Abhinandan Kunder

Domain: Polymer Manufacturing Facility Project: Reduction of Quality inspection time by eliminating inprocess checks Process: Once the material has been produced in the reactor, it is transferred to the next vessel known as blowdown vessel. In this vessel, basically chemical addition is done to remove the volatile organic component (VOC) of the batch. Next the same material is transferred to a blender vessel. In this vessel, pH, Viscosity and Solid content is adjusted by adding water and some chemicals. In both vessels we can track the Agitator Amperes which correlates to viscosity, the weight of the material in the vessel which can be correlated with solid content. Present Scenario: We have successfully achieved the results by making models and implementing the same using a tool known as TrendMiner. It is an advanced AI analytics tools which is linked to DCS to directly take the data and predict the future outcomes based on past data. It has also been mapped batch wise. Using this tool, continuous SPC are plotted and in real time we have a visibility on how the data will vary. Challenges: The challenge here is over the period of time The data will start to map based on present trends and information of the past is lost. There are some physical challenges where for example weight of the blender and blowdown increases due to accumulation over time. This will alter the data trend. Which will lead to wrong correlation impacting the results ultimately resulting in failure of the project. Action to be taken: When using AI, it is not a one time approach. A continuous monitoring of the process and models have to be done. The equations have to be tested to see if they represent the real world scenario. Accordingly inputs have to be given inorder to modify the models. Role of the process owner becomes critical. In the Control Plan of the project this has to be taken into account. New parameters have to be taken into account to adjust the models.

Domain: Polymer Chemical Industry Project: To increase the over all plant utilisation from 80% to 90% Process Explanation: The plant basically contains two kinds of polymer reactions. SBR-Styrene-Butadiene Reactions and ABR-Acrylic Based reactions. The same reactor is used for both and both are batch process. The reactor size is 40 KL. The average batch size is 33 KL per batch. This is to accommodate a the pressure build up in the reactor. For ABR the pressure build up goes upto 2 bar incase of a stable reaction. For SBR the pressure build up actually goes upto 6 Bar under stable conditions. The batches can be taken back to back with a mild steam flush at 3 Bar and 120oC. There are 3 ways to increase the output. Increase the batch size keeping the cycle time Constant: To keep the cycle time constant, we need to increase the flow rate of chemicals into the reactor. Increase in the flow will result in lower reaction time for the chemicals. This might lead to unreacted chemicals, lower heat dissipation and vapour formation which leads to pressure build up. Reduce the cycle time by increasing the flow keeping the batch size constant: This results in the same above problems but pressure build is less. Reduce the cycle time by increasing the flow and also increasing the batch size : This results in all the issues and is a more riskier process. We produced around 27 odd batches which involved a total of 350 Raw materials. The ratio also varies across each batch. Varying the ratio impacts the quality and application. It also affects downstream processes. For SBR the minimum time is 4.5 hrs and for ABR the minimum time is 3.5 hrs. Time cannot be reduced below this. MBB Vs AI AI: AI is able to analyse the data much faster than the MBB. All parameters can be studied in one go. Simulations can be made. Hypothesis can be designed. It can even design a DOE. MBB: The role of MBB is to recheck the hypothesis. Ensure that limits are taken into account. Verify the designed DOE. Get the practical experience into the system. In simple words, AI helps to speed up the process which we as MBB would do manually and also increase the number of data points to get accurate results. In manual process we may miss out data options. But AI can be used to analyse all the data at once. But causations have to be rechecked.

Context: Construction Chemical Industry Project: Increase the OEE from 50% across 5 factories in India to 65%. With AI in the picture, the way DMAIC is seen changes dramatically. AI helps in faster identification of problem, faster measurement system analysis, Analyse the result and also provide improvement plans. It can also provide with control plans-renewed Poka Yoke methods to ensure the problem doesn't persist. Although AI can do all these, it still lacks the ability to understand the limitations of data, understand the meaning of metrics and understand the real time issues being faced. Going with the mentioned project, the goal is to increase the OEE of 5 factories across india from 50% to 65% in order to avoid capex investment for expansion projects for the the next 5 years and to meet the increasing demand. Define: AI can analyse the factory capacity vs present demand to tell us that we lack in capacity. But an MBB or an improvement specialist is needed to define the project strategy. AI tells us the gap but an MBB will tell us where we need to act on. Measure: An AI will read and analyse the data faster. Much faster than a human can. Incase we have IOT sensors in the plant, the AI can actually provide the best analysis. An MBB's role is to understand the data first. For the AI, Garbage in = Garbage Out. So we basically need someone who can maintain the quality of data. Set correct system which actually gives the right data to AI. If the data is manual, then someone has to check the data before feeding into AI. Analyse: This is where the MBB plays a major role. Now we have massive amount of Data. AI can analyse and give an output in seconds. But it cannot tell us where to act. What an MBB would do is Financial Loss Analysis-Analyse from results as to what are the losses, how practical are the losses and can we actually work on them? What is the First pass rate? Does is show too high? or is it too low that there is lot or re-work Time Loss Analysis-Where are we lagging? are our cycle times higher than other counters? Are our cycle times different across different factories for same product? Is there a standardisation or not? Is the mixing time too high? is the QC time too high? Value Stream Mapping-AI cannot decide what processes are needed and what can be eliminated. You need an MBB to map the process and actually look at what is the value. Creation of a Defect Pareto-Problems have been identified and now you need to see the feasibility of eliminating these problems. For example, Quality check time is a non value added but essential activity. But can we actually eliminate it? What is the role of AI? AI basically takes away the headache of choosing which graph or plots to be used. Is it a P-chart, NP-Chart etc., Still it has to verified by the MBB. Is the data normal or not? its done within seconds by AI provided the data is authentic which is checked by BB All the data available can be evaluated, models can be run and simulated. AI reduces the analysis time by a great extent. Improve: You have analysed the data. You have the entire report ready. The role of an MBB is to identify what are actionable and what are not. For example: lets say one factory has a license limitation for its capacity. Although your capacity is very high, you cannot legally achieve it. So the question arises-Do i implement the improvement measures in this factory or not? instead of focusing on OEE, Do i focus on improving the productivity? Or do i work on increasing the license capacity? These questions can only be answered by an BB and the management. Control: The AI can give you methodologies to implement the control mechanism but we need an MBB to actually build and implement the system. For example, AI gives the suggestion of implementing a Visual digital board on the shop floor. But for a construction chemical company is it actually possible to do this? is it cost optimum? is it safety compatible looking at the dust in the shop floor? An BB can take a call to rather have a manual Daily Work Management board which is cost effective and also dust proof. After all this let's assume the project has been implemented and we have successfully increased the OEE. But one big question only an MBB can answer is Although we avoided the capex for the next 5 years, did we actually sacrifice the flexibility of fulfilling customer demand to increase utilisation of our plants? will we be able to fulfil ad-hoc demands with present new systems implemented? An MBB or a BB would be able to answer this question in accordance with the management before the project even began.

Domain: Construction Chemical Industry Problem: Raw material, Production and Dispatch planning activity in an industry with very low forecast accuracy The present situation involves the production planner getting the product requirement from SAP or any ERP that has been posted by sales. The delivery day count starts once the order has been placed by sales and this is the key KPI for the plant teams on how soon the material can be delivered. Also, there are over 2000 different product mixes which have to be carefully planned to meet all customer demands and in between there are also ad hoc orders based on customer priority and weather conditions. The production planner has to check his capacity, look at the changeover matrix, check the space available at the factory, plan for the manpower and much more. The production planner then checks, if the we already have stock of these material with us. If yes, the plan is shared with the dispatch planner who is responsible for working out the logistics details which includes but is not limited to logistics planning, freight value negotiations, selection of transporter based on area to where the material has to be delivered etc., The dispatch planner has the main role of optimising cost for both the company and the customer. They have to look at full truck load as well as part truck load and see the impact on cost. They have to take a call on lead time also. The question arises what if i wait for another order from the same customer to same location to fulfill a full truck load which is considerably cheaper? Will this increase my lead time? is this okay with the customer? The role is a basic day to day transactional role but with n number of permutations and combinations. If the product is not available, the production planner reaches to raw material planner. They check the availability of Raw material and packing goods. They check what is the present stock? what is the safety stock? What is the re-order level? What is the lead time level? For example; The system data shows lead time is 2 days for a particular material. But because of the proximity with vendor location, practically the Raw material planner orders on call basis. This saves on inventory management. The Raw material planner has to have a close coordination with the central purchase team to plan accordingly. Due to the forecast accuracy being only around 20%, this becomes extremely challenging to analyse and plan. This is a pure state where one has to prioritize Flexibility vs Productivity. All these challenges leads to the system being more dependent on personal experiences and people rather than Data alone. We can feed years together of data along with curated experience collections from subject matter experts into AI. But where the MBB actually comes into place is in: Defining the actual problem-Is it forecast accuracy? is it the nature of business? Checking the accuracy of the data- "Garbage in is equal to Garbage Out". Checking the data quality, ensuring frequency of updates, having a connected system. Identifying the key performance indicators to be considered. Defining the trade offs-Do i lose my productivity or do i focus on being flexible? Defining the legal bounds for the AI-Interpretation of the Legal system becomes challenging and it also varies from location to location. Keeping all these into account, the role of an MBB improvement lead doesn't end but it rather starts. An AI MBB lead moves from a doer to a strategist Understand what is the base state the system was actually designed for and try to achieve it. Understand the data. Look at the source and its consistency. Ensure quality of the data. Next, define the frameworks within which the AI has to operate. This includes feeding the capacity data, changeover times, product compatibility. Identify the legal bounds. Ensuring the system is being followed after the system implementation and hasn't reverted back to base state. The AI can only analyse the data provided and give us results, but It is the MBB lead who owns the process and ensures the right implementation of AI.

Domain: Construction Chemical Plant To answer the question, "what new habits should AI-driven organisations and leaders should adopt?", i will state the example of Construction chemical industry. The main challenges generally faced in construction chemical industry is Low forecast, lower margins, higher transportation costs, continuously changing plans to name a few. To give an example, during the beginning of the week, generally on a monday, Supply and Demand planning calls are held. The topic of discussion in these calls involves the below points: What was the previous week dispatch? did it match the plan? If not, why did it vary and what better could have been done. What is this week's plan? What volume has been distributed across which factory. What is the accuracy of plan. Do we have sufficient Raw materials and packing materials to support this? Are there sufficient vehicles been assigned for dispatch to customer? will the vehicles be placed by the customer or by us? Do we have sufficient manpower deployed? is the crew plan done according to order? What is the capacity vs actual order for this week? was there a carry forward from last week? if yes, how will it be handled. Are there any urgent orders? And many other queries. In general the answers to these queries are based on personal experience rather than on Data. The approach an AI driven Organisation or Leader needs to take is the Establishment of an integrated real time data module which directly collects information in the field Asking data driven questions rather than gut feel Development of system which can take faster decisions Working on building a team with a data first mindset Coaching teams to utilise AI but at the same time not blindly trust the output Going point by point: Establishment of an integrated real time data module which directly collects information in the field: Use of systems like IOT to connect directly to data platforms can help us to get real time data. for example, an automated packing line connected to an MES can help us to monitor the actual packing time, the tolerance of packing, if packing is slowed then at what juncture, which product etc., This would eventually allow us to map the process of packing end to end and convey clearly the capacity per batch, per material. It can also help us to project real time based on the material parameters as to how long the batch would take. Asking data driven questions rather than gut feel: With this data we will be able to analyse patterns. We can ask questions like if i fulfill an urgent order of "X" material, then "Y" would be my downtime, "W" would be my cleaning time, "Z" would be changeover, "V" would be my total loss. Is this loss okay for sales? this will also allow me to prioritize which batch to be taken first, set the cleaning matrix etc., Development of system which can take faster decisions: with this data in hand, the system can calculate and redo the the entire production plan based on an GO or NO-GO input from the plant team. What would generally take a production manager 2 hours can be adjusted by AI in seconds. Development of a right decision making framework is needed. Working on building a team with a data first mindset: AI is only as good as the people who use it. It is necessary to invest in training of teams to build their capability of asking the write questions. They should understand the logic behind the output. Coaching teams to utilise AI but at the same time not blindly trust the output: Along with training, it is necessary for the teams to understand the limitations of AI and also understand the data accuracy. For example, The production calculations can be possible if we check what kind of data is considered. Has the manpower needed for the activity been considered? if yes, what hours have they already worked. Is it needed for them to work Overtime? Can we deploy the same manpower or a different manpower is needed? Are we complying with the Overtime Time restrictions? Basically we need to check if the right parameters are set. It is necessary for the leader to let go of the bias and move towards a data driven approach.