3.1 Conventional Ion Characteristics of Water Samples
Test data of PW and UW water samples (Table 1).
Table 1
Routine hydrochemical analysis data of water samples
No.
|
pH
|
K++Na+
|
Ca2+
|
Mg2+
|
CO32−
|
HCO3−
|
Cl−
|
SO42−
|
(mg/L)
|
(mg/L)
|
(mg/L)
|
(mg/L)
|
(mg/L)
|
(mg/L)
|
(mg/L)
|
PW1
|
7.46
|
826.30
|
44.01
|
28.25
|
8.43
|
307.23
|
1020.24
|
305.33
|
PW 2
|
7.52
|
864.71
|
35.41
|
18.53
|
9.30
|
341.95
|
1007.83
|
314.39
|
PW 3
|
7.91
|
778.71
|
46.77
|
29.78
|
11.16
|
283.62
|
987.64
|
295.85
|
PW 4
|
7.63
|
912.81
|
28.1
|
12.08
|
312.74
|
11.95
|
943.13
|
365.14
|
PW 5
|
7.53
|
835.01
|
36.35
|
14.91
|
309.71
|
17.92
|
943.13
|
300.5
|
PW 6
|
7.49
|
732.84
|
37.51
|
16.64
|
252.02
|
11.95
|
933.29
|
272.7
|
PW 7
|
8.37
|
794.88
|
46.62
|
15.33
|
5.54
|
329.00
|
841.47
|
371.5
|
PW8
|
8.55
|
841.16
|
32.79
|
13.03
|
282.38
|
17.92
|
918.77
|
290.00
|
PW9
|
8.31
|
732.84
|
37.51
|
16.64
|
252.02
|
11.95
|
933.29
|
272.70
|
PW10
|
8.36
|
719.80
|
38.06
|
18.52
|
224.69
|
8.96
|
946.62
|
268.80
|
PW11
|
8.31
|
735.59
|
37.43
|
17.69
|
218.62
|
8.96
|
931.96
|
267.70
|
PW12
|
8.24
|
835.93
|
41.71
|
20.51
|
18.45
|
225.16
|
821.44
|
268.12
|
PW13
|
8.23
|
868.37
|
29.11
|
15.54
|
11.16
|
243.91
|
1046.90
|
343.05
|
PW14
|
8.40
|
835.01
|
36.35
|
14.91
|
309.71
|
17.92
|
943.13
|
300.50
|
PW15
|
8.35
|
825.55
|
39.65
|
16.08
|
297.56
|
11.95
|
953.95
|
300.50
|
UW1
|
8.31
|
377.17
|
70.00
|
22.48
|
19.53
|
360.20
|
322.58
|
354.50
|
UW2
|
7.80
|
415.17
|
68.40
|
30.12
|
22.37
|
357.57
|
358.32
|
573.63
|
UW3
|
7.59
|
470.10
|
35.28
|
12.51
|
21.35
|
245.08
|
449.51
|
248.34
|
UW4
|
8.91
|
383.34
|
43.22
|
14.38
|
33.48
|
280.78
|
3403.44
|
325.30
|
UW5
|
8.39
|
504.53
|
47.09
|
15.40
|
5.56
|
326.16
|
347.39
|
374.6
|
UW6
|
8.29
|
506.95
|
55.31
|
20.67
|
5.58
|
278.25
|
231.88
|
278.25
|
UW7
|
8.30
|
460.52
|
28.03
|
6.34
|
38.41
|
473.52
|
216.26
|
249.76
|
UW8
|
8.31
|
459.78
|
25.99
|
6.42
|
76.89
|
281.45
|
225.44
|
325.67
|
UW9
|
8.23
|
417.27
|
14.63
|
14.96
|
43.41
|
250.18
|
259.08
|
109.69
|
UW10
|
8.34
|
464.93
|
25.26
|
11.46
|
39.53
|
337.51
|
221.96
|
333.72
|
UW11
|
8.37
|
434.88
|
46.62
|
15.33
|
5.58
|
329.00
|
341.47
|
371.50
|
UW12
|
8.39
|
514.53
|
47.09
|
15.40
|
5.58
|
326.16
|
347.39
|
374.60
|
UW13
|
7.52
|
371.98
|
82.86
|
28.30
|
25.47
|
357.80
|
237.70
|
445.60
|
UW14
|
8.29
|
306.95
|
55.31
|
20.67
|
257.51
|
278.25
|
331.88
|
278.25
|
UW15
|
7.98
|
415.38
|
40.88
|
13.54
|
321.91
|
253.97
|
271.26
|
266.75
|
According to the mathematical statistical analysis of conventional ion content in water samples, it was found that the average contents of K++Na+, Ca2+, Mg2+, CO32−, HCO3−, Cl− and SO42− in PW water samples were 809.3mg/L, 37.83mg/L, 17.9mg/L, 168.237mg/L, respectively UW water samples were 433.57mg/L,45.73mg/L, 16.5 mg/L, 61.48 mg/L, 315.73 mg/L, 504.37 mg/L and 327.34 mg/L. On the average, the order of anion content in the two water samples is Cl−> SO42−> HCO3−> CO32−, and the order of cation change is K++Na+> Ca2+> Mg2+. However, the contents of K++Na+, CO32−, HCO3− and Cl− in the two kinds of water samples are quite different, especially the contents of K++Na+ and Cl− in PW are twice as high as those in UW, which may be due to the use of sodium hypochlorite in tap water disinfection by waterworks, and the residual sodium hypochlorite causes the excessive contents of Cl−and Na+ (Wang et al.2018).
In order to determine the hydrochemical type of water sample, draw Piper three-line discrimination diagram of water sample (Figure 2). According to Fig. 2, on Piper diagram, the hydrochemical types of PW water samples are mainly concentrated on Na-Cl·SO4, and those of UW water samples are mainly concentrated on Na-Cl.
3.2 Sample selection for conventional ion discrimination
In this study, tap water samples from buried pipes and shallow groundwater samples were collected, and PW1-PW8 and UW1-UW were used to build the model, and PW9-PW15 and UW9-UW15 were used to verify the model.
In order to quickly and effectively establish a discriminant model based on conventional ions in cities, it is necessary to find the characteristic components that can represent PW and UW, that is, the characteristic ions of PW and UW, which can be processed by box statistical graph and statistical software classification principal component analysis, respectively, so as to provide a basis for selecting discriminant factors when further studying the discriminant model of underground pipe leakage.
The comparison of conventional ion content characteristics between PW1-PW8 and UW1-UW8 can be visually expressed by box diagram (A diagram shows conventional water chemical composition of PW and B diagram shows conventional water chemical composition of UW). It can be seen from Fig. 3 that PW and UW show different characteristics, which explain the importance of water-rock interaction in conventional water chemical composition of water samples to varying degrees, mainly reflected in K++Na+ and Cl−. PW shows high K++Na+ and high Cl−, UW shows low K++Na+ and low Cl−, which can be used to distinguish PW from UW(Craig 1964).
The Minitab software is used to calculate the discriminant results of the best subset of conventional ions in each aquifer. Table 2 shows that the subsets of K++Na+, Ca2+, CO32− and SO42− have the largest R-Sq(adj) value, so these ions are selected as recognition factors(Berry 2011).
Table 2
Calculation results of optimal subset of conventional ions for each water sample
|
|
Mallows
|
|
|
|
|
|
|
|
|
|
|
Vars
|
R-Sq
|
R-Sq(adj)
|
C-p
|
S
|
pH
|
Na++K+
|
Ca2+
|
Mg2+
|
CO32−
|
HCO3−
|
SO42−
|
Cl−
|
1
|
8.1
|
7.1
|
3.8
|
0.82751
|
|
▀
|
|
|
|
|
|
|
1
|
7.5
|
6.5
|
4.5
|
0.83025
|
|
|
|
|
|
▀
|
|
|
2
|
11.3
|
9.4
|
2.5
|
0.81744
|
|
▀
|
▀
|
|
|
|
|
|
2
|
11.2
|
9.2
|
2.7
|
0.81810
|
|
▀
|
|
|
|
|
▀
|
|
3
|
13.3
|
10.4
|
2.5
|
0.81259
|
|
▀
|
▀
|
|
|
|
▀
|
|
3
|
13.1
|
10.3
|
2.6
|
0.81333
|
|
|
▀
|
|
|
▀
|
|
▀
|
4
|
14.7
|
10.9
|
3.1
|
0.81064
|
|
▀
|
▀
|
|
▀
|
|
▀
|
|
4
|
14.6
|
10.8
|
3.1
|
0.81083
|
|
▀
|
▀
|
▀
|
|
|
▀
|
|
5
|
15.9
|
11.2
|
3.8
|
0.80922
|
|
▀
|
▀
|
▀
|
▀
|
|
▀
|
|
5
|
14.8
|
10.0
|
5.0
|
0.81467
|
|
▀
|
▀
|
|
▀
|
▀
|
▀
|
|
6
|
16.0
|
10.3
|
5.7
|
0.81344
|
|
▀
|
▀
|
▀
|
▀
|
|
▀
|
▀
|
6
|
15.9
|
10.2
|
5.8
|
0.81380
|
|
▀
|
▀
|
▀
|
▀
|
▀
|
▀
|
|
7
|
16.4
|
9.7
|
7.3
|
0.81597
|
|
▀
|
▀
|
▀
|
▀
|
▀
|
▀
|
▀
|
7
|
16.0
|
9.2
|
7.7
|
0.81808
|
▀
|
▀
|
▀
|
▀
|
▀
|
|
▀
|
▀
|
8
|
16.7
|
8.9
|
9.0
|
0.81949
|
▀
|
▀
|
▀
|
▀
|
▀
|
▀
|
▀
|
▀
|
3.3 Discrimination model of buried pipe water leakage based on conventional ions
In further analysis, factor analysis of ions extracted from the optimal subset was performed using SPSS. The results obtained are shown in Table 3. From the table, we can see that three principal factors (PC1, PC2, and PC3) are obtained under the premise that the eigenvalue is greater than 1, and the overall information interpretation rate reaches 69.43%. After orthogonal rotation, the information interpretation rates of VF1, VF2 and VF3 are 28.43%, 25.43% and 15.57%, respectively. Among them, Ca2+ and SO42− have higher positive load on VF1, while Mg2+ has medium positive load. According to previous studies, VF1 is mainly related to the dissolution of carbonates and sulfates, while K++Na+ and HCO3− have higher positive load on VF2, Cl− has higher positive load on VF3, and CO32− has medium positive load, indicating that VF2 and VF3 mainly represent the weathering of feldspathic silicates and the dissolution of chloride salts(Wu et al.2004).
Table 3
Calculation and analysis results of SPSS factors for conventional ions
Comp-onent
|
Characteristic value
|
Before rotation
|
After rotation
|
Total
|
Variance /%
|
Cumul-ative variance /%
|
Total
|
Variance /%
|
Cumul-ative variance /%
|
Total
|
Variance /%
|
Cumulative varianc/%
|
1
|
2.35
|
29.35
|
29.35
|
2.35
|
29.35
|
29.35
|
2.28
|
28.43
|
28.43
|
2
|
2.00
|
24.94
|
54.29
|
2.00
|
24.94
|
54.29
|
2.03
|
25.43
|
60.86
|
3
|
1.21
|
15.14
|
69.43
|
1.21
|
15.14
|
69.43
|
1.25
|
15.57
|
76.43
|
4
|
0.96
|
11.99
|
81.42
|
|
|
|
|
|
|
5
|
0.66
|
8.19
|
89.61
|
|
|
|
|
|
|
6
|
0.45
|
5.68
|
95.29
|
|
|
|
|
|
|
7
|
0.37
|
4.57
|
99.86
|
|
|
|
|
|
|
8
|
0.01
|
0.14
|
100.00
|
|
|
|
|
|
|
|
FC1
|
FC2
|
FC3
|
|
|
VF1
|
VF2
|
VF3
|
|
pH
|
-0.66
|
-0.51
|
-0.03
|
|
pH
|
-0.78
|
-0.26
|
-0.18
|
|
K++Na+
|
0.12
|
0.94
|
0.05
|
|
K++Na+
|
0.41
|
0.85
|
0.16
|
|
Ca2+
|
0.79
|
-0.09
|
-0.04
|
|
Ca2+
|
0.71
|
-0.35
|
0.08
|
|
Mg2+
|
0.64
|
-0.32
|
0.20
|
|
Mg2+
|
0.46
|
-0.52
|
0.27
|
|
CO32−
|
-0.01
|
0.08
|
0.03
|
|
CO32−
|
-0.07
|
0.06
|
0.50
|
|
HCO3−
|
-0.47
|
0.81
|
0.03
|
|
HCO3−
|
-0.18
|
0.92
|
0.03
|
|
SO42−
|
0.28
|
0.06
|
0.84
|
|
Cl−
|
0.13
|
-0.07
|
0.88
|
|
Cl−
|
0.75
|
0.25
|
-0.46
|
|
SO42−
|
0.86
|
0.00
|
-0.31
|
|
The pH, K++Na+, Ca2+, Mg2+, CO32−, HCO3−, SO42− and Cl− are determined as characteristic ions, and classified and analyzed by SPSS. the model is established according to the obtained equations. the final discriminant equations are shown in Tables 4 and 5. It can be concluded from Table 5 that the established discriminant equation is:
F1=2.283×pH+0.015×Ca2+-0.004×CO32−+0.002×HCO3−+0.001×SO42-20.136 (1)
F2=0.516×pH+0.003×Ca2++0.012×CO32−-0.004×HCO3−+0.002×SO42−5.014 (2)
Where F1 and F2 were the discriminant functions, and pH, Ca2+, CO32-, HCO3-, and SO42- respectively represented their contents.
Table 4
Interpretation of conventional ion model information of each water sample
Function
|
Characteristic value
|
Variance /%
|
Cumulative variance /%
|
Canonical correlation
|
1
|
2.49
|
51.30
|
51.30
|
0.81
|
2
|
1.50
|
30.50
|
84.00
|
0.75
|
Table 5
Discrimination equation based on conventional ions
Criterion function
|
Function 1
|
Function 2
|
Criterion function
|
Function 1
|
Function 2
|
pH
|
2.283
|
0.516
|
HCO3−
|
0.002
|
-0.004
|
Ca2+
|
0.015
|
0.003
|
SO42−
|
0.001
|
0.002
|
CO32−
|
-0.004
|
0.012
|
Constant
|
-20.136
|
-5.014
|
In order to investigate whether the above judgment method was excellent or not, samples PW9-PW15 and UW9-UW15 (altogether 16) from known sources were substituted into the judgment function formula for regression analysis. The regression results are shown in Table 6, where Ⅰ is the tap water sample from the underground buried pipe, and Ⅱ is the shallow underground water sample near the underground buried pipe.
Table 6
Results of Back Judgment Test
Sample No.
|
Discriminating result
|
Actual type
|
Sample No
|
Discriminating result
|
Actual type
|
PW9
|
Ⅰ
|
Ⅰ
|
UW9
|
Ⅰ
|
Ⅱ
|
PW10
|
Ⅰ
|
Ⅰ
|
UW10
|
Ⅱ
|
Ⅱ
|
PW11
|
Ⅰ
|
Ⅰ
|
UW11
|
Ⅱ
|
Ⅱ
|
PW12
|
Ⅰ
|
Ⅰ
|
UW12
|
Ⅱ
|
Ⅱ
|
PW13
|
Ⅰ
|
Ⅰ
|
UW13
|
Ⅱ
|
Ⅱ
|
PW14
|
Ⅰ
|
Ⅰ
|
UW14
|
Ⅰ
|
Ⅱ
|
PW15
|
Ⅰ
|
Ⅰ
|
UW15
|
Ⅱ
|
Ⅱ
|
It can be seen from Table 6 that the overall discrimination effect is good, reaching 87.50%, and shallow groundwater samples near two underground pipes are discriminated as tap water samples of underground pipes. However, the discriminant model based on conventional ions is difficult to accurately describe the leakage of buried tap water, that is, the amount of tap water entering shallow groundwater.
3.4 Distribution of hydrogen and oxygen isotopes in PW and UW and its influencing factors
Atmospheric precipitation is the ultimate source of groundwater, but the composition of hydrogen and oxygen isotopes in atmospheric precipitation in different regions is different due to different natural geographical environments. Craig(1961) found that there is a close correlation between δD and δ18O in atmospheric precipitation, and put forward the correlation formula: δD=8 × δ18O+10, because the δ D-δ18O graph is a straight line, which is called the atmospheric precipitation line (MWL)(Craig 1964). Dansgaard first proposed the concept of deuterium surplus, which was used to evaluate the degree of deviation of hydrogen and oxygen isotopes in regional precipitation from the global precipitation line. The formula is d = δD-8δ18. Atmospheric precipitation lines in different regions are slightly different due to different geographical environments (Dansgaard 1964). Rozanskiet al. (1993) analyzed 206 samples from all over the world through global LAEA network stations, and obtained the arithmetic mean value (r2=0.99): δ D = (8.17±0.06) × δ18O+(10.35±0.65)(Rozanskiet et al.1982). According to the statistics of Zheng Shuhui et al. (1983), the relationship of hydrogen and oxygen isotopic composition of modern atmospheric precipitation in China is δD=7.9 × δ18O+8.2(Zheng et al.1983).
The distribution of hydrogen and oxygen isotopes in water samples is shown in Figure 4. It can be seen from Fig. 4 that the precipitation line in Hefei is basically parallel to the global precipitation line (due to the lack of precipitation isotope data in Hefei area, this study selects precipitation data in Anhui Province instead(Zhang et al.2017) ) and slightly higher than the global precipitation line, and the hydrogen and oxygen isotope values in water samples all fall below the atmospheric precipitation line (LMWL) in China, indicating that the recent sources of PW and UW are atmospheric precipitation, and that atmospheric precipitation experienced strong evaporation before being recharged underground(Zhang et al.2020). At the same time, it is found that the water in the pipe network is closer to the precipitation line than the shallow groundwater near the pipe network. It may be that the shallow groundwater near the pipe network is irradiated by sunlight and evaporates strongly, which makes the hydrogen and oxygen isotopes in the water relatively enriched(Wang et al.2013). Therefore, it is possible to distinguish the water samples of the pipe network and the water samples of the leakage points of the pipe network according to the difference of deuterium values in the water body of the tap water network and its nearby shallow groundwater(Liu et al.1997).
It can also be seen from Fig. 4 that both kinds of water samples show the characteristics of hydrogen and oxygen isotope drift. Different reasons lead to isotope drift in different layers of groundwater. The change of δ18O is mainly the exchange between groundwater and surrounding rocks, mainly carbonate rocks. The change of δD is mainly caused by the exchange of groundwater with minerals containing OH radical (gypsum, clay minerals) or H2S (Wicks et al.1994;A et al.2010).
In a word, because there are few samples and scattered water samples at present, no obvious isotope drift characteristics can be seen, but the shallow groundwater has the characteristics of δ18O drift and δD drift due to the rich silicate minerals in the soil.
3.5 Establishment of Water Leakage Discrimination Model Based on Hydrogen and Oxygen Isotope
Select conventional ions Ca2+, Mg2+, K++Na+, CO32−, HO3−, Cl−, SO42− and isotopes δD and δ18O in water samples, and draw the correlation diagram between hydrogen and oxygen isotopes of PW and UW and conventional ions by SPSS software, as shown in Fig. 5.
It can be seen from Fig. 6 that δ18O of PW has a good correlation with K++Na+, and δD has a good correlation with K++Na+, which indicates that the δ value of hydrogen and oxygen isotope of buried tap water sample has a great relationship with the content of K++Na+. Ions with great correlation with δ18O in shallow groundwater samples include Ca2+, Mg2+ and CO32−, while ions with great correlation with δD are Cl−. This result may be due to the difference of hydrogen and oxygen isotopes between PW and UW, which mainly comes from the different selectivity between ions in the process of water-rock interaction. It can be seen that hydrogen and oxygen isotopes basically have the function of water source discrimination, so the hydrogen and oxygen isotope characteristics of water samples can be used to discriminate water leakage in urban buried pipes.
The hydrogen and oxygen isotopes of water samples are shown in Table 7. The stable hydrogen and oxygen isotopes of PW and UW are different. It is found that the stable oxygen isotope δ18O in PW ranges from -7.05‰ to -7.38‰, with an average value of-7.233‰. Hydrogen isotope δ d ranges from -47.37‰ to -49.64‰, with an average value of -48.38‰. The stable oxygen isotope δ18O in UW ranges from -5.75‰ to -6.74‰, with an average value of -6.350‰. Hydrogen isotope δD Dranges from -39.36‰ to -47.48‰, with an average value of -44.239‰. Comparison shows that the stable hydrogen and oxygen isotopes in PW are lower than those in UW, and the hydrogen and oxygen isotopes in UW are more abundant. By calculating the deuterium value of each water sample, the deuterium value in PW ranges from 9.34‰ to 9.65‰, with an average value of 9.490‰, and the deuterium value in UW ranges from 6.41‰ to 6.71‰, with an average value of 6.545‰. The deuterium value in UW is obviously lower than that in PW, which indicates that the difference of deuterium value between PW and PW can be used to discriminate urban buried pipe leakage.
Table 7
Hydrogen and oxygen isotope contents of PW and UW
No.
|
PW
|
编号
|
UW
|
δ18O/‰
|
δD/‰
|
d/‰
|
δ18O/‰
|
δD/‰
|
d/‰
|
PW1
|
-7.26
|
-48.52
|
9.56
|
UW1
|
-6.54
|
-45.76
|
6.56
|
PW 2
|
-7.05
|
-46.81
|
9.59
|
UW2
|
-5.87
|
-40.34
|
6.62
|
PW 3
|
-7.38
|
-49.64
|
9.40
|
UW3
|
-5.75
|
-39.36
|
6.64
|
PW 4
|
-7.09
|
-47.37
|
9.35
|
UW4
|
-6.17
|
-42.91
|
6.45
|
PW 5
|
-7.23
|
-48.20
|
9.64
|
UW5
|
-6.34
|
-44.30
|
6.42
|
PW 6
|
-7.37
|
-49.31
|
9.65
|
UW6
|
-6.56
|
-45.77
|
6.71
|
PW 7
|
-7.35
|
-49.43
|
9.37
|
UW7
|
-6.74
|
-47.48
|
6.44
|
PW8
|
-7.14
|
-47.62
|
9.50
|
UW8
|
-6.44
|
-45.11
|
6.41
|
PW 9
|
-7.22
|
-48.27
|
9.49
|
UW9
|
-6.61
|
-46.22
|
6.66
|
PW10
|
-7.24
|
-48.58
|
9.34
|
UW10
|
-6.46
|
-45.14
|
6.54
|
PW 11
|
-7.34
|
-48.24
|
10.48
|
UW11
|
-7.24
|
-49.31
|
8.62
|
PW 12
|
-7.08
|
-47.58
|
9.06
|
UW12
|
-6.80
|
-46.87
|
7.54
|
PW13
|
-7.41
|
-46.97
|
12.31
|
UW13
|
-6.63
|
-45.89
|
7.18
|
PW14
|
-7.38
|
-46.37
|
12.67
|
UW14
|
-6.71
|
-46.34
|
7.34
|
PW15
|
-7.24
|
-47.26
|
10.66
|
UW15
|
-6.64
|
-45.39
|
7.73
|
In order to study the influence of water from pipe network on hydrogen and oxygen isotope of shallow groundwater nearby, and the relationship between tap water leakage and deuterium value in mixed water, this study measured the hydrogen and oxygen isotope abundance ratio of mixed water (MW) after mixing PW and UW with different proportions (0%, 5%, 10%, 20%, 40%, 80%, 100%). See Table 8 for the hydrogen and oxygen isotope abundance ratio in mixed water.
Table 8
Hydrogen and oxygen isotopes in mixed water
Proportion of PW /%
|
Mean value
|
|
δ18O/‰
|
δD/‰
|
d/‰
|
0
|
-6.48
|
-45.47
|
6.41
|
5
|
-6.52
|
-45.66
|
6.48
|
10
|
-6.56
|
-45.86
|
6.64
|
20
|
-6.68
|
-46.22
|
7.21
|
40
|
-6.83
|
-46.96
|
7.63
|
80
|
-7.16
|
-48.42
|
8.88
|
100
|
-7.35
|
-49.15
|
9.65
|
It can be seen from Fig. 6 that after PW leaks into shallow underground aquifer, the abundance ratio of hydrogen and oxygen isotopes in mixed water will obviously decrease, and there is a good positive correlation between PW leakage and deuterium value in MW, and the equation is y = 0.0321x+6.3875, R2=0.9945. The prediction model of deuterium value in MW (d MW) and water leakage in pipe network (VPW) is obtained:
$${v}_{\text{P}\text{W}}=({d}_{\text{M}\text{W}}-6.3875)÷0.0321$$
3
To verify the accuracy of the above model, the pipe network tap water samples and underground water samples with known hydrogen and oxygen isotope abundance ratio were prepared into mixed water samples and added into a 10mL centrifuge tube with plugs, in which the proportion of pipe network tap water in the mixed water samples was 0%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% and 100% in turn See Table 9 for the actual and predicted values of water ratio of pipe network in mixed water.
Table 9
Actual value and predicted value of water ratio in pipe network
True value of PW scale /%
|
d/‰
|
True value of PW scale /%
|
True value of PW scale /%
|
d/‰/%
|
True value of PW scale /%
|
0
|
6.40
|
0.39
|
60
|
8.32
|
60.20
|
10
|
6.73
|
10.67
|
70
|
8.67
|
71.11
|
20
|
7.05
|
20.64
|
80
|
9.01
|
81.70
|
30
|
7.38
|
30.92
|
90
|
9.24
|
88.86
|
40
|
-6.83
|
-46.96
|
100
|
9.64
|
101.32
|
50
|
8.02
|
50.86
|
|
|
|
The calculated MW is 0.39% and the error is 0.39% when PW is not added. When PW is 10%, the calculated MW is 10.67%, and the error is 0.67%. When PW is 100%, the calculated MW is 101.32%, and the error is 1.32%. The error between the calculated water content and the actual water content in the whole gradient dilution experiment is about 2.0%. In practical application, it is of great guiding significance to accurately determine the leakage of tap water in urban underground pipes by measuring deuterium value in MW.