制精度、响应速度和负荷参与需求响应公平性等方面的综合性能较优。
以下结合具体实施例和附图对本发明内容具体说明如下:
一种集群温控负荷系统灵活调频的方法,包括以下步骤:
步骤1:采集集群温控负荷初始化参数,初始化参数包括当前N台温控负荷所处开关状态、室内温度、室外温度等;
步骤2:建立2D状态仓集群温控负荷模型;
所述步骤2包含以下具体步骤:
21)建立单体空调负荷模型:等效热参数模型的二阶离散化微分方程为:
Figure PCTCN2022137165-appb-000016
其中,Δt表示仿真步长,θ a表示空调负荷的室内气体温度;θ m表示空调负荷的室内物质温度;θ s表示空调负荷所处的环境温度;R a、R m分别为室内空气和室内物质的等值热阻;C a、C m分别为空气温度和物质温度的等值热容;当空调负荷开启时,Q a为空调额定功率;关闭时,Q a等于0。
图2所示为单体空调热运行特性,其中θ +_ETP、θ -_ETP分别表示空调负荷设定温度的调节上限和下限;θ s_ETP表示空调负荷的温度设定值;τ on为空调设备的开启时间;τ off为空调设备的开启时间。
22)建立状态仓:根据集群温控负荷当前开关状态将其分为关闭群和开启群;对于关闭群,根据用户舒适度室内空气温度上下限值
Figure PCTCN2022137165-appb-000017
和室内物质温度的上下限值
Figure PCTCN2022137165-appb-000018
将温度区间等长度分为N i/2个室内空气温度小区间和N m/2个室内物质温度小区间,形成N a*N m/4个状态仓;对于开启群,采用同样做法。形成如图3所示2D状态仓转移模型。
23)确定当前集群温控负荷可调度容量:取2D状态仓中所有当期关闭仓,温度从低到高,除去第一个最接近容许温度下限状态仓所有的空调设备额定功率之和为最大可上调功率容量,加上当前已经开启空调负荷功率之和即为可调度调节的最大功率容许值P max;取2D状态仓中所有当前开启仓,温度从低到高,除去最后一个最接近容许温度上限的状态仓所有的空调设备额定功率之和为最大可下调功率容量,即为集群空调群组可调度调节的最小功率容许值P min
步骤3:基于马尔科夫链求解状态仓转移概率;所述步骤3包含以下具体步骤:
31)建立集群温控负荷状态空间模型:在步骤2所述的2D状态仓转移模型的基础上,进一步用时变离散状态空间方程表示:
x(k+1)=A(k)x(k)+B(k)u(k)
y(k)=C(k)x(k)      (2)
其中,x(k)表示第k时刻系统状态向量,表达式如下:
x(k)=[x 1,off(k)x 2,off(k)…x N/2,off(k)x 1+N/2,on(k)x 2+N/2,on(k)…x N,on(k)] T   (3)
式中N表示状态仓总数,N=N a*N m/2,下同。其元素x i,off(k)表示第k时刻关闭群状态仓i内的空调负荷个数除以总负荷数比值,i=1,2,…,N/2;x j,on(k)表示第k时刻开启群状态仓j内的空调负荷个数除以总负荷数比值,j=N/2+1,N/2+2,…,N。x(k+1)表示第k+1时刻系统状态向量。
A(k)表示第k时刻的系统矩阵,其元素A ij(k)表示在第k个时间步长内,状态仓j的空调负荷转移到状态仓i的转移概率。
u(k)表示第k时刻的控制信号,即当前时刻k各状态仓内的空调负荷需要切换的百分比;当该信号为正值时表示开启动作,为负值时表示关闭动作。
B(k)表示第k时刻的输入矩阵,其元素B ij(k)表示在u(k)作用下,需要进行开关动作的状态仓j的空调负荷转移到状态仓i的转移概率,表示如下:
Figure PCTCN2022137165-appb-000019
其中,diag表示对角矩阵,diag sub表示副对角矩阵。
C(k)表示第k时刻的输出矩阵,即当前时刻k各状态仓空调负荷的平均功率向量,表示如下:
C(k)=mP ave(k)*S      (5)
其中,
Figure PCTCN2022137165-appb-000020
m表示空调负荷总个数,P ave(k)表示第k时刻开启群各状态仓空调负荷的平均功率,P agg(k)表示第k时刻空调负荷群聚合输出功率的观测值。S表示各状态仓的开关函数向量,S i表示状态仓i的开关状态,w i表示第i个负荷的开关状态,0为关状态,1为开状态。
y(k)表示第k时刻集群空调负荷模型输出功率。
32)随机选取初始时刻的空间状态作为仿真初始状态,设定所有空调负荷初始温度在θ -_ETP和θ +_ETP间均匀分布;
33)基于式(7)对m个空调负荷进行模特卡洛随机模拟,得到每个空调负荷各个仿真时刻(仿真时长为一天,采样间隔为1分钟,共1440个点)的时间-温度热运行曲线;
Figure PCTCN2022137165-appb-000021
式中:x s表示空调温度的设定值;x in,t表示t时刻的室内温度;Δx表示空调温度设定值允许的偏移值;s AC,t表示t时刻空调的工作状态,值为0表示空调关闭,值为1表示开启。
34)根据每个空调负荷在不同仿真时刻的2D温度,依次进行状态仓编号;
35)统计在相邻仿真时刻k到k+1间,状态仓i的空调负荷转移到状态仓j的个数(i,j=1,2,…,N);
36)计算一阶马尔可夫链第k时段状态仓i的空调负荷转移到状态仓j的状态转移概率:
Figure PCTCN2022137165-appb-000022
式中,n i,j(k)表示第k时段状态仓i的空调负荷转移到状态仓j的个数;n i(k)表示第k时段状态仓i中发生状态转移的总负荷数;N表示状态仓总数。
37)由步骤31)可知,根据i和j的取值不同,可得各个状态转移概率p i,j(k),从而得到状态转移矩阵P(k)。
由于A(k)的任意一列j表示当前第k时段状态仓i的空调负荷转移到状态仓1到N的状态转移概率,而P(k)的任意一行i表示当前第k时段状态仓i的空调负荷转移到状态仓1到N的状态转移概率,故A(k)=P T(k).
步骤4:建立集群温控负荷一次调频模型;所述步骤4包含以下具体步骤:
41)设置集群温控负荷一次调频系数K AC *:集群温控负荷的频率调节特性如图4所示,根据下式整定集群温控负荷一次调频系数K AC *
Figure PCTCN2022137165-appb-000023
Figure PCTCN2022137165-appb-000024
Figure PCTCN2022137165-appb-000025
式中,K g *表示区域火电机组统一的一次调频系数标么值,Δf *表示系统频率偏差的标幺值,K AC *表示集群空调群组一次调频系数的标幺值,ΔK AC *表示集群空调群组总功率变化的标幺值。
42)建立集群温控负荷一次调频系统模型:集群温控负荷的单区域一次调频仿真模型如图5所示。如图5所示,集群温控负荷一次调频系统模型分为传统单区域调频系统模型、集群空调群组调频模块两部分。传统单区域调频系统模型为具有积分调节系统的闭环系统,功率调整信号分别通过系统二次调频传递函数、发电机调速器传递函数、原动机传递函数转换为汽轮机输入功率调整变量,和负荷波动变量一起参与系统调频。集群空调群组调频模块由空调群组的调频死区、调频系数、可调度潜力上下限外,和空调响应时间延迟组成。
需要说明的是,系统二次调频传递函数为
Figure PCTCN2022137165-appb-000026
发电机调速器传递函数为
Figure PCTCN2022137165-appb-000027
原动机传递函数为
Figure PCTCN2022137165-appb-000028
负荷波动变量为ΔP K(s),调频系数为K AC,空调响应时间延迟为
Figure PCTCN2022137165-appb-000029
其中,K n为调速器静态增益,T n为调速器时间常数,R为调速器的调差系数;K n为汽轮机静态增益,T T为汽轮机时间常数,K r为再热系数,T r为再热时间常数;K P为二次调频比例调节系数,K I为二次调频积分调节系数。T AC为温控负荷响应时间延迟,取值为0.1s~0.5s之间。
43)计算集群温控负荷一次调频功率变化值:在32)所述模型基础上,每台空调设备监测到频率偏差信号后,根据式(9)到(11)计算得到集群温控负荷总功率变化值ΔP AC
步骤5:建立集群温控负荷二次调频模型;所述步骤5包含以下具体步骤:
51)建立集群温控负荷二次调频系统模型:在42)中所述模型的基础上,建立集群温控负荷的二次调频仿真模型如图6所示。
在集群空调一次调频模型的基础上,集群空调群组接受的控制信号包括系统频率偏差Δf(一次调频)和区域AGC计算出的系统功率偏差信号的整定值(二次调频),区域AGC功率偏差整定值由整定倍率r决定。其余结构与集群空调一次调频相同。
其中,r表示区域AGC功率偏差在空调群组上的整定倍率,由下式进行计算:
Figure PCTCN2022137165-appb-000030
式中,ΔP′ g(s)表示为火电机组二次调频功率偏差信号,ΔP′ AC(s)表示为集群空调群组二次调频功率偏差信号,ΔP c(s)为区域AGC计算出的二次调频功率偏差信号。
52)计算集群温控负荷二次调频功率变化值:在41)所述模型基础上,每台空调设备监测到的区域AGC二次调频功率偏差信号后,根据式(12)计算得到集群温控负荷总功率变化值ΔP AC′(s)。
步骤6:基于MPC的滚动优化集群温控负荷控制;所述步骤6包含以下具体步骤:
如图7所示,集群温控负荷控制方法主要包括步骤:
步骤一:建立集群温控负荷2D状态仓转移模型;
步骤二:建立集群温控负荷时变离散状态空间方程;
步骤三:基于模型预测算法得到k时刻控制模型;
步骤四:多尺度优先级排序指标进行对象遴选;
步骤五:执行模型预测最优控制信号;
在步骤五之后,判断k<k max是否成立,若为是,则继续针对k+1时刻从步骤二开始依次执行各步骤,若为否,则结束。
经过步骤2和步骤3建模之后,通过步骤4或步骤5得到一次调频或一次及二次调频信号,集群温控负荷将通过以下步骤进行调控。
61)建立预测系统状态方程:设预测时长为p,设当前k时刻条件下,第k+p时刻的预测状态为x(k+ζ|k),(ζ=1,2,…,p)。根据步骤31)中式(2),可推导写出第k+1时刻到k+p时刻的状态方程:
X(k)=A P(k)x(k|k)+B P(k)U(k)     (13)
此处,
Figure PCTCN2022137165-appb-000031
Figure PCTCN2022137165-appb-000032
A P内分块矩阵A p=[A(k+p-1)…A(k+1)A(k)]表示当前k时刻下,系统第k+p时刻状态转移矩阵的预测值,其元素A p(i,j)表示仅已知当前时刻k各状态仓负荷个数向量x(k|k),系统在第k+p个时间步长内,状态仓j的空调负荷转移到状态仓i的转移概率预测值。B P完全类似,不再赘述。
62)建立集群温控负荷优化控制模型:采用最小化跟踪误差作为目标函数,其优化控制目标函数为:
Figure PCTCN2022137165-appb-000033
其中,W err表示模型输出与实际目标值的跟踪误差权重系数矩阵,本发明设为单位矩阵;D(k)=diag{C(k+1)C(k+2)…C(k+p)},R(k)=[r(k+1)r(k+2)…r(k+p)] T,而r(k+ζ)表示在k+ζ时刻输出目标轨迹值。
基于上文的2D状态仓转移模型,控制变量U(k)取值范围应在-1到1之间;状态变量X(k)取值范围应在0到1之间。因此,通过以集群输出跟踪误差最小为目标,可将集群空调负荷的优化控制模型转化为如下所示的二次规划问题:
Figure PCTCN2022137165-appb-000034
滚动优化求解:式(17)约束条件中的0和1均是向量形式,同时,该二次规划模型可以通过调用MATLAB优化工具箱提供的二次规划函数进行求解。求解之后便得控制时域p*Δt内状态仓空调负荷开关个数构成的优化控制序列,在当前调度时刻仅下发该优化序列的第一个分量u *(k|k)。等待下一个调度周期到来,重复上述滚动优化过程。
步骤7:基于多尺度优先级排序进行可控负荷遴选;所述步骤7包含以下具体步骤:
71)建立基于归一化温度距离的排序指标:
Figure PCTCN2022137165-appb-000035
其中,NTD i,k是第i个空调负荷在当前k时刻的归一化温度距离,为取值范围在0和1之间的无量纲系数。δ表示温度死区,即用户舒适度温度上、 下限值θ high和θ low之差。θ i,t表示第i个空调负荷在当前k时刻的温度,O k和C k分别表示在当前k时刻的开启群和关闭群,m是空调负荷总个数。
72)建立基于功率相似度的排序指标:
Figure PCTCN2022137165-appb-000036
其中,SIM i, (p,q)为状态仓(p,q)中空调负荷功率与所需调整功率的相似度指数,P i为第i个空调负荷的额定功率,P aim,(p,q)为状态仓(p,q)需要响应的目标功率。N (p,q)表示状态仓(p,q)中的空调负荷个数。
73)建立基于累计控制次数的排序指标:
NC i,k=(C i,k-C k,min)/(C k,max-C k,max)      (20)
其中,C i,k为空调负荷i在k时刻的累计控制次数,C k,min和C k,max表示当前时刻k负荷已被控次数的最小值和最大值。
74)基于多尺度优先级进行负荷排序:可控开状态群的空调负荷i在时刻k的综合排序参考值Γ open如下式所示:
Figure PCTCN2022137165-appb-000037
K T,K S和K C分别为相应的权重系数,K T值越大,用户的热舒适度越能得到保证;K S越大,负荷响应的精度越高;K C越大,负荷响应的公平性越好,分别设为0.3、0.3、0.4。
根据多尺度优先级综合指数Γ open的取值大小对开状态群对应的各2D状态仓内空调负荷进行排序,Γ open值越小表示该空调负荷在该状态仓内的优先级越高。按优先级从高到低,对温控负荷进行控制,实现调频功能。
如图8所示,本发明提供一种集群温控负荷系统调频系统,包括:
获取单元10,配置为获取集群温控负荷初始化参数;
负荷模型建立单元11,配置为根据单体空调负荷模型建立2D状态仓集群温控负荷模型,并确定当前集群温控负荷可调度容量;
空间模型建立单元12,配置为根据2D状态仓集群温控负荷模型和当前集群温控负荷可调度容量,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵;
调频模型计算单元13,配置为根据集群温控负荷一次调频模型计算集群温控负荷一次调频功率变化值,根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值,进而得到集群温控负荷总功率变化值;
控制模型优化单元14,配置为根据状态转移矩阵建立集群温控负荷优化控制模型,集群温控负荷总功率变化值采用滚动优化求解得到集群温控负荷控制方案;
可控负荷遴选单元15,配置为基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴选,并输出可控负荷遴选结果。
如图9所示,本发明第三个目的是提供一种电子设备,包括存储器20、处理器21以及存储在所述存储器20中并可在所述处理器21上运行的计算机程序,所述处理器执行所述计算机程序时实现所述集群温控负荷系统调频方法的步骤。此外,如图9所示,电子设备还可以包括通讯接口22,用于与外部其他设备进行通讯,例如,进行数据传输等,本发明实施例不作限定。
所述集群温控负荷系统调频方法,包括:
获取集群温控负荷初始化参数;
根据单体空调负荷模型建立2D状态仓集群温控负荷模型,并确定当前集群温控负荷可调度容量;
根据2D状态仓集群温控负荷模型和当前集群温控负荷可调度容量,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵;
根据集群温控负荷一次调频模型计算集群温控负荷一次调频功率变化值,根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值,进而得到集群温控负荷总功率变化值;
根据状态转移矩阵建立集群温控负荷优化控制模型,集群温控负荷总功率变化值采用滚动优化求解得到集群温控负荷控制方案;
基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴选,并输出可控负荷遴选结果。
本发明第四个目的是提供一种计算机可读存储介质,所述计算机可读存储介质存储有计算机程序,所述处理器执行所述计算机程序时实现所述集群温控负荷系统调频方法的步骤。
所述集群温控负荷系统调频方法,包括:
获取集群温控负荷初始化参数;
根据单体空调负荷模型建立2D状态仓集群温控负荷模型,并确定当前集群温控负荷可调度容量;
根据2D状态仓集群温控负荷模型和当前集群温控负荷可调度容量,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵;
根据集群温控负荷一次调频模型计算集群温控负荷一次调频功率变化值,根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值,进而得到集群温控负荷总功率变化值;
根据状态转移矩阵建立集群温控负荷优化控制模型,集群温控负荷总功率变化值采用滚动优化求解得到集群温控负荷控制方案;
基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴 选,并输出可控负荷遴选结果。
本领域内的技术人员应明白,本发明的实施例可提供为方法、系统、或计算机程序产品。因此,本发明可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且,本发明可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
最后应当说明的是:以上实施例仅用以说明本发明的技术方案而非对其限制,尽管参照上述实施例对本发明进行了详细的说明,所属领域的普通技术人员应当理解:依然可以对本发明的具体实施方式进行修改或者等同替换,而未脱离本发明精神和范围的任何修改或者等同替换,其均应涵盖在本发明的权利要求保护范围之内。
工业实用性
本发明提供了一种集群温控负荷系统灵活调频的方法,基于2D状态仓时变马尔可夫链的集群温控负荷建模方法,在此基础上建立集群温控负荷电力系统一次和二次调频模型,最后创新性提出一种基于归一化温度距离、功率相似度和累计控制次数的多尺度优先级遴选负荷的优化控制方法。给出了集群空调群组在负荷聚合商组织下参与电力系统一次及二次调频的控制流程,验证了集群空调群组参与系统调频的有效性,并且在负荷波动时可改善电网的频率特性,增强系统运行安全性。本发明提出的方法可以在 负荷波动时有效改善电网频率特性,且相比较于传统方法,温控负荷建模精度更高,其控制方法在控制精度、响应速度和负荷参与需求响应公平性等方面的综合性能较优。

Claims (11)

  1. 一种集群温控负荷系统调频方法,包括:
    获取集群温控负荷初始化参数;
    根据单体空调负荷模型建立2D状态仓集群温控负荷模型,并确定当前集群温控负荷可调度容量;
    根据2D状态仓集群温控负荷模型和当前集群温控负荷可调度容量,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵;
    根据集群温控负荷一次调频模型计算集群温控负荷一次调频功率变化值,根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值,进而得到集群温控负荷总功率变化值;
    根据状态转移矩阵建立集群温控负荷优化控制模型,集群温控负荷总功率变化值采用滚动优化求解得到集群温控负荷控制方案;
    基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴选,并输出可控负荷遴选结果。
  2. 根据权利要求1所述的方法,其中,
    所述根据单体空调负荷模型建立2D状态仓集群温控负荷模型,并确定当前集群温控负荷可调度容量,具体包含以下步骤:
    根据等效热参数模型的二阶离散化微分方程,建立单体空调负荷模型;
    根据集群温控负荷当前开关状态将其分为关闭群和开启群;分别根据用户舒适度室内空气温度上下限值
    Figure PCTCN2022137165-appb-100001
    和室内物质温度的上下限值
    Figure PCTCN2022137165-appb-100002
    将温度区间等长度分为N i/2个室内空气温度小区间和N m/2个室内物质温度小区间,形成N a*N m/4个关闭群和开启群的状态仓;进而形成2D状态仓转移模型;
    取2D状态仓中所有当期关闭仓,温度从低到高,除去第一个最接近容许温度下限状态仓所有的空调设备额定功率之和为最大可上调功率容量,加上当前已经开启空调负荷功率之和作为可调度调节的最大功率容许值P max;取2D状态仓中所有当前开启仓,温度从低到高,除去最后一个最接近容许温度上限的状态仓所有的空调设备额定功率之和为最大可下调功率容量,得到集群空调群组可调度调节的最小功率容许值P min
  3. 根据权利要求2所述的方法,其中,
    等效热参数模型的二阶离散化微分方程为:
    Figure PCTCN2022137165-appb-100003
    θ m(k+1)=bθ m(k)+(1-b)θ a(k)
    Figure PCTCN2022137165-appb-100004
    其中,Δt表示仿真步长,θ a表示空调负荷的室内气体温度;θ m表示空调负荷的室内物质温度;θ s表示空调负荷所处的环境温度;R a、R m分别为室内空气和室内物质的等值热阻;C a、C m分别为空气温度和物质温度的等值热容;当空调负荷开启时,Q a为空调额定功率;关闭时,Q a等于0。
  4. 根据权利要求1所述的方法,其中,
    所述根据2D状态仓集群温控负荷模型和当前集群温控负荷可调度容量,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵,包含以下步骤:
    在2D状态仓集群温控负荷模型的基础上,建立集群温控负荷状态空间模型,用时变离散状态空间方程表示:
    x(k+1)=A(k)x(k)+B(k)u(k)
    y(k)=C(k)x(k)    (2)
    其中,x(k)表示第k时刻系统状态向量,表达式如下:
    x(k)=[x 1,off(k)x 2,off(k)…x N/2,off(k)x 1+N/2,on(k)x 2+N/2,on(k)…x N,on(k)] T  (3)
    式中,N表示状态仓总数,N=N a*N m/2,其元素x i,off(k)表示第k时刻关闭群状态仓i内的空调负荷个数除以总负荷数比值,i=1,2,…,N/2;x j,on(k)表示第k时刻开启群状态仓j内的空调负荷个数除以总负荷数比值,j=N/2+1,N/2+2,…,N;x(k+1)表示第k+1时刻系统状态向量;
    A(k)表示第k时刻的系统矩阵,其元素A ij(k)表示在第k个时间步长内,状态仓j的空调负荷转移到状态仓i的转移概率;
    u(k)表示第k时刻的控制信号;
    B(k)表示第k时刻的输入矩阵,其元素B ij(k)表示在u(k)作用下,需要进行开关动作的状态仓j的空调负荷转移到状态仓i的转移概率,表示如下:
    Figure PCTCN2022137165-appb-100005
    其中,diag表示对角矩阵,diag sub表示副对角矩阵;
    C(k)表示第k时刻的输出矩阵,即当前时刻k各状态仓空调负荷的平均功率向量,表示如下:
    C(k)=mP ave(k)*S  (5)
    其中,
    S=[S 1 S 2 … S N]
    Figure PCTCN2022137165-appb-100006
    Figure PCTCN2022137165-appb-100007
    Figure PCTCN2022137165-appb-100008
    Figure PCTCN2022137165-appb-100009
    m表示空调负荷总个数,P ave(k)表示第k时刻开启群各状态仓空调负荷的平均功率,P agg(k)表示第k时刻空调负荷群聚合输出功率的观测值;S表示各状态仓的开关函数向量,S i表示状态仓i的开关状态,w i表示第i个负荷的开关状态,0为关状态,1为开状态;
    y(k)表示第k时刻集群空调负荷模型输出功率;
    随机选取初始时刻的空间状态作为仿真初始状态,设定所有空调负荷初始温度在θ -_ETP和θ +_ETP间均匀分布;
    基于式(7)对m个空调负荷进行模特卡洛随机模拟,得到每个空调负荷各个仿真时刻的时间-温度热运行曲线;
    Figure PCTCN2022137165-appb-100010
    式中:x s表示空调温度的设定值;x in,t表示t时刻的室内温度;Δx表示空调温度设定值允许的偏移值;s AC,t表示t时刻空调的工作状态,值为0表示空调关闭,值为1表示开启;
    根据每个空调负荷在不同仿真时刻的2D温度,依次进行状态仓编号;
    统计在相邻仿真时刻k到k+1间,状态仓i的空调负荷转移到状态仓j的个数,i,j=1,2,…,N;
    计算一阶马尔可夫链第k时段状态仓i的空调负荷转移到状态仓j的状态转移概率:
    Figure PCTCN2022137165-appb-100011
    式中,n i,j(k)表示第k时段状态仓i的空调负荷转移到状态仓j的个数;n i(k)表示第k时段状态仓i中发生状态转移的总负荷数;N表示状态仓总数;
    根据i和j的取值不同,可得各个状态转移概率p i,j(k),从而得到状态转移矩阵P(k)。
  5. 根据权利要求1所述的方法,其中,
    所述根据集群温控负荷一次调频模型计算集群温控负荷一次调频功率变化值,具体包括以下步骤:
    根据集群温控负荷的频率调节特性设置集群温控负荷一次调频系数;
    建立集群温控负荷一次调频系统模型;所述集群温控负荷一次调频系统模型包括单区域调频系统模型、集群空调群组调频模块;单区域调频系统模型为具有积分调节系统的闭环系统,单区域调频系统模型的功率调整信号分别通过系统二次调频传递函数、发电机调速器传递函数及原动机传递函数转换为汽轮机输入功率调整变量,输入功率调整变量和负荷波动变量参与系统调频;集群空调群组调频模块包括空调群组的调频死区、调频系数、可调度潜力上下限外和空调响应时间延迟;
    在所述集群温控负荷一次调频系统模型基础上,每台空调设备监测到频率偏差信号后,计算得到集群温控负荷一次调频功率变化值ΔP AC
  6. 根据权利要求1所述的方法,其中,
    所述根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值,具体包括以下步骤:
    在集群温控负荷一次调频系统模型的基础上,建立集群温控负荷的二次调频仿真模型;集群温控负荷的二次调频仿真模型在集群空调一次调频模型的基础上,集群空调群组接收的控制信号包括一次调频系统频率偏差和二次调频的区域AGC计算出的系统功率偏差信号的整定值,区域AGC功率偏差整定值由整定倍率r决定;
    其中,r表示区域AGC功率偏差在空调群组上的整定倍率,由下式进行计算:
    Figure PCTCN2022137165-appb-100012
    式中,ΔP g′(s)表示为火电机组二次调频功率偏差信号,ΔP AC′(s)表示为集群空调群组二次调频功率偏差信号,ΔP c(s)为区域AGC计算出的二次调频功率偏差信号;
    在每台空调设备监测到的区域AGC二次调频功率偏差信号后,根据式(12)计算得到集群温控负荷总功率变化值ΔP AC′(s)。
  7. 根据权利要求1所述的方法,其中,
    所述根据状态转移矩阵建立集群温控负荷优化控制模型,集群温控负荷总功率变化值采用滚动优化求解得到集群温控负荷控制方案,具体包括以下步骤:
    设预测时长为p,设当前k时刻条件下,第k+p时刻的预测状态为x(k+ζ|k),ζ=1,2,…,p;根据集群温控负荷状态空间模型,建立第k+1时刻到k+p时刻的状态方程:
    X(k)=A P(k)x(k|k)+B P(k)U(k)   (13)
    其中,
    A P(k)=[A(k) A(k+1)A(k) … A(k+p-1)…A(k)] T
    Figure PCTCN2022137165-appb-100013
    X(k)=[x(k+1|k) x(k+2|k) … x(k+p|k)] T
    U(k)=[u(k|k)u(k+1|k)…u(k+p-1|k)] T  (15)
    A P内分块矩阵A p=[A(k+p-1)…A(k+1)A(k)]表示当前k时刻下,系统第k+p时刻状态转移矩阵的预测值,其元素A p(i,j)表示仅已知当前时刻k各状态仓负荷个数向量x(k|k),系统在第k+p个时间步长内,状态仓j的空调负荷转移到状态仓i的转移概率预测值;
    采用最小化跟踪误差作为目标函数,其优化控制目标函数为:
    Figure PCTCN2022137165-appb-100014
    其中,W err表示模型输出与实际目标值的跟踪误差权重系数矩阵,D(k)=diag{C(k+1)C(k+2)…C(k+p)},R(k)=[r(k+1)r(k+2)…r(k+p)] T,而r(k+ζ)表示在k+ζ时刻输出目标轨迹值;
    通过以集群输出跟踪误差最小为目标,将集群空调负荷的优化控制模型转化得到二次规划函数:
    Figure PCTCN2022137165-appb-100015
    并对二次规划函数进行滚动优化求解,求解之后便得控制时域p*Δt内状态仓空调负荷开关个数构成的优化控制序列,在当前调度时刻仅下发该优化序列的第一个分量u *(k|k);等待下一个调度周期到来,重复上述滚动优化过程,得到集群温控负荷控制方案。
  8. 根据权利要求1所述的方法,其中,
    所述基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴选,并输出可控负荷遴选结果,包含以下步骤:
    建立基于归一化温度距离的排序指标:
    Figure PCTCN2022137165-appb-100016
    其中,NTD i,k是第i个空调负荷在当前k时刻的归一化温度距离;δ表示温度死区,用户舒适度温度上、下限值θ high和θ low之差;θ i,t表示第i个空调负荷在当前k时刻的温度,O k和C k分别表示在当前k时刻的开启群和关闭群,m是空调负荷总个数;
    建立基于功率相似度的排序指标:
    Figure PCTCN2022137165-appb-100017
    其中,SIM i, (p,q)为状态仓(p,q)中空调负荷功率与所需调整功率的相似度指数,P i为第i个空调负荷的额定功率,P aim,(p,q)为状态仓(p,q)需要响应的目标功率;N (p,q)表示状态仓(p,q)中的空调负荷个数;
    建立基于累计控制次数的排序指标:
    NC i,k=(C i,k-C k,min)/(C k,max-C k,max)  (20)
    其中,C i,k为空调负荷i在k时刻的累计控制次数,C k,min和C k,max表示当前时刻k负荷已被控次数的最小值和最大值;
    基于基于归一化温度距离的排序指标、建立基于功率相似度的排序指标及建立基于累计控制次数的排序指标,得到多尺度优先级综合指数Γ open如下式所示:
    Figure PCTCN2022137165-appb-100018
    其中,K T,K S和K C分别为相应的权重系数;
    根据多尺度优先级综合指数Γ open的取值大小对开状态群对应的各2D状态仓内空调负荷进行排序,按优先级从高到低进行可控负荷遴选,得到可控负荷遴选结果。
  9. 一种集群温控负荷系统调频系统,包括:
    获取单元,配置为获取集群温控负荷初始化参数;
    负荷模型建立单元,配置为根据单体空调负荷模型建立2D状态仓集群温控负荷模型,并确定当前集群温控负荷可调度容量;
    空间模型建立单元,配置为根据2D状态仓集群温控负荷模型和当前集群温控负荷可调度容量,建立集群温控负荷状态空间模型,基于马尔科夫链求解状态仓转移概率,从而得到状态转移矩阵;
    调频模型计算单元,配置为根据集群温控负荷一次调频模型计算集群温控负荷一次调频功率变化值,根据集群温控负荷二次调频模型计算集群温控负荷二次调频功率变化值,进而得到集群温控负荷总功率变化值;
    控制模型优化单元,配置为根据状态转移矩阵建立集群温控负荷优化控制模型,集群温控负荷总功率变化值采用滚动优化求解得到集群温控负荷控制方案;
    可控负荷遴选单元,配置为基于多尺度优先级排序对所述集群温控负荷控制方案进行可控负荷遴选,并输出可控负荷遴选结果。
  10. 一种电子设备,包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序,所述处理器执行所述计算机程序时 实现权利要求1-8任一项所述集群温控负荷系统调频方法的步骤。
  11. 一种计算机可读存储介质,所述计算机可读存储介质存储有计算机程序,所述计算机程序被处理器执行时实现权利要求1-8任一项所述集群温控负荷系统调频方法的步骤。
PCT/CN2022/137165 2022-02-25 2022-12-07 一种集群温控负荷系统调频方法、系统、电子设备及存储介质 WO2023160110A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210182642.2A CN114512994A (zh) 2022-02-25 2022-02-25 一种集群温控负荷系统调频方法、系统、设备及介质
CN202210182642.2 2022-02-25

Publications (1)

Publication Number Publication Date
WO2023160110A1 true WO2023160110A1 (zh) 2023-08-31

Family

ID=81553954

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/137165 WO2023160110A1 (zh) 2022-02-25 2022-12-07 一种集群温控负荷系统调频方法、系统、电子设备及存储介质

Country Status (2)

Country Link
CN (1) CN114512994A (zh)
WO (1) WO2023160110A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117033894A (zh) * 2023-10-10 2023-11-10 北京英沣特能源技术有限公司 一种闭式蓄冷罐瞬间失水防负压抽真空检测系统
CN117291363A (zh) * 2023-09-08 2023-12-26 国网山东省电力公司营销服务中心(计量中心) 一种基于异构温控负荷模型的负荷调控方法及系统
CN117434989A (zh) * 2023-12-20 2024-01-23 福建省力得自动化设备有限公司 一种电气柜内环境调控系统及调控方法

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114512994A (zh) * 2022-02-25 2022-05-17 中国电力科学研究院有限公司 一种集群温控负荷系统调频方法、系统、设备及介质
CN115268536B (zh) * 2022-08-02 2024-05-14 阳光电源股份有限公司 一种储能系统的温度控制方法及相关装置
CN116093978B (zh) * 2023-03-17 2023-08-18 天津大学 一种面向电网调频的自备电厂和负载协调控制方法、系统、电子设备和可读介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108287477A (zh) * 2018-02-09 2018-07-17 福建和盛高科技产业有限公司 基于模型预测和多尺度优先级的集群温控负荷控制方法
CN109103893A (zh) * 2018-09-27 2018-12-28 华北电力大学 一种集群温控负荷参与电网agc的辅助调频方法
US20190293774A1 (en) * 2018-03-20 2019-09-26 Institute Of Electronics, Chinese Academy Of Sciences Method and device for designing and optimizing multi-degree-of-freedom frequency-modulation signal
CN112994036A (zh) * 2021-02-02 2021-06-18 中南大学 一种基于模型预测的温控负荷参与微电网调控方法及系统
CN114512994A (zh) * 2022-02-25 2022-05-17 中国电力科学研究院有限公司 一种集群温控负荷系统调频方法、系统、设备及介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108287477A (zh) * 2018-02-09 2018-07-17 福建和盛高科技产业有限公司 基于模型预测和多尺度优先级的集群温控负荷控制方法
US20190293774A1 (en) * 2018-03-20 2019-09-26 Institute Of Electronics, Chinese Academy Of Sciences Method and device for designing and optimizing multi-degree-of-freedom frequency-modulation signal
CN109103893A (zh) * 2018-09-27 2018-12-28 华北电力大学 一种集群温控负荷参与电网agc的辅助调频方法
CN112994036A (zh) * 2021-02-02 2021-06-18 中南大学 一种基于模型预测的温控负荷参与微电网调控方法及系统
CN114512994A (zh) * 2022-02-25 2022-05-17 中国电力科学研究院有限公司 一种集群温控负荷系统调频方法、系统、设备及介质

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117291363A (zh) * 2023-09-08 2023-12-26 国网山东省电力公司营销服务中心(计量中心) 一种基于异构温控负荷模型的负荷调控方法及系统
CN117033894A (zh) * 2023-10-10 2023-11-10 北京英沣特能源技术有限公司 一种闭式蓄冷罐瞬间失水防负压抽真空检测系统
CN117033894B (zh) * 2023-10-10 2023-12-22 北京英沣特能源技术有限公司 一种闭式蓄冷罐瞬间失水防负压抽真空检测系统
CN117434989A (zh) * 2023-12-20 2024-01-23 福建省力得自动化设备有限公司 一种电气柜内环境调控系统及调控方法
CN117434989B (zh) * 2023-12-20 2024-03-12 福建省力得自动化设备有限公司 一种电气柜内环境调控系统及调控方法

Also Published As

Publication number Publication date
CN114512994A (zh) 2022-05-17

Similar Documents

Publication Publication Date Title
WO2023160110A1 (zh) 一种集群温控负荷系统调频方法、系统、电子设备及存储介质
Wang et al. A distributed control scheme of thermostatically controlled loads for the building-microgrid community
Meng et al. Coordinated dispatch of virtual energy storage systems in smart distribution networks for loading management
CN104698843B (zh) 一种基于模型预测控制的数据中心节能控制方法
Wang et al. Operating a commercial building HVAC load as a virtual battery through airflow control
WO2022127652A1 (zh) 一种空调聚合控制方法、系统和调控装置
Xia et al. Hierarchical control of thermostatically controlled loads oriented smart buildings
CN109149648A (zh) 一种自适应宽度动态规划智能发电控制方法
Zhao et al. Model predictive based frequency control of power system incorporating air-conditioning loads with communication delay
Zhou et al. A dynamic water-filling method for real-time HVAC load control based on model predictive control
CN104699051A (zh) 一种温控设备需求响应控制方法
Wang et al. Coordinated load shedding control scheme for recovering frequency in islanded microgrids
Malik et al. Optimal control based on scheduling for comfortable smart home environment
CN108224692B (zh) 考虑室外气温预测误差的空调柔性控制响应能力预测方法
Ma et al. Control and Communication for Demand Response with Thermostatically Controlled Loads
CN105958513B (zh) 一种基于hvac系统的区域电网功率动态平衡的方法
CN109301817A (zh) 一种考虑需求响应的多时间尺度源网荷协调调度方法
Zhang et al. Distributed intelligent load management and control system
Ma et al. Controller design and parameter optimization of aggregated thermostatically controlled loads for frequency regulation
CN113078629A (zh) 一种用于集群温控负荷聚合商功率调控的聚合体功率分配模型及分布式一致性控制方法
Yu et al. Research on Intelligent Air Conditioning Optimization Control Algorithms Based on Neural Networks and Heuristic Algorithms
Zheng et al. A Distributed Coordination Strategy for Heterogeneous Building Flexible Thermal Loads in Responding to Smart Grids
Li et al. Control of fuzzy discrete event systems and its application to air conditioning system
CN117646978A (zh) 一种模型预测控制的智能空调负荷聚合调控系统及方法
Yazdkhasti et al. Design of a smart controller agent for demand-side management with low payback effect

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22928370

Country of ref document: EP

Kind code of ref document: A1