本文是留学生计算机专业的作业格式范例，题目是“Big Data in the Aerospace and Defense Industries（航空航天和国防工业的大数据）”，以历史为例，今天最新的飞机平台将延伸到未来50年。这对航空航天和国防工业的大数据可持续性来说是个好消息，因为大数据已经在当今最新的飞机上找到了一席之地。作为一个严重依赖大量传感器信息的行业，航空航天和国防行业是大数据分析和决策支持应用的早期采用者。本报告研究了大数据在当今航空航天和国防工业中所扮演的角色，以及当前的挑战、差距和机遇。此外，还讨论了该行业的效益、可持续性和发展方向。
If history is to serve as an example, today’s newest aircraft platforms will extend 50 years into the future. This is good news for the sustainability of big data in the aerospace and defense industry since big data has already found a home on today’s newest aircraft. As an industry that is heavily dependent on huge amounts of information from a wide array of sensors, the aerospace and defense industry were early adopters of big data analytics and decision support applications. This report examined the role that big data plays in the aerospace and defense industry today, as well as the challenges, gaps and opportunities that are present. The benefits, sustainability and directions for advancement in the industry were discussed as well.
Big data was found to be most widely used in the industry for predictive maintenance on aircraft. The industry faces challenges with meaningfully using the data collected, as well as with the infrastructure to stream the data. Nevertheless, there are opportunities to utilize big data to expand predictive maintenance, improve data analytics, as well as optimize flight plans and improve air traffic control. One of the gaps in the industry is the lack of streaming critical data, such as data from the flight data recorder, back to ground control using existing technology. Recommendations for the industry included investing in the infrastructure for big data both on and off the aircraft, as well as investing in research to develop better models, analytics and decision making applications. The benefits of big data will allow for more automated aircraft while increasing safety, reliability as well as profit margins. The next generation of aircraft will be designed utilizing big data, therefore today’s investors in the technology will be the leaders of the future aerospace and defense industry.
In 2014, the disappearance of Malaysia Airlines Flight 370 (MH370) with all 239 people on board captured the world’s attention as investigators scrambled to figure out what had happened to the aircraft (MacLeod, Winter, & Gray, 2014). As theories swirled around the cause of the disappearance, pundits asked why aircraft that record terabytes of data do not transmit the data back to the ground, making it immediately evident what the cause of an accident was. With 100 thousand flights a day worldwide, the infrastructure to support this kind of data transfer simply does not exist, but, perhaps the better question is: should this infrastructure exist? (IATA, 2018). The answer to this question requires an analysis of big data and its applications in the aerospace and defense industry (ADI).
2014年，马来西亚航空公司370航班(MH370)的失踪，机上所有239人引起了全世界的关注，因为调查人员紧急查明飞机发生了什么(MacLeod, Winter， & Gray, 2014)。随着人们对飞机失踪原因的猜测甚热，专家们问道，为什么记录了数据的飞机不把数据传回地面，从而立即证明事故的原因是什么。全球每天有10万次航班，支持这种数据传输的基础设施根本不存在，但是，也许更好的问题是:这种基础设施应该存在吗?(IATA,2018)。要回答这个问题，需要对大数据及其在航空航天和国防工业(ADI)中的应用进行分析。
The ADI has long been on the leading edge of technology. When it comes to big data aerospace is once again a pioneer in the field. According to one definition, big data differentiates itself from traditional data by having the 5 Vs, which are “huge Volume, high Velocity, high Variety, low Veracity, and high Value” (Jin, Wah, Cheng, & Wang, 2015, p. 59). Consequently, with big data comes big challenges. This report examines the current state of big data in the ADI and provides an analysis of the gaps, challenges and opportunities of big data in aerospace. The benefits provided by big data, as well as their sustainability, are discussed as well. To provide the context of how the industry got to where it is today, first, a brief history of the industry is presented.
2.Industry History and Overview行业历史及概述
The study of flight goes back millennia, however, it was not until the first successful flight by the Wright brothers in 1903 that the aeronautical industry was born. By 1911, aircraft were being used to deliver mail and by 1915 multi-engine aircraft were being used for commercial passenger transport. Technology in the field rapidly progressed and by World War II aircraft were the defining factor of a country’s military might. Coming out of the war, the jet engine allowed for an explosion in commercial air travel, driving the development of larger and more sophisticated aircraft. In parallel, the space race was taking form creating the field of astronautics, which combined with aeronautics to form the aerospace industry. The thousands of sensors and complex systems needed to control aircraft and spacecraft presented some of the earliest challenges of how to handle mass amounts of data.
By the 1990s, the fall of the Soviet Union and the corresponding drop in military spending, along with an economic crisis on the commercial side of the industry, led to a major consolidation among aerospace companies. The consolidations and dropping margins indicated that the industry had entered the mature phase of its lifecycle. Today, most aerospace companies are involved in both the commercial and defense sides of the industry. The modern ADI designs and manufactures airplanes, spaceships, helicopters, missiles, satellites and defense products. Suppliers provide components ranging from sensors, computers and integrated systems, to engines, structures and interiors. Also, maintenance, repair and overhaul play a big role in the industry (Longo, 2017).
Revenue growth for the industry is expected to occur at a rate of 3% through 2023 (Longo, 2017). The ADI is a highly cyclical industry and is currently in a growth phase. The industry generally follows the global economy and has key drivers such as air traffic forecasts, plane orders and plane deliveries (Corridore & Chuah, 2018). While there are hundreds of players in the industry, revenue is strongly consolidated among a few large players, namely, Boeing, Airbus, Lockheed Martin, United Technologies, BAE Systems and Raytheon (Corridore & Chuah, 2018). On the other end, the primary customers of the industry are governments and airlines.
The factors in the general environment that influence the ADI include economic, global, political, demographic and technological components (Corridore & Chuah, 2018). Due to the nature of the ADI, it is unsurprising that it is characterized by high levels of technology change, globalization and regulations (Longo, 2017). Additionally, consolidation has created high barriers to entry and a medium level of capital intensity (Longo, 2017). From Porter’s Five Forces perspective, rivalry is strong due to shrinking margins and fierce competition. The power of the buyer is strong since customers have options, while the power of the seller is moderate as consolidation continues. Finally, the threat of new entrants and the threat of substitute are low due to the high barriers to entry and the lack of realistic alternatives to flying, respectively.
Powerful buyers, high technology demands and expensive regulation place a lot of pressure on ADI companies to find competitive advantages to maintain their margins. The potential applications of big data in the ADI are enormous, and successful implementation can provide a company the edge it needs. The industry has not ignored this potential nor remained idle to the opportunities of big data. In the next section, the current state of big data in the aerospace industry is discussed.
3.Status Quo of the Aerospace Industry航空航天工业的现状
The ADI is well entrenched in the utilization of big data. The current amount of data produced by the industry is mind-boggling. Take, for example, the typical Boeing 737, the most widely flown commercial aircraft. The engines alone on a 737 record over 240 terabytes of data over a 6-hour flight (Badea, Zamfiroiu, & Boncea, 2018). Figure 1 shows the vast scope of this amount of data when it is extrapolated to all US commercial air traffic.
ADI在大数据利用方面根深蒂固。目前该行业产生的数据量令人难以置信。以典型的波音737为例，这是飞行最广泛的商用飞机。仅737飞机的引擎在6小时的飞行中就记录了超过240tb的数据(Badea, Zamfiroiu， & Boncea, 2018)。图1显示了这一数据量的巨大范围，当它被推断到所有美国商业空中交通。
The applications of big data analysis in the industry include areas such as optimizing flight plans, modeling weather effects on flight, determining customer patterns and providing predictive maintenance (Badea et al., 2018). Big data is used in the field to optimize performance and in the lab as feedback in design optimization (Sethi, 2015). Examples include monitoring engine pressure and temperature to increase fuel efficiency, as well as monitoring stress levels and temperature exposure of parts to predict when they will fail (Sethi, 2015). Furthermore, through the use of Internet of Things (IoT) manufacturers are able to track the performance of their operations and improve efficiencies (How Big Data Is Transforming The Aerospace Industry, n.d.).
Of all the applications of big data noted, predictive maintenance currently has the most widespread use. Predictive maintenance utilizes big data to create models that determine the conditions that precede the failure of an aircraft part (Ezhilarasu, Skaf, & Jennions, 2019). The models then recommend to the operator to replace the part prior to a failure occurring, thereby increasing safety and reducing unplanned downtime costs. These systems, called Integrated Vehicle Health Monitoring (IVHM), are currently used on avionics, aircraft engines, unmanned aerial vehicles (UAV), fuel systems, satellites and spacecraft (Ezhilarasu et al., 2019). SAP’s Predictive Maintenance program is an example of one application commonly used by airlines for IVHM (Badea et al., 2018).
While the aerospace industry is ahead of most industries in the use of big data, there are a number of challenges and opportunities the industry has yet to address. In some areas, there are some serious gaps that have drawn widespread criticism of the industry. The analysis section examines these challenges, opportunities and gaps in detail.
The determination to build the Airbus A380, the largest passenger jet in the world, emanated from a joint market study completed by Airbus and Boeing (EASA, 2017; Reuters, 1995). Airbus determined that the market was looking for a high-capacity jet and developed the A380 while Boeing determined the market was looking for medium jets and developed the 787 Dreamliner. In 2019, Airbus announced it would shut down the production of its A380 jets after producing just 30 aircraft at an estimated loss of $25 billion (Airbus, 2019; West, 2014). Meanwhile, Boeing has produced 840 of the 787 aircraft and continues to ramp up production (Ostrower, 2014). Despite looking at identical data sets, the companies drew different conclusions from the data. Unfortunately for Airbus, their misinterpretation cost them $25 billion and 21 years of development.
空客和波音完成的一项联合市场研究(EASA, 2017;Reuters,1995)。空客认为市场正在寻找大容量喷气式飞机，于是开发了A380，而波音认为市场正在寻找中型飞机，于是开发了787梦幻客机。2019年，空客宣布将关闭其A380客机的生产，此前该公司仅生产了30架飞机，估计亏损250亿美元(Airbus, 2019; West, 2014)。与此同时，波音公司已经生产了840架787飞机，并继续提高产量(Ostrower, 2014)。尽管研究的是相同的数据集，但两家公司从数据中得出了不同的结论。不幸的是，空客公司的错误解读让他们损失了250亿美元和21年的研发时间。
Raw data on its own is meaningless. The data must be processed, analyzed and correctly interpreted. Yet correctly interpreting the data is just one component of the challenges that big data presents to the aerospace industry. The analysis that follows digs into the details of the challenges, opportunities and gaps of big data in aerospace, and provides resulting recommendations for advancement.
As with industries everywhere, big data is still a developing technology in aerospace (Urbinati, Bogers, Chiesa, & Frattini, 2019). Huge amounts of data are collected in the ADI raising common challenges with data complexity, computational complexity and system complexity (Jin et al., 2015). The largest challenges of big data in the aerospace industry are limitation on models to interpret data, data streaming limitation and realtime use of data.
与世界各地的行业一样，大数据在航空航天领域仍然是一项发展中的技术(Urbinati, Bogers, Chiesa， & Frattini, 2019)。ADI收集了大量的数据，带来了数据复杂性、计算复杂性和系统复杂性等共同挑战(Jin et al.， 2015)。航空航天行业的大数据面临的最大挑战是解释数据的模型的局限性、数据流的限制以及数据的实时使用。
There is no doubt that the ADI generates astronomical amounts of data, yet the majority of the data goes unused (Badea et al., 2018). Creating models and tools to make accurate use of the data has proved to be a difficult engineering challenge (Taylor, & Waldron, 2019). There is still plenty of research to be done to create models to determine how stress, fatigue, wear and temperature affect the useful life of a part. Until then, predictive maintenance is limited to well-understood areas and components on an aircraft (Ezhilarasu et al., 2019).
When a plane flies over an ocean, satellite streaming is its primary way of transferring data. Satellite streaming is both expensive and limited in bandwidth (Taylor, & Waldron, 2019). This presents a significant challenge for realtime big data use, limiting applications of the data to the capabilities on the aircraft. One resulting limitation is that it prevents the optimization of flight paths based on weather conditions and flight data (Badea et al., 2018). Furthermore, even when the cost investment is made into streaming flight data, concerns over the cybersecurity of the data are a major consideration due to the sensitive security nature of flight (Taylor, & Waldron, 2019). Yet, with challenges come opportunities, which are discussed next.
The wide range of opportunities for the advancement and use of big data in aerospace paints a promising picture for the industry. Predictive maintenance models and algorithms are still in the infancy stage of application; their potential for widespread use on the aircraft is enormous. To date, a large focus for predictive maintenance has been on the aircraft engines since they are the single largest cost items on the aircraft (Ezhilarasu et al., 2019). According to Ezhilarasu et al., the expansion of predictive maintenance from a component level to the overall aircraft level is necessary to accomplish the goal of preventing unexpected downtime (Ezhilarasu et al., 2019). To get IVHM to the aircraft level requires an installation of sensors throughout the aircraft, upgraded analytical capabilities on the aircraft and improvements of in-air data streaming.
大数据在航空航天领域的广泛发展和应用为该行业描绘了一幅前景广阔的图景。预测性维修模型和算法仍处于应用的初级阶段;它们在飞机上广泛应用的潜力是巨大的。迄今为止，预测性维修的重点一直集中在飞机发动机上，因为它们是飞机上最大的单一成本项目(Ezhilarasu et al., 2019)。根据Ezhilarasu等人的观点，为实现防止意外停机的目标，有必要将预测性维修从部件级扩展到整个飞机级(Ezhilarasu et al., 2019)。要将IVHM提升到飞机水平，需要在整个飞机上安装传感器，升级飞机上的分析能力，并改进空中数据流。
An additional benefit of increasing the sensing capabilities throughout the aircraft is that, over time, the data collected will allow for better characterization of failure modes thereby increasing the effectiveness of the IVHM models. This brings up another opportunity – the majority of the world’s aircraft fleet is still made up of aging aircraft with limited built-in capability to handle big data (Taylor, & Waldron, 2019). Proactively upgrading aircraft that may have another 15 years of life ahead of them will help with downtime reductions while improving data gathering for IVHM capability improvements.
Another opportunity is to utilize the data output from the aircraft, combined with ground control data, to reduce air traffic congestion and thereby increase airport turnaround times for aircraft (Badea et al., 2018). Better turnaround times result in higher profitability (Schlesinger, 2011). A similar combination of live flight data with weather data can be used to optimize flight plans inflight, thereby increasing aircraft efficiency (Badea et al., 2018). Both of these opportunities, once again, utilize existing data systems but require new programs to implement the optimizations.
With the amassed amount of data already collected by the ADI, there is a huge opportunity to put that data to use. The first opportunity this presents is determining what parameters of the data collected is useful. If certain parameters are deemed unneeded, manufacturers can optimize the investment in big data capabilities on the aircraft (Sethi, 2015). The second opportunity presented by this treasure trove of data is using analytics to design the next generation of efficient airplanes (Sethi, 2015). The first-to-market that utilizes this opportunity will likely obtain a new competitive advantage.
Gaps exist in the industry as was demonstrated after the loss of flight MH370. Streaming of flight recorder data can be done with existing technology and is an improvement the ADI should implement. Big data technology in its current form would not have prevented the tragic loss of flight MH370, but it could have shed light on why the aircraft crashed, as well as provide the location of the crash, thereby providing closure for the families who lost loved ones. Big data will design tomorrow’s aircraft. The ADI should invest in big data infrastructure on and off the aircraft, as well as invest in research to create better models and decision making capabilities. The results will allow for more automated aircraft while increasing safety, reliability as well as profit margins. Big data has a sustainable and promising future in the ADI. Therefore, the early adopters of big data today will be the industry leaders of tomorrow.