CN103617563B - A kind of water supply network theoretical based on geo-statistic spatial analysis is without monitoring node pressure determination statement - Google Patents

Abstract

The invention discloses a kind of water supply network theoretical based on geo-statistic spatial analysis without monitoring node pressure determination statement, choose water supply network Water Source Pumping Station node and other with the pressure of the node of pressure monitoring devices as known pressure, being then based on known node pressure uses general kriging analysis method to carry out interpolation calculation to without monitoring node pressure, thus obtains the water supply network node pressure without monitoring point；During interpolation calculation, by the Euclidean distance during change substitutes general kriging analysis method relatively of pressure between node.The present invention uses general kriging analysis method, in conjunction with system-head curve, Euclidean distance is substituted with relative pressure change, use the relative change of the calculation of design parameters pressure of water supply network, choose the variation function model being suitable for grid hydraulic characteristic(s), water supply network is calculated without the node pressure of monitoring point and predicted, computational accuracy is high, and feasibility is strong.

Description

A kind of water supply network theoretical based on geo-statistic spatial analysis is true without monitoring node pressure Determine method

Technical field

The present invention relates to water supply network without monitoring node pressure determination statement, contribute to hydraulic pipeline is supervised by this method Survey the data in the case of data deficiencies to expand, can be pipe network Optimized Operation and soil's rigidity offer technical support, belong to city The scheduling of water supply network hydraulic analysis, pipe network and soil's rigidity field.

Background technology

Water supply network safe and stable operation is related to the normal production of the normal life of city dweller, city enterprise, public The safe operation of facility and urban fire control are safely etc..In recent years, along with developing rapidly of computer technology, the most progressively set up confession The hydraulic pressure of grid or flow automatic checkout system, i.e. SCADA (Supervisory Control And Data Acquisition) system, this system utilizes sensor that hydraulic pressure and the flow of some important node of water supply network are carried out automatic reality Time monitoring, and use load mode timing wirelessly or non-wirelessly to pass hydraulic pressure and flow signal back control centre and be used for analyzing pipe network Hydraulic regime.But due to reasons such as economy and maintenances, the pressure monitoring point quantity of laying is generally all far fewer than the joint of pipe network Point quantity, so grasping and simulate the hydraulic regime of pipe network accurately except depending on hydraulic pressure limited in pipe network or flow monitoring Outside Dian, use effective hydraulic regime to analyze method and the data obtained are expanded the most particularly important.

Hydraulic pipeline state analysis method can be divided into microcosmic method and macroscopic method two class.Wherein, based on " black box theory " grand Sight method is always the main method of hydraulic pipeline state analysis because of its simple possible, and the method does not consider answering of water distribution system structure Polygamy, several main input/output variables in only consideration system, such as the pressure of pumping plant output, pressure of supply water and pressure tap Deng, on the basis of these data, the method using statistical modeling, water supply network running status is carried out Macrovision analog.The grandest Sight method model specifically includes that ratio period pipe net leakage rate and ratio water supply network macromodel, distribution system of water supply equivalent network at times Network model and the pipe network macromodel etc. set up based on neural network.

Ratio period pipe net leakage rate supposes that water supply network output of each time period in water supply cycle accounts for the total water of pipe network The constant rate of amount, and the discharge pressure of any one water factory and pipe network all nodes total water consumption and other water factories in pipe network Relevant.Mixing supply owing to China's major part urban industry supplies water to supply water with resident, node water consumption is closed with total water consumption ratio Being uncertain, ratio period pipe net leakage rate is unsuitable for the public supply mains of China.

Ratio water supply network macromodel is the improvement of comparative example period pipe net leakage rate at times, model using one day as one Individual water supply cycle, and a water supply cycle is divided into several period, it is assumed that each period interior nodes flow proportional is certain.At times Ratio pipe network macromodel solves node water consumption and accounts for always

The ratio problems of water consumption, but set up the model of different periods, increase the complexity of workload and model, and And have ignored the change in season and impact festivals or holidays water consumption rate fluctuations produced, divide the ratio macroscopic view pipe network of period Model needs constantly to check and revise, and model is relatively complicated.

Distribution system of water supply equivalent network model not only allows for pressure of supply water and the output of each water factory of pipe network, and is also contemplated for The distribution of pipe network each node pressure and mutual impact thereof and inner link, utilize the prison of measured response pipeline distribution situation Measuring point and pumping plant node, construct an equivalent network simplified.Model introduces the real time information of monitoring point in water supply network, Can reflect and predict the actual motion state of pipe network, model accuracy increases.

Neural network model is a kind of simulation animal nerve network behavior feature, carries out the number of distributed parallel information processing Learn model.The pipe network macromodel set up based on neural network is to supply with water supply pumping plant pressure of supply water, pipe network in water supply network In the water yield and pipe network, pressure and other pipe network attribute variables that can obtain of pressure monitoring point as its input and export, and pass through The a large amount of inputs obtained and output data, train a recessive network, this invisible web can numerically simulate from Known variables obtains known variables, after i.e. inputting output and other known variables, can export monitoring point corresponding thereto Pressure.It is relatively strong that the advantage of neural network model is to simulate non-linear mapping capability, but water supply network carries out macroscopic view operating mode mould During plan, avoid water system pipe network structure completely, cause its simulation precision the highest.

Summary of the invention

For deficiencies of the prior art, it is an object of the invention to provide a kind of computational accuracy high, feasibility The strong water supply network theoretical based on geo-statistic spatial analysis is without monitoring node pressure determination statement.

To achieve these goals, the technical solution used in the present invention is as follows:

A kind of water supply network theoretical based on geo-statistic spatial analysis, without monitoring node pressure determination statement, chooses feed pipe Net Water Source Pumping Station node and other with the pressure of the node of pressure monitoring devices as known pressure, be then based on known node Pressure uses general kriging analysis method to carry out interpolation calculation to without monitoring node pressure, thus obtains water supply network without monitoring point Node pressure；During interpolation calculation, by the Euclidean distance during change substitutes general kriging analysis method relatively of pressure between node.

It concretely comprises the following steps:

1) sign of relation size between node:

Utilize the relative change of pressure between all nodes of calculation of design parameters of water supply network, relatively change by design pressure Size characterizes the relation between two node spaces；

2) the asking for of general kriging analysis method variation function:

With Water Source Pumping Station monitoring node as starting point, according to the 1st) step gained, other monitoring node is monitored with Water Source Pumping Station Internodal relative pressure variable quantity sorts from small to large, and asks for variation function value by variation function definition；With relative distance For abscissa, variation function value is ordinate, describes the relation curve between both in rectangular coordinate system；With variation function Theoretical model figure compares, and selects similar theoretical model and is fitted by least square method, so that it is determined that variation letter Number；

3) by the 1st) pressure of step relatively change substitute into the 2nd) variation function that step obtains, try to achieve the change between all nodes Different functional value；

4) during general kriging analysis, the pressure of each node is with the relative pressure between this node with Water Source Pumping Station node Power change is relevant, if the pressure function of this node is: h (x)=α₀+α₁x；

Wherein: x is the relative pressure change of required node and Water Source Pumping Station node；α₀And a₁For function coefficients, wherein For required weight coefficient；K is known monitoring node number；

5) general kriging analysis equation Λ X is determined_t=B_t’Variation function matrix Λ in (t=0,1) and variation function to Amount B_t；Wherein the variation function value between each node in variation function matrix Λ is by the 3rd) step can obtain；

6) matrix equation is calculated；Inverse matrix Λ of changes persuing different Jacobian matrix Λ^-1, take t=0 and 1 and carry out matrix computations respectively X_t=Λ^-1B_t；(t=0,1), obtains factor alpha₀And α₁Weight coefficient vector

7) by weight coefficient vectorSubstitute intoTry to achieve node pressure function coefficients α₀And α₁Value, then node pressure function determines；

8) by step 1) each relative pressure change x without monitoring node with Water Source Pumping Station node can be obtained in water supply network, X is substituted into node pressure functionIn, can be calculated each without monitoring node force value.

Wherein step 1) between interior joint the change relatively of pressure obtain as follows: suppose pipe network always using in the T moment The water yield, carries out assignment of traffic by area ratio discharge method and obtains in pipe network each node in the simulation water consumption in T moment；With feed pipe In net, each node simulation water consumption, design pipe range, caliber, coefficient of pipe friction are known conditions, use finite difference calculus to ask Obtain matrix Q

Wherein,The change impact on its all node pressures for pipe network node flow；

The diagonal entry that each row in matrix Q are respectively divided by its correspondence can be converted into relative pressure transformation matrices V,

I-th row of matrix V, jth column element represent the relative pressure change between node i and node j；

The present invention is according to the relation curve depicted and the comparison of Variogram Theory Model figure and water supply network water Force characteristic, the actual Variogram Theory Model chosen is spherical model, and its covariance variation function is

$\mathrm{\&gamma;}\left(h\right)=\left\{\begin{array}{cc}0,& h=0\\ c_0+c(\frac{3h}{2a}-\frac{h^3}{2a^3}),& 0<h\&le;a\\ c_0+c,& h>a\end{array}\right.$

Wherein: h is the relative pressure change in water supply network between known pressure node；

c₀Block gold number for nested model；

A is the range of nested model；

C is the inclined base station value of nested model；

c₀+ c is the base station value of nested model.

According to the pressure of known monitoring node, use least square method can draw the coefficient in above-mentioned function.

Compared to existing technology, there is advantages that

The present invention uses general kriging analysis method, in conjunction with system-head curve, substitutes Euclidean distance with relative pressure change, uses The relative change of the calculation of design parameters pressure of water supply network, chooses the variation function model being suitable for grid hydraulic characteristic(s), Water supply network is calculated without the node pressure of monitoring point and predicted, computational accuracy is high, and feasibility is strong.

Accompanying drawing explanation

Fig. 1-embodiment of the present invention for water supply network schematic diagram.

Fig. 2-present invention general kriging analysis result and pipeline network simulation results contrast schematic diagram.

Detailed description of the invention

Geo-statistic spatial analysis theory is incorporated into and meets randomness and structural water supply network node pressure by the present invention Analysis with calculate, choose water supply network Water Source Pumping Station node and other be with the pressure conduct of the node of pressure monitoring devices Known pressure, is then based on known node pressure and uses general Krieger (Universal Kriging) interpolation method to without monitoring joint Point pressure carries out interpolation calculation, thus obtains the water supply network node pressure without monitoring point.During interpolation calculation, with pressure between node The Euclidean distance during relatively change substitutes general kriging analysis method；Choose the variation letter being suitable for grid hydraulic characteristic(s) simultaneously Digital-to-analogue type.

Concretely comprise the following steps:

1) sign of relation size between node:

Utilize the relative change of pressure between all nodes of calculation of design parameters of water supply network, relatively change by design pressure Size characterizes the relation in space between two nodes；

Certain city's network topology known, numbering nodes and pipe sections and each pipeline section caliber, pipe range and hydraulic parameters.Assume T moment pipe network total water consumption, uses pipe network state estimate to obtain each node pressure of pipe network, and is numbered (due in city Water supply network actual cannot measure flow and the pressure obtaining each node in running, and generally uses pipe network state estimate to obtain Pipeline network simulation pressure as with reference to pressure, its specific practice assume that pipe network the accurate water consumption of certain moment all nodes and with Design water consumption is different, and carries out hydraulic analogy as known conditions, is calculated at each node of this moment relatively accurate Pressure condition).

Between node, the change relatively of pressure obtains as follows: assuming that pipe network is at the total water consumption in T moment, by area ratio Discharge method carries out assignment of traffic and obtains in pipe network each node in the simulation water consumption in T moment；With each node in water supply network Simulation water consumption, design pipe range, caliber, coefficient of pipe friction are known conditions, use finite difference calculus to try to achieve matrix Q,

Wherein,The change impact on its all node pressures for pipe network node flow；

Each row in matrix Q are respectively divided by the diagonal entry of its correspondence, relative pressure change square can be converted into Battle array V,

$V=\left[\begin{array}{cccc}\frac{\&part;h_1}{\&part;q_1}/\frac{\&part;h_1}{\&part;q_1}& \frac{\&part;h_1}{\&part;q_2}/\frac{\&part;h_2}{\&part;q_2}& \mathrm{\&Lambda;}& \frac{\&part;h_1}{\&part;q_n}/\frac{\&part;h_n}{\&part;q_n}\\ \frac{\&part;h_2}{\&part;q_1}/\frac{\&part;h_1}{\&part;q_1}& \frac{\&part;h_2}{\&part;q_2}/\frac{\&part;h_2}{\&part;q_2}& \mathrm{\&Lambda;}& \frac{\&part;h_2}{\&part;q_n}/\frac{\&part;h_n}{\&part;q_n}\\ M& M& O& M\\ \frac{\&part;h_n}{\&part;q_1}/\frac{\&part;h_1}{\&part;q_1}& \frac{\&part;h_n}{\&part;q_2}/\frac{\&part;h_2}{\&part;q_2}& \mathrm{\&Lambda;}& \frac{\&part;h_n}{\&part;q_n}/\frac{\&part;h_n}{\&part;q_n}\end{array}\right]=\left[\begin{array}{cccc}\frac{\&part;h_1}{\&part;h_1}& \frac{\&part;h_1}{\&part;h_2}& \mathrm{\&Lambda;}& \frac{\&part;h_1}{\&part;h_n}\\ \frac{\&part;h_2}{\&part;h_1}& \frac{\&part;h_2}{\&part;h_2}& \mathrm{\&Lambda;}& \frac{\&part;h_2}{\&part;h_n}\\ M& M& O& M\\ \frac{\&part;h_n}{\&part;h_1}& \frac{\&part;h_n}{\&part;h_2}& \mathrm{\&Lambda;}& \frac{\&part;h_n}{\&part;h_n}\end{array}\right],$

I-th row of matrix V, jth column element represent the relative pressure change between node i and node j；

2) the asking for of general kriging analysis method variation function:

With Water Source Pumping Station monitoring node as starting point, according to the 1st) step gained, other monitoring node is monitored with Water Source Pumping Station Internodal relative pressure variable quantity sorts from small to large, and asks for variation function value by variation function definition；With relative distance For abscissa, variation function value is ordinate, describes the relation curve between both in rectangular coordinate system；With variation function Theoretical model figure compares, and selects similar theoretical model and is fitted by least square method, so that it is determined that variation letter Number；

In water supply network, represent that the stochastic variable of node pressure is designated as h (x), then mathematical definition γ (s) of variation function =Var [h (x)-h (x+s)]/2=E [h (x)-h (x+s)]²/2-{E[h(x)]-E[h(x+s)]}²/2.It is carried out in general gram During lattice (Kriging) interpolation, there is precondition: regionalized variable (i.e. representing stochastic variable h (x) of node pressure) meets two Rank stationary hypothesis.So, to any distance s, there is E [h (x)]=E [h (x+s)].Then variation function γ (s)=E [h (x)-h (x+s)]²/ 2, when the measured value only one of which of node pressure, γ (s)=[h (x)-h (x+s)] can be directly substituted into²/2。

According to the relation curve depicted and the comparison of Variogram Theory Model figure and water supply network hydraulic characteristic(s), Described Variogram Theory Model chooses spherical model, and its covariance variation function is

$\mathrm{\&gamma;}\left(h\right)=\left\{\begin{array}{cc}0,& h=0\\ c_0+c(\frac{3h}{2a}-\frac{h^3}{2a^3}),& 0<h\&le;a\\ c_0+c,& h>a\end{array}\right.---\left(1\right)$

Wherein:

H is the relative pressure change in water supply network between known pressure node；

c₀Block gold number for nested model；

A is the range of nested model；

C is the inclined base station value of nested model；

c₀+ c is the base station value of nested model.

Pressure (can obtain pressure relatively to change) according to known monitoring node, uses least square method can draw above-mentioned function In coefficient (c₀,a,c).To its detailed process carrying out least square fitting it is:

In formula, h is the relative pressure change in water supply network between known pressure node, is independent variable.When relative distance 0 During ＜ h≤a, have:

$\mathrm{\&gamma;}\left(s\right)=c_0+c(\frac{3h}{2a}-\frac{h^3}{2a^3})$

It is used least square fitting, Ke Yiji: y=γ (s), b₀=c₀, b₁=3c/2a, b₂=-c/2a³, x₁= H, x₂=h³, above formula can be changed into:

Y=b₀+b₁x₁+b₂x₂ (2)

When water supply network node pressure being carried out kriging analysis and estimating, it is known that relative between destination node in pipe network Distance and pressure thereof, by x₁=h, x₂=h³, can in the hope of the independent variable x of above formula₁、x₂, dependent variable y=γ (s), by variation function The mathematical definition of γ (s):

$\mathrm{\&gamma;}\left(s\right)=\frac{1}{2}Var\&lsqb;h\left(x\right)-h(x+s)\&rsqb;=\frac{1}{2}E{\&lsqb;h\left(x\right)-h(x+s)\&rsqb;}^2-\frac{1}{2}{\{E\&lsqb;h\left(x\right)\&rsqb;-E\&lsqb;h(x+s)\&rsqb;\}}^2$

Understand, after all node pressure data measurement obtained are screened, use above formula to seek its covariance, counted Calculate result γ (s) and be the value of y.

Now, it is known that condition is some groups of variable x₁、x₂, y, use least square method can obtain each of formula (2) with matching Coefficient: b₀、b₁、b₂.Solve equation below group, the block gold number c of the variation function asked₀, range a and partially base station value c.

$\left\{\begin{array}{c}{b}_0=c_0\\ b_1=3c/2a\\ b_2=-c/2a^3\end{array}\right.$

Result substitutes into, and obtains the variation function formula of matching.

3) acquiring method of covariance matrix (seeking covariance by variation function) in general kriging analysis method:

By the 1st) pressure of step relatively change substitute into the 2nd) variation function that step obtains, try to achieve the variation between all nodes Functional value；

4) during general kriging analysis, the pressure of each node is with the relative pressure between this node with Water Source Pumping Station node Power change is relevant, if the pressure function of this node is: h (x)=α₀+α₁x；

Wherein: x is the relative pressure change of required node and Water Source Pumping Station node；α₀And α₁For function coefficients, wherein For required weight coefficient；K is known node (monitoring point) quantity；

5) general kriging analysis equation Λ X is determined_t=B_t, (t=0,1 ..., P) in variation function matrix Λ and variation Functional vector B_t, wherein:

$\mathrm{\&Lambda;}=\left[\begin{array}{cccccccc}\mathrm{\&gamma;}(x_1-x_1)& \mathrm{\&gamma;}(x_1-x_2)& L& \mathrm{\&gamma;}(x_1-x_k)& 1& h_1\left(x_1\right)& L& h_P\left(x_1\right)\\ \mathrm{\&gamma;}(x_2-x_1)& \mathrm{\&gamma;}(x_2-x_2)& L& \mathrm{\&gamma;}(x_2-x_k)& 1& h_1\left(x_2\right)& L& h_P\left(x_2\right)\\ M& M& O& M& M& M& O& M\\ \mathrm{\&gamma;}(x_k-x_1)& \mathrm{\&gamma;}(x_k-x_2)& L& \mathrm{\&gamma;}(x_k-x_k)& 1& h_1\left(x_k\right)& L& h_P\left(x_k\right)\\ 1& 1& L& 1& 0& 0& L& 0\\ h_1\left(x_1\right)& h_1\left(x_2\right)& L& h_1\left(x_k\right)& 0& 0& L& 0\\ h_2\left(x_1\right)& h_2\left(x_2\right)& L& h_2\left(x_k\right)& 0& 0& L& 0\\ M& M& O& M& M& M& O& M\\ h_P\left(x_1\right)& h_P\left(x_2\right)& L& h_P\left(x_k\right)& 0& 0& L& 0\end{array}\right]$

For weight coefficient vector,

$B_t^T=\&lsqb;0,0,...,0,\mathrm{\&delta;}(0,t),\mathrm{\&delta;}(1,t),...,\mathrm{\&delta;}(P,t)\&rsqb;;$

The necessary condition that the estimation of pressure function coefficient should meet unbiasedness is:

Wherein:

In formula, P is the quantity of pressure function multinomial different item, P=1 in the method；γ(x_i-x_ii) it is node x_iWith joint Point x_iiBetween variation function value, by step 3) can obtain；For Lagrange's multiplier, represent that required object is pressure function t Term coefficient；h_l(x_i) it is the i-node variable x of pressure function multinomial l item_iCall by value, remaining mathematic sign meaning refers to General Krieger basic skills；

6) matrix equation is calculated；Inverse matrix Λ of changes persuing different Jacobian matrix Λ^-1, take t=0 and 1 and carry out matrix computations respectively X_t=Λ^-1B_t；(t=0,1), obtains factor alpha₀And α₁Weight coefficient vector

7) by weight coefficient vectorSubstitute intoTry to achieve node pressure function coefficients α₀And α₁Value, then node pressure function determines；

8) by step 1) each relative pressure change x without monitoring node with Water Source Pumping Station node can be obtained in water supply network, X is substituted into node pressure function h (x)=α₀+α₁In x, can be calculated each without monitoring node force value.

The present invention provides new research direction in water supply network macroscopic analysis field, geo-statistic spatial analysis principle is drawn Enter in inquiring into pipe network state, take into full account pipe network structure characteristic and hydraulic parameters randomness, define a kind of new pipe network State macroscopic analysis method.Meanwhile, for pipeline section character between the architectural characteristic of pipe network, node (caliber, pipe range, pipeline section resistance system Number etc.) interrelated degree, it is proposed that be suitable for pipe network state analysis relative distance substitute Euclidean distance, by geo-statistic side Method principle combines with system-head curve, and the research and development for pipe network state analysis method has reference and directive significance.

Below in conjunction with a specific embodiment, this method is described in detail.

Fig. 1 is a water supply network schematic diagram.Known network topology, numbering nodes and pipe sections and each pipeline section caliber, Pipe range and hydraulic parameters.

In addition to Water Source Pumping Station node 1, pipe network is respectively arranged with pressure monitoring device, then at node [6,9,14,20] totally 4 Can record the T moment comprise at the 5 of Water Source Pumping Station node node gross head be respectively [82.15,67.17,35.78, 31.22,21.28], unit is 10kPa.

Use general kriging analysis method that the pressure of other nodes of water supply network is carried out interpolation.The steps include:

1) suppose this pipe network total water consumption 338L/s in the T moment, carry out assignment of traffic by area ratio discharge method and obtain pipe In net, each node is in the simulation water consumption in T moment.With each node of water supply network simulation water consumption, design pipe range, caliber, pipe Road friction coefficient is known conditions, uses finite difference calculus try to achieve matrix Q and matrix Q is converted into relative pressure transformation matrices V, I-th row of matrix V, jth column element represent relative pressure change (table 1) between node i and node j.

The relative pressure change of each node of table 1

Continued

2) with Water Source Pumping Station node 1 as starting point, check in and the phase of other known monitoring point in relative pressure transformation matrices V Pressure is changed.Relative pressure change is sorted from small to large, and seek two nodes of correspondence respectively by variation function definition Variation function value (table 2).

Table 2 presses relative pressure change sequence and corresponding variation function value

3) variation function is sought.By matched curve and the comparison of model, selected suitable theoretical model, use least square method Definitive variation function coefficients.Selecting the shrink-fit structure based on spherical model in this example, fitting result is:

$\mathrm{\&gamma;}\left(h\right)=\left\{\begin{array}{c}0,h=0\\ -397.07+1208.1h-26{.25h}^3,0<h<3.92\\ 2778.4,h>3.92\end{array}\right.---\left(1\right)$

4) during general kriging analysis, the relative pressure between the pressure of node and node and Water Source Pumping Station node becomes Change relevant, if the pressure function form of node is: h (x)=α₀+α₁x。

Wherein: x is the relative pressure change to pumping plant of the required node； Required weight coefficient；

5) general kriging analysis equation Λ X is calculated_t=B_t’The variation function matrix Λ of (t=0,1) and variation function vector B_t；Relative pressure change between each node checks in relative pressure transformation matrices V, substitutes into variation function and calculates；

6) matrix equation is calculated.Inverse matrix Λ of changes persuing different Jacobian matrix Λ^-1Go forward side by side respectively when taking t=0 and 1 row matrix meter Calculate X_t=Λ^-1B_t；(t=0,1), obtains factor alpha₀And α₁Weight coefficient vectorResult of calculation is as follows:

Table 3 weight coefficient result of calculation

7) weight coefficient vectorSubstitute intoThe node pressure letter asked Number factor alpha₀And α₁Value be respectively 78.26 and-31.69.Then estimate that node pressure function is: h (x)=78.26-31.69x；

8) from relative pressure transformation matrices V poor water supply network the relative pressure of each node and Water Source Pumping Station node become Change, and substituted into above formula calculates each node pressure value.

The result of calculation of general kriging analysis method is listed in Fig. 2 and compares analysis, figure can be seen that analog result Satisfactory, simulation precision is higher, and maximum relative error is less than 2%.

The above embodiment of the present invention is only for example of the present invention is described, and is not the enforcement to the present invention The restriction of mode.For those of ordinary skill in the field, can also be made other not on the basis of the above description Change and variation with form.Here cannot all of embodiment be given exhaustive.Every belong to technical scheme That is amplified out obviously changes or changes the row still in protection scope of the present invention.

Claims (2)

1. a water supply network based on geo-statistic spatial analysis theory is without monitoring node pressure determination statement, it is characterised in that: Choose water supply network Water Source Pumping Station node and other with the pressure of the node of pressure monitoring devices as known pressure, then base Use general kriging analysis method to carry out interpolation calculation to without monitoring node pressure in known node pressure, thus obtain water supply network Node pressure without monitoring point；During interpolation calculation, with the Europe during change substitutes general kriging analysis method relatively of pressure between node Family name's distance；

Water supply network is without concretely comprising the following steps that monitoring node pressure determines:

1) sign of relation size between node:

Utilize the relative change of pressure between all nodes of calculation of design parameters of water supply network, relatively change size by design pressure Characterize the relation between two node spaces；

2) the asking for of general kriging analysis method variation function:

With Water Source Pumping Station monitoring node as starting point, according to the 1st) step gained, by other monitoring node and Water Source Pumping Station monitoring node Between relative pressure variable quantity sort from small to large, and by variation function definition ask for variation function value；With relative distance as horizontal stroke Coordinate, variation function value is ordinate, describes the relation curve between both in rectangular coordinate system；Theoretical with variation function Model figure compares, and selects similar theoretical model and is fitted by least square method, so that it is determined that variation function；

3) by the 1st) pressure of step relatively change substitute into the 2nd) variation function that step obtains, try to achieve the variation letter between all nodes Numerical value；

4) during general kriging analysis, the pressure of each node becomes with the relative pressure between this node and Water Source Pumping Station node Change relevant, if the pressure function of this node is: h (x)=α₀+α₁x；

Wherein: x is the relative pressure change of required node and Water Source Pumping Station node；α₀And α₁For function coefficients, wherein For required weight coefficient；K is known monitoring node number；

5) general kriging analysis equation Λ X is determined_t=B_t, the variation function matrix Λ in (t=0,1) and variation function vector B_t； Wherein the variation function value between each node in variation function matrix Λ is by the 3rd) step can obtain；

6) matrix equation is calculated；Inverse matrix Λ of changes persuing different Jacobian matrix Λ^-1, take t=0 and 1 and carry out matrix computations X respectively_t= Λ^-1B_t；(t=0,1), obtains factor alpha₀And α₁Weight coefficient vector

7) by weight coefficient vectorSubstitute intoTry to achieve node pressure function coefficients α₀With α₁Value, then node pressure function determines；

8) by step 1) each relative pressure change x without monitoring node with Water Source Pumping Station node can be obtained in water supply network, by x generation Ingress pressure function h (x)=α₀+α₁In x, can be calculated each without monitoring node force value；

Step 1) between interior joint the change relatively of pressure obtain as follows: assuming that pipe network is at the total water consumption in T moment, by face Long-pending specific discharge method carries out assignment of traffic and obtains in pipe network each node in the simulation water consumption in T moment；With in water supply network each Node simulation water consumption, design pipe range, caliber, coefficient of pipe friction are known conditions, use finite difference calculus to try to achieve matrix Q,

Wherein,The change impact on its all node pressures for pipe network node flow；

The diagonal entry that each row in matrix Q are respectively divided by its correspondence can be converted into relative pressure transformation matrices V,

I-th row of matrix V, jth column element represent the relative pressure change between node i and node j.

Water supply network the most according to claim 1 is without monitoring node pressure determination statement, it is characterised in that: according to depicting The comparison of relation curve and Variogram Theory Model figure and water supply network hydraulic characteristic(s), described variation function theory mould Spherical model chosen by type, and its covariance variation function is

$\mathrm{\&gamma;}\left(h\right)=\left\{\begin{array}{cc}0,& h=0\\ c_0+c(\frac{3h}{2a}-\frac{h^3}{2a^3}),& 0<h\&le;a\\ c_0+c,& h>a\end{array}\right.$

Wherein: h is the relative pressure change in water supply network between known pressure node；

c₀Block gold number for nested model；

A is the range of nested model；

C is the inclined base station value of nested model；

c₀+ c is the base station value of nested model；

According to the pressure of known monitoring node, least square method is used to draw the coefficient c in above-mentioned covariance variation function₀, a and c。