3.1. DP Values of Cellulose Paper
Figure 1 illustrates the chemical structure of cellulose polymers. The cellulose consists of linear, polymeric chains of cyclic, β-D-glucopyranose units that are composed of C, H, and O elements [
14]. Cellulose paper mainly comprises cellulose, and the DP value of cellulose is the average amount of glucose for the cellulose chain. Existing works have shown that DP values are closely related to the condition of the cellulose paper, such as mechanical strength and tenacity [
9]. Thus, DP values are used to mirror the aging extent of cellulose paper, and the DP values of the cellulose paper in Groups A–D were tested and are shown in
Figure 2.
The DP values of the cellulose paper rapidly decline in the initial aging period, and then decrease at a low rate due to the structure of cellulose paper. The cellulose paper can be divided into crystalline and amorphous zones in a microcosm, and the two types of zones are alternately joined to form cellulose paper. The molecular arrangement in amorphous areas is loose and disorderly compared with that in crystalline zones. Thus, amorphous areas are more likely to be broken due to the effects of heat, oxygen, and moisture [
4]. Given that the crystalline and the amorphous areas generally alternate, the destruction of the amorphous areas leads to a rapid decrease in the DP values of the cellulose paper. However, the performance of the crystallization areas is relatively stable. Thus, the decrease rate of the DP values slows down in the later aging period.
The comparison between Groups A and B shows that the decreasing rate of DP values in Group B is apparently slower than that in Group A. This phenomenon confirms that the three-element mixed insulation oil can retard the aging rate of cellulose paper. The retardation effect was evident from the beginning of the aging experiment. On the 90th day, the DP values of cellulose paper impregnated by mineral oil dropped to 228, whereas those of cellulose paper impregnated by three-element mixed insulation oil remained above 400.
Groups C and D had oil replacement, and the cellulose paper exhibited different aging trends due to the different oils used. The aging rate of the cellulose paper was almost unchanged after oil replacement with unused mineral oil, and the DP values of cellulose paper in Group C almost coincided with those in Group A. By contrast, the aging rate of cellulose paper was apparently reduced after oil replacement with the three-element mixed insulation oil, and the DP values of the cellulose paper in Group D were considerably higher than those in Groups A and C. When the DP values of the cellulose paper in Group A dropped to 228 on the 90th day, those of the cellulose paper in Group D were 370. The comparison of the aging experiments shows that the retardation ability of mixed oil in terms of the aging rate of the cellulose paper is not only useful for new cellulose papers, but also effective for aged cellulose paper whose further aging rate is delayed.
The mixed oil delayed the aging rate of the cellulose paper due to the vegetable oils in the three-element mixed insulation oil. Vegetable oil comprises polar triglycerides; thus, mixed oil can hold additional water, and saturated-water content is 3–4 times that of mineral oil [
6]. Based on the moisture equilibrium between oil and cellulose paper [
15], some moisture in paper diffuses into insulation oil, reducing the moisture content in cellulose paper. Simultaneously, ester bonds of vegetable oil may hydrolyse [
16], thereby consuming water and decreasing moisture content. Although this process generates macromolecular acids, they slightly affect the aging rate of cellulose paper [
17]. Moisture is the number-one enemy of cellulose polymer aging [
4], and the retardation of the aging rate of cellulose paper by mixed oil can be attributed to the reduction in moisture content in the cellulose polymer.
Infrared spectrometry is always used to qualitatively analyse changes in molecular functional groups. Attenuated total reflectance Fourier-transform infrared spectrometry (ATR-FTIR) of new paper and 90-day-aged cellulose paper was measured, and the results are shown in
Figure 3. It is worth noting that cellulose paper needs to be degreased with n-hexane before an ATR-FTIR test to avoid the influence of insulation oil.
Figure 3 shows that the infrared spectrum of the cellulose paper in Group B was the same as that in the new cellulose paper and that in Group A, although some peak areas were different, which was due to the different aging extents of cellulose paper. For cellulose papers with different aging extents, the destruction extents of a chemical bond are different, thus exhibiting different characteristic peak areas. There was no new characteristic peak in Group B, indicating that no new chemical bond was massively produced. This phenomenon supports that the mixed oil retards the aging rate of cellulose paper by decreasing the moisture content in cellulose paper.
A zero-order kinetic equation (Equation (1)) is usually adapted to fit the variation of DP values during the aging process of cellulose paper to analyse the aging rate of cellulose paper [
18]. Here is a simple algorithm to understand Equation (1). Cellulose paper is mainly made of cellulose, whose molecular formula is (C
6H
10O
5)
n. Understandably, the number of chain scissions is approximately equal to the increased number of cellulose chains because each breakage of a cellulose chain means that the number of molecules increases by one. Supposing m is the mass of the used cellulose paper, the number of chain scissions (N) during aging time 0–t can be calculated, as shown in Equation (2), in which 162 is the molar mass of the cellulose monosaccharide. Equation (3) is a variant of Equation (2), which suggests that 1/DP
t−1/DP
0 is a multiple of the number of chain scissions. Comparing Equations (1) and (3), k
0 represents the relative chain-scission rate of cellulose over time t, as shown in Equation (4), in which N/t is the real chain-scission rate. The chain scission of cellulosic polymers means the aging of the polymer; thus, k
0 can also be considered as a relative aging rate.
Here, 1/DP
t−1/DP
0 was calculated according to the data in
Figure 2. The DP values of cellulose paper impregnated with mineral oil (Groups A and C) had fallen below 250 on the 90th day, so results before 90 days were fitted with a zero-order kinetic equation. For comparison, the DP values in Groups B and D conducted the same fitting, although there were still two points (120th and 150th day) not included; the fitting results are shown in
Figure 4 and
Table 5.
Group A showed good fitting effect with a zero-order kinetic equation before the 90th day, which was mainly because the relative chain-scission rate calculated at each aging period was around the average values (k
0) during 0–90 days. Based on the above analysis, the relative chain-scission rate of cellulose in each aging period could be calculated with Equation (5), which evolved from Equation (1). The difference is that Equation (1) is for the entire aging process, while Equation (5) is for each aging period. The calculation results with Equation (5) for Groups A and B are presented in
Figure 5a. In fact, the calculated results in
Figure 5a are exactly the slope of the adjacent two points for Groups A and B, as shown in
Figure 5b. The relative chain-scission rates of different aging periods in Group A fluctuated around the average value of 3.95 before 90 days, suggesting that the chain-scission rate was nearly the same. However, the chain-scission rate of cellulose paper impregnated with mineral oil decreased significantly after 90 days, which resulted from the leveling-off DP [
19]. This indicates the chain-scission rate is very slow when DP value decreases to a low value, which is around 200 for cellulose paper impregnated by mineral oil [
19]. For Group B (
Figure 5a), however, not only was the chain-scission rate lower than that in Group A in each aging period, but the average chain-scission rate was also significantly decreased in the late aging period, especially for days 52–150, during which DP values were below 500. This result suggests that the mixed oil retarded the aging of the cellulose paper mainly at the late aging period, which can also be found in
Figure 4.
where
DP0 denotes the initial DP value of cellulose paper;
t is aging time (days);
DPt is DP value at time
t;
k0 is a constant, which indicates the relative aging rate of cellulose paper in Equation (1);
N is the number of chain scissions, in mol;
m is the mass of the cellulose paper, in g; and 162 is the molar mass of the cellulose monosaccharide, in g/mol.
Insulation oil was replaced in Groups C and D, and data in the late aging period (30–90 days) were fitted with Equation (1), and fitting results were satisfactory (
Figure 4). Similarly, results in Group D showed that the mixed oil can significantly delay the aging rate of cellulose paper in the late period.
Fitting results in
Table 5 show that the aging rate of cellulose paper in Group B is apparently slower than that in Group A. This phenomenon confirms that the three-element mixed-insulation oil strongly retards the aging rate of cellulose paper. The oil in Group C was replaced with unused mineral oil, thereby decreasing the aging rate of cellulose paper to 0.83 times of that in Group A. The mineral oil in Group A underwent a long aging time and produced additional aging byproducts such as acid, which further accelerate the aging of the cellulose paper [
17]. As a result, the cellulose paper in Group C slowly slightly degraded. The oil in Group D was replaced with three-element mixed-insulation oil, and the aging rate of cellulose paper decreased to approximately 0.29 times the aging rate of that in Group A, which shows the delaying effect of the mixed oil on the further aging rate of aged cellulose paper.
Cellulose paper loses most of its mechanical strength when DP values decrease below 250. Assuming that DP values drop to 250 as the life endpoint of cellulose paper, then the thermal-aging life of cellulose paper in Groups A and B could be calculated with Equation (6), which is a variation of Equation (1), and the results are listed in
Table 6.
Considering the change in aging rate in Groups C and D, the lifetime is divided into two parts. Thermal-aging life in Group C after the 30th day was calculated with the fitting results for Group C in
Table 5, and then total aging life can be calculated, as shown in
Table 6. For Group D, DP value only reduced to 294.4 after 150 days of aging; thus, its accurate aging life could not be obtained based on the experiment results. Similar to Group C, aging life after 30 days was estimated based on the fitting results in
Table 5, and the result was 171.9 days; thus, the total lifespan of cellulose paper in Group D could be recorded as 201.9 days.
Comparing the results in Groups A and B, the aging life of cellulose paper was significantly improved by the mixed oil, which increased by 86.8% when compared with the cellulose paper in Group A. Supposing that mineral oil is replaced with mixed oil when the DP is approximately 600 (Group D), the aging life of cellulose paper is extended to 2.5 times. The retarding-aging effect of mixed oil is mainly reflected in the period of low DP values, especially those below 500, during which aging rate was considerably reduced. Therefore, oil replacement with mixed oil in the later aging stage could also delay the further aging rate of the aged cellulose paper.
An interesting phenomenon is that cellulose paper life in Group B was significantly lower than that in Group D, which was mainly due to the different methods used in the fitting. In fact, the aging rate of cellulose paper in Group B should have been near that in Group D in the late aging period due to the consistent action of the mixed oil, which was proved by that the slope of the partial linear fitting of Group B was close to that in Group D in the late aging period in
Figure 4. Supposing the aging life of cellulose paper after the 90th day in Group B is roughly estimated using the slope of partial linear fitting result, the remaining life after 90 days would approximately be 114.4 days. This suggests that the total life of cellulose paper in Group B was 204.4 days, which is closed to the estimated lifespan in Group D. When oil replacement was conducted, the DP values of cellulose paper in Group B were higher than those in Group D. However, the oil in Group B underwent longer aging time, resulting in more byproducts that accelerate the aging rate of cellulose paper [
17], so the final estimated life was nearly the same as that in Group D.
Someone may doubt the abovementioned estimation method for Groups B and D. In fact, after 90 days, there are two more points that were tested during the experiment, which could be used to verify the correctness of the estimation. The k
0 used in the estimation in Groups B and D was 1.347 * 10
–5 and 1.157 * 10
–5, respectively. Assume k
0 remains unchanged after 90 days, the DP values of cellulose paper in Groups B and D at 120 and 150 days could be calculated with Equation (7). Take the 120th day in Group B as an example: if t1 equals 120 and t2 equals 90, then Equation (8) can be obtained. The measured DP
90 is 406.7, so the DP value on the 120th day can be calculated with k
0 (1.347 * 10
–5). All calculated results are shown in
Table 7. The results show that the calculated results are very close to the measured DP values, which proves that partial fitting in Groups B and D can be used to predict the subsequent development trend of DP values.
where
t1,
t2, 90 and 120 denote aging time, in days.
Although lifespan estimation in Groups B and D is not absolutely precise, they showed apparently longer lifespan than that in Groups A and C, because the DP values of cellulose paper immersed in the mixed oil (Groups B and D) stayed around 300 at 150th day, which far exceeded the 81-day life of cellulose paper immersed in mineral oil (Group A). The results demonstrate that the three-element mixed insulation oil can retard the aging rate of cellulose paper, which is hopeful to replace aged mineral oil in old transformers to extend the aging life of cellulose paper in field transformers.