CN102354111B - Working fluid level prediction-based optimization method for stroke frequency of submersible reciprocating oil pumping unit - Google Patents

Abstract

The invention relates to a stroke frequency method of a submersible reciprocating oil pumping unit, more particularly to a working fluid level prediction-based optimization method for a stroke frequency of a submersible reciprocating oil pumping unit. Therefore, problems that there is no reasonable stroke frequency optimization control method for a submersible reciprocating oil pumping unit and a linear motor can not be adjusted according to an oil well state can be solved. The method comprises the following steps that: step one, normalization processing is carried out on a working fluid level time sequence; step two, a sample space of the working fluid level time sequence that has been processed by the normalization processing is reconstructed; step three, working fluid level prediction model supporting a vector machine is established and regression prediction on the working fluid level time sequence is carried out; step four, an economic target function of stroke frequency optimization is determined; step five, a stroke frequency optimization model is determined; step six, with utilization of the stroke frequency optimization model obtained in the step five, an optimization method for the stroke frequency is obtained by employing section division and time interval division methods. According to the invention, the method is applied to an oil well of a submersible reciprocating oil pumping unit.

Description

Latent oily reciprocating beam-pumping unit jig frequency optimization method based on the working fluid level prediction

Technical field

The present invention relates to a kind of latent oily reciprocating beam-pumping unit jig frequency method.

Background technology

The oily reciprocating beam-pumping unit of diving is a kind of new pumping unit that drives with linear electric motors, at present, the auxiliary facility of this oil pumper that each elephant uses is still incomplete, especially linear motor control system lacks effectively optimal control method, most oily reciprocating beam-pumping units of diving are in uncontrolled running status, this situation can cause the waste of oil-field development resource and the raising of cost of winning, and the oily reciprocating beam-pumping unit failure rate that may cause diving raises.

Oily reciprocating beam-pumping unit has adopted the novel artificial lifting way that linear electric motors drive owing to dive, so that oil pumper has good controllability and energy-saving potential.But at present, the control system of latent oily reciprocating beam-pumping unit is not considered from energy-conservation and angle High-efficient Production when design, linear electric motors jig frequency (being linear electric motors per minute reciprocal time) adjustment cycle long (one is more than 1 month), can't in time adapt to the variation of down-hole working fluid level, oil pumper very likely is in sky and takes out state before next jig frequency adjustment cycle arrives, so both wasted electric energy and also the oil pumper life-span itself has been impacted.

At present, the duty of oily reciprocating beam-pumping unit of diving is manually adjusted beam-pumping unit jig frequency after by manual measurement well fluid level high and low position, adjustment cycle is long, and rely on artificial experience fully, do not possess the theory support that jig frequency is optimized, working fluid level changes frequently pumping unit of well and may be in for a long time high jig frequency sky and take out state and waste electric energy, when the working fluid level position is low, the oil well that jig frequency is lower may flow slow because of down-hole liquid again, cause the faults such as wax deposition, therefore, the oily reciprocating beam-pumping unit of diving needs a kind of rational jig frequency optimal control method, in time adjust linear electric motors according to the oil well state, reach energy-conservation, the purpose of High-efficient Production.

Summary of the invention

The objective of the invention is does not have rational jig frequency optimal control method in order to solve the oily reciprocating beam-pumping unit of diving, and can not adjust according to the oil well state problem of linear electric motors, and a kind of latent oily reciprocating beam-pumping unit jig frequency optimization method based on the working fluid level prediction is provided.

The inventive method is to realize by following steps:

Step 1, working fluid level seasonal effect in time series normalized;

Step 2, the working fluid level seasonal effect in time series sample space after the step 1 normalized is reconstructed;

Step 3, set up support vector machine working fluid level forecast model, carry out working fluid level seasonal effect in time series regression forecasting;

Step 4, determine the economy objective function that jig frequency is optimized

Take the parameter relevant with the economy of the oily reciprocating beam-pumping unit operation of diving or function as parameter, set up the function model of jig frequency and economic index, to the power of the influence degree of function, simplify the economy function according to each parameter in the economy function, obtain objective function;

Step 5, determine the jig frequency Optimized model:

Jig frequency upper and lower bound value in conjunction with concrete oil pumper, and predefined working fluid level prediction fate, provide the constraint condition that jig frequency is optimized, in conjunction with constraint condition and the economy objective function that jig frequency is optimized, obtain the jig frequency Optimized model of this oily reciprocating beam-pumping unit of diving;

Step 6, the jig frequency Optimized model that utilizes step 5 to obtain, and employing by stages, method at times obtains the optimization method of jig frequency: the jig frequency interval at the oily reciprocating beam-pumping unit of diving is divided into some sub-ranges, correspondingly the interval with the working fluid level of this oil well is divided into corresponding sub-range, the hydrodynamic face amount that prediction is obtained, jig frequency value in the jig frequency sub-range corresponding with the working fluid level sub-range at its place is complementary, total predicted time is divided into the some time section, respectively within each time period, the jig frequency value substitution spurt Optimized model that will be complementary with the hydrodynamic face amount that prediction obtains calculates, and obtains the jig frequency optimum results.

The present invention has following beneficial effect: one, the present invention does not need the hydrodynamic face amount of manual measurement oil well, can obtain the variation tendency of following working fluid level by the historical data of working fluid level, utilize the working fluid level Forecasting Methodology, be applied to the jig frequency optimization of latent oily reciprocating beam-pumping unit with predicting the outcome, in time adjust linear electric motors according to the oil well state, so that the raising of oil pumper auto-control degree, and energy-saving effect is remarkable.Two, the algorithm that the present invention relates to is through simplifying, and algorithm is simple, is easy to on-the-spot realization of recovering the oil.

Description of drawings

Fig. 1 is based on the process flow diagram of the latent oily reciprocating beam-pumping unit jig frequency optimization method of working fluid level prediction, and Fig. 2 is the comparative graph (the SVM method is that SVM prediction method, ARMA method are linear prediction method among the figure) of SVM prediction method and linear prediction method precision of prediction; Fig. 3 is the comparative graph (the SVM method is the SVM prediction method among the figure, and improving the SVM method is the SVM prediction method of inputting with jig frequency) of common SVM prediction method and the SVM prediction method precision of prediction of inputting with jig frequency; Fig. 4 is that jig frequency is optimized error curve diagram.

Embodiment

Embodiment one: in conjunction with Fig. 1 present embodiment is described, present embodiment realizes by following steps:

Step 1, working fluid level seasonal effect in time series normalized;

Step 2, the working fluid level seasonal effect in time series sample space after the step 1 normalized is reconstructed;

Step 3, set up support vector machine working fluid level forecast model, carry out working fluid level seasonal effect in time series regression forecasting;

Step 4, determine the economy objective function that jig frequency is optimized

Take the parameter relevant with the economy of the oily reciprocating beam-pumping unit operation of diving or function as parameter, set up the function model of jig frequency and economic index, to the power of the influence degree of function, simplify the economy function according to each parameter in the economy function, obtain objective function;

Step 5, determine the jig frequency Optimized model:

Jig frequency upper and lower bound value in conjunction with concrete oil pumper, and predefined working fluid level prediction fate, provide the constraint condition that jig frequency is optimized, in conjunction with constraint condition and the economy objective function that jig frequency is optimized, obtain the jig frequency Optimized model of this oily reciprocating beam-pumping unit of diving;

Step 6, the jig frequency Optimized model that utilizes step 5 to obtain, and employing by stages, method at times obtains the optimization method of jig frequency: the jig frequency interval at the oily reciprocating beam-pumping unit of diving is divided into some sub-ranges, correspondingly the interval with the working fluid level of this oil well is divided into corresponding sub-range, the hydrodynamic face amount that prediction is obtained, jig frequency value in the jig frequency sub-range corresponding with the working fluid level sub-range at its place is complementary, total predicted time is divided into the some time section, respectively within each time period, the jig frequency value substitution spurt Optimized model that will be complementary with the hydrodynamic face amount that prediction obtains calculates, and obtains the jig frequency optimum results.

Embodiment two: in the step 1 of present embodiment to working fluid level seasonal effect in time series normalized process:

Adopt formula

The sample data of working fluid level and jig frequency is changed in the same order of magnitude scope, the raw data of working fluid level and linear electric motors jig frequency is normalized between [1,1],

In the formula:

Be normalization data, x _iBe raw sample data, x _MaxBe the maximal value in the time series, x _MinBe the minimum value in the time series.

Because when supporting vector machine model is predicted, stress to revise the corresponding weights of the large variable of numerical value, can impact by little Weight Training corresponding to variable of logarithm value, therefore, the variable that numerical value is large is larger on the output bias impact, and working fluid level data variation scope is large, differ more with the order of magnitude of jig frequency data, when working fluid level and jig frequency data during jointly as support vector machine training input, cause easily flooding of jig frequency information, therefore before being trained, must process data on network.Other step is identical with embodiment one.

Embodiment three: the restructuring procedure in working fluid level timed sample sequence space in the step 2 of present embodiment: take the normalization time series of working fluid level and jig frequency as the basis, structure support vector machine training sample, and input the reconstruct of sample space, input, output matrix are

Formula one:

Formula two: $\left(\begin{array}{c}{y}_1\\ y_2\\ .\\ .\\ .\\ y_{n-m}\end{array}\right)=\left(\begin{array}{c}{x}^{(m+1)}\\ x^{(m+2)}\\ .\\ .\\ .\\ x^{\left(n\right)}\end{array}\right)$

In the formula: x _iAnd y _iBe respectively i day input and output vectors (i=1,2 ..., n-m), x ⁽ⁱ⁾Sample value for i day in the time series, n is the total fate of seasonal effect in time series, m is the fate of i before day, also cry and embed dimension, the working fluid level data that make i day are L (i), can predict i day working fluid level by front m days the working fluid level historical data of i day, in actual oil recovery process, the work jig frequency of latent oil piston pump is as the oil recovery process controlled quentity controlled variable, its variation is relatively slow, therefore, and when introducing jig frequency data are predicted i day working fluid level as another input message, quote known proxima luce (prox. luc) jig frequency K (i-1) as input, have

Formula three:

In the formula:

Be Nonlinear Mapping

In restructuring procedure, the choosing method that embeds dimension m adopts pseudo-neighborhood method.Other step is identical with embodiment one.

Embodiment four: set up in the step 3 of present embodiment in the support vector machine working fluid level forecast model, the kernel function of choosing is radial basis function:

Set up support vector machines working fluid level forecast model, carry out working fluid level seasonal effect in time series regression forecasting, utilize supporting vector machine model to predict and be by Nonlinear Mapping the sample of the input space to be mapped to high-dimensional feature space to do linear regression, namely have Make SVM that corresponding regression function is

Formula four: y (x)=ω Φ (x)+b

In the formula:

Represent former sample space,

Sample space after the expression mapping, m ₀For shining upon the dimension of rear sample space, n ₀Be the dimension in former sample space, ω is the support vector machine weight vector, and Φ (x) is mapping function, and the vector that ω and Φ (x) are corresponding is m ₀Dimensional vector, b is amount of bias, ω and b can be obtained by following formula:

Formula five: $\underset{\mathrm{\ω},b,\mathrm{\ξ},{\mathrm{\ξ}}^{*}}{\min }\frac{1}{2}{\mathrm{\ω}}^T-\mathrm{\ω}+C\underset{i=1}{\overset{N}{\mathrm{\Σ}}}({\mathrm{\ξ}}_i+{\mathrm{\ξ}}_i^{*})$

$s.t.\begin{array}{}\left\{\begin{array}{c}{y}_i-{\mathrm{\ω}}^T\mathrm{\Φ}\left(x_i\right)-b\≤\mathrm{\ϵ}+{\mathrm{\ξ}}_i\\ {\mathrm{\ω}}^T\mathrm{\Φ}\left(x_i\right)+b-y_i\≤\mathrm{\ϵ}+{\mathrm{\ξ}}_i^{*}\\ {\mathrm{\ξ}}_i,{\mathrm{\ξ}}_i^{*}\≥0,i=1,\·\·\·,N\end{array}\right.\end{array}$

In the formula, C is the penalty coefficient of support vector machine, and N is input sample total, ξ _iWith

For lax, find the solution ω in the formula formula five at higher dimensional space, choosing radial basis function is kernel function, obtains the ω expression formula to be

Formula six: $\mathrm{\ω}=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}({\mathrm{\α}}_i-{\mathrm{\α}}_i^{*})\mathrm{\Φ}\left(x_i\right)$

In the formula, α _iWith

Be Lagrange multiplier, can try to achieve amount of bias b according to optimal condition, then regression function is

Formula seven: $y\left(x\right)=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}({\mathrm{\α}}_i-{\mathrm{\α}}_i^{*})K(x_i,x)+b$

In the formula, K (x _i, x) be radial basis function, can be in the hope of the Lagrange multiplier α in seven in the formula _iWith

Corresponding vector is support vector, with the input of the working fluid level sample data after the reconstruct as the support vector model, the working fluid level time series is done regression forecasting, is predicted the outcome.Other step is identical with embodiment one.

Embodiment five: the economy objective function in the step 4 of present embodiment

Determine:

Take the parameter relevant with the economy of the oily reciprocating beam-pumping unit operation of diving or function as parameter, the function model of setting up jig frequency and economic index is:

Formula eight: S (k)=1440[ova (l) ek-dc (l) k]

S (k) is the economy objective function of jig frequency k, and in the formula: o is crude oil price, and e is the fluid ratio, v is the oil-immersed pump useful volume, and d is the unit electricity price of Power Consumption of Pumping Units, and a (l) is the function of oil-immersed pump degree of filling and working fluid level relation, c (l) is the function of every jig frequency power consumption and working fluid level relation, o in the above parameter, e, v, d is constant, the degree of filling of oil-immersed pump is the function of suction pressure, and suction pressure is closely related with working fluid level, has

Formula nine: p _f=p _x+ ρ ₀Gl

In the formula: p _fBe pump intake pressure, p _cBe surface casing pressure, ρ ₀For fluid density g is acceleration of gravity, l is dynamic liquid level height,

According to each parameter in the economy function to the power of the influence degree of function, simplify the economy function, ignoring casing pressure, under the condition of oil pressure, a (l) regards the function take dynamic liquid level height as independent variable as, in like manner, under the different dynamic liquid level height, the linear electric motors difference of exerting oneself, its power consumption is also different, c (l) also regards the function of working fluid level as, and in optimizing process, working fluid level is again the function of jig frequency k in the forecast model of working fluid level, this shows, S (k) is the nonlinear function of jig frequency k, and a (l) and c (l) are constant function, economy function S (k) are approximately the linear function of jig frequency k, the economic worth of unit interval oil pump capacity is reduced to the economy optimization aim function in a period of time greater than the oil pumper electric energy loss

Formula ten: $Y=\max \left(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{k}_i\right)$

In the formula: Y is the economy objective function after simplifying, and n is total fate, k _iIt is i days jig frequency.Other step is identical with embodiment one.

Embodiment six: determine the jig frequency Optimized model in the step 5 of present embodiment: in conjunction with jig frequency higher limit and lower limit and the predefined working fluid level prediction fate of oil pumper, provide the constraint condition that jig frequency is optimized, in conjunction with constraint condition and the economy objective function that jig frequency is optimized, obtain the jig frequency Optimized model of this oily reciprocating beam-pumping unit of diving

Formula 11: $\max \left(\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{k}_i\right)$

$s.t.\left\{\begin{array}{cc}{k}_i\≥k_{i-1}\≥\·\·\·\≥k_{i-a}\≥K_U& (a\≤T_U)\\ k_i\≤k_{i-1}\≤\·\·\·\≤k_{i-b}\≤K_D& (b\≤T_D)\end{array}\right.$

In the formula: n is for optimizing total fate of time period, k _iBe i days jig frequency, K _UBe the jig frequency higher limit of linear electric motors continuous working, K _DBe the jig frequency lower limit of linear electric motors continuous working, k _I-aBe a days linear electric motors jig frequency before the prediction day, k _I-bBe b days linear electric motors jig frequency before the prediction day, T _UBe the higher limit of the high jig frequency of oil-immersed pump and low jig frequency continuous working fate, T _DLower limit for the high jig frequency of oil-immersed pump and low jig frequency continuous working fate.Other step is identical with embodiment one.

Embodiment seven: adopt the by stages in the step 6 of present embodiment, method at times obtains the optimization method of jig frequency: the jig frequency interval at the oily reciprocating beam-pumping unit of diving is divided into some sub-ranges, correspondingly the interval with the working fluid level of this oil well is divided into corresponding sub-range, jig frequency value in the hydrodynamic face amount that obtains of the prediction jig frequency sub-range corresponding with the working fluid level sub-range at its place is complementary, total predicted time is divided into the some time section, respectively within each time period, calculate in the jig frequency value substitution spurt Optimized model formula 11 that will be complementary with the hydrodynamic face amount that prediction obtains, obtain the jig frequency optimum results.Other step is identical with embodiment one.

Embodiment eight: when the division jig frequency sub-range of present embodiment and working fluid level sub-range, can have overlapping between the jig frequency sub-range, the working fluid level sub-range cannot have overlapping, and each working fluid level sub-range must have well-determined jig frequency sub-range corresponding with it.Other step is identical with embodiment seven.

Embodiment nine: the time period that present embodiment is divided in total predicted time is the interval of isometric and non-overlapping copies, and the effect of optimization that how much affects optimized algorithm and arithmetic speed that the time period divides can be set according to the concrete condition of oil pumper is on-the-spot.Other step is identical with embodiment seven.

Application example of the present invention:

1, with a bite dive working fluid level historical data and the linear electric motors jig frequency data instance of oily reciprocating pump oil well, 1181.70 meters of this well depths, within in June in 2008 on May 31st, 1 day 1 this time period, the operation area, place is measured once this well fluid level every day, do during sampling and repeatedly measure the processing of averaging, the work jig frequency of the latent oil piston pump of record, and rule of thumb reject obviously unusual data, by the adjacent data replacement of averaging, the jig frequency variation range is per minute 0～12 time.

At first the working fluid level time series is carried out normalized, reconstruct working fluid level timed sample sequence space; Then, set up support vector machine working fluid level forecast model, carry out working fluid level seasonal effect in time series regression forecasting, the working fluid level that obtains this oil well predicts the outcome; Determine the economy objective function that jig frequency is optimized, take the parameter relevant with the economy of the oily reciprocating beam-pumping unit operation of diving or function as parameter, set up the function model of jig frequency and economic index, according to each parameter in the economy function to the power of the influence degree of function, simplify the economy function, obtain objective function; At last, determine the jig frequency Optimized model, jig frequency upper and lower bound value in conjunction with concrete oil pumper, and predefined working fluid level prediction fate, provide the constraint condition that jig frequency is optimized, constraint condition and economy objective function in conjunction with jig frequency optimization, obtain the jig frequency Optimized model of this oily reciprocating beam-pumping unit of diving, adopt the by stages, method at times obtains the optimization method of jig frequency, jig frequency interval at the oily reciprocating beam-pumping unit of diving is divided into some sub-ranges, correspondingly the interval with the working fluid level of this oil well is divided into corresponding sub-range, the hydrodynamic face amount that prediction is obtained, jig frequency value in the jig frequency sub-range corresponding with the working fluid level sub-range at its place is complementary, total predicted time is divided into the some time section, within each time period, the jig frequency value substitution spurt Optimized model that will be complementary with the hydrodynamic face amount that prediction obtains calculates, and obtains the jig frequency optimum results respectively.

2, for investigating the precision of working fluid level prediction, introduce three error precision prediction index, and set up linear prediction model and supporting vector machine model compares, and make simulation analysis.

(1) average relative error:

In the formula, y _i(i) and y _d(i) be respectively actual working fluid level and the prediction hydrodynamic face amount of i day

(2) root-mean-square error:

(3) in the test sample book, Relative Error | the sample point of e|≤10% accounts for the number percent of total sample number.

Table 1 working fluid level predicated error is analyzed

The key of working fluid level prediction is, jig frequency generation acute variation causes when working fluid level changes, predict the outcome and still can follow the tracks of well the variation of working fluid level, carry out emulation experiment to linear method, support vector machine method with the support vector machine working fluid level Forecasting Methodology of jig frequency input, obtain working fluid level predicated error curve as shown in Figures 2 and 3, contrast respectively the predicated error of 3 kinds of methods, can find out that the support vector machine working fluid level Forecasting Methodology of precision of prediction aspect band jig frequency input is better than common support vector machine method and linear prediction method.

3, take this test pit as example, choosing the test sample book fate is 50 days, and jig frequency Optimized model parameter is chosen T _U=3, T _D=1, K _U=11, K _D=2, the jig frequency sub-range is divided as shown in table 2.

Table 2 jig frequency sub-range table

In 50 days, under the continuous non-stop run condition of reciprocating pump, jig frequency summation when using artificial experience adjustments is 288, jig frequency summation when using optimization method to adjust is 315, curve as shown in Figure 4, the jig frequency adjustment that can find out optimization method is more mild than experience adjustments, and concentrates on the interval stage casing of jig frequency, has therefore guaranteed the stationarity of machine operation.Total jig frequency had improved 8.7% more originally, prove the optimization jig frequency after, output is improved significantly.

Claims (3)

1. latent oily reciprocating beam-pumping unit jig frequency optimization method based on working fluid level prediction, it is characterized in that: described method realizes by following steps:

Step 1, working fluid level seasonal effect in time series normalized:

Adopt formula

The sample data of working fluid level and jig frequency is changed in the same order of magnitude scope, the raw data of working fluid level and linear electric motors jig frequency is normalized between [1,1],

In the formula:

Be normalization data, x _iBe raw sample data, x _MaxBe the maximal value in the time series, x _MinBe the minimum value in the time series;

The restructuring procedure in the working fluid level timed sample sequence space after step 2, the step 1 normalized:

Take the normalization time series of working fluid level and jig frequency as the basis, structure support vector machine training sample, and input the reconstruct of sample space, input, output matrix are

Formula one:

Formula two:

In the formula: x _iAnd y _iBe respectively i day input and output vectors (i=1,2 ..., n-m), x ⁽ⁱ⁾Sample value for i day in the time series, n is the total fate of seasonal effect in time series, m is the fate of i before day, also cry and embed dimension, the working fluid level data that make i day are L (i), can predict i day working fluid level by front m days the working fluid level historical data of i day, when introducing jig frequency data are predicted i day working fluid level as another input message, quote known proxima luce (prox. luc) jig frequency K (i-1) as input, have

Formula three:

In the formula: Be Nonlinear Mapping

In restructuring procedure, the choosing method that embeds dimension m adopts pseudo-neighborhood method;

Step 3, set up support vector machine working fluid level forecast model, carry out working fluid level seasonal effect in time series regression forecasting, the kernel function of choosing is radial basis function:

Set up support vector machines working fluid level forecast model, carry out working fluid level seasonal effect in time series regression forecasting, utilize supporting vector machine model to predict and be by Nonlinear Mapping the sample of the input space to be mapped to high-dimensional feature space to do linear regression, Φ is namely arranged:

Make SVM that corresponding regression function is

Formula four: y (x)=ω Φ (x)+b

In the formula:

Represent former sample space,

Sample space after the expression mapping, m ₀For shining upon the dimension of rear sample space, n ₀Be the dimension in former sample space, ω is the support vector machine weight vector, and Φ (x) is mapping function, and the vector that ω and Φ (x) are corresponding is m ₀Dimensional vector, b is amount of bias, ω and b can be obtained by following formula:

Formula five:

In the formula, C is the penalty coefficient of support vector machine, and N is input sample total, ξ _iWith

For lax, find the solution ω in the formula formula five at higher dimensional space, choosing radial basis function is kernel function, obtains the ω expression formula to be

Formula six:

In the formula, α _iWith

Be Lagrange multiplier, can try to achieve amount of bias b according to optimal condition, then regression function is

Formula seven:

In the formula, K (x _i, x) be radial basis function, can be in the hope of the Lagrange multiplier α in seven in the formula _iWith

Corresponding vector is support vector, with the input of the working fluid level sample data after the reconstruct as the support vector model, the working fluid level time series is done regression forecasting, is predicted the outcome;

Step 4, determine the economy objective function that jig frequency is optimized

Take the parameter relevant with the economy of the oily reciprocating beam-pumping unit operation of diving or function as parameter, the function model of setting up jig frequency and economic index is:

Formula eight: S (k)=1440[ova (l) ek-dc (l) k]

S (k) is the economy objective function of jig frequency k, and in the formula: o is crude oil price, and e is the fluid ratio, v is the oil-immersed pump useful volume, and d is the unit electricity price of Power Consumption of Pumping Units, and a (l) is the function of oil-immersed pump degree of filling and working fluid level relation, c (l) is the function of every jig frequency power consumption and working fluid level relation, o in the above parameter, e, v, d is constant, the degree of filling of oil-immersed pump is the function of suction pressure, and suction pressure is closely related with working fluid level, has

Formula nine: p _f=p _c+ ρ ₀Gl

In the formula: p _fBe pump intake pressure, p _cBe surface casing pressure, ρ ₀For fluid density g is acceleration of gravity, l is dynamic liquid level height, a (l) regards the function take dynamic liquid level height as independent variable as, in like manner, under the different dynamic liquid level height, the linear electric motors difference of exerting oneself, its power consumption is also different, c (l) also regards the function of working fluid level as, in optimizing process, working fluid level is again the function of jig frequency k in the forecast model of working fluid level, this shows, S (k) is the nonlinear function of jig frequency k, a (l) and c (l) are constant function, economy function S (k) is approximately the linear function of jig frequency k, and the economic worth of unit interval oil pump capacity is reduced to the economy optimization aim function in a period of time greater than the oil pumper electric energy loss

Formula ten:

In the formula: Y is the economy objective function after simplifying, and n is total fate, k _iIt is i days jig frequency;

Step 5, determine the jig frequency Optimized model:

Jig frequency higher limit and lower limit and predefined working fluid level prediction fate in conjunction with oil pumper provide the constraint condition that jig frequency is optimized, and in conjunction with constraint condition and the economy objective function that jig frequency is optimized, obtain the jig frequency Optimized model of this oily reciprocating beam-pumping unit of diving

Formula 11:

In the formula: n is for optimizing total fate of time period, k _iBe i days jig frequency, K _UBe the jig frequency higher limit of linear electric motors continuous working, K _DBe the jig frequency lower limit of linear electric motors continuous working, k _I-aBe a days linear electric motors jig frequency before the prediction day, k _I-bBe b days linear electric motors jig frequency before the prediction day, T _UBe the higher limit of the high jig frequency of oil-immersed pump and low jig frequency continuous working fate, T _DLower limit for the high jig frequency of oil-immersed pump and low jig frequency continuous working fate;

Step 6, the jig frequency Optimized model that utilizes step 5 to obtain, and employing by stages, method at times obtains the optimization method of jig frequency: the jig frequency interval at the oily reciprocating beam-pumping unit of diving is divided into some sub-ranges, correspondingly the interval with the working fluid level of this oil well is divided into corresponding sub-range, the hydrodynamic face amount that prediction is obtained, jig frequency value in the jig frequency sub-range corresponding with the working fluid level sub-range at its place is complementary, total predicted time is divided into the some time section, respectively within each time period, the jig frequency value substitution spurt Optimized model that will be complementary with the hydrodynamic face amount that prediction obtains calculates, and obtains the jig frequency optimum results.

2. described latent oily reciprocating beam-pumping unit jig frequency optimization method based on working fluid level prediction according to claim 1, it is characterized in that: when dividing jig frequency sub-range and working fluid level sub-range, can have overlapping between the jig frequency sub-range, the working fluid level sub-range cannot have overlapping, and each working fluid level sub-range must have well-determined jig frequency sub-range corresponding with it.

3. described latent oily reciprocating beam-pumping unit jig frequency optimization method based on working fluid level prediction according to claim 1, it is characterized in that: the time period of being divided in total predicted time is the interval of isometric and non-overlapping copies.

Images