Industrial Router Crystal 3.2*2.5mm 3225 26M (26.000MHZ) 12PF 10PPM 20PPM 30PPM

Photocoupler

With the outbreak and escalation of the Japanese nuclear accident, the decline of the nuclear power industry, which once dominated the new energy sector, reflects the direction of progress. Is another critical new energy industry, the photovoltaic industry, set to enjoy better development opportunities?

Power Supply Circuit

China’s photovoltaic market presents both opportunities and challenges.

In this regard, industry experts offer a deeper perspective: although the nuclear power industry is no longer as strong as it once was, the photovoltaic industry is also facing severe tests. Since China became the world's largest producer of photovoltaic products in 2007, progress has been steady across various sectors. However, it is crucial to transform the current market reliance on government subsidies into market-driven forces. Depending solely on government subsidies poses significant risks; any change or termination of policies could severely impact the industry. Establishing a genuine photovoltaic market system as soon as possible is essential to allow the photovoltaic industry to operate under market principles. This approach would not only leverage government subsidies effectively but also mitigate potential damage caused by policy shifts.

All participants in the photovoltaic industry—especially leaders—must recognize this reality: regardless of market booms, product prices across the industry must continue to decrease until the cost of photovoltaic power generation is genuinely lower than that of conventional thermal power generation. Cost reduction remains the true driving force behind the growth of the photovoltaic industry.

So, how can costs be reduced? Many immediately think of cutting raw material costs like polysilicon. However, after 2009, much of the potential for cost savings in polysilicon has already been realized. New and real competition is now emerging at the level of company processes and quality management. Similar to the traditional semiconductor electronics industry, the one-time yield of photovoltaic products determines the robustness of the production process and also impacts production costs.

Drawing lessons from foreign photovoltaic giants like First Solar, Ege Solar, and JA Solar, it becomes evident that yield control cannot be achieved without meticulous data-based analysis and decision-making. This is a technical task, and it is precisely where domestic photovoltaic companies lag behind. Of course, quantitative analysis of advanced manufacturing processes does not mean blindly pursuing overly complex data warehouses or data modeling; instead, there is growing emphasis on practicality. In recent years, a new method of data analysis—interactive visual data analysis—has been adopted by leading companies to identify inferior costs, analyze their causes, develop improvement plans, and optimize production processes, ultimately enhancing one-time yields and overall quality while reducing costs. Interactive visual data analysis promotes the use of more graphical tools to enable two-way interaction between analysts and data via interactive graphics, thereby reducing reliance on traditional statistical analysis tools (or using advanced statistical analysis methods behind simple graphs). This lowers the barrier to use and provides better insights into critical information hidden within the data. Below, we use a professional Six Sigma and quality management statistical analysis software, JMP, which is widely used in the photovoltaic industry, to briefly explain how interactive visual analysis can improve the one-time yield of solar cells and enhance productivity while reducing costs.

The basic process of simple solar cell production can be roughly divided into eight main steps, as shown in the following flowchart:

Overall Process of Simple Solar Cell

First, after collecting historical data, we aim to quickly identify the best areas for improvement and determine which processes have the greatest impact on improving the overall throughput rate. From JMP’s predictive descriptors, we can see some clues: on the linear relationship between the actual yield of the eight steps and the overall actual throughput rate, it is evident that the actual diffusion yield and the actual etching yield have a significantly larger slope compared to the actual yields of other steps. Further adjustments to the actual diffusion yield and the actual etching yield result in the most noticeable changes in the overall throughput rate, allowing us to intuitively conclude that the diffusion and etching processes are the most influential, making them key steps in the production process.

Predictive Morph Analyzer Analysis

Next, we naturally want to understand the causes of low etching yields. It is straightforward to identify these using JMP’s Pareto chart. The main defects in the diffusion process are rework and internal debris in the diffusion process. The primary defects in the etching process are fragments during loading and fragments inside the machine. Thus, we need to focus on and control these four defects.

Pareto Chart Analysis

Then, we begin using more advanced analytical tools like regression modeling and decision trees to uncover the root causes. Of course, when it comes to advanced analysis tools, many might feel discouraged due to their perceived abstraction and complexity. However, this concern is unwarranted. In JMP, all analysis tools—whether simple or complex—can be visualized through various statistical graphs. For instance, when using JMP’s decision tree function for factor analysis, we only need to click the split button on the analysis interface to progressively minimize within-group differences while maximizing between-group differences, grouping data to uncover valuable insights:

1. Since the decision tree selects four variables—silicon wafer manufacturer, date, wafer lot, and shift—from numerous candidate variables during the grouping process, it can be concluded that these are key factors influencing the defect rate.

2. Focusing on the silicon wafer manufacturer factor, it is easy to notice that the quality issues of two factories are minimal, whereas the other two factories have more significant quality problems. This can be seen from the left-side graph of the decision tree, where the green area representing high yield rates is extensive, corresponding to classification levels A and D. Conversely, the right-side graph shows a large red area representing low yields, corresponding to classification levels B and C.

3. Regarding the date factor, the situation on November 1 was particularly poor, with none of the four silicon wafer manufacturers meeting standards. Similarly, the two days of November 2 and 7 also encountered significant issues, with all production batches of Class B being substandard, warranting a thorough on-site investigation.

Decision Tree Analysis

In practice, technicians at solar cell companies can employ more interactive and visual data analysis methods to delve deeper into technical reasons and optimize improvement plans. Due to space constraints, further details cannot be elaborated here. However, one thing is certain: mastering interactive visual data analysis will be a crucial issue for domestic photovoltaic companies aiming to achieve dual improvements in quality management and cost control.