Long-term durability prediction system and long-term durability prediction method

Abstract

This technology provides a method for estimating the long-term creep strength of a resin pipe using the results of physical property tests, thereby predicting its long-term durability. [Solution] The long-term durability prediction system for resin pipes comprises: a physical property data acquisition unit 10 that acquires physical property data of resin pipes; a model generation unit 20 that generates a creep strength estimation model that estimates the full-circumference notch creep strength in an arbitrary stress range using statistical analysis with the full-circumference notch creep strength included in the physical property data as the objective variable and multiple physical properties included in the physical property data other than the full-circumference notch creep strength as explanatory variables; and a creep strength prediction unit 50 that predicts the full-circumference notch creep strength of the target resin pipe by applying the prediction physical properties of the target resin pipe to the creep strength estimation model.

Description

[Technical Field]

[0001] This invention relates to a long-term durability prediction system and a long-term durability prediction method for predicting the long-term durability of resin pipes. [Background technology]

[0002] The method for evaluating the physical properties of polyethylene resin according to Patent Document 1 involves the steps of deriving the maximum stress at which permanent deformation does not occur in a polyethylene resin test piece, measuring the content of linking molecules and the entanglement molecular weight of the test piece, and deriving a significant factor for the full-notch creep test time from the relationship between the content ratio of linking molecules and the entanglement molecular weight. The full-notch creep test time value of the polyethylene resin molded product is then predicted based on the maximum stress and the significant factor for the full-notch creep test time.

[0003] In the long-term durability prediction method described in Patent Document 2, the full-circumference notch creep test time value is predicted using a prediction formula that utilizes the content of linking molecules and entanglement molecular weight of polyethylene resin. In this process, in order to obtain training data, a full-notch creep test based on ISO 16770 is performed on test specimens with various physical property data at a stress of 4.0 MPa and a temperature of 80°C. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Patent No. 6685544 [Patent Document 2] Patent No. 6942245 [Overview of the Initiative] [Problems that the invention aims to solve]

[0005] The method for evaluating the physical properties of polyethylene resin according to Patent Document 1 derives values ​​from data obtained by melting the resin, and therefore does not take into account the effects of tubular molding. As a result, the durability of molded pipes cannot be properly evaluated. For example, it is difficult to evaluate the differences in durability between pipes molded under different conditions (pipes molded by different molding companies) or to evaluate the deterioration of durability due to use. Similarly, the long-term durability prediction method according to Patent Document 2 also makes it difficult to evaluate the performance of polyethylene resin material in pipe form.

[0006] In view of the above circumstances, the object of the present invention is to provide a technology for predicting the long-term durability of a resin pipe by estimating the long-term creep strength of a full-circumference notch type using the test results of a physical property test. [Means for solving the problem]

[0007] The long-term durability prediction system for resin pipes according to the present invention comprises: a physical property data acquisition unit that acquires physical property data of the resin pipe; a model generation unit that generates a creep strength estimation model that estimates the full-circumference notch creep strength at stress in an arbitrary stress range using statistical analysis with the full-circumference notch creep strength included in the physical property data as the objective variable and a plurality of physical properties included in the physical property data other than the full-circumference notch creep strength as explanatory variables; and a creep strength prediction unit that predicts the full-circumference notch creep strength of the resin pipe to be predicted by applying predictive physical properties of the resin pipe to be predicted to the creep strength estimation model.

[0008] According to this configuration, a creep strength estimation model is generated that estimates the full-circumference notch creep strength of a resin pipe in an arbitrary stress range by statistically analyzing the full-circumference notch creep strength of a resin pipe, which is included in the material property data of the resin pipe prepared in advance through material property tests, as the dependent variable, and multiple material properties included in the material property data other than the full-circumference notch creep strength as independent variables. By applying the material properties of the resin pipe to be predicted to this creep strength estimation model, the model calculates the full-circumference notch creep strength of the target resin pipe, and this calculated full-circumference notch creep strength corresponds to the rupture time of the target resin pipe. The longer this rupture time, the higher the long-term durability is considered to be. Furthermore, by calculating and comparing the full-circumference notch creep strength of various resin pipes, it becomes possible to compare the strength (long-term durability) of various resin pipes. In this invention, full-circumference notch creep strength refers to the creep strength in a full-circumference notch creep test.

[0009] To generate a creep strength estimation model (regression equation) that calculates the full-circumference notch creep strength (dependent variable) based on statistical analysis, it is first necessary to select material property data that have a strong correlation with the full-circumference notch creep strength as explanatory variables. However, using many types of material properties as explanatory variables leads to the problem of increased burden in obtaining those properties. For this reason, considering the accurate calculation of the full-circumference notch creep strength and the burden of obtaining material properties, this invention proposes that yield stress, tensile elongation, Charpy impact value, and density be used as the explanatory variables.

[0010] In order to properly calculate the creep strength of the full-circumference notch type, it is sometimes preferable to add further physical property data. For this reason, in this invention, it is also proposed that oxidation induction time, melt mass flow rate, and degree of crystallinity may be used as explanatory variables.

[0011] Furthermore, if the creep strength of the full-circumference notch type is strongly influenced by the number of years of use of the resin pipe and the diameter of the resin pipe, it is proposed in the present invention that the number of years of use and the diameter of the resin pipe may be used as explanatory variables.

[0012] Although it varies depending on the type of resin, in a full-circumference notch creep test of polyethylene pipe material, the stress at which slow crack propagation occurs, which is a failure mode due to creep degradation, is generally considered to be 6 MPa or less. However, since the test takes longer the smaller the test stress, it is practically preferable to perform the full-circumference notch creep test at approximately 3 to 6 MPa. For this reason, in the present invention, it is also proposed that the resin pipe be a polyethylene pipe and that the arbitrary stress range be set to 3 to 6 MPa, which is the slow crack propagation range for the polyethylene pipe.

[0013] It is known that regression analysis is used as the statistical analysis to accurately and easily generate a statistical analysis model that takes explanatory variables as input and outputs a dependent variable. In this case, it is preferable to use the best statistical analysis selected from partial least squares regression analysis, multiple regression analysis, principal component regression analysis, etc., depending on the statistical characteristics of each variable used as the explanatory and dependent variables, especially their correlations. Therefore, in this invention, it is proposed that one of the following statistical analysis methods—partial least squares regression, multiple regression analysis, or principal component regression analysis—is selected according to the statistical characteristics of the variables used as the dependent and explanatory variables.

[0014] The present invention is not only directed to a long-term durability prediction system, but also directed to a long-term durability prediction method. Such a long-term durability prediction method comprises: a physical property data acquisition step of acquiring physical property data of the resin pipe; and a model generation step of generating a creep strength estimation model for estimating full-notch creep strength in an arbitrary stress range by using statistical analysis that takes the full-notch creep strength contained in the physical property data as a dependent variable and a plurality of physical property values contained in the physical property data other than the full-notch creep strength as independent variables; and a creep strength prediction step of predicting the full-notch creep strength of the resin pipe to be predicted by applying the physical property values for prediction of the resin pipe to be predicted to the creep strength estimation model. Such a long-term durability prediction method also has various functions and effects described in the above long-term durability prediction system.

[0015] Other features, functions and effects of the present invention will be clarified by the following description of the present invention with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] [Figure 1] It is a schematic diagram showing the schematic configuration of a long-term durability prediction system. [Figure 2] It is a graph showing the relationship between the stress range that causes slow crack growth and the rupture time. [Figure 3] It is a graph showing estimated values and actually measured values of rupture time obtained with a creep strength estimation model generated by multiple regression analysis using four independent variables. [Figure 4] It is a graph showing estimated values and actually measured values of rupture time obtained with a creep strength estimation model generated by multiple regression analysis using seven independent variables. [Figure 5] It is a graph showing estimated values and actually measured values of rupture time obtained with a creep strength estimation model generated by multiple regression analysis using nine independent variables. [Figure 6] It is a graph showing estimated values and actually measured values of rupture time obtained with a creep strength estimation model generated by partial least squares regression analysis. [Figure 7] This graph shows the estimated and measured fracture times obtained from a creep strength estimation model generated by principal component regression analysis. [Modes for carrying out the invention]

[0017] Figure 1 is a schematic diagram showing the general configuration of the long-term durability prediction system according to the present invention and the processing flow of the long-term durability prediction method. The basic functions of the long-term durability prediction system will be explained using Figure 1.

[0018] The long-term durability prediction system in this embodiment predicts the long-term durability of resin pipes, particularly polyethylene pipes. The long-term durability prediction system comprises, as its functional units, a physical property data acquisition unit 10, a model generation unit 20, and a durability prediction management unit 30.

[0019] The physical property data acquisition unit 10 acquires physical property data of the resin pipe. This physical property data includes not only known physical properties of the resin pipe, but also physical properties measured using a test specimen of the resin pipe, in particular its full-circumference notch creep strength (creep strength in the full-circumference notch creep test). The full-circumference notch creep strength here is a value obtained by material testing, and is the logarithm of the fracture time (h) at an arbitrary stress of 3 to 6 MPa applied to the test specimen, obtained by the full-circumference notch tensile creep test. Other physical properties include tensile yield stress (%), tensile elongation (%), and Charpy impact value (kJ / m) related to the resin pipe. 2 ), density (g / cm 3 These include the oxidation induction time (min), melt mass flow rate (g / 10min), and degree of crystallinity (%), as well as characteristic values ​​that are considered physical properties here, such as the number of years the resin pipe has been used (years) and the pipe diameter (A). The physical property data acquisition unit 10 stores the above-mentioned physical property data, obtained through manual input via a terminal or extraction from a database, in a predetermined memory.

[0020] The model generation unit 20 uses the full-circumference notch type creep strength (rupture time) included in physical property value data as an objective variable, and uses a plurality of physical property values included in the physical property value data other than the full-circumference notch type creep strength as explanatory variables to perform statistical analysis, for example multiple regression analysis, thereby generating a creep strength estimation model 40 that estimates and outputs the full-circumference notch type creep strength in any stress range.

[0021] The estimated value estimated by this creep strength estimation model 40 is the rupture time (h) at any stress between 3 to 6 MPa, that is, the rupture time (h) in a so-called full-circumference notch tensile creep test. In the full-circumference notch tensile creep test, this stress range of 3 to 6 MPa is in most cases evaluated as the slow crack growth range for test pieces taken from polyethylene pipes, and it is considered that the larger the estimated value (rupture time) in the stress range of 3 to 6 MPa, the higher the long-term durability. The relationship between such a stress range that causes slow crack growth and the rupture time (lifetime) is shown in Figure 2. The slow crack growth range is located between ductile fracture and brittle fracture caused by the environment.

[0022] In this embodiment, the creep strength estimation model 40 is substantially a regression equation (estimation equation) as shown below. log(t)=a0+a1·Z1+a2·Z2+···+a n ·Z n Here, t is the rupture time (h) in a full-circumference notch tensile creep test, Z i is a standardized explanatory variable, a i is the regression coefficient for each explanatory variable. In this case, Z i is represented by the following formula. Z i =(X i -μ i ) / σ i , Here, X i is the explanatory variable, μ i is the average value of the explanatory variable, σ i is the invariant standard deviation of the explanatory variable.

[0023] Next, we will explain the first, second, and third regression equations, which are multiple regression models generated using the following three groups of material properties included in the material property data of the resin pipe—namely, the first combination example, the second combination example, and the third combination example—as explanatory variables. The objective variable is the logarithm of the fracture time (h) at any stress in the range of 3 to 6 MPa applied to the specimen in a full-circumference notch tensile creep test. Note that the material properties used as explanatory variables are standardized.

[0024] (1) First regression equation The first example combination of explanatory variables is: tensile yield stress (MPa): Z1, tensile elongation (%): Z2, Charpy impact value (kJ / m 2 ):Z3, density (g / cm 3 ):Z4 is included as an explanatory variable. The regression coefficients obtained from the multiple regression analysis are: a0=3.34, a1=-0.34, a2=-0.22, a3=-0.02, a4=0.15, The coefficient of determination was "0.879". Figure 3 shows the relationship between the measured (logarithmic) and estimated (logarithmic) fracture time in the full-circumference notch tensile creep test in this first regression equation. In Figure 3, the horizontal axis represents the measured value and the vertical axis represents the estimated value.

[0025] (2) Second regression equation A second example combination of explanatory variables includes: tensile yield stress (MPa): Z1, tensile elongation (%): Z2, Charpy impact value (kJ / m 2 ):Z3, density (g / cm 3 In addition to (Z4), the following are included as explanatory variables: oxidation induction time (min): Z5, melt mass flow rate (g / 10min): Z6, and crystallinity (%): Z7. The regression coefficients obtained from the multiple regression analysis are: a0=3.34, a1=-0.29, a2=-0.25, a3=-0.17, a4=0.16, a5=0.26, a6=0.07, a7=0.28, And, The coefficient of determination was "0.941". Figure 4 shows the relationship between the measured (logarithmic) and estimated (logarithmic) fracture time in the full-circumference notch tensile creep test in this second regression equation. In Figure 4, the horizontal axis represents the measured value and the vertical axis represents the estimated value.

[0026] (3) Third regression equation A second example combination of explanatory variables includes: tensile yield stress (MPa): Z1, tensile elongation (%): Z2, Charpy impact value (kJ / m 2 ):Z3, density (g / cm 3 In addition to the following explanatory variables: ):Z4, oxidation induction time (min):Z5, melt mass flow rate (g / 10min):Z6, crystallinity (%):Z7, the number of years of resin use (years):Z8, and pipe diameter (A):Z9 are included. The regression coefficients obtained from the multiple regression analysis are: a0=3.34, a1=-0.34, a2=-0.27, a3=-0.25, a4=0.19, a5=0.30, a6=0.12, a7=0.35, a8=0.01, a9=-0.08, And, The coefficient of determination was "0.980". Figure 5 shows the relationship between the measured (logarithmic) and estimated (logarithmic) fracture time in the full-circumference notch tensile creep test in this third regression equation. In Figure 5, the horizontal axis represents the measured value and the vertical axis represents the estimated value.

[0027] The durability prediction management unit 30 includes a creep strength estimation model 40 and a creep strength prediction unit 50. Specifically, the durability prediction management unit 30 generates the creep strength estimation model 40 based on regression coefficients as statistical analysis values ​​generated by the model generation unit 20. The creep strength prediction unit 50 applies the prediction physical properties of the target resin pipe to the creep strength estimation model 40 to estimate the full-circumference notch creep strength (breaking time indicating long-term durability) of the target resin pipe (polyethylene pipe).

[0028] In this embodiment, as described above, the model generation unit 20 measures the physical properties of the polyethylene pipe and the full-circumference notch creep strength at several points, and applies this physical property data to multiple regression analysis to generate a regression equation which is a creep strength estimation model 40 that estimates the full-circumference notch creep strength of an unknown polyethylene resin (corresponding to the rupture time in a full-circumference notch tensile creep test). In this case, the model generation unit 20 can use other statistical analyses, such as partial least squares regression analysis or principal component regression analysis, instead of multiple regression analysis.

[0029] Next, we will explain the analysis results when the model generation unit 20 employs partial least squares regression analysis. Partial least squares regression analysis has the advantage of eliminating weighted fit and multicollinearity of variables. The explanatory variables used are the nine physical properties described above, which are standardized (by subtracting the mean value from each variable and dividing by the standard deviation). The dependent variable is the fracture time in the full-circumference notch tensile creep test, and it is logarithmic. Using the least squares regression analysis method, four principal components were generated from the nine physical properties, and regression analysis was performed, resulting in a coefficient of determination of "0.932". The relationship between the measured values ​​(logarithmic) and estimated values ​​(logarithmic) in this regression equation is shown in Figure 6. In Figure 6, the horizontal axis represents the measured values ​​and the vertical axis represents the estimated values.

[0030] Furthermore, the analysis results when the model generation unit 20 employs principal component regression analysis will be explained. Principal component regression analysis can effectively reduce the number of explanatory variables by aggregating many explanatory variables to create new principal components (latent variables). In this case, there are four latent variables, meaning that the original nine explanatory variables have been aggregated into four. The coefficient of determination for this creep intensity estimation model 40 was "0.78". Figure 7 shows the relationship between the observed values ​​(logarithmic values) and estimated values ​​(logarithmic values) in this regression equation. In Figure 7, the horizontal axis represents the observed values ​​and the vertical axis represents the estimated values.

[0031] As described above, the model generation unit 20 can generate creep intensity estimation models 40 using various statistical analyses, but since each has different characteristics, it may be equipped with a function to select one of partial least squares regression, multiple regression analysis, or principal component regression analysis according to the statistical characteristics (coefficient of determination, coefficient of determination, test statistic, information quantity, etc.) of the variables used as the dependent and independent variables.

[0032] [Another embodiment] (1) In the embodiments described above, the long-term durability of polyethylene pipes was predicted, but the present invention is not limited to polyethylene pipes and can be applied to other resin pipes as well.

[0033] (2) In the above-described embodiment, the durability prediction management unit 30 predicted the long-term durability of one resin pipe, but it may be configured to predict the long-term durability of multiple resin pipes having different shape characteristics, perform relative comparisons of each, and output data for screening.

[0034] (3) In the embodiments described above, the explanatory variables used were tensile yield stress, tensile elongation, Charpy impact value, density, oxidation induction time, melt mass flow rate, degree of crystallinity, years of resin use, and pipe diameter. However, other physical properties may be added or replaced with these other physical properties.

[0035] Furthermore, the configurations disclosed in the above embodiments (including other embodiments, the same applies hereinafter) can be applied in combination with configurations disclosed in other embodiments, as long as no inconsistencies arise. Moreover, the embodiments disclosed herein are illustrative, and the embodiments of the present invention are not limited thereto, and can be modified as appropriate without departing from the object of the present invention. [Industrial applicability]

[0036] This invention is applicable to predicting the long-term durability of various resin pipes. [Explanation of Symbols]

[0037] 10: Physical property data acquisition unit 20: Model generation unit 30: Durability Prediction Management Department 40: Creep Strength Estimation Model 50: Creep strength prediction unit

Claims

1. A system for predicting the long-term durability of resin pipes, A physical property data acquisition unit that acquires physical property data of the resin pipe, A model generation unit generates a creep strength estimation model that estimates the full-circumference notch creep strength in an arbitrary stress range using statistical analysis with the full-circumference notch creep strength included in the aforementioned physical property data as the dependent variable and multiple physical property values ​​other than the full-circumference notch creep strength included in the aforementioned physical property data as independent variables. A creep strength prediction unit predicts the full-circumference notch creep strength of the target resin pipe by applying the predictive physical properties of the target resin pipe to the creep strength estimation model, A long-term durability prediction system equipped with the following features.

2. The long-term durability prediction system according to claim 1, wherein yield stress, tensile elongation, Charpy impact value, and density are used as explanatory variables.

3. The long-term durability prediction system according to claim 2, wherein oxidation induction time, melt mass flow rate, and degree of crystallinity are further used as explanatory variables.

4. The long-term durability prediction system according to claim 3, wherein the number of years of use and the pipe diameter are further used as explanatory variables.

5. The long-term durability prediction system according to claim 1, wherein the resin pipe is a polyethylene pipe, and the arbitrary stress range is set to 3 to 6 MPa, which is the low-speed crack propagation range of the polyethylene pipe.

6. The long-term durability prediction system according to any one of claims 1 to 4, wherein the statistical analysis is selected from partial least squares regression, multiple regression analysis, and principal component regression analysis, depending on the statistical characteristics of the variables used as the dependent variable and independent variables.

7. A method for predicting the long-term durability of resin pipes, A physical property data acquisition step for acquiring physical property data of the resin pipe, A model generation step to generate a creep strength estimation model that estimates the full-circumference notch creep strength in an arbitrary stress range using statistical analysis with the full-circumference notch creep strength included in the aforementioned physical property data as the dependent variable and multiple physical property values ​​other than the full-circumference notch creep strength included in the aforementioned physical property data as independent variables; A creep strength prediction step involves applying predictive physical properties of the target resin pipe to the creep strength estimation model to predict the full circumference notch creep strength of the target resin pipe, A long-term durability prediction method that includes the following features.

Images