CN103532142B - Refined and stable control method and system of electric system - Google Patents

Abstract

The invention provides a refined and stable control method of an electric system, which comprises the steps of acquiring a power grid system model, sheddable load and unit information to be subjected to stable control, generating a stable power grid control model, obtaining control strategies of predicted faults, acquiring real-time fault information of a power grid system, enquiring the corresponding control strategy according to the fault information, and assigning a load shedding command to sheddable load equipment in the power grid system according to the load shedding command in the control strategy. The invention further provides a corresponding system which can be refined to various load equipments to perform refined load shedding control on the electric system.

Description

Power system becomes more meticulous stable control method and system

Technical field

The present invention relates to electricity load control technology field, particularly relate to a kind of power system and to become more meticulous stable control method, and a kind of power system becomes more meticulous stabilizing control system.

Background technology

At present, electrical network industry is more and more paid attention to electric power safety emergency management, and the ratio that Main Basis causes electrical network to subtract for load in the industry divides incident classification.Such as, for the province of load more than 2,000 ten thousand kilowatts, autonomous region's electrical network, subtract for load less than more than 13% 30%, then form major accident; Subtract for load more than 30%, then form special major accident.At present, China's single extra-high voltage alternating current-direct current power transmission engineering transmission capacity is comparatively large, and such as, the specified transmission power of bright and beautiful Su Zhiliu reaches 7,200,000 kilowatts, and Changzhi-Nanyang-Jingmen specified transmission power of extra-high-voltage alternating current demonstration project reaches 5,000,000 kilowatts.Before extra-high voltage alternating current-direct current transmission of electricity forms network structure, after single UHV transmission channel failure, amphidromic stream transfering channel, causes greater impact to two ends electrical network, unavoidably causes the machine of cutting of larger proportion, the consequence of cutting load.

At present, the cutting load device of electric power system stability control system is generally arranged on 110kV and above transformer station, draws and stop outlet switch during cutting load, there is the feature of " a black a slice ", the mode of cutting load belongs to extensive, and to different load equipment non-selectivity, fine degree is inadequate.Therefore, how for feature and the current electric power thus supplied of different load equipment, cut off load pointedly, significant for electric power system stability control.

Summary of the invention

Based on this, the invention provides a kind of power system and to become more meticulous stable control method and system, all kinds of load equipment can be refine to, to the cutting-off controlling that the load of power system becomes more meticulous.

A kind of power system becomes more meticulous stable control method, comprises the steps:

Acquisition need carry out stability contorting network system model, can cutting load and unit information;

According to described network system model, can cutting load and unit information, generate network stability control model, obtain the control strategy of each forecast failure;

Obtain the real time fail information of described network system;

According to the corresponding described control strategy of fault message inquiry;

According to the cutting load order in control strategy, by cutting load call allocation to the cut off load equipment in described network system;

Described according to described network system model, can cutting load and unit information, generate network stability control model, the step obtaining the control strategy of each forecast failure comprises:

Step 1:

Load aggregation modeling: be polymerized modeling again by after load classification;

Step 2:

Controlling model is set up:

With in n machine system with n-th machine for reference to machine, transient process is in t=t ₀start, set up following merit angular integral type index:

$J_s=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{J}_{\mathrm{si}}=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt}---\left(1\right)$

As power system transient stability quantizating index, in formula (1)

$J_{\mathrm{si}}={\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt},i=\mathrm{1,2},...,n-1---\left(2\right)$

δ in formula (2) _in(t) for t i-th (i=1,2 ..., n-1) platform machine relative to the angle with reference to machine, obtained by time-domain-simulation; t _ffor interval terminal observing time; J _sfor the severe degree of system transient modelling process work angular oscillation; J _s≤ M=π ²(t _f-t ₀) be power system transient stability constraints;

Set up following network stability control model:

min c ^TU (3)

$s.t.\stackrel{\·}{X}=F(X,Y,U)---\left(4\right)$

0＝G(X,Y,U) (5)

$J_s=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt}\≤M---\left(6\right)$

0≤U≤U _max(7)

Wherein c=[c ₁, c ₂..., c _n] ^tfor the control cost of unit controlled quentity controlled variable, described unit controlled quentity controlled variable comprises can cut unit and can cutting load; U=[u ₁, u ₂..., u _n] ^tfor cutting machine, cutting load dominant vector; with the subordination principle that 0=G (X, Y, U) is the transient state process of electric power system preset, X is state vector, and Y is algebraically vector, for merit angular integral type Transient Stability Constraints, M is the constant preset, M=π ²(t _f-t ₀); 0≤U≤U _maxfor the constraint of controlled quentity controlled variable value.

A kind of power system becomes more meticulous stabilizing control system, comprise on-line pre-decision system subsystem, real-time matching subsystem and control executive subsystem, described on-line pre-decision system subsystem comprises the first acquisition module and generation module, and described real-time matching subsystem comprises measures checkout equipment and decision machine;

Described first acquisition module, for obtain need carry out stability contorting network system model, can cutting load and unit information;

Described generation module, for according to described network system model, can cutting load and unit information, generate network stability control model, obtain the control strategy of each forecast failure;

Described measurement checkout equipment, for obtaining the real time fail information of described network system;

Described decision machine, for inquiring about corresponding described control strategy according to fault message;

Described control executive subsystem is used for according to the cutting load order in control strategy, by cutting load call allocation to the cut off load equipment in described network system;

Wherein, described generation module also comprises:

First module, for:

Load aggregation modeling: be polymerized modeling again by after load classification;

Second module, for:

Controlling model is set up:

With in n machine system with n-th machine for reference to machine, transient process is in t=t ₀start, set up following merit angular integral type index:

$J_s=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{J}_{\mathrm{si}}=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt}---\left(1\right)$

As power system transient stability quantizating index, in formula (1)

$J_{\mathrm{si}}={\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt},i=\mathrm{1,2},...,n-1---\left(2\right)$

δ in formula (2) _in(t) for t i-th (i=1,2 ..., n-1) platform machine relative to the angle with reference to machine, obtained by time-domain-simulation; t _ffor interval terminal observing time; J _sfor the severe degree of system transient modelling process work angular oscillation; J _s≤ M=π ²(t _f-t ₀) be power system transient stability constraints;

Set up following network stability control model:

min c ^TU (3)

$s.t.\stackrel{\·}{X}=F(X,Y,U)---\left(4\right)$

0＝G(X,Y,U) (5)

$J_s={\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}\≤M---\left(6\right)$

0≤U≤U _max(7)

Wherein c=[c ₁, c ₂..., c _n] ^tfor the control cost of unit controlled quentity controlled variable, described unit controlled quentity controlled variable comprises can cut unit and can cutting load; U=[u ₁, u ₂..., u _n] ^tfor cutting machine, cutting load dominant vector; with the subordination principle that 0=G (X, Y, U) is the transient state process of electric power system preset, X is state vector, and Y is algebraically vector, for merit angular integral type Transient Stability Constraints, M is the constant preset, M=π ²(t _f-t ₀); 0≤U≤U _maxfor the constraint of controlled quentity controlled variable value.

Above-mentioned power system becomes more meticulous stable control method and system, according to network system model, cutting load and unit information can set up the precisely controlled strategy of network stability control model, cutting load order accurately can be assigned to each can end loads equipment, and the cutting load that realizes selectively becoming more meticulous controls; Wherein according to network system model, can cutting load and unit information, generate the control strategy that network stability control model obtains each forecast failure in advance, thus can quick search corresponding described control strategy when obtaining real time fail information, and then distribute cutting load order rapidly.

Accompanying drawing explanation

Fig. 1 is that power system of the present invention becomes more meticulous stable control method schematic flow sheet in one embodiment.

Fig. 2 is the structural representation of 10 machine 39 node systems in an embodiment.

Fig. 3 is without controlling generator relative angle response schematic diagram in Fig. 2.

Fig. 4 be in Fig. 2 generator 2-9 relative to the angle schematic diagram of generator 1.

Fig. 5 is that power system of the present invention becomes more meticulous stabilizing control system structural representation in one embodiment.

Detailed description of the invention

Below in conjunction with embodiment and accompanying drawing, the present invention is described in further detail, but embodiments of the present invention are not limited thereto.

As shown in Figure 1, be that a kind of power system of the present invention becomes more meticulous stable control method, comprise the steps:

S11, obtain need carry out stability contorting network system model, can cutting load and unit information;

Real-time grid model and can cutting load and unit information can be obtained from regional power grid EMS and user side EMS; Described network system model is the physical connection model of each cutting load equipment and unit in network system.

S12, according to described network system model, can cutting load and unit information, generate network stability control model, obtain the control strategy of each forecast failure;

According to the network system model that step S11 gets, and each can cutting load equipment can cutting load, each unit information, generate network stability control model, according to each fault in advance, solve control strategy;

In a preferred embodiment, described according to described network system model, can cutting load and unit information, generate network stability control model, the step obtaining the control strategy of each forecast failure is:

According to described network system model, cutting load and unit information can set up following network stability control model:

min c ^TU

$s.t.\stackrel{\·}{X}=F(X,Y,U)$

0＝G(X,Y,U)

$J_s={\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}\≤M$

0≤U≤U _max

Wherein c=[c ₁, c ₂..., c _n] ^tfor the control cost of unit controlled quentity controlled variable, described unit controlled quentity controlled variable comprises can cut unit and can cutting load; U=[u ₁, u ₂..., u _n] ^tfor cutting machine, cutting load dominant vector; with

0=G (X, Y, U) is the subordination principle of the transient state process of electric power system preset, and X is state vector, and Y is algebraically vector, for merit angular integral type Transient Stability Constraints, M is the constant preset, M=π ²(t _f-t ₀); 0≤U≤U _maxfor the constraint of controlled quentity controlled variable value.

S13, obtain the real time fail information of described network system;

S14, according to fault message inquiry corresponding described control strategy;

S15, according to the cutting load order in control strategy, by cutting load call allocation to the cut off load equipment in described network system;

The control strategy of each fault in advance obtaining this network system from S12, again by Real-Time Monitoring power system, if receive fault message, then according to the corresponding described control strategy of fault message inquiry, according to the cutting load order in control strategy, by cutting load call allocation to the cut off load equipment in described network system, realize the load cutting-off controlling become more meticulous fast.

In a preferred embodiment, describedly set up network stability control model, the step obtaining the control strategy of each forecast failure can be:

Step 1:

Load aggregation modeling.In principle, the load modeling of the stabilizing control system that becomes more meticulous should arrive each load in detail, but due to load thousands of, the scale of the stability contorting Mathematical Modeling that becomes more meticulous after the detailed modeling of each load will become too large.For simplifying the Mathematical Modeling of the stability contorting that becomes more meticulous; by load classification and then modeling can be polymerized; such as; connection and all air conditioner loads equivalences under certain 110kV bus are become the air-conditioning that large; all water heater equivalences are become the water heater etc. that large; this polymerization modeling is based on precise informations such as the type of each load and parameters, and comparatively conventional load polymerization modeling method is high for modeling accuracy.

Step 2:

Controlling model is set up.

If in n machine system with n-th machine for reference to machine, transient process is in t=t ₀start, be defined as follows merit angular integral type index

$J_s=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{J}_{\mathrm{si}}=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt}---\left(1\right)$

As power system transient stability quantizating index, in formula (1)

$J_{\mathrm{si}}={\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt},i=\mathrm{1,2},...,n-1---\left(2\right)$

δ in formula (2) _in(t) for t i-th (i=1,2 ..., n-1) platform machine relative to the angle with reference to machine, obtained by time-domain-simulation; t _ffor interval terminal observing time.J _scharacterize the severe degree of system transient modelling process work angular oscillation, the J in the Transient Instability situation of merit angle _sbe far longer than the J in transient rotor angle stability situation _s.Can J in rough estimate transient stability situation like this _sthe upper limit: assuming that t ₀~ t _fin time on arbitrary time point | δ _in(t) | (i=1,2 ..., n-1) be π, so by J in invention _s≤ M=π ²(t _f-t ₀) as power system transient stability constraints.Due to J under transient stability and Transient Instability situation _sdiffer greatly, M can select in very wide scope.

Based on merit angular integral type index J _semergent control problem arises be following Nonlinear programming Model

min c ^TU (3)

$s.t.\stackrel{\·}{X}=F(X,Y,U)---\left(4\right)$

0＝G(X,Y,U) (5)

$J_s={\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}\≤M---\left(6\right)$

0≤U≤U _max(7)

C=[c in formula (3) ₁, c ₂..., c _n] ^tfor the control cost of unit controlled quentity controlled variable (cutting machine, cutting load), U=[u ₁, u ₂..., u _n] ^tfor cutting machine, cutting load dominant vector; Formula (4), (5) are the differential-Algebraic Equation set describing transient state process of electric power system, X is state vector, and Y is algebraically vector, and formula (6) is merit angular integral type Transient Stability Constraints, M is the constant preset, M=π in invention ²(t _f-t ₀); Formula (7) is the constraint of controlled quentity controlled variable value.Above-mentioned Mathematical Modeling can use the gradient method solving optimal control problem to solve.

Step 3:

Augmentation index constructs.

Be constructed as follows augmentation index:

$J=c^{T}U+{\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T(F(X,Y,U)-\stackrel{\·}{X})\mathrm{dt}+{\∫}_{t_0}^{t_f}{\mathrm{\ξ}}^{T}G(X,Y,U)\mathrm{dt}+\mathrm{\μ}\·\max [0,{\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}-M]---\left(8\right)$

Wherein λ, ξ are Lagrange multiplier vector, for penalty function item, μ is penalty factor.

Step 4:

Calculate augmentation index to the gradient of dominant vector.

If when controlling initial value U=0, system transient modelling instability namely have

$J=c^{T}U+{\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T(F(X,Y,U)-\stackrel{\·}{X})\mathrm{dt}+{\∫}_{t_0}^{t_f}{\mathrm{\ξ}}^{T}G(X,Y,U)\mathrm{dt}+\mathrm{\μ}\·({\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}-M)---\left(9\right)$

Right application integration by parts, has

${\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T\stackrel{\·}{X}\mathrm{dt}={\mathrm{\λ}}^{T}X{|}_{t_0}^{t_f}-{\∫}_{t_0}^{t_f}{\stackrel{\·}{\mathrm{\λ}}}^T\mathrm{Xdt}$

Therefore

$J=c^{T}U-{\mathrm{\λ}}^{T}X{|}_{t_0}^{t_f}+{\∫}_{t_0}^{t_f}[\mathrm{\μL}\left(X\right)+{\mathrm{\λ}}^{T}F(X,Y,U)+{\mathrm{\ξ}}^{T}G(X,Y,U)+{\stackrel{\·}{\mathrm{\λ}}}^{T}X]\mathrm{dt}-\mathrm{\μM}---\left(10\right)$

The increment of J can be expressed as

$\begin{array}{c}\mathrm{\ΔJ}=c^T\mathrm{\ΔU}-{\mathrm{\λ}}^T\mathrm{\ΔX}{|}_{t_0}^{t_f}+\\ {\∫}_{t_0}^{t_f}[(\mathrm{\μ}\frac{\∂L}{\∂X}+{\mathrm{\λ}}^T\frac{\∂F}{\∂X}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂X}+{\stackrel{\·}{\mathrm{\λ}}}^T)\mathrm{\ΔX}+({\mathrm{\λ}}^T\frac{\∂F}{\∂Y}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂Y})\mathrm{\ΔY}+({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{\ΔU}]\mathrm{dt}\end{array}---\left(11\right)$

Choose λ, ξ to make

$\mathrm{\μ}\frac{\∂L}{\∂X}+{\mathrm{\λ}}^T\frac{\∂F}{\∂X}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂X}+{\stackrel{\·}{\mathrm{\λ}}}^T=0,{\mathrm{\λ}}^T\left(t_f\right)=0---\left(12\right)$

${\mathrm{\λ}}^T\frac{\∂F}{\∂Y}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂Y}=0---\left(13\right)$

Consider again Δ X (t ₀)=0, then

$\mathrm{\ΔJ}=c^T\mathrm{\ΔU}+{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{\ΔUdt}=[c^T+{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{dt}]\mathrm{\ΔU}---\left(14\right)$

Therefore augmented objective function J (U) to the gradient of dominant vector U is

${\▿}_{U}J=c+{\left[{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{dt}\right]}^T---\left(15\right)$

Step 5:

Put k=0, setup control vector U _k=0.

Step 6:

Solve the subordination principle shown in formula (4), (5) with numerical integration method and obtain X _k, Y _k, then bring formula (12), (13) into, obtain multiplier vector λ _k, ξ _k.By X _k, Y _k, λ _k, ξ k brings formula (15) into, obtains the gradient of augmented objective function to dominant vector

Step 7:

Put α is step factor.If U _k+1(i)≤0, then U _k+1(i)=0; If U _k+1(i)>=U _max(i), then U _k+1(i)=U _max(i).U _k+1i () represents U _k+1i-th element, i=1,2 ..., N.

Step 8:

If ε is given convergence precision, then stop calculating, U _k+1for the dominant vector of finally trying to achieve; Otherwise, put k=k+1, go to step 6.

By above-mentioned steps, can realize can the accurate control of cutting load in power system; In fact the present embodiment is considered, experiences for user power utilization, and various kinds of equipment is different to the real-time of power supply, the requirement of interruptibility.The present embodiment solves, become more meticulous the problem of cutting load just, instead of the outlet switch of non-selectivity ground directly in La Ting transformer station i.e. " directly by cutting load call allocation to each interruptible load equipment, the cutting load that realizes selectively becoming more meticulous controls ".Such as, for the water heater of shower, if only have a power failure a few minutes, it is imperceptible that user uses during water heater; Again such as, mansion central air-conditioning has a power failure a few minutes, and the temperature in mansion can't obviously rise, and Consumer's Experience can't be affected.Electric power system stability control system mainly considers power system transient stability usually, its time scale is level second (being generally within 5 to 10s), therefore, the interruptible load such as water heater, air-conditioning that reaches minute level has great function for electric power system stability control system the interruptible price time of a large number of users side.

Next the process of control strategy corresponding to above-mentioned each fault of acquisition power system is elaborated again by a specific embodiment.

10 machine systems as shown in Figure 2, during t=0s there are three relative ground circuits in bus 3 place, and t=0.2s short circuit is removed, without system unstability when controlling, as shown in Figure 3.If air conditioner load accounts for 30% in the load at bus 3,4,7,18 place, and with the cutting load composition of proportions cutting load dominant vector u at these bus places _li(i=1,2,3,4), if machine dominant vector u is cut in machine of the cutting composition of proportions of generator 8 ~ 10 _gi(i=1,2,3), each controlled quentity controlled variable unit controls cost and is 1, is controlled in 0.3s and applies.Given solving precision ε=0.01, gradient method step factor α=0.05, solves the machine of cutting/cutting load amount by gradient method, calculates cut machine dominant vector U _g=[0.132,0,0] ^t, cutting load dominant vector U _l=[0,0.151,0,0] ^t, cut machine dominant vector and cutting load dominant vector is above-mentioned control strategy, can cutting load order, generator 8 ratio of excising is 13.2%, and load 4 ratio of excising is 15.1%, applies power system transient stability after this control, as shown in Figure 4.

The present invention also provides a kind of power system to become more meticulous stabilizing control system, be illustrated in figure 5 the structural representation of this system, comprise on-line pre-decision system subsystem 51, real-time matching subsystem 52 and control executive subsystem 53, described on-line pre-decision system subsystem 51 comprises the first acquisition module 511 and generation module 512, and described real-time matching subsystem 52 comprises measures checkout equipment 521 and decision machine 522;

Described first acquisition module 511, for obtain need carry out stability contorting network system model, can cutting load and unit information;

In a preferred embodiment, described network system model is the physical connection model of each cutting load equipment and unit in network system.

Described generation module 512, for according to described network system model, can cutting load and unit information, generate network stability control model, obtain the control strategy of each forecast failure;

Described measurement checkout equipment 521, for obtaining the real time fail information of described network system;

Described decision machine 522, for inquiring about corresponding described control strategy according to fault message;

Described control executive subsystem 53 for according to the cutting load order in control strategy, by cutting load call allocation to the cut off load equipment in described network system;

In a preferred embodiment, described generation module 512 also for:

According to described network system model, cutting load and unit information can set up following network stability control model:

min c ^TU

$s.t.\stackrel{\·}{X}=F(X,Y,U)$

0＝G(X,Y,U)

$J_s={\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}\≤M$

0≤U≤U _max

Wherein c=[c ₁, c ₂..., c _n] ^tfor the control cost of unit controlled quentity controlled variable, described unit controlled quentity controlled variable comprises can cut unit and can cutting load; U=[u ₁, u ₂..., u _n] ^tfor cutting machine, cutting load dominant vector; with the subordination principle that 0=G (X, Y, U) is the transient state process of electric power system preset, X is state vector, and Y is algebraically vector, for merit angular integral type Transient Stability Constraints, M is the constant preset, M=π ²(t _f-t ₀); 0≤U≤U _maxfor the constraint of controlled quentity controlled variable value.

Above-mentioned power system becomes more meticulous stabilizing control system, comprises on-line pre-decision system subsystem, real-time matching subsystem and controls executive subsystem.On-line pre-decision system system obtains real-time grid model and can cutting load and unit information from regional power grid EMS and user side EMS, by solving the control strategy of network stability control model generation for each forecast failure, the decision machine then control strategy calculated being sent to real-time matching subsystem stores.The amount of calculation calculating control strategy when forecast failure is more, power system is larger is larger, usually forecast failure is divided into groups, adopts multiple computing node parallel computation.Normally 15 ~ 30 minutes cycle that on decision machine, control decision table upgrades.The fault message of measurement device moment detection power transmission network, once there be fault to occur, fault message is delivered to decision machine, decision machine inquires about corresponding control strategy according to fault message, and give outlet port unit control strategy, outlet port unit is delivered to control executive subsystem by dispatch data net control strategy.

The difference of stability contorting and extensive stability contorting of becoming more meticulous is to control executive subsystem.After the control executive subsystem of stabilizing control system of becoming more meticulous receives cutting load order, by user side EMS by cutting load call allocation to each interruptible load equipment, the cutting load that realizes selectively becoming more meticulous controls.

In a preferred embodiment, described generation module 512 also comprises:

First module, for:

Load aggregation modeling: be polymerized modeling again by after load classification;

Second module, for:

Controlling model is set up:

With in n machine system with n-th machine for reference to machine, transient process is in t=t ₀start, set up following merit angular integral type index:

$J_s=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{J}_{\mathrm{si}}=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt}---\left(1\right)$

As power system transient stability quantizating index, in formula (1)

$J_{\mathrm{si}}={\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt},i=\mathrm{1,2},...,n-1---\left(2\right)$

δ in formula (2) _in(t) for t i-th (i=1,2 ..., n-1) platform machine relative to the angle with reference to machine, obtained by time-domain-simulation; t _ffor interval terminal observing time.J _sfor the severe degree of system transient modelling process work angular oscillation; J _s≤ M=π ²(t _f-t ₀) be power system transient stability constraints;

Network stability control model is:

min c ^TU (3)

$s.t.\stackrel{\·}{X}=F(X,Y,U)---\left(4\right)$

0＝G(X,Y,U) (5)

$J_s={\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}\≤M---\left(6\right)$

0≤U≤U _max(7)

C=[c in formula (3) ₁, c ₂..., c _n] ^tfor the control cost of unit controlled quentity controlled variable (cutting machine, cutting load), U=[u ₁, u ₂..., u _n] ^tfor cutting machine, cutting load dominant vector; Formula (4), (5) are the differential-Algebraic Equation set describing transient state process of electric power system, X is state vector, and Y is algebraically vector, and formula (6) is merit angular integral type Transient Stability Constraints, M is the constant preset, M=π in invention ²(t _f-t ₀); Formula (7) is the constraint of controlled quentity controlled variable value;

3rd module, for:

Augmentation index constructs:

Be constructed as follows augmentation index:

$J=c^{T}U+{\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T(F(X,Y,U)-\stackrel{\·}{X})\mathrm{dt}+{\∫}_{t_0}^{t_f}{\mathrm{\ξ}}^{T}G(X,Y,U)\mathrm{dt}+\mathrm{\μ}\·\max [0,{\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}-M]---\left(8\right)$

Wherein λ, ξ are Lagrange multiplier vector, for penalty function item, μ is penalty factor.

Four module, for: calculate augmentation index to the gradient of dominant vector:

If when controlling initial value U=0, system transient modelling instability namely have

$J=c^{T}U+{\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T(F(X,Y,U)-\stackrel{\·}{X})\mathrm{dt}+{\∫}_{t_0}^{t_f}{\mathrm{\ξ}}^{T}G(X,Y,U)\mathrm{dt}+\mathrm{\μ}\·({\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}-M)---\left(9\right)$

Right application integration by parts, has

${\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T\stackrel{\·}{X}\mathrm{dt}={\mathrm{\λ}}^{T}X{|}_{t_0}^{t_f}-{\∫}_{t_0}^{t_f}{\stackrel{\·}{\mathrm{\λ}}}^T\mathrm{Xdt}$

Therefore

$J=c^{T}U-{\mathrm{\λ}}^{T}X{|}_{t_0}^{t_f}+{\∫}_{t_0}^{t_f}[\mathrm{\μL}\left(X\right)+{\mathrm{\λ}}^{T}F(X,Y,U)+{\mathrm{\ξ}}^{T}G(X,Y,U)+{\stackrel{\·}{\mathrm{\λ}}}^{T}X]\mathrm{dt}-\mathrm{\μM}---\left(10\right)$

The increment of J can be expressed as

$\begin{array}{c}\mathrm{\ΔJ}=c^T\mathrm{\ΔU}-{\mathrm{\λ}}^T\mathrm{\ΔX}{|}_{t_0}^{t_f}+\\ {\∫}_{t_0}^{t_f}[(\mathrm{\μ}\frac{\∂L}{\∂X}+{\mathrm{\λ}}^T\frac{\∂F}{\∂X}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂X}+{\stackrel{\·}{\mathrm{\λ}}}^T)\mathrm{\ΔX}+({\mathrm{\λ}}^T\frac{\∂F}{\∂Y}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂Y})\mathrm{\ΔY}+({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{\ΔU}]\mathrm{dt}\end{array}---\left(11\right)$

Choose λ, ξ to make

$\mathrm{\μ}\frac{\∂L}{\∂X}+{\mathrm{\λ}}^T\frac{\∂F}{\∂X}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂X}+{\stackrel{\·}{\mathrm{\λ}}}^T=0,{\mathrm{\λ}}^T\left(t_f\right)=0---\left(12\right)$

${\mathrm{\λ}}^T\frac{\∂F}{\∂Y}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂Y}=0---\left(13\right)$

Due to Δ X (t ₀)=0, then

$\mathrm{\ΔJ}=c^T\mathrm{\ΔU}+{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{\ΔUdt}=[c^T+{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{dt}]\mathrm{\ΔU}---\left(14\right)$

Therefore augmented objective function J (U) to the gradient of dominant vector U is

${\▿}_{U}J=c+{\left[{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{dt}\right]}^T---\left(15\right);$

5th module, for:

Put k=0, setup control vector U _k=0.

6th module, for:

Solve the subordination principle shown in formula (4), (5) with numerical integration method and obtain X _k, Y _k, then bring formula (12), (13) into, obtain multiplier vector λ _k, ξ _k.By X _k, Y _k, λ _k, ξ _kbring formula (15) into, obtain the gradient of augmented objective function to dominant vector

7th module, for:

Put α is step factor.If U _k+1(i)≤0, then U _k+1(i)=0; If U _k+1(i)>=U _max(i), then U _k+1(i)=U _max(i).U _k+1i () represents U _k+1i-th element, i=1,2 ..., N.

Step 8:

If ε is given convergence precision, then stop calculating, U _k+1for the dominant vector of finally trying to achieve; Otherwise, put k=k+1, turn the 6th module.

Power system of the present invention becomes more meticulous stable control method and system, according to network system model, cutting load and unit information can set up the precisely controlled strategy of network stability control model, cutting load order accurately can be assigned to each can end loads equipment, and the cutting load that realizes selectively becoming more meticulous controls; Wherein according to network system model, can cutting load and unit information, generate the control strategy that network stability control model obtains each forecast failure in advance, thus can quick search corresponding described control strategy when obtaining real time fail information, and then distribute cutting load order rapidly.

Direct La Ting transformer station outlet switch during current electric power system stability control system excision load, fail to consider the different requirements of all kinds of load equipment to power supply real-time, interruptibility, to different load equipment non-selectivity, fine degree is inadequate.

Along with the development of technology of Internet of things, by being installed on the smart jack with radio communication, telemetry and telecommand function on each type load and wireless network, user side EMS (UEMS) can telemonitoring and a large amount of customer charge equipment of manipulation.Based on U-EMS, electric power system stability control system directly controls to each load equipment of user side, realizes the cutting load control that selectively becomes more meticulous, become possibility according to the impact experienced user power utilization.

Not affecting under the prerequisite that user power utilization experiences, compared with current electric power system stability control system, the become more meticulous control ability of stability contorting of power system will significantly promote.Meanwhile, due to power system become more meticulous stability contorting in principle Direct Modeling to each load equipment, the difficulty of the load modeling that conventional electric power system stability control system can be avoided to a certain extent to run into.

The present invention realizes selectively becoming more meticulous cutting load function by user side EMS, namely preferentially excises in short-term and does not affect load that user power utilization experiences as water heater, central air-conditioning etc.User side EMS is based on technology of Internet of things, and by being installed on the smart jack with radio communication, telemetry and telecommand function on each type load and wireless network, user side ENERGY CONTROL CENTERS can telemonitoring and a large amount of customer charge equipment of manipulation.

The above embodiment only have expressed several embodiment of the present invention, and it describes comparatively concrete and detailed, but therefore can not be interpreted as the restriction to the scope of the claims of the present invention.It should be pointed out that for the person of ordinary skill of the art, without departing from the inventive concept of the premise, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with claims.

Claims (6)

1. power system becomes more meticulous a stable control method, it is characterized in that, comprises the steps:

Acquisition need carry out stability contorting network system model, can cutting load and unit information;

According to described network system model, can cutting load and unit information, generate network stability control model, obtain the control strategy of each forecast failure;

Obtain the real time fail information of described network system;

According to the corresponding described control strategy of fault message inquiry;

According to the cutting load order in control strategy, by cutting load call allocation to the cut off load equipment in described network system;

Described according to described network system model, can cutting load and unit information, generate network stability control model, the step obtaining the control strategy of each forecast failure is:

Step 1:

Load aggregation modeling: be polymerized modeling again by after load classification;

Step 2:

Controlling model is set up:

With in n machine system with n-th machine for reference to machine, transient process is in t=t ₀start, set up following merit angular integral type index:

$J_s=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{J}_{\mathrm{si}}=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt}---\left(1\right)$

As power system transient stability quantizating index, in formula (1)

$J_{\mathrm{si}}={\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt},i=\mathrm{1,2},...,n-1---\left(2\right)$

δ in formula (2) _in(t) for t i-th (i=1,2 ..., n-1) platform machine relative to reference to the angle of machine, obtained by time-domain-simulation; t _ffor interval terminal observing time; J _sfor the severe degree of system transient modelling process work angular oscillation; J _s≤ M=π ²(t _f-t ₀) be power system transient stability constraints;

Set up following network stability control model:

min c ^TU (3)

$s.t.\stackrel{\·}{X}=F(X,Y,U)---\left(4\right)$

0＝G(X,Y,U) (5)

$J_s={\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt}\≤M---\left(6\right)$

0≤U≤U _max(7)

Wherein c=[c ₁, c ₂..., c _n] ^tfor the control cost of unit controlled quentity controlled variable, described unit controlled quentity controlled variable comprises can cut unit and can cutting load; U=[u ₁, u ₂..., u _n] ^tfor cutting machine, cutting load dominant vector; with the subordination principle that 0=G (X, Y, U) is the transient state process of electric power system preset, X is state vector, and Y is algebraically vector, for merit angular integral type Transient Stability Constraints, M is the constant preset, M=π ²(t _f-t ₀); 0≤U≤U _maxfor the constraint of controlled quentity controlled variable value.

2. power system according to claim 1 becomes more meticulous stable control method, it is characterized in that, describedly sets up network stability control model, and the step obtaining the control strategy of each forecast failure also comprises:

Step 3:

Augmentation index constructs:

Be constructed as follows augmentation index:

$J=c^{T}U+{\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T(F(X,Y,U)-\stackrel{\·}{X})\mathrm{dt}+{\∫}_{t_0}^{t_f}{\mathrm{\ξ}}^{T}G(X,Y,U)\mathrm{dt}+\mathrm{\μ}\·\max [0,{\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}-M]---\left(8\right)$

Wherein λ, ξ are Lagrange multiplier vector, for penalty function item, μ is penalty factor;

Step 4: calculate augmentation index to the gradient of dominant vector:

If when controlling initial value U=0, system transient modelling instability namely have

$J=c^{T}U+{\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T(F(X,Y,U)-\stackrel{\·}{X})\mathrm{dt}+{\∫}_{t_0}^{t_f}{\mathrm{\ξ}}^{T}G(X,Y,U)\mathrm{dt}+\mathrm{\μ}\·({\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}-M)---\left(9\right)$

Right application integration by parts, has

${\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T\stackrel{\·}{X}\mathrm{dt}={\mathrm{\λ}}^{T}X{|}_{t_0}^{t_f}-{\∫}_{t_0}^{t_f}{\stackrel{\·}{\mathrm{\λ}}}^T\mathrm{Xdt}$

Therefore

$J=c^{T}U-{\mathrm{\λ}}^{T}X{|}_{t_0}^{t_f}+{\∫}_{t_0}^{t_f}[\mathrm{\μL}\left(X\right)+{\mathrm{\λ}}^{T}F(X,Y,U)+{\mathrm{\ξ}}^{T}G(X,Y,U)+{\stackrel{\·}{\mathrm{\λ}}}^{T}X]\mathrm{dt}-\mathrm{\μM}---\left(10\right)$

The increment of J can be expressed as

$\begin{array}{c}\mathrm{\ΔJ}=c^T\mathrm{\ΔU}-{\mathrm{\λ}}^T\mathrm{\ΔX}{\∫}_{t_0}^{t_f}+\\ {\∫}_{t_0}^{t_f}[(\mathrm{\μ}\frac{\∂L}{\∂X}+{\mathrm{\λ}}^T\frac{\∂F}{\∂X}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂X}+{\stackrel{\·}{\mathrm{\λ}}}^T)\mathrm{\ΔX}+({\mathrm{\λ}}^T\frac{\∂F}{\∂Y}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂Y})\mathrm{\ΔY}+({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{\ΔU}]\mathrm{dt}\end{array}----\left(11\right)$

Choose λ, ξ to make

$\mathrm{\μ}\frac{\∂L}{\∂X}+{\mathrm{\λ}}^T\frac{\∂F}{\∂X}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂X}+{\stackrel{\·}{\mathrm{\λ}}}^T=0,{\mathrm{\λ}}^T\left(t_f\right)=0---\left(12\right)$

${\mathrm{\λ}}^T=\frac{\∂F}{\∂Y}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂Y}=0---\left(13\right)$

Due to Δ X (t ₀)=0, then

$\mathrm{\ΔJ}=c^T\mathrm{\ΔU}+{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T+\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{\ΔUdt}=[c^T+{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{dt}]\mathrm{\ΔU}---\left(14\right)$

Therefore augmented objective function J (U) to the gradient of dominant vector U is

${\▿}_{U}J=c+{\left[{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T+\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{dt}\right]}^T---\left(15\right);$

Step 5:

Put k=0, setup control vector U _k=0;

Step 6:

Solve the subordination principle shown in formula (4), (5) with numerical integration method and obtain X _k, Y _k, then bring formula (12), (13) into, obtain multiplier vector λ _k, ξ _k; By X _k, Y _k, λ _k, ξ _kbring formula (15) into, obtain the gradient of augmented objective function to dominant vector

Step 7:

Put α is step factor; If U _k+1(i)≤0, then U _k+1(i)=0; If U _k+1(i)>=U _max(i), then U _k+1(i)=U _max(i); U _k+1i () represents U _k+1i-th element, i=1,2 ..., N;

Step 8:

If ε is given convergence precision, then stop calculating, U _k+1for the dominant vector of finally trying to achieve; Otherwise, put k=k+1, go to step 6.

3. power system according to claim 1 becomes more meticulous stable control method, and it is characterized in that, described network system model is the physical connection model of each cutting load equipment and unit in network system.

4. a power system becomes more meticulous stabilizing control system, it is characterized in that, comprise on-line pre-decision system subsystem, real-time matching subsystem and control executive subsystem, described on-line pre-decision system subsystem comprises the first acquisition module and generation module, and described real-time matching subsystem comprises measures checkout equipment and decision machine;

Described first acquisition module, for obtain need carry out stability contorting network system model, can cutting load and unit information;

Described generation module, for according to described network system model, can cutting load and unit information, generate network stability control model, obtain the control strategy of each forecast failure;

Described measurement checkout equipment, for obtaining the real time fail information of described network system;

Described decision machine, for inquiring about corresponding described control strategy according to fault message;

Described control executive subsystem is used for according to the cutting load order in control strategy, by cutting load call allocation to the cut off load equipment in described network system;

Wherein, described generation module comprises:

First module, for:

Load aggregation modeling: be polymerized modeling again by after load classification;

Second module, for:

Controlling model is set up:

With in n machine system with n-th machine for reference to machine, transient process is in t=t ₀start, set up following merit angular integral type index:

$J_s=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{J}_{\mathrm{si}}=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt}---\left(1\right)$

As power system transient stability quantizating index, in formula (1)

$J_{\mathrm{si}}={\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt},i=\mathrm{1,2},...,n-1---\left(2\right)$

δ in formula (2) _in(t) for t i-th (i=1,2 ..., n-1) platform machine relative to reference to the angle of machine, obtained by time-domain-simulation; t _ffor interval terminal observing time; J _sfor the severe degree of system transient modelling process work angular oscillation; J _s≤ M=π ²(t _f-t ₀) be power system transient stability constraints;

Set up following network stability control model:

min c ^TU (3)

$s.t.\stackrel{\·}{X}=F(X,Y,U)---\left(4\right)$

0＝G(X,Y,U) (5)

$J_s={\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}=\frac{1}{n-1}\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}{\∫}_{t_0}^{t_f}{\mathrm{\δ}}_{\mathrm{in}}^2\left(t\right)\mathrm{dt}\≤M---\left(6\right)$

0≤U≤U _max(7)

Wherein c=[c ₁, c ₂..., c _n] ^tfor the control cost of unit controlled quentity controlled variable, described unit controlled quentity controlled variable comprises can cut unit and can cutting load; U=[u ₁, u ₂..., u _n] ^tfor cutting machine, cutting load dominant vector; with the subordination principle that 0=G (X, Y, U) is the transient state process of electric power system preset, X is state vector, and Y is algebraically vector, for merit angular integral type Transient Stability Constraints, M is the constant preset, M=π ²(t _f-t ₀); 0≤U≤U _maxfor the constraint of controlled quentity controlled variable value.

5. power system according to claim 4 becomes more meticulous stabilizing control system, and it is characterized in that, described generation module also comprises:

3rd module, for:

Augmentation index constructs:

Be constructed as follows augmentation index:

$J=c^{T}U+{\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T(F(X,Y,U)-\stackrel{\·}{X})\mathrm{dt}+{\∫}_{t_0}^{t_f}{\mathrm{\ξ}}^{T}G(X,Y,U)\mathrm{dt}+\mathrm{\μ}\·\max [0,{\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}-M]---\left(8\right)$

Wherein λ, ξ are Lagrange multiplier vector, for penalty function item, μ is penalty factor;

Four module, for: calculate augmentation index to the gradient of dominant vector:

If when controlling initial value U=0, system transient modelling instability namely have

$J=c^{T}U+{\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T(F(X,Y,U)-\stackrel{\·}{X})\mathrm{dt}+{\∫}_{t_0}^{t_f}{\mathrm{\ξ}}^{T}G(X,Y,U)\mathrm{dt}+\mathrm{\μ}\·({\∫}_{t_0}^{t_f}L\left(X\right)\mathrm{dt}-M)---\left(9\right)$

Right application integration by parts, has

${\∫}_{t_0}^{t_f}{\mathrm{\λ}}^T\stackrel{\·}{X}\mathrm{dt}={\mathrm{\λ}}^{T}X{|}_{t_0}^{t_f}-{\∫}_{t_0}^{t_f}{\stackrel{\·}{\mathrm{\λ}}}^T\mathrm{Xdt}$

Therefore

$J=c^{T}U-{\mathrm{\λ}}^{T}X{|}_{t_0}^{t_f}+{\∫}_{t_0}^{t_f}[\mathrm{\μL}\left(X\right)+{\mathrm{\λ}}^{T}F(X,Y,U)+{\mathrm{\ξ}}^{T}G(X,Y,U)+{\stackrel{\·}{\mathrm{\λ}}}^{T}X]\mathrm{dt}-\mathrm{\μM}---\left(10\right)$

The increment of J can be expressed as

$\begin{array}{c}\mathrm{\ΔJ}=c^T\mathrm{\ΔU}-{\mathrm{\λ}}^T\mathrm{\ΔX}{\∫}_{t_0}^{t_f}+\\ {\∫}_{t_0}^{t_f}[(\mathrm{\μ}\frac{\∂L}{\∂X}+{\mathrm{\λ}}^T\frac{\∂F}{\∂X}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂X}+{\stackrel{\·}{\mathrm{\λ}}}^T)\mathrm{\ΔX}+({\mathrm{\λ}}^T\frac{\∂F}{\∂Y}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂Y})\mathrm{\ΔY}+({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{\ΔU}]\mathrm{dt}\end{array}----\left(11\right)$

Choose λ, ξ to make

$\mathrm{\μ}\frac{\∂L}{\∂X}+{\mathrm{\λ}}^T\frac{\∂F}{\∂X}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂X}+{\stackrel{\·}{\mathrm{\λ}}}^T=0,{\mathrm{\λ}}^T\left(t_f\right)=0---\left(12\right)$

${\mathrm{\λ}}^T=\frac{\∂F}{\∂Y}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂Y}=0---\left(13\right)$

Due to Δ X (t ₀)=0, then

$\mathrm{\ΔJ}=c^T\mathrm{\ΔU}+{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T+\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{\ΔUdt}=[c^T+{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{dt}]\mathrm{\ΔU}---\left(14\right)$

Therefore augmented objective function J (U) to the gradient of dominant vector U is

${\▿}_{U}J=c+{\left[{\∫}_{t_0}^{t_f}({\mathrm{\λ}}^T+\frac{\∂F}{\∂U}+{\mathrm{\ξ}}^T\frac{\∂G}{\∂U})\mathrm{dt}\right]}^T---\left(15\right);$

5th module, for:

Put k=0, setup control vector U _k=0;

6th module, for:

Solve the subordination principle shown in formula (4), (5) with numerical integration method and obtain X _k, Y _k, then bring formula (12), (13) into, obtain multiplier vector λ _k, ξ _k; By X _k, Y _k, λ _k, ξ _kbring formula (15) into, obtain the gradient of augmented objective function to dominant vector

7th module, for:

Put α is step factor; If U _k+1(i)≤0, then U _k+1(i)=0; If U _k+1(i)>=U _max(i), then U _k+1(i)=U _max(i); U _k+1i () represents U _k+1i-th element, i=1,2 ..., N;

Step 8:

If ε is given convergence precision, then stop calculating, U _k+1for the dominant vector of finally trying to achieve; Otherwise, put k=k+1, turn the 6th module.

6. power system according to claim 4 becomes more meticulous stabilizing control system, and it is characterized in that, described network system model is the physical connection model of each cutting load equipment and unit in network system.