<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Rushikesh Kardule on Medium]]></title>
        <description><![CDATA[Stories by Rushikesh Kardule on Medium]]></description>
        <link>https://medium.com/@rushikeshkardule98900?source=rss-8650cd47a33a------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*h1ikx5cGI600-076</url>
            <title>Stories by Rushikesh Kardule on Medium</title>
            <link>https://medium.com/@rushikeshkardule98900?source=rss-8650cd47a33a------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 01:19:49 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rushikeshkardule98900/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Yield Loss in Sulphation Process during Startup and Stoppage]]></title>
            <link>https://medium.com/@rushikeshkardule98900/yield-loss-in-sulphation-process-during-startup-and-stoppage-50f8cd750df7?source=rss-8650cd47a33a------2</link>
            <guid isPermaLink="false">https://medium.com/p/50f8cd750df7</guid>
            <category><![CDATA[yield-optimization]]></category>
            <category><![CDATA[quality-assurance]]></category>
            <category><![CDATA[kaizen]]></category>
            <category><![CDATA[cost-optimization]]></category>
            <category><![CDATA[lean-startup]]></category>
            <dc:creator><![CDATA[Rushikesh Kardule]]></dc:creator>
            <pubDate>Tue, 20 Oct 2026 18:04:43 GMT</pubDate>
            <atom:updated>2026-10-20T18:04:43.456Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*F4-38JBhsJepB6JbVpUlIg.png" /></figure><p>When people talk about improving yield in a continuous plant, they usually look at the big things: reactor design, raw material quality, impurities. In my experience, some of the biggest losses hide in the moments when the plant is <em>not</em> in steady state i.e., startup and shutdown.</p><p>I work in surfactants manufacturing, where sulphation is a continuous process. Every time it starts or stops, we can lose a little or lose a lot. This is the story of how a small change in our shutdown procedure saved about 100 kg of organic alcohol per continuous run.</p><p><strong>A quick look at the reactor</strong></p><p>Our sulphation technology comes from Ballestra, a company that supplies equipment and consulting for this process. The heart of the plant is a falling-film reactor: a bundle of 72 vertical tubes. The organic feed (in this case, alcohol) enters from the top and spreads as a thin film along the inner wall of each tube. The SO₃ gas flows through the same tubes, and the gas-liquid reaction happens on that film.</p><p>The thin film gives excellent contact and heat transfer, but it is also unforgiving. The reaction is fast and exothermic, and it works well only when the gas and organic flows are in the right balance.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/388/1*9bFsM8_MobJEs-PUGN8t7g.png" /><figcaption><em>Ballestra thin film reactor schematics</em></figcaption></figure><p><strong>The startup problem</strong></p><p>When we start the plant, the reaction is not steady at first. We cannot send the material generated during this time period to the next equipment, so it has to be diverted and collected separately. It is easy to recognise dark, often nearly black. When SO₃ and the organic react and the product is not neutralised within the right time, it begins to degrade. Over-sulphation takes place, and the result is a dark, high-organic-load material that is not desired.</p><p>So, the real question at startup is how long do we keep diverting?</p><ul><li>Divert for too short a time, and off-spec product contaminates the good material.</li><li>Divert for too long, and we throw away perfectly good product.</li></ul><p>Either way, we have consumed raw material and energy and got less saleable product out of it. That pushes up our specific energy consumption, the kWh per metric ton of finished product.</p><p>This balance depended largely on operator skill and judgement. Good operators get it right, but a process that relies only on individual skill will vary from shift to shift.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*3lwg8hlzAoGTJiNjz5Eh2Q.png" /><figcaption><em>Real world thin film reactor where gas liquid reaction happens</em></figcaption></figure><p><strong>The shutdown problem</strong></p><p>When we cut off the SO₃ gas, the organic feed must not stop at the same moment. If the organic stops first, the thin film dries up while SO₃ is still in the tubes, and the tubes get a charred, blackish deposit.</p><p>This hurts in two ways. Charring shortens the life of the reactor tubes. And when we restart, the old deposit affects the new run, so product quality suffers until things settle again.</p><p>The organic has to keep flowing long enough after the gas is cut to flush the tubes. But how long is long enough? That was another judgement call, again dependent on whoever was on shift.</p><p><strong>Turning a skill into a standard</strong></p><p>We decided to stop leaving this to individual judgement. We wrote an SOP for plant shutdown in which the organic keeps feeding for <strong>60 seconds</strong> after the SO₃ gas is cut off.</p><p>The result is less charring, less off-spec generation, and about 100 kg of organic alcohol saved per continuous run.</p><p>That saving matters for a reason beyond the raw material itself. Every kgs of off-spec has to be dealt with. Either it goes back for rework, which costs time, energy and handling, or it is sold at a price well below what good product earns. Both are losses on top of the yield loss.</p><p><strong>What others can take from this</strong></p><p>If you run a sulphation plant, or any continuous process, here is what I would suggest:</p><ol><li><strong>Look for where off-spec material is generated.</strong> Startup and shutdown are the usual suspects, but every plant has its own.</li><li><strong>Capture operator skill instead of only relying on it.</strong> If your best operator handles a transition better than others, find out what they do and turn it into an SOP.</li></ol><p><strong>Final thought</strong></p><p>Yield improvement doesn’t always come from big projects. Sometimes it comes from noticing a small detail, like the first few minutes of a startup or the last 60 seconds of a shutdown, and deciding it deserves a proper standard</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=50f8cd750df7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Hidden Loss Behind Batch Plant’s Missing Production]]></title>
            <link>https://medium.com/@rushikeshkardule98900/the-hidden-loss-behind-batch-plants-missing-production-818fb3262345?source=rss-8650cd47a33a------2</link>
            <guid isPermaLink="false">https://medium.com/p/818fb3262345</guid>
            <category><![CDATA[training]]></category>
            <category><![CDATA[tpm]]></category>
            <category><![CDATA[process-improvement]]></category>
            <category><![CDATA[manufacturing]]></category>
            <category><![CDATA[cost-optimization]]></category>
            <dc:creator><![CDATA[Rushikesh Kardule]]></dc:creator>
            <pubDate>Sun, 04 Oct 2026 18:25:42 GMT</pubDate>
            <atom:updated>2026-10-04T18:25:42.335Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*udlrN2sqnFP5yZBsKvWrZg.png" /><figcaption><em>AI generated bank of batch reactors, wide shot</em></figcaption></figure><p>Picture ten reactors are running and two operators are looking after them. One operator is at the far end of the floor, taking a sample. A reactor behind him has just reached temperature and needs the next step. Also, a new batch is ready to start, and the other operator is busy with something that cannot wait either. Two people simply cannot be in ten places at once. So, one chooses, and whatever is not chosen waits.</p><p>At the end of the month, production is lower than it should have been. The reactors were healthy; the raw material was ready. The only thing missing was a free pair of hands at the right moment.</p><p>Total Productive Maintenance (TPM) has a name for this. It is called line organization loss.</p><p><strong>What is line organization loss?</strong></p><p>TPM groups plant losses into 16 major losses. Eight are equipment losses, five are manpower losses, and three are losses of resources like yield and energy. Line organization loss belongs to the manpower group.</p><p>It is the time a line waits because the people and resources it needs are not available when required. You can recognize it with three questions:</p><ul><li><strong>Is the equipment healthy?</strong> If something broke, it is a breakdown loss.</li><li><strong>Is the process ready for its next step?</strong> If the chemistry is still running, it is normal process time loss</li><li><strong>Is the step not happening because nobody is free to do it?</strong> Then it is line organization loss.</li></ul><p><strong>How it spreads</strong></p><p>When two reactors need attention together, the operator has to prioritize. The reactor that waits loses time, and that time is the loss.</p><p>The delay does not stay there. The equipment downstream of the waiting reactor also sits idle, so one small wait becomes several. It is worst when a new batch is due to start just as other activities fall due. The operator picks one, and something starts late.</p><p>A single wait looks harmless. But across ten reactors, three shifts and thirty days, it quietly takes a large part of the plant’s potential production.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*7jpZKL6MdV6c8druXBgERQ.png" /><figcaption><em>AI generated- an operator on a reactor platform</em></figcaption></figure><p><strong>Can we make it zero?</strong></p><p>No, and I think we should be honest about that. Removing every wait would mean automating every activity on every reactor. In batch manufacturing, that is rarely economical. The cost goes into the product price, and next to a competitor’s price that gap can hurt sales. I described the same trade-off in my earlier article on OEE.</p><p><strong>Start by listening</strong></p><p>The best place to start is not a spreadsheet but the shop floor. Rank the reactors by hours lost to line organization, then go and ask the operators what is slowing them down. They know. In my experience, the answer is almost always one of three things:</p><ol><li><strong>Too many activities at the same time.</strong> The workload is more than one person can handle. <em>Automation helps.</em></li><li><strong>Working conditions.</strong> Comfort, access or ergonomics make a task harder than it needs to be. <em>Better conditions help.</em></li><li><strong>Competency gaps.</strong> The person is unsure of something and nobody has had time to teach them. <em>Training helps.</em></li></ol><p>It is worth saying that the second and third are not failures of the person. They are signs that the system has not given the person what they need.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*6c-rQQayfkVjJN7k6Ig4Zw.png" /><figcaption><em>Simple flow diagram. Reactor ranked by lost hours → three root causes → three fixes</em></figcaption></figure><p><strong>Where automation makes sense</strong></p><p>Let me go deeper on the first cause, because that is where I have seen the biggest change.</p><p>If a reactor is the bottleneck reactor, automating its activities pays off directly. Every hour it gains is an hour the whole plant gains. If it is not the bottleneck, full automation is not a fruitful investment. It only needs to be automated enough that it never becomes the bottleneck.</p><p>I have seen this work. At one plant, we reached the highest production in its history just by reducing line organization losses.</p><p><strong>Recipe management: the final level</strong></p><p>The most complete form of this is called recipe management. Each process step is built as a sequence of its own. These sequences are then linked one after another with trigger points in between. A trigger point is a condition such as temperature reached, level achieved or a quality.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*m_CyG4306NSZo6qb7pocmw.png" /><figcaption><em>Recipe flow diagram. Phase 1 → trigger → Phase 2 → trigger → Phase 3 → trigger → completion</em></figcaption></figure><p>This is the highest level of automation, so it should be a deliberate choice. Safety, quality and productivity should all be considered, and it is justified only if it helps all three, not just for the sake of automating.</p><p>It also depends on standardization. A recipe can only be built on a process that is already well understood and consistent. Once it is in place, the batch no longer depends on who is on shift:</p><ul><li><strong>Quality</strong> stays in range, because every batch follows the same sequence.</li><li><strong>Safety</strong> improves, because interlocks are part of the sequence and there is less manual intervention.</li><li><strong>Productivity</strong> reaches the highest the process allows, because the reactor no longer waits for anyone.</li></ul><p>For the operator, the effect is just as good. The person who used to run between reactors, choosing which one to let down, can now watch the plant and use their judgment where it matters.</p><p><strong>The takeaway</strong></p><p>Line organization loss looks like a quiet wait, and it is easy to miss unless you have stood on the floor.</p><p>Find out which reactors lose the most hours and why, then match the fix to the cause: automation for overload, better conditions for comfort, training for knowledge gaps. Start with the bottleneck reactor. Done well, the plant produces more with the same reactors and the same people.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=818fb3262345" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Insulation Selection Impacts Heat Loss: A Waste Heat Recovery Case Study]]></title>
            <link>https://medium.com/@rushikeshkardule98900/how-insulation-selection-impacts-heat-loss-a-waste-heat-recovery-case-study-3a3666c04446?source=rss-8650cd47a33a------2</link>
            <guid isPermaLink="false">https://medium.com/p/3a3666c04446</guid>
            <category><![CDATA[chemical-engineering]]></category>
            <category><![CDATA[sulphation]]></category>
            <category><![CDATA[process-improvement]]></category>
            <category><![CDATA[project-management]]></category>
            <category><![CDATA[boiler]]></category>
            <dc:creator><![CDATA[Rushikesh Kardule]]></dc:creator>
            <pubDate>Tue, 29 Sep 2026 16:54:30 GMT</pubDate>
            <atom:updated>2026-09-29T16:54:30.631Z</atom:updated>
            <content:encoded><![CDATA[<p>Choosing the right type and thickness of insulation for high-temperature lines (below 500 °C), demonstrated on a hot air line feeding a waste heat recovery boiler</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*n_i1GNSbeB0GNVgWW59dTw.png" /><figcaption><em>AI generated image showing heat losses through insulation</em></figcaption></figure><p>What role does insulation play when temperatures are in the higher range, below 500 °C? How is insulation selected for a particular temperature and fluid? What happens if the proper type of insulation is applied? We will take one example to demonstrate how insulation selection affects energy losses and, in turn, the opportunity cost savings.</p><h3>The Process: Where the Hot Air Comes From</h3><p>Sulphur is burnt in a furnace at around 670 °C to give SO2. This SO2 is then passed to the catalytic converter, where it is converted into SO3. The conversion rate of SO2 into SO3 is at its maximum at 430 °C. The conversion is also exothermic in nature, meaning that the temperature of the gas mixture increases as SO3 is formed. To achieve higher conversion, we need to cool the gas mixture across four different beds. The cooling medium provided for this is atmospheric air, which absorbs the heat and cools the gas mixture in a shell and tube heat exchanger.</p><p>The temperature of the atmospheric air then rises to around 370 °C, and from here onwards we will call it hot air. At that temperature, insulation is obviously needed on those lines.</p><h3>Where the Hot Air Goes: The Waste Heat Recovery Boiler</h3><p>The hot air from the heat exchangers is passed on to the Waste Heat Recovery Boiler (WHRB). The hot air flows through the tube side of the boiler, and there is water on the shell side; at that temperature, the hot air converts the water into steam. The pressure of the steam produced can be adjusted, but we will consider a pressure of 7 kg/cm2 for this purpose, because the steam pressure requirement does not exceed that level.</p><h3>The Question: Two Insulation Options</h3><p>We are going to consider two examples: one with a thermal conductivity of 0.07 W/m·K and the other with 0.35 W/m·K. How is this going to impact the heat losses?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/516/1*eErQ9beWc_dD2-78tRh8Qg.png" /><figcaption>Heat loss equation through an insulation considering convection losses</figcaption></figure><p>Well, the equation says it all: the higher the thermal conductivity (k) of the insulation, the higher the losses through a pipe. It is not just the thermal conductivity that matters; the thickness of the insulation is equally important.</p><h3>Why Not Simply Add More Insulation?</h3><p>Why can’t we just increase the insulation thickness to such an extent that the heat loss disappears, so that we retain the maximum enthalpy? When we increase the insulation thickness, the heat loss will no doubt decrease, but the cost of insulation increases too, as shown in Figure 1. We need to look for an optimal point that balances heat loss against insulation cost.</p><h3>Practical Example: Heat Loss and Cost Across Insulation Thicknesses</h3><p>The example below demonstrates this relationship. It is a practical case in which waste heat from a heat exchanger is used to produce steam.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/901/1*bTPCMXqYy7OX9r-miJLZSg.png" /><figcaption><em>Heat loss and cost of insulation at different insulation thicknesses (insulation k = 0.07 W/m·K)</em></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/922/1*vMeTN0GM-d6KBgtU2GjTaw.png" /><figcaption><em>Heat loss and cost of insulation versus insulation thickness</em></figcaption></figure><h3>Conclusion</h3><p>Simply by selecting the right type of insulation and the proper thickness, we can produce steam from heat losses that were previously not considered. In a waste heat recovery system, this retained heat is particularly valuable because it remains available for steam generation rather than being lost to the surroundings. Thus, insulation can be viewed not only as a means of reducing heat loss, but also as a way of protecting the energy available for waste heat recovery.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3a3666c04446" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Do You Do with a Byproduct You’re Required to Make?]]></title>
            <link>https://medium.com/@rushikeshkardule98900/what-do-you-do-with-a-byproduct-youre-required-to-make-862fda2efc21?source=rss-8650cd47a33a------2</link>
            <guid isPermaLink="false">https://medium.com/p/862fda2efc21</guid>
            <category><![CDATA[process-improvement]]></category>
            <category><![CDATA[sustainability]]></category>
            <category><![CDATA[cost-optimization]]></category>
            <category><![CDATA[chemical-engineering]]></category>
            <category><![CDATA[manufacturing]]></category>
            <dc:creator><![CDATA[Rushikesh Kardule]]></dc:creator>
            <pubDate>Mon, 21 Sep 2026 17:29:51 GMT</pubDate>
            <atom:updated>2026-09-21T17:29:51.286Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Environmental norms make some byproducts unavoidable. Here is where the cost goes, and one practical way to bring it down.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/851/0*Pp2fzrG8JpN1dyKl" /><figcaption>Example of scrubbers in series</figcaption></figure><p>In chemical manufacturing, we are trained to think about byproducts in one way and that is to reduce them. Less byproduct usually means better yield, and better yield means less raw material wasted.</p><p>But there is a situation where that instinct doesn’t quite hold. What if the byproduct is something the plant has to produce, because environmental norms leave no other option?</p><p>Then the question changes. It is no longer “how do we get rid of this byproduct?” It becomes “can we manage it in a way that lowers the overall cost of manufacturing?”</p><p><strong>When a byproduct isn’t optional</strong></p><p>Government guidelines exist to keep the environment safe, and plants are expected to stay within them. Doing that takes equipment, technology and people to run it, and all of that has a cost. So where does that cost go? It gets loaded onto the product cost.</p><p><strong>An example from sulphation process</strong></p><p>Take a sulphation process. An alcohol is reacted with sulphur trioxide (SO₃), and the result is neutralised with sodium hydroxide. No matter how well we run it, the reaction doesn’t convert everything, and some unreacted SO₃ leaves the process.</p><p>We can’t release that gas into the atmosphere. SO₃ is hazardous, and in air it forms acidic mist that is harmful to people and the environment. So, it has to be treated before leaving the plant.</p><p>The usual answer is a scrubber. The off-gas, carrying SO₃, some unreacted sulphur dioxide (SO₂) and acid mist, is scrubbed with sodium hydroxide solution. What we get is a solution containing sodium sulphate, sodium sulphite and water. That is our byproduct.</p><p>Even the scrubber doesn’t remove 100%. The target is to bring what is left below the limit set by environmental norms, not down to zero.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YyoT6ROhbf3rD5IX400WDg.png" /><figcaption>An example from sulphation</figcaption></figure><p><strong>So, what do we do with it?</strong></p><p>Making the byproduct is only half of the story. The other half is deciding what happens to it next. There are two options.</p><p>Option 1: Give it to a third party. The byproduct goes out along with a processing charge per kg. It works, but it is one more cost added to the product.</p><p>Option 2: Blend it back into the product. This one needs a bit of explanation. The finished product already contains some sodium sulphate, because unreacted SO₃ gets neutralised along the way and ends up as sodium sulphate. How much depends on the conversion rate and the separation efficiency. So, a certain amount of sodium sulphate is already a normal part of the product.</p><p>If the customer’s specification allows it, some of the scrubber byproduct can be added as a formulation filler. The plant still pays to scrub it, but it no longer pays someone else to process it. The feasibility has to be checked first, meaning we need to find a way to blend it that stays within the quality specification limits set by customers, and doesn’t harm the mechanical integrity of the equipment.</p><p><strong>Something in between</strong></p><p>Byproduct handling is not only an environmental topic; it is a cost topic too. Whatever is spent on treating and disposing of it ends up in the product cost.</p><p>We often sort material streams into two boxes: product and waste. A byproduct like this one sit somewhere in between. Something with no use in one part of the process may still be useful in another, as long as we understand its composition and what it does to the final product.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EGt5MmNced_u16Vn67JlHQ.png" /><figcaption>AI generated image representation</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=862fda2efc21" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Beyond kWh per Tonne: What Specific Energy Consumption Really Tells You About a Continuous Plant]]></title>
            <link>https://medium.com/@rushikeshkardule98900/beyond-kwh-per-tonne-what-specific-energy-consumption-really-tells-you-about-a-continuous-plant-c3f3c0c3ab2f?source=rss-8650cd47a33a------2</link>
            <guid isPermaLink="false">https://medium.com/p/c3f3c0c3ab2f</guid>
            <category><![CDATA[energy]]></category>
            <category><![CDATA[energy-efficiency]]></category>
            <category><![CDATA[process-improvement]]></category>
            <category><![CDATA[operational-excellence]]></category>
            <category><![CDATA[chemical-engineering]]></category>
            <dc:creator><![CDATA[Rushikesh Kardule]]></dc:creator>
            <pubDate>Sat, 19 Sep 2026 02:43:00 GMT</pubDate>
            <atom:updated>2026-09-19T02:43:00.422Z</atom:updated>
            <content:encoded><![CDATA[<p>Specific energy consumption is one of the easiest numbers to put on a dashboard. Take the units consumed, divide by the tonnes produced, and you have a single figure that looks comparable across days or months.</p><p>But in a continuous plant, that single figure hides at least as much as it reveals. This piece is my attempt to unpack it, using a real scenario: a continuous sulphation plant that made 90 MT yesterday and drew 12,000 kWh doing it. That’s 133 kWh/MT. My question, and I mean it seriously — is that good?</p><p><strong>A Quick Refresher</strong></p><p>Specific energy consumption = energy consumed ÷ product manufactured.</p><p>12,000 ÷ 90 = 133 kWh/MT. Simple enough. What it buys you is comparability; it strips out the volume effect so a 90 MT day and a 60 MT day can sit next to each other on the same chart.</p><p>Except it doesn’t strip that effect out cleanly at all, and that’s one of the four things this number needs before it becomes a genuine KPI. The four are: a defined boundary, a bottom-up benchmark, a normalisation rule, and a breakdown by section. Take any one away and you’re reporting energy rather than managing it.</p><p>Notice one more thing before we go further. This ratio has two moving parts, energy on top, and production at the bottom. When it moves, either one could be responsible.</p><p><strong>The Boundary Problem: What’s Actually Inside That 12,000?</strong></p><p>Before you can say whether 133 is good, you have to say what it includes.</p><p>The boundary has to be a written document, not a mental model. What’s in, what’s out, how shared utilities are apportioned and on what basis. Just write the basis down and never change it silently. If you do change it, restate the history, or you’ll spend a quarter explaining a step change that was an accounting error rather than an engineering one.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/670/1*xdYrSrauZ_Zna_ECVR4G-Q.png" /><figcaption><em>Boundary example</em></figcaption></figure><p><strong>The Benchmark: Build It from Equipment, Not from History</strong></p><p>Once the boundary is fixed, you need something to compare against. And the benchmark has to come from the equipment list, not from last year’s average.</p><p>The reason is straightforward. Benchmark against history and you’ve enshrined last year’s inefficiency as this year’s target. You’ll hit it comfortably and learn nothing. A bottom-up benchmark tells you what the plant <em>should</em> draw if everything were healthy and loaded as designed and the gap between that and reality is usually wider than anybody expects.</p><p>One caveat that genuinely matters: the benchmark is throughput-specific. A number computed at 100% design rate is not a fair target when the plant is running at 70%.</p><p><strong>Why the Number Drifts Up</strong></p><p>Specific consumption creeps upward over time. Everywhere, always. And it creeps for reasons that are almost entirely mechanical.</p><ul><li><strong>Fouled heat transfer surfaces</strong>- scale in the exchanger, and deposits on the reactor jacket. The approach temperature widens, the chiller compensates by working harder, and the kW climbs while nothing looks wrong on the DCS.</li><li><strong>A choking blower</strong>- increase in pressure drop, impeller clearances opening, a belt slipping slightly. Same airflow delivered, more power drawn. Expensive here, because the air package is one of the largest single loads in a sulphation plant.</li><li><strong>Worn pump internals</strong>- clearances open, efficiency falls, more kW for the same duty.</li></ul><p>What all of this share is that none of them trip an alarm. None of them stop production. They just quietly cost you two, three, five kWh per tonne and they accumulate.</p><p><strong>Reading a Spike: Numerator or Denominator?</strong></p><p>So, the number jumps. 133 becomes 151. What happened? There are exactly two possibilities, and separating them is the first move.</p><p>The denominator moved, you simply made less product. This pushes kWh/MT up even with every machine in perfect health, because a chunk of your load is fixed. Lighting, AC, PDB, standby circulation, idling utilities: all of it exists whether you make 90 MT or 60 MT. Spread across less output, it shows up as higher intensity. A continuous plant running at 65% rate will look inefficient while being operated perfectly.</p><p>Or the numerator moved for same production, more energy. <em>That</em> is the one you investigate. Equipment health, a drifting operating parameter, a setpoint someone changed three weeks ago and forgot.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/908/1*-3byHheT2XvBenjSW4O7lA.png" /><figcaption><em>Example of comparison of MT vs kWh/MT</em></figcaption></figure><p><strong>Stoppages: What to Exclude, and Why</strong></p><p>A startup consumes a great deal of energy against almost no production. You’re heating sulphur, running the air blower, circulating material, and making nothing. A shutdown is the same in reverse; a changeover sits somewhere between. Include those days in your trend will not make any sense.</p><p>But don’t discard the stoppage days entirely instead track them separately, because the energy burned on startups is itself a saving theme, and often a large one. Six changeovers a month, each costing roughly half a day’s power, turns campaign planning into an energy project with a hard number attached to it. That’s a conversation production planning will actually engage with.</p><p>A few other things worth normalising for, especially in surfactants: product mix, because different actives carry genuinely different loads; ambient conditions, because the air package and chilling plant work measurably harder in a Mumbai May than a Mumbai January; and feedstock quality, which shows up in the reaction section.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/685/1*Bi7zCpiONAp38I1-SyxykQ.png" /><figcaption>Significant energy consumer in sulphation process- air blower</figcaption></figure><p><strong>The Takeaway</strong></p><p>So, 133 kWh/MT on its own doesn’t tell us much. But put that number in context with a benchmark of 124 kWh/MT, it points to a specific deviation, a specific area, and potentially, a specific action.</p><p>The real value of specific energy consumption lies not in making the number smaller, but in understanding why it changes. When you chase the underlying causes rather than the figure itself, significant energy consumption becomes what it was meant to be, a simple indicator of a much more complex story happening inside the plant.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/555/1*-Dnb89trOSPNfzYIO5KAGA.png" /><figcaption><em>PLC operators troubleshooting the process</em></figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c3f3c0c3ab2f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Beyond OEE: What Really Determines the Performance of a Batch Chemical Plant]]></title>
            <link>https://medium.com/@rushikeshkardule98900/beyond-oee-what-really-determines-the-performance-of-a-batch-chemical-plant-a76bec3faba6?source=rss-8650cd47a33a------2</link>
            <guid isPermaLink="false">https://medium.com/p/a76bec3faba6</guid>
            <category><![CDATA[performance]]></category>
            <category><![CDATA[batch-processing]]></category>
            <category><![CDATA[oee]]></category>
            <category><![CDATA[process]]></category>
            <category><![CDATA[process-improvement]]></category>
            <dc:creator><![CDATA[Rushikesh Kardule]]></dc:creator>
            <pubDate>Sun, 30 Aug 2026 01:51:33 GMT</pubDate>
            <atom:updated>2026-08-30T01:51:33.641Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*cFrYIbzxsqIX6g28_Xcwlw.png" /><figcaption><em>A modern batch chemical manufacturing facility where multiple reactors operate simultaneously</em></figcaption></figure><p>Overall Equipment Effectiveness (OEE) is one of the most quoted numbers on any plant manager’s dashboard. It’s clean, it’s a single percentage, and it invites comparison across shifts, lines, even sister plants. But in a batch chemical plant — where ten reactors might be running at once — that single number hides more than it reveals unless you know exactly what’s feeding it.</p><p>This piece is my attempt to unpack that, using a real scenario: a batch plant running 10 reactors simultaneously, where the plant is under constant pressure to trim labour cost.</p><p><strong>A Quick Refresher</strong></p><p>OEE = Availability Rate × Performance Rate × Quality Rate</p><p>On paper, it’s simple. In a batch environment, each of these three components is shaped by a different set of pressures — and each can be gamed, misread, or genuinely improved depending on <em>why</em> it moved.</p><p><strong>Availability Rate: The Cost of Planning Badly</strong></p><p>Availability is hit by planned activities (shutdowns, changeovers, maintenance windows) and by the dead time between two successive batches. The moment a planned activity overruns its scheduled window, availability — and with it, OEE — takes a direct hit.</p><p>The fix isn’t complicated in concept: activities need to be planned properly, using MSP or dedicated scheduling software, with manpower mapped against the job list in advance. I’ve seen this play out in industry — walking into a shutdown without a plan is like playing football without goalposts: you’re just running up and down a 120-metre rectangle, busy but scoreless.</p><p><strong>Performance Rate: Where the Real Story Lives</strong></p><p>Performance rate = actual batch cycle time ÷ standard batch cycle time (considering the batch size remains same throughout).</p><p>This is the ratio that tells you the truth about how the shop floor actually runs, because actual cycle time stretches for very specific reasons:</p><ul><li>Shift changeover losses</li><li>An operator prioritising another reactor’s activity over this one</li><li>Equipment breakdowns and unplanned stoppages mid-batch</li><li><strong>Line organisation losses</strong> — the accumulated drag of manual activities across a bank of reactors</li></ul><p>That last one is the crux of it. With 10 reactors running together, every manual activity that competes for the same one or two operators becomes a queue. The more manual steps in the sequence, the more often a batch sits waiting not because the chemistry is slow, but because a person is stretched thin.</p><p>The opportunity here is real: automating even a few of these activities frees manpower for the next task, directly lifting performance rate. That could mean adopting new technology that genuinely shortens process time versus the conventional route, or connecting two process phases so manual intervention (and the wait it creates) disappears altogether.</p><p><strong>Quality Rate: What “Right First Time” Really Costs</strong></p><p>Quality rate is actual first-pass output over theoretical output. If a process is standardized, a small deviation from it is usually what dents this number.</p><p>The clearest way to see this: getting the product right in one go <em>is</em> what quality rate measures. Take something like water recovery — one operator runs it for two hours, another runs the “same” step for one hour. Different durations, different outcomes, and over time, a measurable dent in quality rate. Standardizing the process — and automating it where possible — is what closes that gap.</p><p><strong>The Labour-Cost Trap</strong></p><p>Here’s where it gets political. Every employer wants to optimize labour cost — that’s not wrong. But optimizing it for the wrong reasons can quietly drag OEE down.</p><p>Cutting manpower without addressing the manual-activity load doesn’t remove the work; it just redistributes it onto fewer people, across ten reactors that don’t wait for anyone. That shows up as line organisation loss in the performance rate — a self-inflicted wound dressed up as a cost-saving.</p><p>The counter-argument is just as real, though: in batch manufacturing, you cannot automate everything. Full automation of every activity across 10 reactors is often not financially viable — the capital required can push up unit product cost enough that, next to a competitor’s price, the gap looks damaging in the market. So, the real skill isn’t “automate more” or “cut heads” — its choosing <em>which</em> manual activities are costing you the most in organisational loss, and automating those first.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/871/1*7D50KVozrbm07OdGtbjcYQ.png" /><figcaption><em>Sample OEE dashboard</em></figcaption></figure><p><strong>Does Improving OEE Actually Mean More Production?</strong></p><p>Mostly, yes — but only when the improvement comes from the right lever. Fixing availability through better shutdown planning genuinely reclaims run time. Fixing performance by removing organisational losses genuinely reclaims cycle time. Both translate to real additional batches.</p><p>Where it gets misleading is when OEE improves on paper without addressing root cause — for instance, if quality issues get compensated for elsewhere, or if data capture itself is inconsistent between shifts. In that case, OEE goes up, but the plant isn’t actually producing more or better.</p><p><strong>Does OEE Tell You the Truth About the Plant?</strong></p><p>Broadly, yes — OEE does point in the right direction. But it’s only as honest as the data behind it. If the losses feeding into Availability, Performance, and Quality rate aren’t all being tracked — and tracked with the <em>right reason</em> attached to each loss — the number becomes a proxy that management trusts more than it should.</p><p>In this 10-reactor scenario, the pattern is fairly clean: low quality rate and low availability rate usually have a traceable, logical explanation — a missed shutdown window, a process deviation. Performance rate is the trickier one — it’s really a referendum on how standardized (and how automated) the process is, filtered through organisational losses that can be optimized once you see them clearly.</p><p><strong>The Honest Caveat</strong></p><p>Is every factor affecting plant efficiency captured inside OEE? Not entirely. OEE was built around discrete, repeatable equipment cycles — it wasn’t designed with the full texture of manual, multi-reactor batch operations in mind. Things like operator fatigue across a long shift, the judgment calls a skilled operator makes between competing reactors, or the true cost-benefit of partial automation don’t show up as a clean OEE input — they show up as unexplained drift in performance rate that takes floor experience, not just the dashboard, to diagnose.</p><p><strong>The Takeaway</strong></p><p>OEE is a good compass for a batch chemical plant. It will tell you <em>that</em> something is off in availability, performance, or quality — but only floor-level understanding will tell you <em>why</em>, and whether the fix is a scheduling tool, a standardization exercise, or a genuinely justified investment in automation. Chase the number for its own sake, and you can improve OEE while quietly making the plant worse. Chase the <em>reasons</em> behind the number, and OEE becomes what it was meant to be: a very useful shorthand for a much more complicated truth.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/680/1*wHaMlF4zs-sm1vsyc5ZXWw.png" /><figcaption><em>AI generated: control room / batch process operator</em></figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a76bec3faba6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[More Quality Checks Don’t Always Mean Better Quality]]></title>
            <link>https://medium.com/@rushikeshkardule98900/more-quality-checks-dont-always-mean-better-quality-8e6e064000fc?source=rss-8650cd47a33a------2</link>
            <guid isPermaLink="false">https://medium.com/p/8e6e064000fc</guid>
            <category><![CDATA[kaizen]]></category>
            <category><![CDATA[quality-assurance]]></category>
            <category><![CDATA[sample]]></category>
            <category><![CDATA[batch-processing]]></category>
            <category><![CDATA[continuous-improvement]]></category>
            <dc:creator><![CDATA[Rushikesh Kardule]]></dc:creator>
            <pubDate>Mon, 24 Aug 2026 18:48:18 GMT</pubDate>
            <atom:updated>2026-08-24T18:48:18.675Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/494/1*fZwGR4yh85RXmBaO5v2t3A.png" /><figcaption>AI generated image- person performing sample analysis</figcaption></figure><p>In chemical manufacturing, “good product” is not a matter of opinion but it’s written down as a finished goods specification. Meet it, and you have a product you can dispatch. Miss it, and you have scrap. That single fact shapes almost everything else in a plant. It is the reason why quality checks exist at every stage — raw material, semi-finished goods, finished goods, and the material to be dispatched. And it’s why, if you’ve ever worked in chemical industry, you already know the tension I’m about to describe.</p><p><strong>The Real Cost of — Just to Be Safe</strong></p><p>Every sample that goes to the lab takes time. Depending on the test, that could be minutes — or it could be a day or two. Scan that across every stage of a batch, and you start to see it show up where it matters most- in total turnaround time for dispatch.</p><p>I’m not writing this to question whether quality checks are necessary. They are, without question, especially in an industry where the finished product’s quality can affect the end user, the applications it goes into, and the company’s credibility in the market. The real issue is different: are we checking things because it’s the optimal way to catch a problem, or because it’s just the way it’s always been done?</p><p>Compliance for the sake is a trap. It feels responsible, but it quietly eats into production capacity without anyone stopping to ask if there’s a better way to get the same or better assurance of quality.</p><p>I saw this play out directly. In one case, simply by optimizing the sampling method at a stage that had quietly become a bottleneck, we freed up enough time to run one full extra batch, every single day. Nothing about the quality standard changed. Nothing about safety was compromised. We just stopped doing the check in a way that was slower than it needed to be.</p><p><strong>Catch It Early, or Chase It Later</strong></p><p>One of the most underrated levers here is intermediate-stage sampling. Testing at that stage is not extra work — it’s often the fastest way to avoid wasted work downstream. Because here’s what happens when something does go out of spec: you don’t just reject a batch and move on. You go backward. Back to the previous stage, sometimes all the way to the raw material, hunting for where things went wrong. And you can’t stop at “we found where it happened” — you have to find <em>why</em> it happened, or you’ll be having this exact same conversation again next week, and the week after.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/593/1*9Owl0wm6gPxkQRCQupvgrg.png" /><figcaption>Person removing sample from a reactor</figcaption></figure><p>The root cause could be almost anything: a manual error somewhere in the process, a technical or equipment issue, an operating deviation, a problem inherited from the preceding material, or even the testing equipment itself giving a false read.</p><p>Meanwhile, out-of-spec material keeps piling up, with no easy option except blending it with in-spec batches — a process that eats time and manpower you didn’t budget for. This is exactly why root cause analysis isn’t a bureaucratic checkbox.</p><p><strong>What Continuous Plants Can Teach Us</strong></p><p>Everything above — the bottleneck, the extra batch we gained, the root cause chase — comes from a batch mindset, where each sample is a discrete decision point. Continuous plants face the same fundamental question, just at a different scale. When you’re generating quality readings continuously, alongside your process parameters, you get trends and not just individual results.</p><p>Looking at that data over time can reveal things a single sample never will: where the process itself could be tightened, where quality is drifting before it becomes a rejection, and — most relevant to everything above — where the sampling methodology itself could be optimized.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*RoCQpj2VvCY-ZMNXDol_Ng.png" /><figcaption>Example: Trend of sample results in continous plant</figcaption></figure><p><strong>Quality checks matter. Full stop.</strong></p><p>Quality checks matter. But necessary and optimal aren’t the same thing — and confusing them is how a plant ends up with inspection routines that protect compliance on paper while quietly costing a production opportunity.</p><p>The better question isn’t <em>should we test this</em> it should be “<em>is this the best way to test it, at the best point in the process, for the same level of assurance?”</em> Ask that enough times across a plant, and you’ll find capacity that was there all along, disguised as just following procedure.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8e6e064000fc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Safety Is a Profit Strategy — Not a Cost Centre]]></title>
            <link>https://medium.com/@rushikeshkardule98900/why-safety-is-a-profit-strategy-not-a-cost-centre-b5c850033cd3?source=rss-8650cd47a33a------2</link>
            <guid isPermaLink="false">https://medium.com/p/b5c850033cd3</guid>
            <category><![CDATA[industrial-safety]]></category>
            <category><![CDATA[process-safety]]></category>
            <category><![CDATA[safety]]></category>
            <category><![CDATA[process-safety-management]]></category>
            <category><![CDATA[environmental-safety]]></category>
            <dc:creator><![CDATA[Rushikesh Kardule]]></dc:creator>
            <pubDate>Tue, 18 Aug 2026 06:47:01 GMT</pubDate>
            <atom:updated>2026-08-18T06:47:01.773Z</atom:updated>
            <content:encoded><![CDATA[<h4>A hard lesson the chemical industry keeps re-learning the expensive way</h4><p>When we ask people to name what drives profitability in chemical manufacturing, and you’ll hear the list which includes — production capacity, raw material costs, market demand, operational efficiency. Safety usually never makes the cut.</p><p>I used to think of it that way too. Safety felt like the department that slowed things down — more paperwork, more inspections, another meeting before you could start up a line. It took me looking closely at what actually happens after a major incident to change my mind.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*5treEzo-dJbPAGWeHNFWoA.png" /><figcaption><em>Chemical industrial section and working personnel</em></figcaption></figure><h4><strong>The morning everything changes</strong></h4><p>Picture a company doing everything right. Targets are being met, demand is strong, investors are happy. Then, on what should have been an ordinary shift, something goes wrong in the process — and someone doesn’t make it home.</p><p>What happens next isn’t a rough patch. It’s a landslide. Emergency shutdown, sometimes for months, sometimes with no clear end date. Investigators on-site before anyone can even talk about restarting. Every day of downtime is money gone. Public trust — the kind that took a decade to earn — can evaporate in a single news cycle, and the stock often reacts before anyone even knows the full story. NGOs start asking pointed questions. Executives can end up facing legal consequences. And customers who need reliable supply quietly start calling other vendors, because they can’t afford to wait.</p><p>This isn’t a thought experiment. In December 1984, a leak of methyl isocyanate gas at the Union Carbide pesticide plant in Bhopal killed thousands — the result of a plant running with half its normal staff, deferred maintenance, and disabled safety systems. Within about a week, Union Carbide had lost roughly 27% of its market value — close to a billion dollars — and the company spent the next several decades in litigation and reputational fallout it never fully recovered from. Every process safety engineer still studies it, for good reason.</p><p>None of that shows up under “safety budget.” Eventually, though, it all shows up somewhere on the balance sheet.</p><h4><strong>Nothing about a major incident is really sudden</strong></h4><p>We like to talk about incidents as freak accidents — bad luck, wrong place, wrong time. In hazardous chemical operations, that’s rarely the honest explanation.</p><p>What actually tends to happen is a slow accumulation of small failures nobody flagged as urgent. A maintenance task pushed back a month. A shortcut that worked fine the last dozen times. An alarm everyone’s learned to tune out. This is basically the idea behind the “Swiss Cheese Model” — every layer of protection has gaps, and disaster strikes when the gaps in several layers happen to line up at once. Bhopal fits the pattern almost exactly: understaffing, skipped maintenance, and disabled alarms didn’t cause the leak on their own. They lined up.</p><p>Which is really the point — you can’t treat safety in this industry as a box to check once a quarter. It has to be something you’re watching constantly, because the chemistry itself doesn’t forgive being ignored.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*mUlew1fVm8G51G9UFNNrmA.png" /><figcaption>Team <em>briefing</em></figcaption></figure><h4><strong>It has to start at the top, not in the EHS office</strong></h4><p>Here’s a pattern I keep noticing: the companies with genuinely strong safety records aren’t necessarily the ones with the newest equipment. They’re the ones where safety isn’t parked in one department.</p><p>Hand safety entirely to an EHS team and it tends to stay a compliance exercise — something to satisfy before an audit. Have leadership actually own it, visibly, and it becomes part of how the whole place runs. A few things that separate companies who mean it from companies that just say it:</p><p><em>Bringing in outside expertise on purpose.</em> Firms like DuPont Sustainable Solutions, DEKRA, and Consultivo have spent decades specializing in process safety and risk assessment — and an outsider will often catch what people inside a plant have simply stopped noticing.</p><p><em>Auditing across your own sites.</em> If a company runs several plants, sending someone from one site to walk through another is a surprisingly underused move. It’s how a near-miss at Plant A actually becomes a lesson at Plant B, instead of staying buried in one site’s incident log.</p><p><em>Making it a value, not a memo.</em> Not a policy binder people sign once and forget. A habit every person on-site — staff or contractor — actually carries with them, because they understand why it matters, not just that it’s the rule.</p><h4><strong>What this actually costs — and saves</strong></h4><p>It’s worth putting real numbers next to this instead of just asserting it. OSHA puts the direct and indirect cost of a single workplace fatality at around $1.4 million once you count investigations, legal fees, fines, and the disruption to operations. The American Society of Safety Professionals has found that every dollar spent on safety tends to return somewhere between two and six dollars, mostly by avoiding injuries, lost work, and the admin overhead that follows them. The Centre for Chemical Process Safety has also found that companies with mature safety programs see productivity gains of up to 5%, largely because operations simply run more smoothly with fewer surprises.</p><p>However, you look at it, prevention is cheaper than the alternative — and that’s before you even factor in reputation, market share, or the part no spreadsheet can capture.</p><h4><strong>Where this leaves us</strong></h4><p>Production earns the profit for this quarter. Safety is what protects the profit for the next ten years. The companies that get this early on tend to be the ones still around — and still growing — a decade later. The ones that only learn it after something goes wrong rarely get their reputation back on the same timeline it took to build it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b5c850033cd3" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cost Improvement in Chemical Manufacturing: Why Small Formulation Changes Beat Big Budget Cuts]]></title>
            <link>https://medium.com/@rushikeshkardule98900/cost-improvement-in-chemical-manufacturing-why-small-formulation-changes-beat-big-budget-cuts-5f11561b9307?source=rss-8650cd47a33a------2</link>
            <guid isPermaLink="false">https://medium.com/p/5f11561b9307</guid>
            <category><![CDATA[cost-improvement]]></category>
            <category><![CDATA[chemical-engineering]]></category>
            <category><![CDATA[process-optimization]]></category>
            <category><![CDATA[continuous-improvement]]></category>
            <dc:creator><![CDATA[Rushikesh Kardule]]></dc:creator>
            <pubDate>Wed, 12 Aug 2026 15:33:00 GMT</pubDate>
            <atom:updated>2026-08-12T15:33:00.727Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/712/1*kdDK3ixCVked57jT0PQ0tQ.png" /><figcaption>AI-generated illustration of a chemical manufacturing facility</figcaption></figure><p>When someone mentions “cost improvement” in a chemical industry, the first thought that comes to mind is slashing manpower, eroding budgets or squeezing department cost centres. But real cost improvement is not about starving the shop floor. It’s all about maximizing value. In chemical manufacturing, most of your money is consumed by two colossi: raw materials and inventory. When unstable supply chains rage across the world, securing your existence means knowing how to deftly manage these two factors. You don’t need a revolutionary, risky process overhaul to improve your bottom line. The most significant financial advances can sometimes be made by implementing a minimal yet calculated alteration in your formulation recipe.</p><p><strong>Strategic Raw Material Substitution</strong></p><p>Suppose you manufacture Product C (your finished good) prepared with equal proportions of Material A and Material B. What if there’s a Material D an alternative raw material that replicates the chemical qualities of Material A, passes every stringent analytical test and costs less? As far as product quality, you can get away with not going for a total replacement; it might jeopardize quality standards or cause regulatory issues. Instead, explore Fractional Blending replacing mere 10% of Material A with Material D.</p><p>The quality of finished goods continues to be optimal. The consumer receives the same within specifications product. But innovatively, the cost equation shines like a diamond in the rough.</p><p><strong>The Example</strong></p><p>Consider a formulation batch at pilot scale that produces 20 kg of Product C using equal quantities of Material A and Material B.</p><p>In the original formulation, the recipe uses 10 kg of Material A priced at ₹157/kg, costing ₹1570, and 10 kg of Material B priced at ₹165/kg, costing ₹1650. This brings the total raw material cost for a 20 kg batch to ₹3220.</p><p>Now consider a small optimization. Instead of completely replacing Material A, only 10% of it is substituted with Material D, which costs ₹148/kg and has been validated to deliver the same performance. The revised formulation uses 9 kg of Material A costing ₹1413, 1 kg of Material D costing ₹148, and 10 kg of Material B costing ₹1650. The total raw material cost for the same 20 kg batch becomes ₹3211.</p><p><strong>Large Scale Impact</strong></p><p>At first glance, saving ₹9 on a 20 kg batch may seem insignificant — just ₹0.45 per kilogram. In a laboratory, that’s easy to overlook. But chemical manufacturing isn’t driven by laboratory batches; it’s driven by continuous, high-volume production. Consider a plant producing 20 tonnes per batch and operating 30 batches per month. The numbers quickly tell a different story:</p><p>Monthly production: 600000 kg of finished product</p><p>Savings per kilogram: ₹0.45</p><p>Monthly cost savings: ₹2.7 lakh</p><p>Annual cost savings: ₹32.4 lakh</p><p><strong>The Golden Rule of Formulation</strong></p><p>Implementing this method entirely depends on keeping that molecule intact. The only way of reducing cost that makes a real difference has only one main guideline. <em>One has to make sure that the change does not affect the final user in any way at all</em>. The product that ends up in the customer’s hands must be of the same as it was earlier; have performance specifications that are identical to the previous ones.</p><p>That’s essentially what we mean by better cost improvement: we are not changing the product; we are changing the process that leads to it; we are cutting the man power cost. It’s a very subtle level of process innovation when chemical manufacturers focus on small raw material substitutions to unlock a lot of hidden profits, and at the same time, they remain true to their product promise made to the market.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5f11561b9307" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>