The seven stages of chemical technology from concept to industrialization (Issue 59/100) – A case study on the para-phenylenediamine project

Thread Content

This post was last edited by xiouxingzhe on 2026-7-5 23:24. The seven stages of chemical technology from concept to industrialization (Issue 59/100) —— Technology refinement: The para-phenylenediamine project as an example. Dear friends: Hello everyone! In the previous issue, we discussed the MDI project – the adaptation and transformation of international mature process packages. In this issue, we will discuss a completely different type of case: the preparation of a process package for the para-phenylenediamine continuous flow process. If the MDI project is about “standing on the shoulders of giants,” the para-phenylene diamine project is about “building one’s own ladder.” This set of processes was developed entirely through independent research and development; there are no existing industrial facilities in China that can serve as a reference, no mature process packages available for use, and no industry practices to follow. All design parameters, operating ranges, and safety interlock setpoints must be established from scratch, based on the data accumulated during pilot, pilot-scale, and full-scale testing. The development of such “original innovative” technologies is fundamentally different from the adoption of mature technologies. In this issue, we will discuss some of the practical problems encountered in this process, as well as the corresponding ways of thinking about them and dealing with them. I. A starting point without any references: When an MDI project receives the international process package, every value on the PFD has a source; the format and depth of the equipment data sheets have been refined over decades through continuous improvement, and the design approach for the protection layers related to interlock logic is clear and well-thought out. The diphenylamine project is different – the pilot and pilot-scale data sets are your only assets. The form and depth to which the process package is developed depend entirely on your understanding of the technical data and your ability to express it in an engineering context. To put it simply, introducing a mature process package is like having the complete set of construction plans for a building; your task is to mark out the layout on site and carry out construction according to those plans. An original innovative process package is like having only a few conceptual sketches from an architect and some material testing data at your disposal; you need to draw up the entire set of construction drawings on your own – and if those drawings are incorrect, the building constructed later might collapse. In this process of establishing technical standards from scratch, the biggest challenge is not the technical aspects themselves, but determining how to avoid either over-designing – which leads to excessive investment costs and a loss of market competitiveness – or under-designing – which creates risks for safe operation in the future – in the absence of any references. There are no preset answers as to how much margin to use in each case or what safety factor to apply; it is necessary to find a reasonable balance between one’s own data and the industry standards. II. Making design decisions based on limited data: The pilot plant data set includes key parameters such as reaction temperature, residence time, conversion rate, and selectivity. But how many hours was the pilot run? How many operating conditions are covered by these data? Have the operating limits of the equipment been determined? Is the degradation pattern of catalyst activity during long-term operation clear? For these issues, there is plenty of historical data in mature processes that can be used to provide answers; however, in projects involving original innovation, the amount of data is often quite limited. Faced with this reality, the project team’s approach is to face uncertainty honestly, without deception. For those processes for which there is sufficient data and clear patterns—such as cases where consistent results are obtained when running multiple batches under optimal conditions—a relatively compact margin is chosen during design, so as not to excessively increase the size of the equipment. For those aspects where the data is insufficient and the patterns are not yet clear—such as performance under certain boundary conditions or changes in performance after long-term operation—relatively conservative margins are adopted during the design phase. Meanwhile, the section on \"Recommendations for Further Actions\" in the process package specifies which parameters need to be closely monitored and verified during the initial operation of the industrial plant. This approach of using “differential margins” is far more reasonable than applying a uniform 20% increase. The former is an active design decision based on risk assessment, while the latter is a passive compromise to conceal cognitive shortcomings. In projects involving original innovation, resources are already limited. If the available margin is distributed evenly across all aspects, the areas that truly require caution may not have enough of it, while investment is wasted in areas that do not need such caution. III. Determination of the operating window: The operating window for a mature process is usually supported by long-term operation data from similar installations – such as the allowable temperature range, pressure fluctuations, and load range; these values are considered “known” within the process package. For the diphenylenediamine project, it is necessary to rely on the limited operational data from the pilot plant stage, along with theoretical analysis, in order to determine the operating window. Take the temperature window as an example. Several sets of experiments were conducted at different temperatures during the pilot scale stage, yielding a relationship between temperature and yield. However, the pilot plant data cover only a limited number of temperature points, and each point was operated for only a limited time. When determining the operating temperature range in the process package, in addition to the experimental data from pilot tests, extrapolation from kinetic models was also utilized – near the measured points, the model was used to estimate the extent of the impact of temperature variations on yield and selectivity, thereby assessing the tolerance of the range. If the model shows a steep change in selectivity around a certain temperature, the operating window becomes narrower, and control precision is improved to ensure yield and selectivity. The finally determined operating window is also expressed in the process package in a manner different from that of mature processes. For mature processes, it may be stated directly as \"the operating temperature range is from X to Y degrees\", as this is a recognized range verified through multiple units. In the process package for diphenylamine, in addition to specifying the ranges, a notation indicating the \"operation window validation status\" has been added: which ranges have been verified through pilot-scale tests and which are based on model extrapolation. The advantage of this is that it enables the subsequent engineering design and operation teams to understand the reliability of each operational boundary, as well as which parameters require closer monitoring during operation and production. IV. Setting of safety interlocks: The interlock design for originally innovative processes faces a challenge that mature processes generally do not encounter: many potential accident scenarios never occur during the pilot stage, so one can only rely on theoretical analysis to predict possible risk situations, and then design interlock protections based on those predictions. It sounds like “defending against an enemy you’ve never seen before.” However, safety cannot be relaxed just because something hasn’t happened – many major chemical accidents occur precisely in the first industrial-scale plants, as the risk scenarios were never identified during the research and development phase. The diphenylamine process involves a diazotization reaction, which is a highly exothermic process that is sensitive to temperature. The project team explored each possible risk scenario in the HAZOP analysis. Can microchannel blockage cause pressure buildup upstream? Can an imbalance in the feed ratio lead to uncontrolled reaction temperature? Will a disruption in the cooling water prevent the reaction heat within the microchannels from being removed in a timely manner? Each scenario followed the logical chain of “cause-consequence-protection measures”. Since it is original innovation, when using HAZOP analysis to determine whether protective measures are sufficient, there is often a lack of historical data from similar installations to rely on – it is not possible to say that “similar installations all handle things this way”. Therefore, for each potential high-risk scenario, a more conservative protection strategy was adopted: it’s better to add an extra layer of independent protection rather than relying on a single protective layer. The process package delivery documents specify the SIL level of each interlock circuit, the sensor configuration scheme, and the requirements for periodic testing. Meanwhile, the “Precautions for Commissioning” section of the process package specifies that during the first commissioning, all SIS interlocks must be in operation and cannot be bypassed ； Which parameters need to be closely monitored while driving? Once abnormal trends are detected, it is necessary to reduce the load or stop the vehicle immediately, rather than adopting a wait-and-see attitude. These requirements, which may be common sense in mature installations, deserve to be emphasized repeatedly for first-industrialized installations. V. Provision for iteration of the process package: A significant difference between the process package of an originally innovative technology and that of a mature technology is that the former should have a \"provision for iteration\". After the mature process package is delivered, in principle no further adjustments to the core technical parameters are required for subsequent engineering design – as these have already been verified across multiple plants. But an original innovative process package cannot be achieved in one step. After the first industrial installation is built and put into operation, certain problems that were not anticipated during the design phase will inevitably arise—such as a heat exchanger having an area that is too large or too small, an operating range that is narrower in actual operation than during the pilot testing phase, or unexpected corrosion of a pipe material after prolonged operation. These problems are not due to mistakes in the design; rather, “first industrialization” itself is exploratory in nature. Therefore, when preparing the para-phenylene diamine process package, space was deliberately left for iterative adjustments in some key stages. For example, in terms of equipment layout, operating space and spare nozzles have been reserved in areas where adjustments might be needed. There may be uncertainty regarding the heat exchanger area – if the fouling coefficient is set on a conservative note, and it is found that there is an excess capacity for heat exchange during actual operation, energy consumption can be optimized by reducing the use of utility resources, without the need to modify the equipment. But if the heat exchange capacity is insufficient, it is necessary to add more heat exchangers or modify the pipelines. In areas where such risks exist, additional space is reserved during the equipment installation, so that if modifications are indeed necessary in the future, no major changes will be required. For example, in terms of control strategies, for those loops for which the optimal strategy has not yet been fully determined, the DCS provides an online switching function for control strategies – after actual operation, the effects of different strategies are compared to select the better one and lock it in. These reserves are not a result of redundant design, but rather a pragmatic approach to the inherent uncertainties of original innovation. They require a bit more initial investment, but they facilitate subsequent optimizations. This concept still holds value in the industrialization of proprietary technologies. VI. Some Insights Looking back at the entire process of preparing the diphenylamine process package, there are a few insights that stand out. First, with the development of truly original technologies, the core challenge is not drawing itself, but making design decisions when data is limited. There are no ready-made references for what values each parameter should take, how much margin to leave, or what safety factor to set. It’s much better to be honest about uncertainties—to know which aspects have sufficient data and which ones remain unclear, and to handle them differently in the design—than to pretend that everything is certain. Forcing oneself to say “everything has been clarified” is equivalent to pushing the issue from the technical design phase to the engineering implementation phase, or even to the actual operation phase, with costs only increasing as a result. Second, the quality of the pilot plant data directly determines the reliability of the process package. This point has been repeated in previous issues, and the experience is particularly evident in the diphenylamine project. In areas with sufficient data, design decisions are more decisive and the margin for flexibility is tighter ； In areas where there is insufficient data, a more conservative approach must be taken to leave room for uncertainty. Spending more time during the pilot-scale phase to clarify the boundaries is a much more proactive approach than relying on conservative estimates made arbitrarily during the process package development stage. If conservative values are used to an excessive extent, it will be found after the installation is completed that there is excess capacity everywhere; production capacity cannot be increased, energy consumption cannot be reduced, and the investment cost exceeds the budget. Attempting to make optimizations at that point comes at a much higher cost than conducting a few additional experiments during the pilot testing phase. Third, the original innovative process package should leave room for subsequent iterations. The first industrial-scale plants served, to a large extent, as a means for \"industrial testing\" – even with comprehensive pilot plant data, the actual performance at an industrial scale can still differ from that observed in pilot tests. When preparing the process package, room for adjustment is reserved in areas where problems may arise – equipment layout, provisions for pipe connections, and flexibility in control schemes. These provisions are not a waste; rather, they represent preparations for future optimizations. This concept is in line with the original innovative process for para-phenylenediamine. Preview for the next issue: Issue 60 – The beginning of engineering translation: from technical language to engineering language. With the case study of the para-phenylenediamine project now completed, Phase 4 \"Technical finalization\" is brought to an end in this issue. Starting from issue 60, we officially enter the fifth phase—engineering translation. The process package is placed on the table; what needs to be done next is to turn it into drawings that the construction team can understand, technical specifications useful for the procurement department, and specialized documents that meet the requirements for safety and environmental assessments. What is the biggest challenge in the translation process from technical language to engineering language? How to avoid information loss in multi-disciplinary collaboration? To be continued in the next issue.