CN115940275A - Distributed controllable resource operation mode optimization matching and coordination switching control method and device under different power shortage - Google Patents
Distributed controllable resource operation mode optimization matching and coordination switching control method and device under different power shortage Download PDFInfo
- Publication number
- CN115940275A CN115940275A CN202310047063.1A CN202310047063A CN115940275A CN 115940275 A CN115940275 A CN 115940275A CN 202310047063 A CN202310047063 A CN 202310047063A CN 115940275 A CN115940275 A CN 115940275A
- Authority
- CN
- China
- Prior art keywords
- mode
- power
- hybrid
- ders
- layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/50—Photovoltaic [PV] energy
- Y02E10/56—Power conversion systems, e.g. maximum power point trackers
Abstract
The invention discloses a distributed controllable resource operation mode optimization matching and coordination switching control method based on a MAS layered framework, which is oriented to different power shortage situations in a power distribution network. Firstly, aiming at the DERs with unique operating characteristics and complex working modes, a hybrid automata model of five DERs under a hybrid system is respectively constructed on the bottom layer based on the hybrid system and the finite automata and combined with power characteristics of photovoltaic power, wind power, an energy storage unit, a fuel cell and a load. Secondly, in the face of complex operation data generated in the working process of the power distribution network, a selective mode-based aggregation single-layer dependency classification algorithm is applied to the middle layer to process the data, and the optimal coordinated switching control command is selected in a control command set by executing classification operation, so that the intelligence of the whole system is effectively improved. Finally, constraints such as load requirements, power balance, DERs power output limit and the like of the system are comprehensively considered at the upper layer, an optimization scheduling model considering system operation cost, environmental benefits and electric energy quality is provided, and flexible scheduling and optimization operation of each unit of the system are achieved by executing an improved random walk fruit fly optimization algorithm.
Description
Technical Field
The invention relates to a distributed controllable resource operation mode optimization matching and coordination switching control method and device under different power shortages, and belongs to the technical field of smart power grids.
Background
With the increasingly prominent global energy crisis and environmental problems, establishing a safer, more efficient and sustainable energy development and utilization mode has become a main bearing form and an important development direction of future energy systems. The shortage of traditional fossil energy and the problem of environmental pollution caused by combustion become major bottlenecks restricting the development of world economy, and the application of distributed power generation technologies such as wind energy, solar energy, energy storage and the like provides an effective way for solving the problems. At present, our country is greatly promoting the utilization of DERs (distributed energy resources), mainly including photovoltaic power generation, wind power generation, storage batteries, fuel cells, and the like. Incorporating a large number of DERs with varying power output characteristics into a power distribution network can create a number of problems, such as power fluctuations, power quality degradation, etc. Therefore, flexible management and control and coordination control of the DERs operating in different working modes are achieved, and the method has important significance for improving the safe and stable economic operation of the power distribution network.
The optimized operation of the power distribution network needs to be realized through multi-mode switching of the DERs. The operation characteristics of the multi-type distributed controllable resources are different, the working mode is complex, and how to add the description of the dynamic operation characteristics of the DERs into the control strategy to realize the discrete and continuous control in the switching process is important for realizing the multi-mode switching of the DERs. In general, a dynamic system consisting of continuous variable and discrete event interaction effects is considered a hybrid system, and a power distribution network is a typical hybrid system. In conclusion, the traditional power distribution network energy management system is single in management object and simple in structure, and flexible management and control on the multiple types of DERs cannot be achieved through the control means.
Disclosure of Invention
The invention aims to overcome the defects in the prior art, provides a method and a device for controlling the operation mode optimization matching and the coordinated switching of distributed controllable resources under different power deficit, can realize the flexible regulation and control of various distributed controllable resources under different power deficit, and has high intelligence degree and high accuracy.
In order to achieve the purpose, the invention is realized by adopting the following technical scheme:
in a first aspect, the present invention provides a method for controlling optimal matching and coordinated switching of distributed controllable resource operation modes under different power shortages, comprising the following steps:
the method comprises the following steps: aiming at a coordination control target of the power distribution network, establishing a MAS (multi-agent system, MAS) -based power distribution network layered framework, and dividing a power distribution network control structure into an upper layer, a middle layer and a bottom layer;
step two: considering that in the control process, the hybrid automatic machine model comprises discrete control commands for switching different distributed power Sources (DERs) and load unit operation modes and continuous control instructions of an execution end inverter, so that continuous variables are added on the basis of the finite automata concept to construct a hybrid automatic machine model represented by a 6-element array;
step three: respectively establishing corresponding hybrid automaton models aiming at the operating characteristics of a photovoltaic power generation unit, a wind power generation unit, an energy storage unit, a fuel cell and a load;
step four: dividing a local control strategy at the bottom layer into an internal switching control strategy and a continuous dynamic management strategy, wherein the internal switching control strategy takes a coordination control command of a middle layer as a reference, and the continuous dynamic management is designed into a double-ring control method based on SPWM;
step five: designing a coordination control strategy in an intermediate-layer multi-agent structure, constructing a system voltage safety evaluation index, carrying out hierarchical processing, and coordinating the working modes of the DERs and the load units by applying event-triggered hybrid control according to a system voltage safety evaluation result;
step six: in order to reduce the switching times of the operation modes of the DERs as much as possible and improve the intelligence of the system and the rapidity and the accuracy of the execution of the control commands, a polymerization single-layer dependence classification algorithm based on a selective mode is utilized, a coordination control command set is used as a label set, and the optimal coordination switching control command is selected in the control command set by executing classification operation;
step seven: the upper-layer intelligent agent considers power balance, the output power of DERs and storage battery capacity constraint aiming at different power shortage, constructs a comprehensive objective function for reducing operation cost, reducing pollutant emission and improving electric energy quality, solves the multi-objective optimization problem by utilizing an improved random walk fruit fly optimization algorithm, and realizes an upper-layer control strategy.
Further, in step1, in order to implement multi-modal switching, a hierarchical structure based on MAS is established, which includes:
step 1.1, upper layer agent-negotiation agent. Energy management is realized through an optimization process, so that the system obtains the maximum economic, environmental and electric energy quality benefits, and an optimization model is designed based on real-time data in a processing module. The time scale of the optimization process is once per hour.
Step 1.2 middle layer agent-negotiation agent. The working modes of the DERs are coordinated to be switched to ensure the voltage safety of the system, and the time scale of switching control is in the order of hours or minutes.
Step 1.3, a bottom layer intelligent agent, namely a hybrid agent, comprises a reaction layer and a negotiation layer. The reactive layer is set to "sense-and-act" with priority to respond quickly to emergency situations. The negotiation layer is highly intelligent and can control or guide the behavior of the intelligent agent, and the negotiation layer is set to be 'belief-desire-intention', and highly intelligent and controls or plans the behavior of the intelligent agent to realize the desire or intention.
Further, step 2 considers that in the process of issuing the control command, the control command contains both discrete values and continuous values, so that on the basis of the finite automata concept, continuous variables are added to construct a hybrid automata model with 6-element array representation.
The automaton is one of modeling methods of the hybrid system and has the advantage of intuitiveness. A generic automaton model is used to handle discrete events, which if based on continuous dynamic features are added, would form a hybrid automaton model. The hybrid automaton mainly consists of a discrete state space, an enabling state and a continuous state. In each discrete state space, there is a corresponding continuous state change, and when a certain enabling state is satisfied, the hybrid automaton has a chance to migrate from one discrete state space to another discrete state space, and the corresponding continuous state also has a transition.
A typical promiscuous automata model can be represented by a 6-element array: h = (D, L, F, S, F, I). Wherein D = { δ 1 ,δ 2 \8230 } represents a collection of discrete state spaces; l represents a set of a series of continuous state spaces; f = { f 1 (δ 1 ),f 2 (δ 2 ),f 3 (δ 3 ) \8230showsthe change rule of a continuous state space under each discrete state space; s = { S = 1 ,S 2 ,S 3 ,…Represents the mapping between discrete state space and continuous state space; f = { F 1 ,F 2 ,F 3 \8230 } represents the condition for state space transfer; i denotes the initial state.
Further, step 3, constructing a hybrid system model for different DERs, comprising:
step 3.1, constructing a hybrid automaton model of the photovoltaic power generation unit:
in the formula, P pv Is the output power of the photovoltaic power generation unit; p mppt Outputting power for the photovoltaic power generation unit to work in a maximum power point mode; l is the illumination intensity; and M is an illumination intensity threshold value and is determined by the performance of the photovoltaic cell. Aiming at the photovoltaic output characteristic, the photovoltaic power generation unit is set into two working modes: the method comprises the following steps of setting two working modes into a discrete state of a hybrid automatic machine model of the photovoltaic power generation unit in an outage mode and an mppt mode; setting the power output value of the photovoltaic power generation unit to be in a continuous state; the change in the illumination intensity is set as a transfer condition.
Step 3.2, constructing a hybrid automaton model of the wind power generation unit:
in the formula, P w Is the output power of the wind power generation unit; p e The rated output power of the wind power generation unit; v is the real-time wind speed; v. of i To cut into the wind speed; v. of r Rated wind speed; v. of o To cut out the wind speed. According to the operating characteristics of wind power generation, the working modes of the wind power generation system are divided into three types: off mode, mppt mode and P e Mode(s). Setting three working modes as discrete states of the hybrid automaton model of the wind power generation unit according to the hybrid automaton model; setting the power output value of the wind power generation unit to be in a continuous state; the wind speed variation is set as a transfer condition.
Step 3.3, constructing an energy storage unit hybrid automaton model:
according to the characteristics of the battery, the battery is divided into five working modes: charge mode, discharge mode, shutdown mode (three). The three shutdown modes comprise normal shutdown, overcharge shutdown and overdischarge shutdown.
According to the hybrid automata model, when SOC is less than or equal to SOC down When the discharge mode of the storage battery is switched to the over-discharge shutdown mode, the storage battery is switched to the charge mode from the over-discharge shutdown mode when relevant switching commands exist.
When SOC is reached up When the SOC is not more than the SOC, the storage battery is switched from the charging mode to the overcharge shutdown mode, and when relevant switching commands exist, the storage battery is switched from the overcharge shutdown mode to the discharge mode.
When the SOC is down ＜SOC＜SOC up At the same time, the storage battery can be switched among a discharging mode, a charging mode and a normal shutdown mode according to the system requirements.
Step 3.4, constructing a fuel cell hybrid automaton model:
according to the hybrid automata model, two operating modes of the fuel cell were designed: off mode, rated output mode.
Step 3.5, constructing a load hybrid automaton model:
the loads are divided into two categories of interruptible loads and non-interruptible loads, and the interruptible loads are higher than the non-interruptible loads in power supply priority. According to the hybrid robot model, two working modes of the load unit are designed: normal operation mode, load shedding mode.
Step 3.6 sets the initial state of the hybrid automata model of all units, activates the discrete state space, sets it to logic "1", and sets the remaining discrete state space to logic "0". Thus, can obtain
A photovoltaic power generation unit: d = { δ 1 ,δ 2 }＝[1,0](ii) a A wind power generator unit: d = { δ 1 ,δ 2 ,δ 3 }＝[1,0,0](ii) a A battery cell: d = { δ 1 ,δ 2 ,δ 3 ,δ 4 ,δ 5 }＝[1,0,0,0,0](ii) a Fuel cell unit：D＝{δ 1 ,δ 2 }＝[1,0](ii) a A load cell: d = { δ 1 ,δ 2 }＝[1,0]。
Further, step 4 divides the local control policy into an internal switching control policy and a continuous dynamic management policy, which are specifically as follows:
step 4.1 for the photovoltaic power generation unit, the internal switching control command is as follows:
for the wind power generation unit, the internal switching control commands are as follows:
the battery cell control commands are as follows:
and 4.2, designing a double-loop control method based on SPWM for the inverter of the DERs, wherein the outer loop is designed to be a controller adopting droop control, and the inner loop is designed to be a controller under a dq rotation coordinate system.
Further, step5, constructing a system voltage safety evaluation index and performing grading processing, and coordinating the DERs and the switching of the load unit working modes by applying event-triggered hybrid control according to a system voltage safety evaluation result, specifically as follows:
step 5.1, extracting the voltage sequence of the ith bus of the power distribution network system by using a wide area signal measurement method, wherein the voltage sequence is expressed as
Then, the average value of the instantaneous voltage of the ith node at the jth time obtained by the voltage sequence is represented as follows:
the voltage deviation value of the ith node at the jth time is expressed as:
in the formula (I), the compound is shown in the specification,
is the actual voltage value of the ith node at the jth moment. />
Thus, the voltage safety evaluation index of the ith node at the jth time can be expressed as:
and 5.2, fusing the voltage safety evaluation indexes of all the nodes by using the information fusion method based on the T-S fuzzy neural network.
Use of
In place of x = (x) 1 ,x 2 ,…,x n ) The output of the fuzzy neural network can be normalized to the required voltage safety evaluation index as the input of the fuzzy neural network. The output of the fuzzy neural network is represented as:
in the formula (I), the compound is shown in the specification,
weights for the fuzzy neural network; theta j Is the product of membership and connectivity. And obtaining the comprehensive voltage safety evaluation index u of each node after information fusion.
Step 5.3, grading the evaluation index u:
and when the u is more than or equal to 0.9 or more than or equal to 1.1, the power distribution network system stops running in a voltage collapse state.
When u is more than 0.95 and more than or equal to 0.9, the power distribution network system supplies insufficient energy to the load, and the fuel cell works or carries out load shedding operation.
When u is more than 0.98 and more than or equal to 0.95, the power distribution network system supplies insufficient energy to the load, and the storage battery performs discharging operation to balance power.
When u is more than 1.02 and more than or equal to 0.98, the system voltage is in a normal fluctuation range.
When u is more than 1.05 and more than or equal to 1.02, the energy supply of the power distribution network system is more than the load demand, and the storage battery is charged to balance the power.
When 1.1 > u ≧ 1.05, the energy supply of the power distribution grid system is much greater than the load demand, the fuel cell switches to shutdown mode or the interruptible load is restored.
Step 5.4 takes into account the trigger duration, the handover sequence and the time interval sufficiently and will
Used as trigger events, u and SOC are used as trigger conditions. To (X)>
The specific design is as follows:
c 2 :H S ((δ 12 ,δ 22 ,δ 23 ,δ 33 ,δ 42 ),(F 12 ,F 25 ,F 26 ,F 33 ,F 42 ))
in the formula, c n For a labelset, n ∈ [1,2, \8230;, 15]；
Is the trigger duration; Δ t is handoverA time interval. And the coordination control command is utilized to flexibly regulate and control the action change of the DERs in the power distribution network.
Further, step 6 uses the aggregation single-layer dependency classification algorithm based on the selective mode to use the coordination control command set as a tag set, and selects the optimal coordination switching control command in the control command set by performing a classification operation, which is as follows:
step 6.1, mining the attributes and attribute values: performing classification operation by adopting a selective mode-based aggregation single-layer dependence classification algorithm, using values of U and SOC in the running process of a power distribution network system as continuous attributes, and using working modes of a storage battery, a fuel cell and a load as discrete attributes, and mining the attributes; the attribute set is as follows: x = { (U, a) 1 ),(SOC,a 2 ),(Battery,a 3 ),(Fuel,a 4 ),(Load,a 5 )}，a 1 ～a 5 Attribute values corresponding to five attributes.
Step 6.2 process data of continuous attributes: the sample space (attribute value) of U is divided into seven continuous intervals, and seven levels are defined as seven set values from low to high with reference to the fuzzy control: "VL", "ML", "L", "Z", "H", "MH", "VH". Similarly, the sample space of the SOC is divided into five consecutive intervals, and five levels are defined from low to high as five set values: "VL", "L", "Z", "H", "VH"; processing discrete attribute data: for the battery cell, the sample space is set to {0,1,2}, where "0" represents the charging mode, "1" represents the discharging mode, and "2" represents the shutdown mode; for the fuel cell unit, the sample space is set to {0,1}, where "0" represents the rated output mode and "1" represents the shutdown mode; for the load cell, the sample space is set to {0,1}, with "0" indicating the normal operation mode and "1" indicating the load shedding mode.
Step 6.3 assume that the dataset D has n attributes and the training instance is denoted X = (a) 1 ,a 2 La n ) Wherein a is i (1. Ltoreq. I.ltoreq.n) is the value of the example X at the ith attribute. The class of the training instance belongs to class C = { C 1 ,c 2 Lc m One of them, the category of the instance is denoted below by c. The probability P (c | X) that the class of X is c can be expressed as:
let the attribute set contained in the set f be { a } 1 ,a 2 ,L,a i And obtaining a Bayesian network:
applying the pattern classification capability to the Bayesian network, aggregating the corresponding conditions of all patterns to obtain a Bayesian probability prediction formula based on selectivity:
step 6.4, in order to weaken the dependency relationship between the attribute of the selective mode and other attributes in the Bayes algorithm, a selective mode-based aggregation single-layer dependency classification algorithm is adopted, and the final probability prediction formula is as follows:
wherein H = { j | i +1 ≦ j ≦ n ^ F (a) j )≥m}，F(a j ) Is that the attribute value contains a j Is limited by the parameter m to achieve the support required for the conditional probability estimation.
And 6.5, selecting corresponding attributes based on a selective mode aggregation single-layer dependence classification algorithm, then mining attribute values according to data generated by the operation of the power distribution network, meanwhile, using a coordination control command set as a label set, and realizing the selection of the optimal coordination switching control command by executing classification operation.
Further, step 7 is to establish a comprehensive objective function for the upper-layer agent, and apply a random walk drosophila optimization algorithm to achieve the objective optimization, specifically as follows:
step 7.1, an optimized scheduling objective function is constructed:
cost objective function:
in the formula: i: the number of DERs; s: the mode of operation of the DERs; alpha is alpha is : when the distribution network system is in the working mode, alpha is =1, otherwise, α is ＝0；r i : fuel cost of ith DERs, for renewable energy, r i ＝0；
The active output of the ith DERs in the s mode; e is : a consumption feature function; m is a group of i : maintenance costs of the ith DERs, and->
Proportionally mixing; c i : the startup cost of the ith DERs; beta is a beta is : when the ith DERs are operated in s mode, β is =1, otherwise, beta is ＝0。
Carbon emission objective function:
in the formula: f 2 : actual carbon emission when the power distribution network operates; pi 1 : the carbon emission intensity of unit active power output of the thermal power generating unit; pi 2 : the carbon emission intensity of the unit active power output of the gas turbine; p is in.t : acquiring electric quantity from the main network in unit time period; p out.t : generating power of the gas turbine in unit time t; the clean energy carbon emission is 0.
Power quality objective function:
in the formula: sigma is And the power quality coefficient of the ith DERs in the s mode. Sigma is Determined by a voltage safety assessment indicator.
Step 7.2, constructing constraint conditions:
and power balance constraint:
P l ＝P pv +P w +P F +(-1) n P bat (14)
in the formula, P l Is the load demand of the system; p is pv Is the output power of the photovoltaic power generation unit; p w Is the output power of the wind power generation unit; p bat Power output/absorbed for the battery; n is an element of [0,1 ]]N =1 when the battery is in the charging mode, and n =0 when the battery is in the discharging mode; p F Is the output power of the fuel cell.
Output power constraints for DERs:
in the formula, P i Output power for the ith DERs;
minimum output power for the ith DERs; />
The maximum output power of the ith DERs.
Battery capacity constraint:
SOC down ＜SOC＜SOC up (16)
in the formula, SOC is the state of charge of the storage battery; SOC down Is the lowest capacity state value of the storage battery; SOC up The highest capacity state value of the storage battery.
And 7.3, based on an improved drosophila optimization algorithm, the method is used for realizing optimized scheduling.
Step 7.4 the multi-objective optimization problem is processed in the form of weight coefficients, and the energy management can be summarized as solving the following objective optimization function that satisfies the constraint condition:
in the formula, omega 1 、ω 2 、ω 3 Weight coefficient, ω, representing each objective function 1 +ω 2 +ω 3 ＝1。
In a second aspect, the present invention provides a distributed controllable resource operation mode optimization matching and coordination switching control device under different power shortages, the device includes:
the architecture building module: the method comprises the steps that a power distribution network layered framework based on the MAS is established aiming at a coordination control target of the power distribution network, and a power distribution network control structure is divided into an upper layer, a middle layer and a bottom layer;
a model construction module: adding continuous variables on the basis of a finite automata concept to construct a hybrid automata model with 6-element array representation;
a model building module: the hybrid automatic machine model is used for respectively establishing corresponding hybrid automatic machine models aiming at the operating characteristics of the photovoltaic power generation unit, the wind power generation unit, the energy storage unit, the fuel cell and the load;
a management policy module: the method comprises the steps that a local control strategy at the bottom layer is divided into an internal switching control strategy and a continuous dynamic management strategy, wherein the internal switching control strategy takes a coordination control command of a middle layer as a reference, and the continuous dynamic management is a double-ring control method based on SPWM;
a coordination control module: the method is used for designing a coordination control strategy in an intermediate-layer multi-agent structure, constructing a system voltage safety evaluation index and carrying out hierarchical processing, and coordinating the DER and the switching of the working modes of the load units by applying event-triggered hybrid control according to a system voltage safety evaluation result;
a classification operation module: the method comprises the steps of using a coordination control command set as a tag set by utilizing an aggregation single-layer dependency classification algorithm based on a selective mode, and selecting the optimal coordination switching control command in the control command set by executing a classification operation;
a multi-objective optimization module: the method is used for an upper-layer intelligent agent, and aims at different power shortages, the power balance, the output power of DERs and the storage battery capacity constraint are considered, a comprehensive objective function for reducing the operation cost, reducing the pollutant emission and improving the electric energy quality is constructed, an improved random walk fruit fly optimization algorithm is utilized, the multi-objective optimization problem is solved, and an upper-layer control strategy is realized.
In a third aspect, the present invention provides a distributed controllable resource operation mode optimization matching and coordination switching control device under different power shortages, which is characterized by comprising a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to the first aspect.
Compared with the prior art, the invention has the following beneficial effects:
the method constructs a power distribution network hierarchical coordination framework based on the MAS, and realizes the optimization matching and the coordination switching control of the distributed controllable resource operation mode according to the coordination optimization strategy formulated by the hierarchical framework under different power requirements;
the hybrid automata model constructed based on the hybrid system and the automata can be matched with the operating characteristics of photovoltaic power, wind power, a storage battery, a fuel cell, loads and the like. Aiming at numerous and complicated operation data generated in the operation process of the power distribution network, the machine learning algorithm can effectively process the operation data, and the optimal coordination switching control of the DERs operation mode is realized through a selective mode-based aggregation single-layer dependence classification algorithm. Comprehensively considering the supply and demand balance constraint of the system and the operation constraint of the DERs, constructing an optimization objective function for reducing the operation cost, reducing the pollutant emission and improving the electric energy quality, and solving by an improved random walk fruit fly algorithm, thereby realizing the safe and economic optimization operation of the system.
Drawings
FIG. 1 is a hybrid robot model of a photovoltaic power generation unit;
FIG. 2 is a hybrid automata model of a wind power generation unit;
FIG. 3 is a battery cell hybrid robot model;
FIG. 4 is a fuel cell hybrid robot model;
FIG. 5 is a load cell hybrid robot model;
FIG. 6 is a hierarchical diagram of a MAS-based power distribution network;
FIG. 7 is a block diagram of an outer loop controller based on P-f and Q-U droop control;
FIG. 8 is a diagram of an inner loop controller architecture in a dq rotation coordinate system;
fig. 9 is a flow chart of a method of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
The first embodiment is as follows:
the embodiment provides a distributed controllable resource operation mode optimization matching and coordinated switching control method under different power shortages, and the applied scene is a power distribution network system. In order to solve the problems that the management object of the conventional power distribution network energy management system is single, the management mode is simple, and the efficient regulation and control of the DERs with different characteristics can not be realized, a power distribution network hierarchical structure based on the MAS is designed and is shown in figure 6, and an optimization coordination control strategy is formulated according to the hierarchical control structure. The photovoltaic power generation unit hybrid robot model established based on the hybrid robot model and the discrete state and transfer conditions thereof are shown in fig. 1, table 1 and table 2; the hybrid automatic machine model of the wind power unit and the discrete state and the transfer condition of the hybrid automatic machine model are shown in fig. 2, table 3 and table 4; the storage battery cell hybrid automata model and the discrete states and the transition conditions thereof are shown in fig. 3, table 5, and table 6; the fuel cell hybrid robot model and its discrete states and transition conditions are shown in fig. 4, table 7, and table 8; the load cell hybrid robot model and its discrete states and transition conditions are shown in fig. 5, table 9, and table 10. Aiming at a large amount of complex data generated during the operation of the power distribution network, a selective mode-based aggregated single-layer dependence classification algorithm and an improved random walk fruit fly optimization algorithm are adopted to effectively process the data, the working state of the power distribution network system is identified through classification operation, and the intelligence of the power distribution network and the accuracy of control command execution are improved.
Table 1 discrete state description of a hybrid automaton model for photovoltaic power generation units
| Discrete states | Description of discrete states |
| δ 11 | Shutdown mode for photovoltaic power generation unit |
| δ 12 | Mppt mode of photovoltaic power generation unit |
Table 2 transfer condition description of hybrid robot model of photovoltaic power generation unit
| Transfer conditions | Description of transfer conditions |
| F 11 | Switching a photovoltaic power generation unit from a shutdown mode to an mppt mode |
| F 12 | Switching photovoltaic power generation unit from mppt mode to shutdown mode |
TABLE 3 discrete State description of hybrid automaton model of wind power plant
Table 4 transfer condition description of hybrid automaton model of wind power generation unit
| Transfer conditions | Description of transfer conditions |
| F 21 | Switching a wind power unit from an off-mode to an mppt-mode as the wind speed increases |
| F 22 | Switching the wind power unit from off-mode to P as the wind speed increases e Output mode |
| F 23 | Switching a wind power generation unit from mppt mode to P e Output mode |
| F 24 | Wind power generation unit slave P e Switching output mode to mppt mode |
| F 25 | Switching a wind power unit from mppt mode to off-stream mode |
| F 26 | Wind power generation unit slave P e Switching output mode to off mode |
| F 27 | Switching a wind power unit from an off-mode to an mppt-mode as the wind speed decreases |
| F 28 | Switching a wind power unit from an off-mode to P as wind speed decreases e Output mode |
TABLE 5 discrete State description of Battery cell hybrid robot model
| Discrete states | Description of discrete states |
| δ 31 | Normal shutdown mode for battery cells |
| δ 32 | Discharge mode of battery cell |
| δ 33 | Charging mode of battery cell |
| δ 34 | Over-discharge shutdown mode of battery cell |
| δ 35 | Overcharge shutdown mode for battery cells |
TABLE 6 transfer condition description of Battery cell hybrid robot model
Table 7 discrete state description of fuel cell hybrid automotive model
| Discrete states | Description of discrete states |
| δ 41 | Shutdown mode of fuel cell unit |
| δ 42 | Rated output mode of fuel cell unit |
Table 8 transfer condition description of fuel cell hybrid robot model
| Transfer conditions | Description of transfer conditions |
| F 41 | Switching a fuel cell unit from an off mode to a rated output mode |
| F 42 | Switching a fuel cell unit from a rated output mode to an off mode |
TABLE 9 discrete State description of load cell hybrid robot model
| Discrete states | Description of discrete states |
| δ 51 | Normal operation mode of load cell |
| δ 52 | Load shedding scheme for load cells |
TABLE 10 transfer Condition description of load cell hybrid robot model
| Transfer conditions | Description of transfer conditions |
| F 51 | Switching the load unit from normal operation mode to load shedding mode |
| F 52 | Switching of load cell from load shedding mode to normal operation mode |
The upper layer, the middle layer and the bottom layer respectively relate to an energy management strategy, a coordination control strategy, an internal switching control strategy and a continuous dynamic management strategy; the bottom layer designs local control as hybrid control, designs an internal switching control strategy based on an event trigger mechanism and a hybrid automaton model, realizes outer loop control by using a P-f and Q-U droop control method as shown in FIG. 7, and designs a structure diagram of an inner loop controller under a dq rotation coordinate system as shown in FIG. 8; the middle layer designs safety indexes aiming at safety evaluation, classifies the safety indexes based on T-S fuzzy control as shown in a table 11, and designs a coordination control command set; the upper layer is provided with a main function, and the single-objective optimization process is realized on the basis of multi-objective optimization by adopting an improved random walk fruit fly optimization algorithm. The optimal matching and coordinated switching control of the distributed controllable resource operation modes under different power shortages are realized through a power distribution network hierarchical control structure and by combining corresponding control strategies.
TABLE 11 grading of voltage safety assessment indicators
| | Grade interval | |
| 1 | 0.9≥u | |
| 2 | 0.95＞u≥0.9 | |
| 3 | 0.98＞u≥0.95 | |
| 4 | 1.02＞u≥0.98 | |
| 5 | 1.05＞u≥1.02 | |
| 6 | 1.1＞u≥1.05 | |
| 7 | u≥1.1 |
Step1, aiming at a coordination control target of a power distribution network, establishing a MAS-based power distribution network layered architecture, and dividing a power distribution network control structure into an upper layer, a middle layer and a bottom layer:
(1) Upper agent-negotiation agent. Energy management is realized through an optimization process, so that the system obtains the maximum economic, environmental and electric energy quality benefits, and an optimization model is designed based on real-time data in a processing module. The time scale of the optimization process is once per hour.
(2) Middle tier agent-negotiation agent. The working modes of the DERs are coordinated to be switched to ensure the voltage safety of the system, and the time scale of switching control is in the order of hours or minutes.
(3) The bottom layer intelligent agent-mixed agent comprises a reaction layer and a negotiation layer. The reactive layer is set to "sense-and-act" with priority to respond quickly to emergency situations. The negotiation layer is set to "belief-desire-intention", highly intelligent, controlling or planning agent behavior to achieve its desire or intention.
Step 2, considering that in the control process, the discrete control commands for switching different distributed power supplies and load unit operation modes are contained, and the continuous control commands of the execution end inverter are also contained, so that on the basis of the finite automata concept, continuous variables are added to construct a hybrid automata model represented by a 6-element array:
the automaton has the advantage of intuitiveness as one of modeling methods of the hybrid system. A common automata model is used to process discrete events, which if based on continuous dynamic features are added, would form a hybrid automata model. The hybrid automaton mainly consists of a discrete state space, an enabling state and a continuous state. In each discrete state space, there is a corresponding continuous state change, and when a certain enabling state is satisfied, the promiscuous automaton has the opportunity to migrate from one discrete state space to another discrete state space, and the corresponding continuous state also has the transition.
A typical hybrid automata model can be represented by a 6-element array: h = (D, L, F, S, F, I). Wherein D = { δ 1 ,δ 2 \8230 } represents a collection of discrete state spaces; l represents a set of a series of continuous state spaces; f = { f 1 (δ 1 ),f 2 (δ 2 ),f 3 (δ 3 ) \8230 } shows the change rule of the continuous state space under each discrete state space; s ={S 1 ,S 2 ,S 3 \8230 } represents the mapping between a discrete state space and a continuous state space; f = { F 1 ,F 2 ,F 3 \8230 } represents the condition for state space transfer; i denotes the initial state.
Step 3, respectively establishing corresponding hybrid automaton models aiming at the operating characteristics of the photovoltaic power generation unit, the wind power generation unit, the energy storage unit, the fuel cell and the load; :
(1) Constructing a hybrid automata model of the photovoltaic power generation unit:
in the formula, P pv Is the output power of the photovoltaic power generation unit; p mppt Outputting power for the photovoltaic power generation unit to work in a maximum power point mode; l is the illumination intensity; and M is an illumination intensity threshold value and is determined by the performance of the photovoltaic cell. For the photovoltaic output characteristics, the photovoltaic power generation unit is set to two working modes: the method comprises the following steps of setting a shutdown mode and an mppt mode as discrete states of a hybrid automaton model of the photovoltaic power generation unit; setting the power output value of the photovoltaic power generation unit to be in a continuous state; the change in the illumination intensity is set as a transfer condition.
(2) Constructing a hybrid automaton model of the wind power generation unit:
in the formula, P w Is the output power of the wind power generation unit; p e The rated output power of the wind power generation unit; v is the real-time wind speed; v. of i To cut into the wind speed; v. of r Rated wind speed; v. of o To cut out the wind speed. According to the operating characteristics of wind power generation, the working modes are divided into three types: off mode, mppt mode and P e Mode(s). Setting three working modes as discrete states of the hybrid automaton model of the wind power generation unit according to the hybrid automaton model; send out wind powerSetting the power output value of the electric unit to be in a continuous state; the wind speed variation is set as a transfer condition.
(3) Constructing a hybrid automaton model of the energy storage unit:
according to the characteristics of the battery, the battery is divided into five working modes: charge mode, discharge mode, shutdown mode (three). The three shutdown modes comprise normal shutdown, overcharge shutdown and overdischarge shutdown.
According to the hybrid automata model, when SOC is less than or equal to SOC down When the discharging mode of the storage battery is switched to the over-discharge shutdown mode, the storage battery is switched to the charging mode from the over-discharge shutdown mode when relevant switching commands exist.
When SOC is reached up When the SOC is not more than the SOC, the storage battery is switched from the charging mode to the overcharge shutdown mode, and when relevant switching commands exist, the storage battery is switched from the overcharge shutdown mode to the discharge mode.
When the SOC is down ＜SOC＜SOC up The storage battery can be switched among a discharging mode, a charging mode and a normal shutdown mode according to system requirements.
(4) Constructing a fuel cell hybrid automaton model:
two operating modes of the fuel cell are designed according to a hybrid automata model: off mode, rated output mode.
(5) Constructing a load hybrid automaton model:
the loads are divided into two categories of interruptible loads and non-interruptible loads, and the interruptible loads are higher than the non-interruptible loads in power supply priority. According to the hybrid robot model, two working modes of the load cell are designed: normal operation mode, load shedding mode.
(6) The initial state of the hybrid automaton model of all cells is set, the discrete state space is activated, set to logic "1", and the remaining discrete state space is set to logic "0". Thus, can obtain
A photovoltaic power generation unit: d = { δ 1 ,δ 2 }＝[1,0](ii) a A wind power generator unit: d = { δ 1 ,δ 2 ,δ 3 }＝[1,0,0](ii) a A battery cell: d = { δ 1 ,δ 2 ,δ 3 ,δ 4 ,δ 5 }＝[1,0,0,0,0](ii) a A fuel cell unit: d = { δ 1 ,δ 2 }＝[1,0](ii) a A load cell: d = { δ 1 ,δ 2 }＝[1,0]。
And 4, dividing the local control strategy at the bottom layer into an internal switching control strategy and a continuous dynamic management strategy, wherein the internal switching control strategy takes the coordination control command of the middle layer as a reference, and the continuous dynamic management is designed into a double-ring control method based on SPWM:
(1) For the photovoltaic power generation unit, the internal switching control command is as follows:
for the wind power generation unit, the internal switching control commands are as follows:
the battery cell control commands are as follows:
(2) A dual-loop control method based on SPWM is designed for the inverter of the DERs, wherein an outer loop is designed to be a controller adopting droop control, and an inner loop is designed to be a controller under a dq rotation coordinate system.
The method of the outer loop controller determines that:
in the formula, f and U are respectively the frequency and amplitude of the output voltage; f. of 0 And U 0 Reference values for voltage frequency and amplitude, respectively; p 0 And Q 0 Reference values for active power and reactive power respectively; p and Q are actual values of active power and reactive power respectively; k p And K q The sag factor.
The inner ring control utilizes a park variable ring to change the original three-phase control problem under an abc natural coordinate system into a two-phase control problem under a dq rotating coordinate system.
Step5, designing a coordination control strategy in the middle-layer multi-agent structure, constructing a system voltage safety evaluation index and carrying out hierarchical processing, and coordinating the DERs and the switching of the working modes of the load units by applying event-triggered hybrid control according to a system voltage safety evaluation result:
(1) Extracting the voltage sequence of the ith bus of the power distribution network system by using a wide area signal measurement method, wherein the voltage sequence is expressed as
Then, the average value of the instantaneous voltage of the ith node at the jth time obtained by the voltage sequence is represented as follows:
the voltage deviation value of the ith node at the jth time is expressed as:
in the formula (I), the compound is shown in the specification,
is the actual voltage value of the ith node at the jth moment.
Thus, the voltage safety evaluation index of the ith node at the jth time can be expressed as:
(2) The information fusion method based on the T-S fuzzy neural network fuses voltage safety evaluation indexes of all nodes.
Use of
In place of x = (x) 1 ,x 2 ,…,x n ) As an input to the fuzzy neural network, the output of the fuzzy neural network may then be normalized to the required voltage safety assessmentAnd (4) indexes. The output of the fuzzy neural network is represented as:
in the formula (I), the compound is shown in the specification,
weights for the fuzzy neural network; theta.theta. j Is the product of membership and connectivity.
In the formula, alpha is the learning rate of the fuzzy neural network; e =0.5 (y) s -y m ) 2 The deviation between the actual output value and the target value is also a performance index in the algorithm learning process; y is s Is the expected value of the output; y is m Is the actual output value.
In the formula (I), the compound is shown in the specification,
is a fuzzy set; />
Membership functions for each input.
in the formula (I), the compound is shown in the specification,
and &>
Respectively the center and width of the membership function.
And the comprehensive voltage safety evaluation index of each node obtained after information fusion is marked as u.
(3) Grading the evaluation index u:
and when u is more than or equal to 0.9 and more than or equal to 1.1, the power distribution network system stops running in a voltage collapse state.
When u is more than 0.95 and more than or equal to 0.9, the power distribution network system supplies insufficient energy to the load, and the fuel cell works or carries out load shedding operation.
When u is more than 0.98 and more than or equal to 0.95, the power distribution network system supplies insufficient energy to the load, and the storage battery performs discharging operation to balance power.
When u is more than 1.02 and more than or equal to 0.98, the system voltage is in a normal fluctuation range.
When u is more than 1.05 and more than or equal to 1.02, the energy supply of the power distribution network system is more than the load demand, and the storage battery is charged to balance the power.
When 1.1 > u ≧ 1.05, the energy supply of the distribution grid system is far greater than the load demand, the fuel cell switches to a shutdown mode or the interruptible load is restored.
(4) Takes full account of the trigger duration, the switching order and the time interval and will
Used as trigger events, u and SOC are used as trigger conditions. To (X)>
The specific design is as follows:
c 2 :H S ((δ 12 ,δ 22 ,δ 23 ,δ 33 ,δ 42 ),(F 12 ,F 25 ,F 26 ,F 33 ,F 42 ))
in the formula, c n Is a set of tags, n ∈ [1,2, \8230;, 15]；
Is the trigger duration; Δ t is the switching time interval. And the coordination control command is utilized to flexibly regulate and control the action change of the DERs in the power distribution network.
Step 6, in order to reduce the switching times of the operation modes of the DERs as much as possible and improve the intelligence of the system and the rapidity and the accuracy of the execution of the control commands, a selective mode-based aggregation single-layer dependency classification algorithm is utilized, a coordination control command set is used as a label set, and the optimal coordination switching control commands are selected in the control command set by executing classification operation:
(1) And (3) mining attributes and attribute values: performing classification operation by adopting a selective mode-based aggregation single-layer dependence classification algorithm, using values of U and SOC in the running process of a power distribution network system as continuous attributes, and using working modes of a storage battery, a fuel cell and a load as discrete attributes, and mining the attributes; the attribute set is as follows: x = { (U, a) 1 ),(SOC,a 2 ),(Battery,a 3 ),(Fuel,a 4 ),(Load,a 5 )}，a 1 ～a 5 The attribute values corresponding to the five attributes.
(2) Processing data of continuous attributes: the sample space (attribute value) of U is divided into seven continuous intervals, and seven levels are defined from low to high as seven set values with reference to fuzzy control: "VL", "ML", "L", "Z", "H", "MH", "VH". Similarly, the sample space of the SOC is divided into five consecutive intervals, and five levels are defined from low to high as five set values: "VL", "L", "Z", "H", "VH"; processing discrete attribute data: for the battery cell, the sample space is set to {0,1,2}, where "0" represents the charging mode, "1" represents the discharging mode, and "2" represents the shutdown mode; for the fuel cell unit, the sample space is set to {0,1}, where "0" represents the rated output mode and "1" represents the shutdown mode; for the load cell, the sample space is set to {0,1}, with "0" indicating the normal operation mode and "1" indicating the load shedding mode.
(3) Assuming that the dataset D has n attributes, the training instance is denoted X = (a) 1 ,a 2 La n ) Wherein a is i (1. Ltoreq. I.ltoreq.n) is the value of the example X at the ith attribute. The class of the training instance belongs to class C = { C 1 ,c 2 Lc m One of them, the category of the instance is denoted below by c. The probability P (c | X) that the class of X is c can be expressed as:
let the attribute set contained in the set f be { a } 1 ,a 2 ,L,a i And obtaining a Bayesian network:
applying the pattern classification capability to the Bayesian network, and aggregating the corresponding conditions of all patterns to obtain a selective Bayesian probability prediction formula:
(4) In order to weaken the dependency relationship between the attribute of the selective mode and other attributes in the Bayesian algorithm, an aggregation single-layer dependency classification algorithm based on the selective mode is adopted, and the final probability prediction formula is as follows:
wherein H = { j | i +1 ≦ j ≦ n ^ F (a) j )≥m}，F(a j ) Is that the attribute value contains a j Is limited by the parameter m to achieve the support required for the conditional probability estimation.
(5) And after selecting corresponding attributes, mining attribute values according to data generated by the operation of the power distribution network, and simultaneously, selecting the optimal coordination switching control command by taking a coordination control command set as a label set and executing classification operation.
And 7, the upper-layer intelligent agent considers power balance, the output power of the DERs and the capacity constraint of the storage battery aiming at different power shortage, constructs a comprehensive objective function for reducing the operation cost, reducing pollutant emission and improving the electric energy quality, solves the multi-objective optimization problem by utilizing a random walk fruit fly optimization algorithm, and realizes an upper-layer control strategy:
(1) Constructing an optimized scheduling objective function:
cost objective function:
in the formula: i: the number of DERs; s: the mode of operation of the DERs; alpha is alpha is : when the distribution network system is in the working mode, alpha is =1, otherwise, α is ＝0；r i : fuel cost of ith DERs, for renewable energy, r i ＝0；
The active output of the ith DERs in the s mode; e is : a consumption feature function; m i : maintenance costs of the ith DERs, and->
Proportionally mixing; c i : the startup cost of the ith DERs; beta is a is : when the ith DERs are operated in s mode, β is =1, otherwise, beta is ＝0。
Carbon emission objective function:
in the formula: f 2 : actual carbon emission when the power distribution network operates; pi 1 : the carbon emission intensity of the unit active power output of the thermal power generating unit; pi 2 : the carbon emission intensity of the unit active power output of the gas turbine; p in.t : acquiring electric quantity from the main network in unit time period; p out.t : generating power by the gas turbine in unit time t; the clean energy carbon emission is 0.
Power quality objective function:
in the formula: sigma is The power quality coefficient of the ith DERs in the s mode is shown. Sigma is Determined by a voltage safety evaluation index.
(2) Constructing a constraint condition:
and power balance constraint:
P l ＝P pv +P w +P F +(-1) n P bat (20)
in the formula, P l Is the load demand of the system; p pv Is the output power of the photovoltaic power generation unit; p w Is the output power of the wind power generation unit; p is bat Power output/absorbed for the battery; n is an element of [0,1 ]]N =1 when the battery is in the charging mode, and n =0 when the battery is in the discharging mode; p F Is the output power of the fuel cell.
Output power constraints for DERs:
in the formula, P i Output power for the ith DERs;
minimum output power for the ith DERs; />
The maximum output power of the ith DERs.
Battery capacity constraint:
SOC down ＜SOC＜SOC up (22)
in the formula, SOC is the state of charge of the storage battery; SOC (system on chip) down Is the lowest capacity state value of the storage battery; SOC up The highest capacity state value of the storage battery.
(3) Based on an improved fruit fly optimization algorithm, the method is used for realizing optimized scheduling. The traditional drosophila optimization algorithm steps are as follows:
position initialization:
in the formula: LR is used to set the location range of the fly population; x is the number of start y start The initial position of the fly mass is determined.
Calculate each drosophila position:
in the formula: v is the range of the fruit fly group, namely the radius of the current fruit fly group.
And (3) distance calculation:
in general, D i Is the distance from fly i to the origin, S i Is the concentration of the food, and the distance D i It is related. Smell i The individual odor concentration of fly i can be calculated by any one of the specific functions Fitness ().
And (3) optimal drosophila identification:
step4:[bestSmellbestIndex]＝max(Smell i )(26)
the position and concentration values of the most concentrated flies were recorded as bestIndex and bestsell, respectively.
Optimal fruit fly retention and fly flock location change:
given the location bestIndex of the fly with the largest concentration value, the location of the bee colony will be updated accordingly. x is the number of axis y axis Representing the best fruit fly location.
And iterating the steps until the requirement of precision or iteration times is met.
In the iterative process, the distance between the current solution and the optimal solution cannot be predicted, so that the characteristic of fruit fly random walk (semi-supervision method) is added, the algorithms step1 to step5 are optimized, and the position weight calculation is added:
in the formula: omega i,j To continuously iterate the positional weights of the two fly populations,
expressing the Euclidean distance between two fly groups, the eta calculation formula is as follows:
(4) The multi-objective optimization problem is processed in the form of weight coefficients, and energy management can be summarized as solving the following objective optimization functions that satisfy constraints:
in the formula, ω 1 、ω 2 、ω 3 Weight coefficient, ω, representing each objective function 1 +ω 2 +ω 3 ＝1。
Example two:
the embodiment provides a distributed controllable resource operation mode optimization matching and coordinated switching control device under different power shortages, which includes:
an architecture building module: the method comprises the steps that a power distribution network layered framework based on the MAS is established aiming at a coordination control target of the power distribution network, and a power distribution network control structure is divided into an upper layer, a middle layer and a bottom layer;
a model construction module: adding continuous variables on the basis of a finite automata concept to construct a hybrid automata model with 6-element array representation;
a model building module: the hybrid automatic machine model is used for respectively establishing corresponding hybrid automatic machine models aiming at the operating characteristics of the photovoltaic power generation unit, the wind power generation unit, the energy storage unit, the fuel cell and the load;
a management policy module: the system comprises a local control strategy and a continuous dynamic management strategy, wherein the local control strategy at the bottom layer is divided into an internal switching control strategy and a continuous dynamic management strategy, the internal switching control strategy takes a coordination control command of a middle layer as a reference, and the continuous dynamic management is a double-loop control method based on SPWM;
a coordination control module: the method is used for designing a coordination control strategy in an intermediate-layer multi-agent structure, constructing a system voltage safety evaluation index and carrying out hierarchical processing, and coordinating the DER and the switching of the working modes of the load units by applying event-triggered hybrid control according to a system voltage safety evaluation result;
a classification operation module: the method comprises the steps of utilizing an aggregation single-layer dependency classification algorithm based on a selective mode to use a coordination control command set as a tag set, and selecting the best coordination switching control command in the control command set by executing classification operation;
a multi-objective optimization module: the method is used for the upper-layer intelligent agent, and aims at different power shortage, the power balance, the output power of the DERs and the storage battery capacity constraint are considered, a comprehensive objective function for reducing the operation cost, reducing the pollutant emission and improving the electric energy quality is constructed, an improved random walk fruit fly optimization algorithm is utilized, the multi-objective optimization problem is solved, and an upper-layer control strategy is realized.
The apparatus of this embodiment can be used to implement the method described in the first embodiment.
Example three:
the embodiment provides a distributed controllable resource operation mode optimization matching and coordination switching control device under different power shortages, which is characterized by comprising a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to embodiment one.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application 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, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, it is possible to make various improvements and modifications without departing from the technical principle of the present invention, and those improvements and modifications should be considered as the protection scope of the present invention.
Claims (10)
1. A distributed controllable resource operation mode optimization matching and coordination switching control method under different power shortage is characterized in that: comprises the following steps:
the method comprises the following steps: aiming at a coordination control target of the power distribution network, establishing a power distribution network layered framework based on the MAS, and dividing a power distribution network control structure into an upper layer, a middle layer and a bottom layer;
step two: on the basis of a finite automata concept, adding continuous variables to construct a hybrid automata model represented by a 6-element array;
step three: respectively establishing corresponding hybrid automaton models aiming at the operating characteristics of a photovoltaic power generation unit, a wind power generation unit, an energy storage unit, a fuel cell and a load;
step four: dividing a local control strategy at the bottom layer into an internal switching control strategy and a continuous dynamic management strategy, wherein the internal switching control strategy takes a coordination control command of a middle layer as a reference, and the continuous dynamic management is a double-ring control method based on SPWM;
step five: designing a coordination control strategy in an intermediate-layer multi-agent structure, constructing a system voltage safety evaluation index, carrying out hierarchical processing, and coordinating the working modes of the DERs and the load units by applying event-triggered hybrid control according to a system voltage safety evaluation result;
step six: using a selective mode-based aggregation single-layer dependency classification algorithm, using a coordination control command set as a tag set, and selecting an optimal coordination switching control command in the control command set by executing classification operation;
step seven: the upper-layer intelligent agent considers power balance, the output power of DERs and storage battery capacity constraint aiming at different power shortage, constructs a comprehensive objective function for reducing operation cost, reducing pollutant emission and improving electric energy quality, solves the multi-objective optimization problem by utilizing an improved random walk fruit fly optimization algorithm, and realizes an upper-layer control strategy.
2. The method for distributed controllable resource operation mode optimization matching and coordinated switching control under different power shortages according to claim 1, wherein: step one, establishing a hierarchical structure based on MAS, comprising:
step 1.1 upper layer agent-negotiation agent. Energy management is realized through an optimization process, so that the system obtains the maximum economic, environmental and electric energy quality benefits, and an optimization model is designed based on real-time data in a processing module. The time scale of the optimization process is once per hour.
Step 1.2 middle layer agent-negotiation agent. The working modes of the DERs are coordinated to be switched to ensure the voltage safety of the system, and the time scale of switching control is in the order of hours or minutes.
Step 1.3, a bottom layer intelligent agent, namely a hybrid agent, comprises a reaction layer and a negotiation layer. The reactive layer is set to "sense-and-act" with priority to respond quickly to emergency situations. The reactive layer has priority to respond quickly to an emergency, and the negotiation layer can control or direct the behavior of the agent.
3. The method for distributed controllable resource operation mode optimized matching and coordinated switching control under different power shortages according to claim 1, wherein: in the second step, the hybrid automaton mainly comprises a discrete state space, an enabling state and a continuous state. In each discrete state space, there is a corresponding continuous state change, and when a certain enabling state is satisfied, the promiscuous automaton has the opportunity to migrate from one discrete state space to another discrete state space, and the corresponding continuous state also has the transition.
The hybrid automata model is represented by a 6-element array: h = (D, L, F, S, F, I). Wherein D = { δ 1 ,δ 2 Of 8230; representing a series of discrete state spacesGathering; l represents a set of a series of continuous state spaces; f = { f 1 (δ 1 ),f 2 (δ 2 ),f 3 (δ 3 ) \8230 } shows the change rule of the continuous state space under each discrete state space; s = { S = { (S) 1 ,S 2 ,S 3 \8230 } represents the mapping between a discrete state space and a continuous state space; f = { F 1 ,F 2 ,F 3 \8230 } represents the condition for state space transition; i denotes the initial state.
4. The method for distributed controllable resource operation mode optimization matching and coordinated switching control under different power shortages according to claim 1, wherein: step three, constructing a hybrid system model aiming at different DERs, comprising the following steps:
step 3.1, constructing a hybrid automaton model of the photovoltaic power generation unit:
in the formula, P pv Is the output power of the photovoltaic power generation unit; p mppt Outputting power for the photovoltaic power generation unit to work in a maximum power point mode; l is the illumination intensity; and M is an illumination intensity threshold value and is determined by the performance of the photovoltaic cell. Aiming at the photovoltaic output characteristic, the photovoltaic power generation unit is set into two working modes: the method comprises the following steps of setting a shutdown mode and an mppt mode as discrete states of a hybrid automaton model of the photovoltaic power generation unit; setting the power output value of the photovoltaic power generation unit to be in a continuous state; the change in the illumination intensity is set as a transfer condition.
Step 3.2, constructing a hybrid automaton model of the wind power generation unit:
in the formula, P w Is the output power of the wind power generation unit; p is e Rated output power for wind power generation unitRate; v is the real-time wind speed; v. of i The wind speed is cut in; v. of r Rated wind speed; v. of o To cut out the wind speed. According to the operating characteristics of wind power generation, the working modes of the wind power generation system are divided into three types: off mode, mppt mode and P e Mode(s). Setting three working modes as discrete states of the hybrid automaton model of the wind power generation unit according to the hybrid automaton model; setting the power output value of the wind power generation unit to be in a continuous state; the wind speed variation is set as a transfer condition.
Step 3.3, constructing an energy storage unit hybrid automaton model:
according to the characteristics of the battery, the battery is divided into five working modes: charging mode, discharging mode and three shutdown modes. The shutdown mode comprises normal shutdown, overcharge shutdown and overdischarge shutdown.
According to the hybrid automata model, when SOC is less than or equal to SOC down When the discharge mode of the storage battery is switched to the over-discharge shutdown mode, the storage battery is switched to the charge mode from the over-discharge shutdown mode when relevant switching commands exist.
When SOC is reached up When the SOC is not more than the SOC, the storage battery is switched from the charging mode to the overcharge shutdown mode, and when relevant switching commands exist, the storage battery is switched from the overcharge shutdown mode to the discharge mode.
When SOC is reached down ＜SOC＜SOC up At the same time, the storage battery can be switched among a discharging mode, a charging mode and a normal shutdown mode according to the system requirements.
Step 3.4, constructing a fuel cell hybrid automaton model:
two operating modes of the fuel cell are designed according to a hybrid automata model: off mode, rated output mode.
Step 3.5, constructing a load hybrid automaton model:
the loads are divided into two categories of interruptible loads and non-interruptible loads, and the interruptible loads are higher than the non-interruptible loads in power supply priority. According to the hybrid robot model, two working modes of the load unit are designed: normal operation mode, load shedding mode.
Step 3.6 sets the initial state of the hybrid automata model of all units, activates the discrete state space, sets it to logic "1", and sets the remaining discrete state space to logic "0". Thus, can obtain
A photovoltaic power generation unit: d = { δ 1 ,δ 2 }＝[1,0](ii) a A wind power generator unit: d = { δ 1 ,δ 2 ,δ 3 }＝[1,0,0](ii) a A battery cell: d = { δ 1 ,δ 2 ,δ 3 ,δ 4 ,δ 5 }＝[1,0,0,0,0](ii) a A fuel cell unit: d = { δ 1 ,δ 2 }＝[1,0](ii) a A load cell: d = { δ 1 ,δ 2 }＝[1,0]。
5. The method for distributed controllable resource operation mode optimization matching and coordinated switching control under different power shortages according to claim 1, wherein: dividing the local control strategy into an internal switching control strategy and a continuous dynamic management strategy, comprising:
step 4.1 for the photovoltaic power generation unit, the internal switching control command is as follows:
for the wind power generation unit, the internal switching control commands are as follows:
the battery cell control commands are as follows:
and 4.2, designing a double-loop control method based on SPWM for the inverter of the DERs, wherein the outer loop is designed to be a controller adopting droop control, and the inner loop is designed to be a controller under a dq rotation coordinate system.
6. The method for distributed controllable resource operation mode optimization matching and coordinated switching control under different power shortages according to claim 1, wherein: step five, establishing a system voltage safety evaluation index, carrying out grading treatment, and coordinating the DERs and the switching of the working modes of the load units by applying event-triggered hybrid control according to a system voltage safety evaluation result, wherein the method comprises the following steps:
step 5.1, extracting the voltage sequence of the ith bus of the power distribution network system by using a wide area signal measurement method, wherein the voltage sequence is expressed as
Then, the average value of the instantaneous voltage of the ith node at the jth time obtained by the voltage sequence is represented as follows: />
The voltage deviation value of the ith node at the jth time is expressed as:
in the formula (I), the compound is shown in the specification,
is the actual voltage value of the ith node at the jth moment.
Thus, the voltage safety evaluation index of the ith node at the jth time can be expressed as:
and 5.2, fusing the voltage safety evaluation indexes of all nodes by using the information fusion method based on the T-S fuzzy neural network.
Use of
In place of x = (x) 1 ,x 2 ,…,x n ) As input to the fuzzy neural network, the fuzzy neural networkThe output of the complex may then be normalized to the required voltage safety assessment index. The output of the fuzzy neural network is represented as:
in the formula (I), the compound is shown in the specification,
weights for the fuzzy neural network; theta j Is the product of membership and connectivity. And the comprehensive voltage safety evaluation index of each node obtained after information fusion is marked as u.
And 5.3, grading the evaluation index u:
and when u is more than or equal to 0.9 and more than or equal to 1.1, the power distribution network system stops running in a voltage collapse state.
When u is more than 0.95 and more than or equal to 0.9, the power distribution network system supplies insufficient energy to the load, and the fuel cell works or carries out load shedding operation.
When u is more than 0.98 and more than or equal to 0.95, the power distribution network system supplies insufficient energy to the load, and the storage battery performs discharge operation to balance power.
When u is more than 1.02 and more than or equal to 0.98, the system voltage is in a normal fluctuation range.
When u is more than 1.05 and more than or equal to 1.02, the energy supply of the power distribution network system is more than the load demand, and the storage battery is charged to balance the power.
When 1.1 > u ≧ 1.05, the energy supply of the distribution grid system is far greater than the load demand, the fuel cell switches to a shutdown mode or the interruptible load is restored.
Step 5.4 takes into account the trigger duration, the handover sequence and the time interval sufficiently and will
Used as trigger events, u and SOC are used as trigger conditions. To (X)>
The design is as follows:
c 2 :H S ((δ 12 ,δ 22 ,δ 23 ,δ 33 ,δ 42 ),(F 12 ,F 25 ,F 26 ,F 33 ,F 42 ))
in the formula, c n Is a set of tags, n ∈ [1,2, \8230;, 15]；
Is the trigger duration; Δ t is the switching time interval. And the coordination control command is utilized to flexibly regulate and control the action change of the DERs in the power distribution network.
7. The method for distributed controllable resource operation mode optimization matching and coordinated switching control under different power shortages according to claim 1, wherein: step six, using a selective mode-based aggregation single-layer dependency classification algorithm to use a coordination control command set as a tag set, and selecting an optimal coordination switching control command in the control command set by executing classification operation, wherein the method comprises the following steps:
step 6.1 digging attribute and attribute value: performing classification operations using an aggregated single-layer dependent classification algorithm based on selective patternsUsing the values of U and SOC in the running process of the power distribution network system as continuous attributes, and using the working modes of a storage battery, a fuel cell and a load as discrete attributes, and mining the attributes; the attribute set is as follows: x = { (U, a) 1 ),(SOC,a 2 ),(Battery,a 3 ),(Fuel,a 4 ),(Load,a 5 )}，a 1 ～a 5 The attribute values corresponding to the five attributes.
Step 6.2 process data of continuous attributes: the sample space (attribute value) of U is divided into seven continuous intervals, and seven levels are defined from low to high as seven set values with reference to fuzzy control: "VL", "ML", "L", "Z", "H", "MH", "VH". Similarly, the sample space of the SOC is divided into five consecutive intervals, and five levels are defined from low to high as five set values: "VL", "L", "Z", "H", "VH"; processing discrete attribute data: for the battery cell, the sample space is set to {0,1,2}, where "0" represents the charging mode, "1" represents the discharging mode, and "2" represents the shutdown mode; for the fuel cell unit, the sample space is set to {0,1}, where "0" represents the rated output mode and "1" represents the shutdown mode; for the load cell, the sample space is set to {0,1}, with "0" indicating the normal operation mode and "1" indicating the load shedding mode.
Step 6.3 assume that the dataset D has n attributes and the training instance is denoted X = (a) 1 ,a 2 L a n ) Wherein a is i (1. Ltoreq. I.ltoreq.n) is the value of the example X at the ith attribute. The class of the training instance belongs to class C = { C 1 ,c 2 L c m One of them, the category of the instance is denoted by c below. The probability P (c | X) that the class of X is c can be expressed as:
let the attribute set contained in the set f be { a } 1 ,a 2 ,L,a i And obtaining a Bayesian network:
applying the pattern classification capability to the Bayesian network, and aggregating the corresponding conditions of all patterns to obtain a selective Bayesian probability prediction formula:
step 6.4, weakening the dependency relationship between the attribute of the selective mode and other attributes in the Bayes algorithm, adopting a selective mode-based aggregation single-layer dependency classification algorithm, wherein the final probability prediction formula is as follows:
wherein H = { j | i +1 ≦ j ≦ n ^ F (a) j )≥m}，F(a j ) Is that the attribute value contains a j Is limited by the parameter m to achieve the support required for conditional probability estimation.
And 6.5, selecting a corresponding attribute based on a selective mode aggregation single-layer dependency classification algorithm, then mining an attribute value according to data generated by the operation of the power distribution network, meanwhile, taking a coordination control command set as a label set, and realizing the selection of the optimal coordination switching control command by executing classification operation.
8. The method for distributed controllable resource operation mode optimization matching and coordinated switching control under different power shortages according to claim 1, wherein: step seven, establishing a comprehensive objective function aiming at the upper-layer intelligent agent, and applying a random walk fruit fly optimization algorithm to realize objective optimization, wherein the method comprises the following steps:
step 7.1, an optimized scheduling objective function is constructed:
cost objective function:
in the formula: i is the number of DERs; s is the operation mode of the DERs; alpha is alpha is As a parameter, when the distribution network system is in the working mode, α is =1, otherwise, α is ＝0；r i Fuel cost for the ith DERs, r for renewable energy sources i ＝0；
The active output of the ith DERs in the s mode; e is Is a consumption feature function; m i Maintenance costs for the ith DERs, and->
Proportioning; c i The startup cost for the ith DERs; beta is a is For cost parameters, when the ith DERs are operated in s-mode, β is =1, otherwise, beta is ＝0。
Carbon emission objective function:
in the formula: f 2 : actual carbon emission when the power distribution network operates; pi 1 : the carbon emission intensity of unit active power output of the thermal power generating unit; pi 2 : the carbon emission intensity of the unit active power output of the gas turbine; p is in.t : acquiring electric quantity from the main network in unit time period; p out.t : generating power by the gas turbine in unit time t; the clean energy carbon emission is 0.
Power quality objective function:
in the formula: sigma is Is the ithThe power quality coefficient of the DERs in the s mode. Sigma is Determined by a voltage safety assessment indicator.
Step 7.2, constructing constraint conditions:
and power balance constraint:
P l ＝P pv +P w +P F +(-1) n P bat (14)
in the formula, P l Is the load demand of the system; p pv Is the output power of the photovoltaic power generation unit; p is w Is the output power of the wind power generation unit; p bat Power output/absorbed for the battery; n is an element of [0,1 ]]N =1 when the battery is in the charging mode, and n =0 when the battery is in the discharging mode; p F Is the output power of the fuel cell.
Output power constraints for DERs:
P i down ＜P i ＜P i up (15)
in the formula, P i Output power for the ith DERs; p i down Minimum output power for the ith DERs; p is i up The maximum output power of the ith DERs.
Battery capacity constraint:
SOC down ＜SOC＜SOC up (16)
in the formula, SOC is the state of charge of the storage battery; SOC (system on chip) down Is the lowest capacity state value of the storage battery; SOC up The highest capacity state value of the storage battery.
And 7.3, based on an improved drosophila optimization algorithm, the method is used for realizing optimized scheduling.
Step 7.4 handles the multi-objective optimization problem in the form of weight coefficients, and energy management can be generalized to solve the following objective optimization function that satisfies the constraint conditions:
in the formula, ω 1 、ω 2 、ω 3 Weight coefficient, ω, representing each objective function 1 +ω 2 +ω 3 ＝1。
9. An apparatus for controlling operation mode optimization matching and coordinated switching of distributed controllable resources under different power shortages, the apparatus comprising:
an architecture building module: the method comprises the steps of establishing a MAS-based power distribution network layered architecture aiming at a coordination control target of the power distribution network, and dividing a power distribution network control structure into an upper layer, a middle layer and a bottom layer;
a model construction module: adding continuous variables on the basis of a finite automata concept to construct a hybrid automata model with 6-element array representation;
a model building module: the hybrid automatic machine model is used for respectively establishing corresponding hybrid automatic machine models aiming at the operating characteristics of the photovoltaic power generation unit, the wind power generation unit, the energy storage unit, the fuel cell and the load;
a management policy module: the method comprises the steps that a local control strategy at the bottom layer is divided into an internal switching control strategy and a continuous dynamic management strategy, wherein the internal switching control strategy takes a coordination control command of a middle layer as a reference, and the continuous dynamic management is a double-ring control method based on SPWM;
a coordination control module: the method is used for designing a coordination control strategy in an intermediate-layer multi-agent structure, constructing a system voltage safety evaluation index and carrying out hierarchical processing, and coordinating the DER and the switching of the working modes of the load units by applying event-triggered hybrid control according to a system voltage safety evaluation result;
a classification operation module: the method comprises the steps of utilizing an aggregation single-layer dependency classification algorithm based on a selective mode to use a coordination control command set as a tag set, and selecting the best coordination switching control command in the control command set by executing classification operation;
a multi-objective optimization module: the method is used for an upper-layer intelligent agent, and aims at different power shortages, the power balance, the output power of DERs and the storage battery capacity constraint are considered, a comprehensive objective function for reducing the operation cost, reducing the pollutant emission and improving the electric energy quality is constructed, an improved random walk fruit fly optimization algorithm is utilized, the multi-objective optimization problem is solved, and an upper-layer control strategy is realized.
10. A distributed controllable resource operation mode optimization matching and coordination switching control device under different power shortages is characterized by comprising a processor and a storage medium;
the storage medium is to store instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310047063.1A CN115940275A (en) | 2023-01-31 | 2023-01-31 | Distributed controllable resource operation mode optimization matching and coordination switching control method and device under different power shortage |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310047063.1A CN115940275A (en) | 2023-01-31 | 2023-01-31 | Distributed controllable resource operation mode optimization matching and coordination switching control method and device under different power shortage |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115940275A true CN115940275A (en) | 2023-04-07 |
Family
ID=86550885
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310047063.1A Pending CN115940275A (en) | 2023-01-31 | 2023-01-31 | Distributed controllable resource operation mode optimization matching and coordination switching control method and device under different power shortage |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115940275A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116581740A (en) * | 2023-04-20 | 2023-08-11 | 国网浙江省电力有限公司嘉兴供电公司 | Distribution network self-healing hybrid control method based on emergency switching of distributed energy operation mode |
-
2023
- 2023-01-31 CN CN202310047063.1A patent/CN115940275A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116581740A (en) * | 2023-04-20 | 2023-08-11 | 国网浙江省电力有限公司嘉兴供电公司 | Distribution network self-healing hybrid control method based on emergency switching of distributed energy operation mode |
| CN116581740B (en) * | 2023-04-20 | 2024-01-12 | 国网浙江省电力有限公司嘉兴供电公司 | Distribution network self-healing hybrid control method based on emergency switching of distributed energy operation mode |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Moghaddam et al. | Multi-objective operation management of a renewable MG (micro-grid) with back-up micro-turbine/fuel cell/battery hybrid power source | |
| CN107994595B (en) | Peak clipping and valley filling control method and system applying control method | |
| CN108471143A (en) | Micro-grid multi-energy method for optimizing scheduling based on positive and negative feedback particle cluster algorithm | |
| De Santis et al. | Genetic optimization of a fuzzy control system for energy flow management in micro-grids | |
| CN103986194B (en) | A kind of self microgrid Optimal Configuration Method and device | |
| Rodriguez et al. | Fuzzy logic-based energy management for isolated microgrid using meta-heuristic optimization algorithms | |
| CN112713618A (en) | Active power distribution network source network load storage cooperative optimization operation method based on multi-scene technology | |
| CN107273968A (en) | A kind of Multiobjective Scheduling method and device based on dynamic fuzzy Chaos-Particle Swarm Optimization | |
| CN104517161A (en) | Virtual power plant distributed power supply combination planning system and method | |
| CN107274085A (en) | A kind of optimum management method of the energy storage device of double electric type ships | |
| CN111786417A (en) | Distributed new energy consumption-oriented active power distribution network multi-target interval optimization scheduling method | |
| Yin et al. | Fuzzy vector reinforcement learning algorithm for generation control of power systems considering flywheel energy storage | |
| CN107528344A (en) | A kind of light storage integrated generating device is incorporated into the power networks control method and system | |
| CN113794199A (en) | Maximum profit optimization method of wind power energy storage system considering electric power market fluctuation | |
| CN115940275A (en) | Distributed controllable resource operation mode optimization matching and coordination switching control method and device under different power shortage | |
| CN114899856A (en) | Method, system, equipment and medium for adjusting power of electric vehicle charging pile | |
| CN113469839A (en) | Smart park optimization strategy based on deep reinforcement learning | |
| Abdelhak et al. | Optimum sizing of hybrid PV/wind/battery using Fuzzy-Adaptive Genetic Algorithm in real and average battery service life | |
| CN117595392A (en) | Power distribution network joint optimization method and system considering light Fu Xiaona and light storage and charge configuration | |
| CN117371755A (en) | Multi-microgrid comprehensive energy system distributed optimization method, device, equipment and medium | |
| CN114285093B (en) | Source network charge storage interactive scheduling method and system | |
| Aghajani et al. | Multi-objective optimal operation in a micro-grid considering economic and environmental goals | |
| CN115622096A (en) | Load characteristic-based capacity configuration method and system for grid-connected micro-grid system | |
| CN113629775B (en) | Fuzzy logic-based hydrogen energy storage system cluster output decision method | |
| Yu et al. | A fuzzy Q-learning algorithm for storage optimization in islanding microgrid |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |