1 Introduction Over the past decade, China's installed power capacity has increased by 10GW annually, greatly alleviating the power shortage. From the surface phenomena of 1997, most parts of the country began to experience power saturation. For example, Shandong Province's installed capacity had reached 16GW, but the peak load for the entire year was less than 11GW. Along with the increase in power supply, the speed of power grid construction has lagged significantly, and network losses have become increasingly prominent. In recent years, the State Power Corporation and provincial and municipal power bureaus have begun to pay attention to this issue. It is widely recognized that reducing network losses is a crucial breakthrough for power supply departments to reduce power supply costs and an important means to increase power supply in the future. Some experts estimate that increasing power supply through loss reduction costs only 1/4 to 1/5 of the cost of building power plants, which is very feasible. The State Power Corporation's Electric Power Research Institute has conducted a comprehensive survey on system network losses, and the analysis report shows the potential and significance of loss reduction [1]. Currently, power grid losses in my country can be divided into three levels: losses at voltage levels of 220kV and above; losses at 110kV and 35kV; and losses at 10kV. The ratio of these three losses is approximately 1.5:1.1:2.5, with the 10kV distribution network showing the greatest potential for loss reduction. The prominent reason for the high losses in 10kV and below distribution networks is the severe lag in distribution network construction, resulting in a weak network structure, aging facilities, long lines, and small wire diameters. Some 10kV lines are 50-70km long, and some even exceed 100km. Furthermore, most distribution transformers are high-energy-consuming transformers. This problem is particularly prominent in rural power grids; at the end of the Eighth Five-Year Plan period, the national average rural power grid loss was 28%. From the perspective of reducing network losses and improving power supply reliability, the length of 10kV distribution lines should be controlled within 10km. Currently, the average length of 10kV distribution lines in urban areas nationwide is about 6km, and in some large cities it has been shortened to within 1-2km, so the loss is not significant (the loss in urban distribution networks is mainly concentrated on the 400V side). However, in the vast rural power grid, due to the limited number of substations, long-distance power supply is a common phenomenon, and because of the small load, it is impossible to choose excessively thick wire diameters, resulting in high line impedance and high losses. The high line impedance causes particularly large voltage drops at the beginning and end of the line, seriously exceeding the requirements of relevant national and local standards, resulting in substandard voltage quality in some areas. The high loss of 10kV distribution lines brings unnecessary losses to the power sector and indirectly increases the electricity price. Coupled with substandard power quality, it seriously affects the enthusiasm of electricity users, especially in rural areas. In the current urban and rural power grid transformation, the problem of power grid loss reduction has attracted widespread attention [2, 3] and has been listed as one of the key issues to be addressed in the power grid transformation. If the loss reduction problem of urban and rural power grids, especially low-voltage power grids, can be well solved, it will greatly tap the potential of rural electricity consumption and is of great significance to solving the problem of power saturation. [b]2 Reactive power compensation is an effective means of reducing losses and increasing voltage in 10kV distribution networks[/b] Figure 1 shows a simplified equivalent circuit of a conductor (the ground admittance of the 10kV line can be ignored). The active power loss ΔPi is [img=112,42]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/gs0401.gif[/img] (1) It is not difficult to see from equation (1) that in order to reduce losses, there are several ways: ① reduce the active current flowing through the conductor; ② reduce the reactive current flowing through the conductor; ③ reduce the equivalent resistance of the conductor, that is, increase the conductor cross-section and shorten the power supply radius. The voltage of the conductor is generally near the rated value and can be treated as a fixed value. [img=169,60]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/0401.gif[/img] Figure 1 is the equivalent circuit diagram of a line. Figure 2 is the equivalent circuit diagram of a transformer. Its active power loss ΔPT is [img=187,39]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/gs0402.gif[/img] （2） [img=141,54]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/0402.gif[/img] Figure 2 Equivalent circuit of transformer There are three methods to reduce transformer loss. Methods ① and ② are the same as the methods for reducing line loss mentioned above. The third method is to reduce GT and RT, that is, to use a low-loss transformer. In addition, taking Figure 1 as an example, the voltage drop of the line is [img=183,39]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/gs0403.gif[/img] (3) The method to reduce the voltage drop is also to reduce the active and reactive current flowing through the conductor and reduce the equivalent impedance of the conductor. In engineering practice, the following measures have been emphasized: (1) Transform the distribution network structure, even increase the voltage level and increase the number of substations, and rationally allocate active and reactive power; (2) Replace high-energy-consuming transformers; (3) Increase the conductor cross-section and shorten the power supply radius; (4) Reactive power compensation. These four measures are related to each other, but they do not conflict. They should be considered mainly based on the actual loss situation and financial situation of the location. The first transformation measure is a good way based on the long-term development of the distribution network. It can rationally transform the imperfect distribution network and provide the distribution network with an efficient and stable operating environment for more than 10 years. However, due to the huge investment and long payback period, most towns and cities cannot currently undertake this work; localized and small-scale renovations are more realistic. The second and third measures have significantly reduced investment, but are still considerable. Regions with ample funds can consider them. The fourth measure requires the least investment. Furthermore, due to the long-standing reactive power shortage in my country's distribution network, the resulting network losses are substantial. Reactive power compensation to reduce network losses and improve voltage is a low-investment, high-return solution. Compensation on the low-voltage side of the distribution transformer can not only reduce line losses but also reduce transformer losses. Some high-energy-consuming transformers can even have their decommissioning delayed due to reduced losses, and voltage quality will also be greatly improved. It is estimated that for 10kV distribution lines with network losses exceeding 10%, losses can be reduced by approximately 5% to 10%, with the investment recoverable within 1 to 2 years. In contrast, the first few measures require 3 to 4 times more investment to achieve the same loss reduction effect. Therefore, in general, using reactive power compensation to reduce losses and increase voltage is an effective way to transform the distribution network. On the basis of reactive power compensation, other transformations should be implemented in a way that is appropriate to local conditions and effective. This should be the preferred solution for most distribution networks, especially rural networks. [b]3 Several existing problems of reactive power compensation in distribution networks[/b] With the increasing attention paid to distribution network construction and the development of reactive power compensation technology, low-voltage side reactive power compensation technology has also begun to be popularized in distribution systems. From static compensation to dynamic compensation, from contact compensation to contactless compensation, rich operational experience has been gained. However, some problems have also been exposed in practice, which must be taken seriously. (1) Problems with compensation methods. At present, many departments still focus on the user side when it comes to reactive power compensation, that is, they only pay attention to compensating the power factor of users, rather than focusing on reducing the losses of the power grid. For example, in order to improve the power factor of a certain power load, adding a compensation box will certainly help reduce losses. However, if effective loss reduction is to be achieved, it is necessary to calculate the reactive power flow and determine the optimal compensation amount and compensation method at each point so that the limited funds can play the greatest role. This is a method of considering problems from the perspective of the power system. For example, the compensation plan for a certain line in Beijing lacked overall consideration, and the investment was three to four hundred thousand yuan, while the line loss reduction was only 1%. However, according to calculations, the loss reduction potential of this line is at least 5% to 6%. If it can be reasonably compensated, this loss reduction target can be fully achieved. (2) Harmonic problem. Capacitors have a certain anti-harmonic capability, but when the harmonic content is too high, it will affect the life of the capacitor and even cause premature damage to the capacitor; and because the capacitor has the amplification effect on harmonics, it will make the harmonic interference of the system more serious. In addition, the control link of the dynamic reactive power compensation cabinet is easily affected by harmonic interference, causing control failure. Therefore, in places with large harmonic interference and reactive power compensation, filtering devices should be considered. This problem is generally ignored, causing some compensation equipment to be damaged inexplicably. Therefore, harmonic control must be considered when designing reactive power compensation. (3) Reactive power backfeed problem. Reactive power backfeed is not allowed in the power system because it will increase the line and transformer losses and increase the line burden. Although manufacturers of reactive power compensation equipment all emphasize that their equipment will not cause reactive power backflow, the actual situation is not so. For contactor-controlled compensation cabinets, the compensation amount is three-phase synchronized; for thyristor-controlled compensation cabinets, although the compensation amount of the three phases can be adjusted separately, many manufacturers only select one phase for sampling and reactive power analysis in order to save money. Therefore, in the case of three-phase load asymmetry, reactive power backflow may occur. As for users who use fixed capacitor compensation, reactive power backflow may also occur during the off-peak load. This should be fully considered when selecting the compensation method. (4) Problems brought about by voltage regulation compensation equipment. Some reactive power compensation equipment determines the reactive power switching amount based on voltage, which helps to ensure the power quality of users, but it is not advisable for the power system. This is because although the fluctuation of line voltage is mainly caused by the change of reactive power, the voltage level of the line is determined by the system conditions. When the line voltage reference is too high or too low, the reactive power switching amount may be far from the actual demand, resulting in over-compensation or under-compensation of reactive power. In summary, reactive power compensation on the low-voltage side of a 10kV distribution network should take into account the characteristics of the system more carefully. The impact of compensation equipment on the system side (including network losses) should not be ignored simply because of the low voltage level and small compensation capacity. If the lines requiring loss reduction can be upgraded based on a comprehensive compensation scheme, the benefits to the power system side will be much greater than those from decentralized, purely user-driven compensation methods. [b]4 Technical Analysis of Reactive Power Compensation[/b] Figure 3 is a schematic diagram of reactive power compensation for 10kV distribution lines. CVC refers to a composite var compensator, which is placed at the output end of the distribution transformer and performs reactive power compensation together with the transformer. [img=244,149]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/0403.gif[/img] Figure 3 Schematic diagram of reactive power compensation for 10 kV distribution lines. In reactive power compensation projects, this paper uses the equal network loss incremental rate to optimize the compensation capacity. Let QC be the total compensation capacity of the line, kvar; K1 be the one-time cost per unit compensation capacity, yuan/kvar; K2 be the operation and maintenance cost per unit compensation capacity per unit time, yuan/kvar*a), T be the time, a; F1 be the total cost within the time period: F1＝K1QC＋K2QC T （4） Let β be the unit electricity price, yuan/kWh; Q1, Q2, …, Qn be the compensation capacity at each point on the line, QC＝Q1＋Q2…＋Qn; τmax be the maximum loss hours per year for the entire network, kW; ΔP be the network loss reduced after line compensation, kW; F2 be the energy cost saved after network loss reduction within the time period T, yuan. Therefore, we have ΔP＝f（Q1,Q2，…，Qn） （5） F2＝βΔPTτmax （6） So the investment return F within the time period T is F＝F2－F1＝βΔPTτmax－K1QC－K2QCT （7） To achieve the best economic benefits, we must make [img=48,39]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/gs0404.gif[/img], so we have [img=231,42]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/gs0405.gif[/img]（8） N equations can be formed from equation (8). Assuming the time is known, the system of equations has N variables (Q1, Q2, ..., Qn), from which Q1, Q2, ..., Qn can be obtained. Let equation (7) = 0, and we can obtain the investment recovery time T0: [img=144,41]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/gs0406.gif[/img] (9) If we first determine the total compensation capacity QC, and take the maximum ΔP as the objective, we can obtain [img=195,36]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/gs0407.gif[/img] (10) Equation (10) can form N-1 equations, and the system of equations has N-1 variables, and we can also obtain Q1, Q2, ..., Qn. Similarly, we can use equation (9) to obtain the investment recovery time T0. The effect of investment compensation can be illustrated by a diagram, as shown in Figure 4. As can be seen from the figure, when cosφ = 1, the line is fully compensated. From zero compensation to full compensation, the slope of the curve decreases, indicating that the reduction in network loss becomes less significant. Therefore, in actual engineering [img=213,158]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/0404.gif[/img] Figure 4, Schematic diagram of investment compensation effect, shows that the compensation degree of the line is generally controlled between a power factor of 0.9 and 0.95. Figure 5 illustrates the relationship between investment returns and time. cosφ1 > cosφ2 > cosφ3, corresponding to the three curves respectively. The compensation degree corresponding to cosφ3 is lower, and the investment payback period is shorter; the compensation degree corresponding to cosφ1 is higher, and the investment payback period is longer, but the returns are greater over a longer period. [img=232,175]http://zszl.cepee.com/cepee_kjlw_pic/files/wx/dwjs/dwjs9906/image1/0405.gif[/img] Figure 5: Relationship between Investment Returns and Time. In summary, 10kV and 380V lines are the intersection of resource and demand sides. Due to their low voltage levels, they have been largely neglected by the power system, resulting in numerous problems, with high losses and poor power quality being particularly prominent. Therefore, the focus of this power grid transformation should be on distribution. However, there is a lack of experience in how to improve the distribution network. Based on my own investigation and research, I believe that comprehensive reactive power compensation for distribution lines plays a crucial role in reducing losses and saving energy, and is a technical direction worthy of attention in current distribution network transformation work. **References** 1. Electric Power Research Institute, Ministry of Energy. Power System Network Loss Investigation and Analysis Report. 1989. 2. Wei Guangyao. The Current Status and Development of Urban Power Grids in my country. Power Supply and Utilization, 1998(1). 3. Bao Xuding. Accelerating Rural Power Grid Transformation and Expanding the Rural Market. 1998, 6.