Automatic DfMA Design Generator (ADAGE): An Internet of Things tool to instill adoption of DfMA by Building Designers

**定义****概念设计：**设计团队对项目简介的初步反应。**详细设计：**根据批准的概念设计开发的详细设计**摘要**由于冠状病毒封锁，英国建筑总活动指数从2月的52.6降至3月的39.3；这是11年来建筑业产出的最大跌幅（IHS-Markit, 2020）。政府希望利用冠状病毒危机作为“建造房屋”的机会，并计划“建造房屋”，作为建筑业和更广泛经济复苏计划的一部分，宣布50亿英镑的基础设施和120亿英镑的可负担新房投资（总理办公室，2020年）。不幸的是，当前/传统的建筑方法效率太低，无法在规定的时间内实现这一目标，因为它们是目前创纪录的400万套住房积压/短缺的主要原因(National housing Federation,2018; BBC housing - briefing,2020；麦肯锡公司，2019年)，并且偏离了到2025年实现国家基础设施计划提出的6500亿英镑项目价值的轨道。目前/传统的建筑方法缓慢、昂贵、质量差，而且相对不安全。与制造业相比，其低下的生产率在20年内使英国经济损失了1400亿英镑（包括税收）（Mace,2018）。针对这个问题，政府支持建筑业通过转向设计制造和组装（DfMA）来模仿制造业（建筑业交易，2018年）。与传统方法相比，DfMA试验项目在持续时间、成本和现场劳动力方面分别减少了60%、44%和70%以上，质量提高了73%，整体生产率提高了80% （RIBA,2008）。尽管取得了许多成果，但对DfMA方法的广泛采用的努力并不成功，只有不到5%的设计师采用该方法。CIOB、RIBA和AECOM的研究表明，缺乏采用率是因为当前的设计师接受过为建筑设计的培训/教育，并且长期以来一直采用这种方法。尝试使用CPD培训只产生了一点点好处，因为设计师们声称太忙了。然而，大流行现在意味着几乎没有学员/设计师注册，培训公司和建筑产出受到冲击。因此，该项目旨在使用数字手段，通过开发BIM软件插件来鼓励更广泛的DfMA采用，该插件可以根据客户/项目简介中的关键建筑设计参数（例如材料选择，建筑用途/目的等）自动生成DfMA概念设计。云计算，人工智能算法（AIA）和大数据分析，包括以下内容：1)**自动DfMA概念设计生成器：**将根据输入参数生成多个概念设计。设计将可编辑，以实现“详细设计”，以满足设计师的喜好。将使用以下方法生成设计：1a)参数化建模ai，它将使用以前DfMA设计的历史数据来生成新的解决方案；1b)生成式设计ai，以改进新的解决方案，产生许多有效的高性能但具有成本效益的选择。2)**DfMA组件顾问：**将建议组件（如格板，壳梁等）可用于编辑已采用的DfMA概念设计，以实现适合设计师偏好的详细设计。3)**DfMA组件可用性和价格检查器：**将提供DfMA组件价格，交货时间，可用性，供应商位置等信息。4)**DfMA设计比较工具：**将根据总成本，估计持续时间等比较选定的生成/编辑设计。

**DEFINITIONS****Concept-design:** Design team's initial response to a project brief.**Detailed-design:** A design with full details, developed from approved concept-design**SUMMARY**Due to coronavirus lockdown, UK Construction Total Activity Index dropped to 39.3 in March from 52.6 in February; the steepest fall in construction output in 11-years (IHS-Markit, 2020). Government wants to use the coronavirus crisis as an opportunity "to build the homes" and plans to "build build build" as part of recovery plans for construction industry and wider economy, announcing £5bn infrastructure and £12billion affordable newbuild homes investments (PrimeMinister's office,2020). Unfortunately, current/traditional construction methods are too inefficient to achieve this in the required time as they are mostly responsible for the current record housing backlog/deficit of 4milliom-homes (National Housing Federation,2018; BBC Housing-Briefing,2020; McKinsey & Company,2019), and were way off-track to achieve national infrastructure programme's proposed £650billion projects worth by 2025Current/traditional construction methods are slow, costly, poor quality and relatively unsafe. Compared to manufacturing industry, its poor productivity has cost UK economy £140billion (including tax) over 20years (Mace,2018). Alluding to the problem, the government backs construction industry in its attempt to emulate manufacturing industry by switching to Design for Manufacture and Assembly (DfMA) (Construction Sector Deal, 2018). DfMA trial projects led to reduction of 60%, 44% and 70%+ in duration, cost and onsite labour respectively, 73% improvement in quality and 80% improvement in overall productivity compared to traditional methods (RIBA,2008).Despite the many gains, efforts towards DfMA approach's wider adoption has been unsuccessful, with less than 5% of designers employing the approach. Research by CIOB, RIBA and AECOM show the lack of adoption is because current designers were trained/taught to design for construction and have practised this method for long. Attempts to use CPD trainings have yielded only slight benefit as designers claim being too busy. However, the pandemic has now meant near zero trainees/designers registering, with the training companies and construction output bearing the bruntThis project thus aims to use digital means to encourage a wider DfMA adoption by developing a BIM software plugin that automatically generates DfMA concept-designs based on key building design parameters from client/project brief (e.g. material choice, building use/purpose, etc.)The proposed plugin will use Internet of Things, Blockchain Technology, cloud computing, artificial intelligence algorithms (AIA) and big data analytics and include the following:1)**Automatic DfMA concept-design generator:** will generate multiple concept-designs based on input parameters. Designs will be editable to achieve 'detailed-design' to suit designers' preference. Designs will be generated using:1a) Parametric modelling AIAs which will use historic data of former DfMA designs to produce new solutions1b)Generative design AIAs to improve new solutions, producing many valid high performance but cost effective options.2)**DfMA component adviser:** will suggest components (e.g. lattice-slab, shell-beams etc.) usable to edit an adopted DfMA concept-design to achieve detailed-design that suits designers' preference.3)**DfMA component availability and price checker:** will provide information on DfMA components prices, delivery times, availability, suppliers' locations, etc.4)**DfMA designs comparison tool:** will compare selected generated/edited designs based on total cost, estimated duration,etc.