石化企业聚焦数智转型 Petrochemical enterprises focus on digital intelligent transformation
问：炼化企业数字化转型的进展如何？主要应用在哪些方面？
以数据为核心、以故障模型为基础、以智能算法为工具，形成“监测—预警—决策—优化”智能化设备全生命周期闭环管理。
■ 独石化副总经理 陆军：
独石化坚持预防在先、抓细抓早，重点关注设备故障诊断。
一是模型依托的诊断预警。独石化从波形、频谱、包络谱特征等17个特征建立故障模型库，实现早期故障诊断和提前预警，通过“智能初筛+专家终审”表决机制进一步提升诊断、预警的准确性，故障定位精准到2个部件以内。同时，公司充分利用运行寿命预估、能效动态管理等功能，动态优化工艺操作并在线调整，合理降低能耗，减少非计划停机。
二是数据驱动的协同决策。实时采集振值、温度等设备运行数据，通过短/长期趋势分析和异常幅值模型综合判定故障，并向特定人员推送处理方案，缩短响应时间。独石化将依照国标设定预警值与通过AI算法自定义报警值相结合，采用“门限+智能”双重报警模式，100%覆盖风险，降低漏检率，同时，通过算法迭代持续优化准确率。
三是效益导向的数字化闭环。通过预测性维护减少突发检修，每年可节约成本150余万元、节约工时4万余小时，设备无故障运行时间超120个月。公司使用数字化工具实现高危泵100%覆盖，健康度A区占比超85%，预知检修率超95%，可避免介质外漏等重大风险，保障生产连续性。（李志强 采访）
设备管理模式由“被动维护、经验驱动”向“主动预测、数据驱动”转变。
■ 乌石化机动设备部副经理 高玉平：
乌石化作为集团公司第二批数智化转型推广企业，今年8月正式迈入实施阶段。
在智能设备全生命周期管理方面，公司机动设备部整合自建系统搭建设备完整性平台，打破专业壁垒，实现关键设备数据自动采集率超90%，覆盖高危泵、压缩机组等核心设备，确保数据实时完整。同步建立数字化档案并关联全周期数据，运用算法预测设备故障与寿命，精准制定维护计划，有效降低库存成本并缩短停机时间。
在监测与运维模式优化上，公司完善在线监测预警机制，实现关键机组和高危泵在线监测全覆盖，新增压缩机组在线监测点位，构建分级预警系统实时推送故障报警，确保隐患及时处置；同时，推行预知性维护减少非计划停机，形成“在线+离线”双重监测网络，显著提升设备状态的感知能力。
在管理机制创新层面，公司建立“业务+技术”双轮驱动机制，明确业务主导、技术支撑的定位，聚焦可靠性与成本优化。乌石化同步推进数字化能力全员培训，提升设备管理人员数据分析与AI工具应用能力，为数字化转型提供人才保障。（胡鑫 采访）
在数字化转型建设中，云南石化共设计7个一级场景和5个二级场景，开发功能700余个，实现重点生产装置全覆盖。
■ 云南石化规划和科技信息部副主任 谢皆群：
近日，云南石化成功通过集团公司炼化板块数字化转型第二批试点单位上线验收，评审结果为优秀，达到智能制造能力成熟度三级标准，正朝着四级目标稳步迈进。
强化智能系统应用。采用智能建模、智能PID控制、多变量智能控制等技术，构建了覆盖全厂22套生产装置、2200多个控制回路的智能模型库，赋予控制回路自学习、自诊断、自适应能力，有效解决了生产装置从基础控制到多变量控制的精准控制难题。项目实施后，厂内重点装置自控率达100%，平稳率达99.96%，显著降低了操作人员的劳动强度及设备的管理成本，今年上半年全厂综合能耗连续5个月下降。
在设备管理上强化智能化研判。在动设备状态监测方面，云南石化应用人工智能技术处理采集的振动波形数据，实现设备运行数据的自动采集、智能分析与自动预警，有效降低事故风险与运营成本，并且还能自动巡检设备的运行状态并推送运维建议，确保设备异常缺陷及时发现、提前处置，降低了人工巡检成本。
云南石化应用知识图谱技术整合设备维修经验、故障案例、维修工艺等，通过AI算法自动匹配并推荐最优维修策略，实现维修方案的智能化与优化。同时，通过建立“故障预测—维修计划—资源调度—执行跟踪—效果评估”全流程智能化维修管理体系，打通了从“监测”到“诊断”再到“维修”的闭环流程，有效消除了系统孤岛现象。（杨勇 唐龙 采访）
问：炼化企业如何调整生产管理模式和技术适配策略，从而抓住转型带来的机遇？
主动对接先进制造，在新建项目和重大改造项目中，优先选用具备深度感知、智能控制、预测诊断能力的国产化智能装备，从源头提升设备的数字化水平。
■ 广西石化机动设备部主任 邓宗华：
管理思维要从“经验主导”转向“数据驱动”，应充分发挥数据作为新型生产要素的作用。广西石化通过设备状态在线监测平台，持续采集设备振动频谱、温度场分布、压力波动、运行效率曲线等关键参数，融合设备物理失效机理与模型算法，构建设备故障预诊模型，精准识别早期异常征兆，动态优化维修策略，提升预防性维修比例，显著降低故障维修比例。
对标“卓越工厂”建设要求，深化设备管理与数智融合。我们必须打破数据壁垒，强化跨部门协作机制。主动对接MES等系统，实现设备状态数据与生产、能耗、安全数据的互联互通与协同分析，着力建设设备完整性管理平台，支撑生产优化与设备管理决策。通过共享数据实现数据的集中管理和深度分析，利用大数据分析技术，识别设备性能劣化趋势，预测潜在故障，为精准决策提供量化的数据支撑，提升管理的精细化、智能化水平。
紧抓智能化机遇，夯实设备数字化根基。《方案》力推机械工业智能装备水平提升，这正是我们优化设备“硬实力”的黄金期。我们要坚持“总体规划、分步实施、注重应用、取得实效”原则，高质量推进数字化转型、智能化发展建设工作。（吴乙博 采访）
要找准数智化投入破局点，摒弃零敲碎打的改良主义，实施系统性转型升级，推动技术落地以实现竞争力突围。
■ 兰州石化总经理助理兼规划和科技信息部部长 张涛涛：
数智化浪潮席卷而来，新质生产力蓬勃发展，5G、人工智能、大模型等新技术不断涌现，传统能源行业正迎来数智化转型的关键时期。
因地制宜合理布局是基础。智能化的起点并非盲目追逐尖端技术，而是源于对自身业务痛点与发展蓝图的深刻洞察。兰州石化将有限的资源投入最能产生效益的领域，紧盯效益最大化目标，构建指向明确、投资回报率高的智能化资产组合，避免陷入“为数字化而数字化”的战略短视。
有效利用数据价值是核心。数智化转型的核心价值，在于让数据成为驱动决策的澎湃动能。这就需要我们进一步提高数据分析能力，实现2大关键突破。一是实现从“事后”到“事前”的转变，通过对设备运行数据的深度学习，精准预判故障，将非计划停工、安全环保事故等扼杀在摇篮中。二是通过对工艺参数实时分析，寻找能耗与收率的“黄金平衡点”，实现生产过程的动态优化，既创造直接经济效益，又为管理者提供“数据说话”的精准支持。
与时俱进挖掘适配场景是关键。真正的数智化转型，是打破企业各部门、各环节、各专业间的壁垒，让数据价值体现在企业的每一个角落。例如，当生产一线的预测性维护和工艺优化取得成功后，可以将这种模式复制并拓展到更广阔的场景，实现从“点”的突破，到“线”的贯通，再到“面”的融合，最终打造具备持续自我进化能力的智慧型企业。
目前，公司已集成24套系统数据，覆盖26万余台套设备数据，业务流程线上率达95%，报表自动生成率达92%，指标自动统计率达63.8%，各层级人员的平台使用比例超95%。
■ 四川石化机动设备部经理 王超：
四川石化工艺流程复杂、设备体量大、安全要求高。因此，提升设备数字化水平，实现精益化管理，是新形势下设备管理工作必须面对的挑战和机遇。就四川石化设备管理而言，重点开展以下几方面工作。
一是向设备完整性管理的理念转变。将“设备运行安全可靠、设备状态预知掌控、设备维修科学经济”作为目标，推动设备管理由“经验驱动”向“数据驱动”转变、由“事后维修”向“主动维护”转变、由“层级管理”向“精准赋能”转变、由“信息孤岛”向“一体化协同”转变。
二是以技术手段为决策提供科学依据。公司积极建设设备完整性管理平台，同时充分运用状态监测和故障诊断等技术手段，实现问题可视化呈现、分析智能化穿透、决策数据化支撑等数智功能，形成以预防为主、主动维护、精准维修的科学方法，同时，显著降低员工的劳动强度。
三是增强管理模式运行效率。公司充分适应当前数字化时代的管理变革，针对设备管理机制现状，系统实施“协同化运营、合作化维保、精准化检修、规范化管理、同向化引领”五化运维工作。同时，发挥公司与设备厂家、检维修单位的专业优势，优化工作流程，做好分工协同，释放设备数字化水平提升带来的巨大潜力。
郑重声明：国能春泉编辑部尽力持续给大家提供价值信息促进交流，本平台发布的行业文章、图片、表格等源自互联网、企业政府官网、公众号等公开渠道整理编辑，我们对文中观点保持中立。转载的内容版权归原作者和机构所有，如有侵权，请联系我们删除，谢谢！本文如对您有帮助，欢迎转发，感谢支持！
Petrochemical enterprises focus on digital intelligent transformation
Question: Progress and Application of Digital Transformation in Refining and Chemical Enterprises
With data as the core, fault models as the foundation, and intelligent algorithms as tools, an intelligent closed-loop full-life-cycle equipment management system featuring "monitoring—early warning—decision-making—optimization" has been formed.
Lu Jun, Deputy General Manager of Dushanzi Petrochemical:
Dushanzi Petrochemical adheres to prevention first and focuses on early and detailed management, with key attention to equipment fault diagnosis.
First, model-based diagnosis and early warning. The company has established a fault model library based on 17 features including waveform, frequency spectrum, and envelope spectrum characteristics, enabling early fault diagnosis and advance warning. A "intelligent preliminary screening + expert final review" voting mechanism further improves the accuracy of diagnosis and warning, with fault location pinpointed to within 2 components. Meanwhile, the company fully utilizes functions such as operational life prediction and dynamic energy efficiency management to dynamically optimize process operations, make online adjustments, reasonably reduce energy consumption, and minimize unplanned shutdowns.
Second, data-driven collaborative decision-making. Real-time collection of equipment operation data such as vibration amplitude and temperature is conducted. Faults are comprehensively determined through short/long-term trend analysis and abnormal amplitude models, and handling plans are pushed to specific personnel to shorten response time. Combining early warning values set in accordance with national standards and custom alarm values via AI algorithms, the company adopts a "threshold + intelligent" dual alarm mode to achieve 100% risk coverage, reduce missed detection rates, and continuously optimize accuracy through algorithm iteration.
Third, benefit-oriented digital closed loop. Predictive maintenance has reduced emergency overhauls, saving over 1.5 million yuan in costs and more than 40,000 working hours annually, with equipment trouble-free operation time exceeding 120 months. The company uses digital tools to achieve 100% coverage of high-risk pumps, with over 85% in health zone A and a predictive maintenance rate of over 95%, avoiding major risks such as medium leakage and ensuring production continuity. (Interviewed by Li Zhiqiang)
The equipment management model has shifted from "passive maintenance and experience-driven" to "proactive prediction and data-driven".
Gao Yuping, Deputy Manager of the Mechanical Equipment Department of Wulumuqi Petrochemical:
As the second batch of digital and intelligent transformation promotion enterprises of the Group, Wulumuqi Petrochemical officially entered the implementation phase in August this year.
In terms of intelligent full-life-cycle equipment management, the Mechanical Equipment Department of the company integrated self-built systems to build an equipment integrity platform, breaking professional barriers, achieving an automatic collection rate of over 90% for key equipment data, covering core equipment such as high-risk pumps and compressor units, and ensuring real-time and complete data. Simultaneously, digital files are established and linked with full-cycle data, and algorithms are used to predict equipment faults and service life, accurately formulate maintenance plans, effectively reduce inventory costs and shorten shutdown time.
In optimizing monitoring and operation and maintenance models, the company improved the online monitoring and early warning mechanism, achieving full coverage of online monitoring for key units and high-risk pumps, adding online monitoring points for compressor units, and building a hierarchical early warning system to push fault alarms in real time to ensure timely disposal of hidden dangers; at the same time, predictive maintenance is implemented to reduce unplanned shutdowns, forming a "online + offline" dual monitoring network, which significantly improves the perception ability of equipment status.
In terms of management mechanism innovation, the company established a "business + technology" dual-drive mechanism, clarifying the positioning of business leadership and technical support, focusing on reliability and cost optimization. Wulumuqi Petrochemical simultaneously promoted full-staff training on digital capabilities to improve equipment managers' data analysis and AI tool application capabilities, providing talent guarantee for digital transformation. (Interviewed by Hu Xin)
In the construction of digital transformation, Yunnan Petrochemical designed 7 first-level scenarios and 5 second-level scenarios, developed more than 700 functions, and achieved full coverage of key production units.
Xie Jiequn, Deputy Director of the Planning, Science, Technology and Information Department of Yunnan Petrochemical:
Recently, Yunnan Petrochemical successfully passed the online acceptance of the second batch of pilot units for digital transformation in the refining and chemical sector of the Group, with an excellent evaluation result, meeting the intelligent manufacturing capability maturity level 3 standard and steadily moving towards the level 4 goal.
Strengthen the application of intelligent systems. Adopting technologies such as intelligent modeling, intelligent PID control, and multivariable intelligent control, an intelligent model library covering 22 production units and more than 2,200 control loops in the whole plant has been built, endowing the control loops with self-learning, self-diagnosis, and self-adaptation capabilities, effectively solving the precise control problems of production units from basic control to multivariable control. After the implementation of the project, the automatic control rate of key units in the plant reached 100%, and the stability rate reached 99.96%, significantly reducing the labor intensity of operators and the management cost of equipment. The comprehensive energy consumption of the whole plant decreased for 5 consecutive months in the first half of this year.
Strengthen intelligent research and judgment in equipment management. In terms of dynamic equipment condition monitoring, Yunnan Petrochemical applies artificial intelligence technology to process the collected vibration waveform data, realizing automatic collection, intelligent analysis, and automatic early warning of equipment operation data, effectively reducing accident risks and operating costs. It can also automatically inspect the operation status of equipment and push operation and maintenance suggestions to ensure timely discovery and advance disposal of equipment abnormal defects, reducing manual inspection costs.
Yunnan Petrochemical applies knowledge graph technology to integrate equipment maintenance experience, fault cases, maintenance processes, etc., and automatically matches and recommends the optimal maintenance strategy through AI algorithms, realizing the intelligence and optimization of maintenance plans. At the same time, by establishing a full-process intelligent maintenance management system of "fault prediction—maintenance planning—resource scheduling—execution tracking—effect evaluation", the closed-loop process from "monitoring" to "diagnosis" and then to "maintenance" is connected, effectively eliminating the phenomenon of system islands. (Interviewed by Yang Yong and Tang Long)
Question:
How Can Refining and Chemical Enterprises Adjust Production Management Models and Technology Adaptation Strategies to Seize Opportunities Brought by Transformation?
Proactively connect with advanced manufacturing, and in new construction projects and major renovation projects, prioritize the selection of localized intelligent equipment with advanced perception, intelligent control, and predictive diagnosis capabilities to improve the digital level of equipment from the source.
Deng Zonghua, Director of the Mechanical Equipment Department of Guangxi Petrochemical:
Management thinking should shift from "experience-led" to "data-driven", and the role of data as a new production factor should be fully exerted. Through the equipment condition online monitoring platform, Guangxi Petrochemical continuously collects key parameters such as equipment vibration spectrum, temperature field distribution, pressure fluctuation, and operation efficiency curve, integrates equipment physical failure mechanism and model algorithms to construct equipment fault pre-diagnosis models, accurately identify early abnormal signs, dynamically optimize maintenance strategies, increase the proportion of preventive maintenance, and significantly reduce the proportion of fault maintenance.
Benchmarking against the construction requirements of "excellent factories", deepen the integration of equipment management with digital intelligence. We must break data barriers and strengthen cross-departmental collaboration mechanisms. Proactively connect with systems such as MES to realize the interconnection and collaborative analysis of equipment status data with production, energy consumption, and safety data, focus on building an equipment integrity management platform, and support production optimization and equipment management decision-making. Through data sharing, realize centralized data management and in-depth analysis, use big data analysis technology to identify equipment performance degradation trends, predict potential faults, provide quantitative data support for precise decision-making, and improve the refinement and intelligence level of management.
Seize the opportunity of intelligence and consolidate the foundation of equipment digitalization. The "Plan" promotes the improvement of the level of intelligent equipment in the machinery industry, which is a golden period for us to optimize the "hard power" of equipment. We should adhere to the principles of "overall planning, phased implementation, focusing on application, and achieving practical results" to promote the construction of digital transformation and intelligent development with high quality. (Interviewed by Wu Yibo)
It is necessary to identify the breakthrough point for digital and intelligent investment, abandon piecemeal reformism, implement systematic transformation and upgrading, and promote technology landing to achieve competitive breakthrough.
Zhang Taotao, General Manager Assistant and Director of the Planning, Science, Technology and Information Department of Lanzhou Petrochemical:
The wave of digitalization is sweeping across, new-quality productive forces are booming, new technologies such as 5G, artificial intelligence, and large models are constantly emerging, and the traditional energy industry is ushering in a critical period of digital and intelligent transformation.
Reasonable layout based on local conditions is the foundation. The starting point of intelligence is not to blindly pursue cutting-edge technologies, but to derive from a profound insight into one's own business pain points and development blueprint. Lanzhou Petrochemical invests limited resources in areas that can generate the most benefits, focuses on the goal of maximizing benefits, builds a targeted and high-return intelligent asset portfolio, and avoids falling into the strategic short-sightedness of "digitalization for the sake of digitalization".
Effective utilization of data value is the core. The core value of digital and intelligent transformation lies in making data a powerful driving force for decision-making. This requires us to further improve data analysis capabilities and achieve 2 key breakthroughs. First, realize the transformation from "post-event" to "pre-event", and through in-depth learning of equipment operation data, accurately predict faults, and eliminate unplanned shutdowns, safety and environmental accidents in the cradle. Second, through real-time analysis of process parameters, find the "golden balance point" between energy consumption and yield, realize dynamic optimization of the production process, not only create direct economic benefits, but also provide precise support for managers to "speak with data".
Keeping pace with the times to explore adaptive scenarios is the key. True digital and intelligent transformation is to break down the barriers between various departments, links, and professions of the enterprise, and let data value be reflected in every corner of the enterprise. For example, after the success of predictive maintenance and process optimization on the production line, this model can be replicated and expanded to broader scenarios, realizing breakthroughs from "points" to "lines" and then to "surface" integration, and finally building a smart enterprise with continuous self-evolution capabilities.
At present, the company has integrated data from 24 systems, covering data of more than 260,000 sets of equipment, with an online business process rate of 95%, an automatic report generation rate of 92%, an automatic index statistics rate of 63.8%, and a platform usage rate of over 95% among personnel at all levels.
Wang Chao, Manager of the Mechanical Equipment Department of Sichuan Petrochemical:
Sichuan Petrochemical has complex technological processes, large-scale equipment, and high safety requirements. Therefore, improving the digital level of equipment and realizing refined management are challenges and opportunities that equipment management work must face under the new situation. For equipment management of Sichuan Petrochemical, the following key work is carried out.
First, transform to the concept of equipment integrity management. Taking "safe and reliable equipment operation, predictable and controllable equipment status, and scientific and economical equipment maintenance" as the goals, promote the transformation of equipment management from "experience-driven" to "data-driven", from "post-event maintenance" to "proactive maintenance", from "hierarchical management" to "precision empowerment", and from "information islands" to "integrated collaboration".
Second, use technical means to provide scientific basis for decision-making. The company actively builds an equipment integrity management platform, and at the same time makes full use of technical means such as condition monitoring and fault diagnosis to realize digital and intelligent functions such as visual presentation of problems, intelligent penetration of analysis, and data-driven decision-making, forming a scientific method focusing on prevention, proactive maintenance, and precise maintenance, while significantly reducing the labor intensity of employees.
Third, enhance the operational efficiency of the management model. The company fully adapts to the current management changes in the digital era, and in response to the current situation of equipment management mechanisms, systematically implements the "five modernizations" operation and maintenance work of "collaborative operation, cooperative maintenance, precise maintenance, standardized management, and same-direction leadership". At the same time, give play to the professional advantages of the company, equipment manufacturers, and inspection and maintenance units, optimize work processes, do a good job in division of labor and collaboration, and release the huge potential brought by the improvement of equipment digitalization level.
Solemn Statement
The editorial department of NESC strives to continuously provide valuable information for communication. The industry articles, pictures, and tables published on this platform are compiled and edited from public channels such as the Internet, official websites of enterprises and governments, and WeChat official accounts. We maintain a neutral stance on the views expressed in the articles. The copyright of the reprinted content belongs to the original authors and institutions. If there is any infringement, please contact us for deletion. Thank you! If this article is helpful to you, please feel free to forward it. Thank you for your support!
