CN105868508A - A Quantitative Productivity Prediction Method Based on Gas Logging Information - Google Patents

Abstract

The invention discloses a capacity quantitative prediction method based on gas logging information, which comprises the following steps: 1, acquiring gas logging and production data; 2, determining an inflow dynamic equation to convert the yield in the sample data into a non-resistance flow; 3, analyzing productivity influence factors and determining the productivity influence factors related to the unimpeded flow; 4, establishing a neural network prediction model, determining input nodes of the neural network according to productivity influence factors during training, and taking the unimpeded flow converted in the step 2 as output nodes; 5, calculating the predicted unimpeded flow by the neural network model according to the well data to be predicted; and 6, calculating the oil yield by combining the inflow dynamic equation and the predicted unimpeded flow. The method is easy to realize and can fully utilize C₁～C₅The content data of various hydrocarbon components (methane-pentane) improves the gas-oil ratio prediction accuracy, and meanwhile, the method can calculate the gas yield according to the oil yield and the gas-oil ratio, namely, the simultaneous prediction of the petroleum and natural gas yields is realized.

Description

A Quantitative Productivity Prediction Method Based on Gas Logging Information

technical field

The invention relates to the field of exploration and development of oil and natural gas, in particular to a production capacity quantitative prediction method based on gas logging information.

Background technique

With the development of the national economy, people are increasingly dependent on oil and natural gas. Oil and gas well reservoir production prediction can guide the formulation of development plans, and is also of great significance for avoiding exploration and development risks.

Gas logging is a natural gas surface testing technology while drilling. It mainly measures and analyzes the composition and content of natural gas in drilling fluid, and uses chromatographic gas logging technology to separate, identify and measure the collected gas through a chromatographic column. , the content of various hydrocarbon components of C ₁ to C ₅ (methane to pentane) can be continuously recorded.

At present, the application of gas logging mainly focuses on identifying oil and gas reservoirs, such as the commonly used chart method, which includes: Pixler chart, hydrocarbon triangle chart, hydrocarbon ratio chart, humidity method chart, etc. The parameters used in these chart methods Less, generally only containing C ₁ ~ C ₄ , and there is no distinction between n-alkanes and iso-alkanes components, gas measurement data has not been fully used. At present, there are many methods for predicting oil field production capacity, and the commonly used oil production capacity prediction methods are relatively cumbersome and inconvenient to use.

Contents of the invention

The purpose of the present invention is to overcome the deficiencies of the prior art and provide a simple and easy-to-use productivity quantitative prediction method based on gas logging information.

The purpose of the present invention is achieved by the following technical solutions: a method for quantitative prediction of production capacity based on gas logging information, which comprises the steps:

① Obtain gas logging and production data;

②Convert the output in the sample data into unimpeded flow according to the inflow dynamic equation;

③Analyze the factors affecting production capacity to determine the factors affecting production capacity related to unimpeded flow;

④ Establish a neural network prediction model, determine the input node of the neural network according to the factors affecting production capacity during training, and determine the output node according to the unimpeded flow in step ②;

⑤ According to the well data to be logged, the neural network model calculates the predicted unimpeded flow;

⑥ Calculate the oil production by combining the inflow dynamic equation and the predicted open flow.

Preferably, the present invention also includes gas-oil ratio calculation, and the gas production is calculated through the gas-oil ratio and oil production.

The invention uses the production test data and the corresponding gas logging data to establish the fitting relationship between the production gas-oil ratio GOR and the area ratio of light and heavy components. Preferably, the area ratio of the light and heavy components is obtained by the ratio of the content ratio of the light components in the spider diagram to the area A _t formed by the center point and the ratio of the content ratio of the heavy components in the spider diagram to the area A _b formed by the center point , that is, A _t /A _b .

Specifically, the spider graph is established through seven ratios of C1/C2, C1/C3, C2/C3, C2/iC4, C3/iC4, iC4/nC4, and iC5/nC5. C1/C2, C1/ The area A _t formed by C3, C2/C3 and the center point, the area A _b formed by C2/C3, C2/iC4, C3/iC4, iC4/nC4, iC5/nC5, C1/C2 and the center point.

Preferably, the inflow dynamic equation is:

${\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}<span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; \&lsqb;</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.2</span>}<span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; f</span>}}}{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.8</span>}{<span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; f</span>}}}{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&rsqb;</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; G</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; 300</span>}\\ \stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}<span\; class="notranslate">\; \&lsqb;</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.2</span>}<span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; f</span>}}}{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.8</span>}{<span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; f</span>}}}{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&rsqb;</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; Q</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; =</span>\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; Q</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; G</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; \&Greater\; Equal;</span>\mathrm{<span\; class="notranslate">\; 300</span>}\end{array}\right.$

In the formula, Q _o —oil flow rate, m ³ /d;

Q _max — unimpeded flow, m ³ /d;

p _{wf —bottomhole} flowing pressure, MPa;

p _r —formation pressure, MPa;

Δ _Q — negative flow increment, m ³ /d;

GOR—production gas-oil ratio, m ³ /m ³ .

In the production capacity prediction process, the present invention establishes a fitting relational expression between the negative flow increment and the production gas-oil ratio according to the existing flow negative increment data and production gas-oil ratio data.

For the well to be predicted, the formation pressure is calculated by establishing a mathematical model with the depth; the bottomhole flow pressure model must be considered in two cases: ① self-spraying interval, through the existing self-spraying interval bottomhole flow The fitting relationship between bottomhole flow pressure and depth was established based on the data of the variation of bottomhole pressure with depth. (2) In the swabbing interval, the relationship between bottomhole flow pressure and bottomhole flow pressure was established based on the existing data of bottomhole flow pressure changing with pumping unit strokes, stroke times, and formation pressure. The fitting relational expression of stroke×stroke times/formation pressure.

Preferably, the neural network is a BP neural network.

Preferably, the single correlation method is used to analyze the influencing factors of the production capacity, and the five parameters of natural gamma, compensated sound wave, compensated density, deep resistivity and effective thickness are determined as the input nodes of the neural network.

The beneficial effects of the present invention are:

1) This method is easy to implement and can make full use of the content data of various hydrocarbon components from C ₁ to C ₅ (methane to pentane), so as to improve the prediction accuracy;

2) The existing gas side logging applications are mainly used to find oil and gas layers, focusing on the distinction between oil and gas layers and water layers. There are very few studies that specifically distinguish oil and gas layers, but this method can be used to distinguish oil and gas layers;

3) Aiming at the problem that some existing production prediction methods can only predict the production of a single fluid (oil or natural gas), and cannot predict the production of oil and natural gas for oil and gas co-production reservoirs (such as solution gas flooding oil and gas reservoirs), this paper The invention proposes a gas-oil ratio calculation method, which can calculate gas production according to oil production and gas-oil ratio, that is, realizes simultaneous prediction of oil and natural gas production;

4) The traditional Vogel method is improved, and corresponding constraints are set, so that the improved inflow dynamic equation is more accurate than the Vogel equation in predicting production capacity;

5) According to the existing well logging data and gas logging data, a parameter equation in line with the actual situation of the target well is fitted, which makes this method more practical and feasible than the traditional fixed parameter equation prediction.

Description of drawings

Fig. 1 is a flowchart of an embodiment of the present invention;

Fig. 2 is the spider diagram that the embodiment of the present invention adopts;

Fig. 3 is the relationship between certain well A _t /A _b and test production gas-oil ratio GOR in the embodiment of the present invention;

Fig. 4 is the unimpeded flow fitting of certain well inflow dynamic equation parameter in the embodiment of the present invention;

Fig. 5 is a certain well inflow dynamic equation parameter fitting in the embodiment of the present invention;

Fig. 6 is a comparative figure of two kinds of equation IPR curves of a certain well in the embodiment of the present invention;

Fig. 7 is a comparison diagram of a well fitting flow rate error in an embodiment of the present invention;

Fig. 8 is a statistical relationship diagram established between gas-oil ratio and flow rate negative increment of a certain well in an embodiment of the present invention;

Fig. 9 is a diagram showing the relationship between formation pressure and depth of a certain well in an embodiment of the present invention;

Fig. 10 is a diagram of the relationship between the bottom hole flow pressure and the formation depth of a certain well self-spraying section in an embodiment of the present invention;

Fig. 11 is a diagram showing the relationship between bottomhole flow pressure and stroke × stroke times/formation pressure in a certain well swabbing interval in an embodiment of the present invention;

Fig. 12 is the relationship between each parameter of the reservoir section and the production capacity in the embodiment of the present invention;

FIG. 13 is a schematic structural diagram of a BP neural network according to an embodiment of the present invention;

Fig. 14 is the comparison chart of predicted oil production and actual oil production obtained after the present invention predicts 8 wells;

Fig. 15 is a comparison chart of predicted gas production and actual gas production obtained after predicting 8 wells according to the present invention.

detailed description

The technical solution of the present invention will be further described in detail below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited to the following description.

As shown in Figure 1, a method for quantitative prediction of productivity based on gas logging information includes the following steps:

① Obtain gas logging and production data, including well testing data, well logging data and test production and oil test data;

②Convert the production in the well test data into open flow according to the inflow dynamic equation;

③According to the well test data, test production and oil test data, analyze the factors affecting production capacity, and determine the factors affecting production capacity related to open flow;

④ Establish a neural network prediction model, determine the input node of the neural network according to the factors affecting production capacity during training, and determine the output node according to the unimpeded flow in step ②;

⑤ According to the logging data of the well to be predicted, the neural network model calculates the predicted open flow;

⑥ Calculate the oil production by combining the inflow dynamic equation and the predicted open flow.

The data required by the present invention are shown in Table 1, wherein the logging data is used for the input node of the neural network, the gas logging data is used for establishing the gas-oil ratio calculation model and predicting the gas-oil ratio, and the production capacity test data is used for calculating the open flow rate And obtain the data of oil and gas production, formation pressure and bottom hole flow pressure to establish corresponding calculation models.

Table 1 Invention Implementation Data Requirements

Note: ○ in Table 1 indicates that there is corresponding data, and × indicates that there is no corresponding data.

The present invention also includes gas-oil ratio calculation, using sample gas logging and oil and gas production data to establish a gas-oil ratio quantitative calculation model, and calculating gas production through the oil-gas ratio and oil production, and the gas-oil ratio, that is, the production gas-oil ratio, is converted to the atmosphere The ratio of the total gas production under atmospheric conditions to the total oil production converted to atmospheric conditions. The total gas production consists of two parts, which include gas flowing into the wellbore in the form of free gas and dissolved gas dissolved in crude oil under formation conditions. The formula for calculating the gas-oil ratio is as follows:

$\mathrm{<span\; class="notranslate">\; G</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; g</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; g</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}}{{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; a</span>}}}$

In the formula, GOR—production gas-oil ratio, m ³ /m ³ ;

q _{ga1 —free} gas production, m ³ ;

q _ga2 - dissolved gas production, m ³ ;

q _oa — oil production, m ³ .

Current prediction methods can only predict the production of a single fluid (oil or natural gas). For some oil and gas co-production reservoirs (such as solution gas flooding reservoirs), existing methods cannot be used to simultaneously predict the production of oil or natural gas. However, the present invention proposes a method for predicting the two-phase productivity of oil and natural gas in view of the characteristics that oil layers, oil and gas co-production layers and gas layers are difficult to distinguish in oil and gas co-production wells, and production capacity cannot be predicted. The invention uses well test (or oil test and production test) data combined with well logging data and mud logging data to establish a prediction model, and finally uses the well logging and mud logging data of the target well to predict the oil and natural gas production of the target well reservoir.

As shown in Figure 2, the present invention utilizes the production test data and the corresponding gas logging data to establish the fitting relationship between the production gas-oil ratio GOR and the area ratio of the light and heavy components, and the area ratio of the light and heavy components is passed through the spider diagram ( The ratio of the content ratio of light components in the star diagram of gas components to the area A _t formed by the center point and the ratio of the content ratio of heavy components in the spider diagram to the area A _b formed by the center point can be obtained, that is, A _t /A _b .

Preferably, the spider graph is established by the seven ratios of C1/C2, C1/C3, C2/C3, C2/iC4, C3/iC4, iC4/nC4, iC5/nC5, C1/C2, C1/ The area A _t formed by C3, C2/C3 and the center point, the area A _b formed by C2/C3, C2/iC4, C3/iC4, iC4/nC4, iC5/nC5, C1/C2 and the center point, as shown in Figure 2 upper half. For layers with greater gas content, A _t is greater relative to A _b , so the area ratio A _t /A _b and the gas-oil ratio GOR can be used for fitting to find the calculation relationship between the two. Combining with Figure 2, the calculation formula of A _t can be obtained as follows: The calculation formula of A _b is as follows:

$\begin{array}{c}{A}_b=\frac{1}{2}\&Center\; Dot;\frac{C_2}{C_3}\&Center\; Dot;\frac{C_2}{{\mathrm{iC}}_4}\&CenterDot;\sin \left(\frac{2\mathrm{\&pi;}}{7}\right)+\frac{1}{2}\&Center\; Dot;\frac{C_2}{{\mathrm{iC}}_4}\&Center\; Dot;\frac{C_3}{{\mathrm{iC}}_4}\&Center\; Dot;\sin \left(\frac{2\mathrm{\&pi;}}{7}\right)+\frac{1}{2}\&CenterDot;\frac{C_3}{{\mathrm{iC}}_4}\&CenterDot;\frac{{\mathrm{iC}}_4}{c_4}\&CenterDot;\sin \left(\frac{2\mathrm{\&pi;}}{7}\right)+\frac{1}{2}\&Center\; Dot;\frac{{\mathrm{iC}}_4}{c_4}\&Center\; Dot;\frac{{\mathrm{iC}}_5}{C_5}\&Center\; Dot;\sin \left(\frac{2\mathrm{\&pi;}}{7}\right)+\\ \frac{1}{2}\&CenterDot;\frac{{\mathrm{iC}}_5}{c_5}\&CenterDot;\frac{C_1}{C_2}\&CenterDot;\sin \left(\frac{2\mathrm{\&pi;}}{7}\right)\end{array}$ ,therefore The lower part of Fig. 2 shows the area ratio A _{t /A b of a certain gas layer, the area ratio A t /A b of a} certain oil and gas layer and the area ratio A _t /A _b of a certain oil layer.

As shown in Fig. 3, the boundary of reservoir fluid properties can be divided according to the gas-oil ratio, and the chart is divided into three regions: the gas-oil ratio GOR≤300 is the oil layer region; the gas-oil ratio 300<GOR≤20000 is the oil-gas same layer region; The gas-oil ratio GOR>20000 is the gas zone. Using the production test data and the corresponding gas logging data to establish the fitting relationship between the production gas-oil ratio GOR and the area ratio of light and heavy components, it can be seen that with the increase of A _t /A _b , the gas-oil ratio GOR increases exponentially, The established fitting relationship is: GOR=3.0455(A _t /A _b ) ^4.3005 , R ² =0.8312, the correlation is good, and R ² is above 0.8.

The present invention establishes an inflow dynamic equation applicable to two situations in which the formation fluid contains free gas and dissolved gas or only contains dissolved gas:

In the formula, Q _o —oil flow rate, m ³ /d;

Q _max — unimpeded flow, m ³ /d;

p _{wf —bottomhole} flowing pressure, MPa;

p _r —formation pressure, MPa;

ΔQ—negative increment of flow rate, m ³ /d;

GOR—production gas-oil ratio, m ³ /m ³ ;

Among them, GOR<300 means that the formation fluid only contains dissolved gas, and GOR≥300 means that the formation fluid contains both dissolved gas and free gas. When GOR<300, the equation can be transformed into:

Y=cX

In the formula, c—constant, Q _max ;

X—independent variable,

Y—dependent variable, Q _o .

The value of c can be fitted by one or more sets of (X, Y) data, that is, the open flow value of the well. Figure 4 shows the fitting of the open flow parameters of the inflow dynamic equation of a certain well. At this time, the open flow is 20.732m ³ /d.

When GOR≥300, the equation can be transformed into:

Y=cX-b

In the formula, c—constant, ;

b—constant, ΔQ;

When there is one or more sets of (X, Y) data, the values of c and b can be fitted. Figure 5 shows the parameter fitting of the inflow dynamic equation of a well, and the open flow value can be calculated as 62.07m ³ /d.

Taking well X as an example, the inflow dynamic curve fitting of the two equations is carried out, as shown in Figure 6, it can be seen that the inflow dynamic relationship proposed by the present invention is more consistent with the test point, and the physical meaning of the negative flow increase can also be seen: with As the flow pressure increases, the production pressure difference decreases, and the output of the two equations begins to decrease from the unobstructed flow rate, while the flow rate of the relationship curve of the equation of the present invention decreases faster. When the flow pressure does not reach the formation pressure, the flow rate reaches 0 in advance, and then the flow rate becomes negative. value. This is because in the case of free gas in the reservoir, the seepage of the gas phase will reduce the relative permeability of the oil phase. On the one hand, the oil flow rate will decrease faster, and on the other hand, the gas will lock the oil under a lower production pressure difference so that The oil phase that could have flowed out could not flow out, and the flow curve that should have flowed out has become a negative value.

According to the well test data of 10 wells, the inflow dynamic curve was fitted by Vogel equation and the equation proposed by the present invention, the production under different flow pressures was calculated and compared with the actual production, and the relative error was calculated. The average error of the Vogel equation is 17.9%, and the average error of the equation proposed by the present invention is 4.7%, indicating that the equation proposed by the present invention is more accurate (as shown in Figure 7).

Preferably, for the well to be predicted, in the process of productivity prediction, to predict the unobstructed flow rate, the production under the actual flow pressure needs to be calculated according to the inflow dynamic equation. If the negative flow increment is unknown, the specific productivity equation cannot be obtained. The negative flow increment is Due to the negative increase in flow rate caused by the decrease of relative permeability of oil phase due to the percolation of free gas phase, the gas-oil ratio GOR is used to establish a statistical relationship with the negative flow rate increase. As shown in Figure 8, according to the existing flow negative increment data and production gas-oil ratio data, a fitting relationship between flow negative increment and production gas-oil ratio is established: ΔQ=4×10 ^-5 GOR ² -0.002GOR+ 2.4591, R ² =0.7448.

Similarly, the formation pressure is calculated by establishing a mathematical model with the depth, as shown in Figure 9, through the existing formation pressure data and depth data, the established mathematical model is: P _r =0.01DEP-0.7548, the correlation coefficient R2 is _0.9931 , where Pr represents the formation pressure, and DEP represents the depth; the ^bottomhole flowing pressure model must be considered in two cases: ① Spontaneous interval, the bottomhole flowing pressure of the existing spontaneous spraying interval The data of depth change establishes the fitting relationship between the bottomhole flowing pressure and the depth, as shown in Fig. 10, the established fitting relationship is: P _wf =0.0081DEP-3.4678, the correlation coefficient R2 is ^0.3991 , where P _wf represents Bottomhole flow pressure, DEP means depth. (2) In the swabbing interval, the fitting relationship between bottomhole flow pressure and stroke × stroke times/formation pressure is established based on the existing data of bottomhole flow pressure changing with formation pressure, as shown in Fig. 11. The established fitting relationship is: P _wf =-1.317(L _st ×T _st /P _r )+7.083, the correlation coefficient R ² is 0.423, where P _wf represents the bottomhole flow pressure, L _st ×T _st /P _r Indicates stroke × stroke times / formation pressure.

Preferably, the single correlation method is used to analyze the influencing factors of the production capacity, and the five parameters of natural gamma, compensated sound wave, compensated density, deep resistivity and effective thickness are determined as the input nodes of the neural network. As shown in Figure 12 (single correlation diagram), it can be seen from Figure 12 that the five parameters of natural gamma, compensated acoustic wave, compensated density, deep resistivity and effective thickness are related to open flow, so the natural gamma, compensated acoustic wave The five parameters of , compensation density, deep resistivity and effective thickness are used as the input nodes of the neural network.

Preferably, the neural network is a BP neural network or a RBF neural network or a support vector machine, and the samples selected during training should be comprehensive, including both high-productivity layers and low-productivity and dry layers, to ensure that the trained The neural network has sufficient generalization ability. Secondly, the selected samples must be representative and typical, and the input parameters and output parameters have a definite implicit relationship to ensure that the network can converge. As shown in Figure 13, select the input and output node data, start training the network after normalization, and adjust the number of hidden nodes and learning efficiency until the error meets the requirements.

The present invention predicts the production capacity of a single well as shown in Table 2. According to the logging data, the open flow rate of each perforation section is predicted through the neural network, and the parameters required for the gas-oil ratio and inflow dynamic relational formula of each perforation section are calculated according to the mathematical model. Therefore, the oil production and gas production are calculated by using the inflow dynamic relationship and the gas-oil ratio. The predicted oil production of this well is 9.9m ^{3 /d, and the gas production is 2692m 3 /} d, while the actual oil production is 7.4m ³ /d, and the gas production is 2653m ³ /d , the predicted output is basically consistent with the actual output.

Table 2 Prediction results of single well productivity

The present invention predicts the productivity of 8 wells (W1～W8), and the results show that the average error of predicted gas-oil ratio is 111m ³ /m ³ , the average error of oil production is 2m ³ /d, and the average error of gas production is 815.1 m ³ /d, the actual production and predicted production of each well are shown in Figure 14 and Figure 15, and the specific data are shown in Table 3. It can be seen that the overall error is within the acceptable range, and it can be judged whether fluid can be produced and The requirements for differentiating the properties of the produced fluids, and the relatively close production ranges indicate that this method can be applied to production capacity prediction.

Table 3 Productivity prediction results of example wells

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. It should be noted that any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention should include Within the protection scope of the present invention.

Claims (10)

1. A productivity quantitative prediction method based on gas logging information, is characterized in that, it comprises the steps: ① Obtain gas logging and production data; ② Determine the inflow dynamic equation to convert the output in the sample data into unimpeded flow; ③Analyze the factors affecting production capacity to determine the factors affecting production capacity related to unimpeded flow; ④ Establish a neural network prediction model, determine the input node of the neural network according to the factors affecting production capacity during training, and use the unimpeded flow converted in step ② as the output node; ⑤ According to the data of the well to be predicted, the neural network model calculates the predicted unimpeded flow; ⑥ Calculate the oil production by combining the inflow dynamic equation and the predicted open flow. 2. A method for quantitatively predicting production capacity based on gas logging information according to claim 1, characterized in that it also includes gas-oil ratio calculation, and the gas production is calculated through the oil-gas ratio and oil production. 3. A kind of production capacity quantitative prediction method based on gas logging information according to claim 2, is characterized in that: utilize production test data and corresponding gas logging data to establish production gas-oil ratio GOR and light-heavy component area ratio Fitting relationship between. 4. A kind of production capacity quantitative prediction method based on gas logging information according to claim 3, it is characterized in that: described light and heavy component area ratio is constituted by light component content ratio and central point in spider diagram The ratio of the area A _t to the content ratio of heavy components in the spider diagram and the area A _b formed by the center point is obtained, that is, A _t /A _b . 5. A kind of production capacity quantitative prediction method based on gas logging information according to claim 4, it is characterized in that: described spider diagram is through C1/C2, C1/C3, C2/C3, C2/iC4, The seven ratios of C3/iC4, iC4/nC4, and iC5/nC5 are established, and the area A _t formed by C1/C2, C1/C3, C2/C3 and the central point, C2/C3, C2/iC4, C3/iC4, The area A _b formed by iC4/nC4, iC5/nC5, C1/C2 and the central point. 6. according to any one of claim 2, 3, 4, 5, a kind of productivity quantitative prediction method based on gas logging information, it is characterized in that: described inflow dynamic equation is: ${\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}<span\; class="notranslate">\; =</span>\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; \&lsqb;</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.2</span>}<span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; f</span>}}}{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.8</span>}{<span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; f</span>}}}{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&rsqb;</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; G</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; 300</span>}\\ \stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}<span\; class="notranslate">\; \&lsqb;</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.2</span>}<span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; f</span>}}}{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.8</span>}{<span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; w</span>}\mathrm{<span\; class="notranslate">\; f</span>}}}{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&rsqb;</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; Q</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; =</span>\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; Q</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; G</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; \&Greater\; Equal;</span>\mathrm{<span\; class="notranslate">\; 300</span>}\end{array}\right.$ In the formula, Q _o —oil flow rate, m ³ /d; Q _max — unimpeded flow, m ³ /d; p _{wf —bottomhole} flowing pressure, MPa; p _r —formation pressure, MPa; ΔQ—negative increment of flow rate, m ³ /d; GOR—production gas-oil ratio, m ³ /m ³ . 7. A kind of production capacity quantitative prediction method based on gas logging information according to claim 6, it is characterized in that: in the process of production capacity prediction, flow rate is established according to existing flow rate negative increment data and production gas-oil ratio data Fitting relationship between negative increment and production gas-oil ratio. 8. A kind of productivity quantitative prediction method based on gas logging information according to claim 7, it is characterized in that: for the well to be predicted, the formation pressure is calculated by establishing a mathematical model with depth; The bottom hole pressure model must be considered in two cases: ① the self-spraying interval, the fitting relationship between the bottomhole flowing pressure and the depth is established through the existing data of the bottomhole flow pressure and depth variation of the self-spraying interval, and ② the swabbing interval , through the existing data of bottom hole flow pressure changing with pumping unit stroke, stroke times and formation pressure, the fitting relational expression of bottom hole flow pressure and stroke×stroke times/formation pressure is established. 9. A method for quantitatively predicting productivity based on gas logging information according to claim 1 or 8, characterized in that: said neural network is a BP neural network. 10. A method for quantitatively predicting production capacity based on gas logging information according to claim 1 or 8, characterized in that: the single correlation method is used to analyze the factors affecting production capacity, and it is determined to combine natural gamma ray, compensated sound wave, and compensated density The five parameters of , deep resistivity and effective thickness are used as the input nodes of the neural network.