Electricity visibility has advanced steadily in Thai factories over the past ten years. Water usage monitoring, by contrast, is often stuck at reading the receiving tank inlet meter once a month. Yet in 2026 the industrial water balance in the EEC has visibly tightened, and industrial estates are beginning to ask tenants to save and reuse water. This article explains why measuring water is technically harder than measuring power, and how to design measurement points in practice.
Why water usage monitoring suddenly matters in Thailand in 2026
“Water is cheap, so there is no need to measure it.” We have heard this many times from managers running factories in Thailand. Compared with the electricity bill, the water invoice really is small, and it almost never reaches the management meeting agenda. In the EEC, that assumption is now breaking down rapidly.
Demand exceeds supply across 41 industrial estates and around 6,000 factories
Thailand’s Eastern Economic Corridor (EEC) covers the three provinces of Chonburi, Rayong and Chachoengsao, and is the country’s densest manufacturing cluster. According to press reports, these three provinces host 41 industrial estates with around 6,000 factories, and industrial water demand has reached 1.5 million cubic metres per day. Against that, the volume the Industrial Estate Authority of Thailand (IEAT) is preparing stands at 1.18 million cubic metres, leaving a shortfall relative to demand.
Simple subtraction gives a gap of 320,000 cubic metres per day. That does not mean supply cuts are a daily occurrence. It describes a structural situation in which the supply plan has not kept pace with growth on the demand side. From the estate operator’s point of view, new tenant applications now have to be screened for whether water can actually be allocated, and there is a clear incentive to ask existing tenants to hold their consumption down.
What matters from the factory’s side is that this shortfall is not “a problem someone will solve eventually”. It arrives at your gate in the form of a request from the estate. And when that request arrives, the first questions are almost always “how much do you use per month?” and “what do you use it for, and how much for each?”. The first can be answered from the invoice. The second cannot be answered without measurement.
Data centres and AI investment are pushing water demand up
Growth on the demand side is being led by data centre and cloud investment. Reports indicate that for the EEC alone there have been 47 data centre and cloud service investment applications, worth a combined 173 billion baht (about USD 5 billion).
Data centres use water for cooling. The larger the facility, the heavier the cooling load, and in Thailand’s high ambient temperatures the share of water-cooled systems rises accordingly. As of April 2026, reporting has also pointed out that the expansion of AI infrastructure poses a threat to water-stressed regions of Thailand.
The point manufacturers need to absorb here is a simple fact of physical infrastructure. Data centres and factories share the same water sources and the same pipelines. When large new consumers connect to the same network, the headroom on the supply side thins out. Even if your own consumption never changes, rising demand around you makes estate-level conservation requests and tariff reviews a realistic prospect.
The real issue is cost and use restrictions, not fear of a dry tap
That said, there is no reason to overstate the crisis. The Metropolitan Waterworks Authority (MWA) has indicated a stable supply outlook through 2026 and has stated that no shutdowns are expected. MWA’s service area is the Bangkok metropolitan region and is separate from the three EEC provinces, so this is worth sharing as a temperature check rather than as proof that the EEC will not be constrained. At the very least, this is not a story about the tap running dry tomorrow.
What a factory should prepare for is these two things.
- Being able to explain where in your own operation a future rise in industrial water cost would bite
- Being able to respond with numbers when the industrial estate asks you to cut consumption or report a reuse rate
Both rest on the same underlying question. Where in your process is your water used, and how much of it? The mechanism that answers this question is water usage monitoring. And awkwardly, that mechanism cannot be built the same way electricity visibility was built.
The decisive difference between electricity monitoring and water usage monitoring is how easily you can retrofit
When people hear “factory energy monitoring”, most picture installing power meters. Electricity visibility genuinely does have a low barrier to entry, and it has spread widely among Japanese-affiliated plants in Thailand over the past decade. Try to extend the same thinking to water, however, and the project usually stalls at the first quotation. The reasons are technical, and those same reasons explain the structural pattern of “electricity moved ahead while water was left behind”.
Electricity is measured with a CT clamp, without cutting any wiring
The current transformer (CT) used for electrical measurement detects current simply by clamping a ring-shaped sensor around the conductor. In principle there is no need to cut wiring and no need to switch off the supply. With a split-core clamp CT, you can retrofit onto an existing cable while it remains live.
How large an advantage this is becomes obvious once you think about work planning. No production stoppage, no weekend shift, no shutdown work, and a new measurement point for one circuit appears within tens of minutes of opening the distribution board door. It is precisely this ease that has made phased rollouts such as “start with the ten main circuits” workable. This is still live work inside an energised panel, so a qualified electrician and proper protective equipment are prerequisites. Even so, the fact that production does not have to stop is the decisive difference from water. For how measurement points are chosen on the electrical side and how the costs are structured, see Energy Monitoring System Cost and Rollout for Thai Factories, which covers the topic in detail.
Water usually means cutting the pipe and inserting a flow meter
Water measurement works differently. The electromagnetic flow meters and paddle wheel (impeller) flow meters widely used in industrial applications are both inline types, with the measuring element sitting inside the pipe. Installing one means cutting the target pipe, assembling flanges or fittings, inserting the meter, reinstating the line and running a leak test.
In other words, every additional point triggers the following work.
- Shutting off water to the relevant system, which means stopping supply to production equipment
- Cutting and machining the pipe, and adding flanges or fittings
- Pressure and leak verification after reinstatement, plus air purging of the system
- Curing time for welding or solvent bonding, depending on the pipe material
Where a CT clamp on the electrical side takes tens of minutes, water takes half a day to a full day, and for some systems weekend work is a precondition. The installation cost per point naturally differs greatly as well. This is exactly why “let us start with ten points” is realistic for electricity, while for water the count gets trimmed back during the study phase.
Ultrasonic clamp-on meters are not a universal answer
Many readers will be thinking that surely water has an externally mounted sensor too. It does. Ultrasonic clamp-on flow meters measure flow with sensors fitted to the outside of the pipe, and in the sense that no pipe is cut, this is the closest equivalent to the electrical CT clamp.
However, this method comes with clearly defined application conditions. The points that most often trip projects up in practice are these.
- Pipe material and wall thickness. Materials that transmit ultrasound poorly, pipes with internal lining, and mortar-lined cast iron pipe all make measurement difficult
- Internal scale and rust tubercles. In pipes that have been in service for years, internal deposits disturb ultrasonic propagation and the reading will not settle
- Straight run. A defined length of straight pipe is required upstream and downstream of the sensor, so you cannot install immediately after an elbow or a valve
- Full pipe condition. Measurement is impossible unless the pipe is running full, which rules out drop pipes and gravity-flow drainage lines
- Minimum flow velocity. Flow that is too slow cannot be detected, so a large pipe carrying a small volume cannot be measured
- Air bubbles and outdoor exposure. Entrained air, direct sunlight and rain exposure all degrade sensor stability over the long term
So ultrasonic clamp-on is an option that may let you avoid cutting the pipe. It is not a universal solution you can install anywhere. You need to confirm the condition of the existing pipework on site and judge feasibility point by point. The mere fact that this site survey step exists is itself a major difference from electricity.
Comparing electricity and water measurement side by side
Putting the differences discussed so far into one view gives the following.
| Aspect | Electrical measurement | Water measurement |
|---|---|---|
| Typical sensor | Split-core clamp CT | Electromagnetic, paddle wheel, ultrasonic clamp-on |
| Retrofit to existing plant | Clamp on while live | Pipe cutting required in most cases |
| Production stoppage needed | Generally not required | System shutdown often required |
| Installation time per point | Tens of minutes | Half a day to a day, weekend work for some systems |
| Advance site survey | Understanding the circuit layout is enough | Pipe material, wall thickness, straight run and internal condition all need checking |
| Adding points later | Easy, the count scales readily | Every additional point means another construction job |
| Conditions that block measurement | Limited. Constrained by cable diameter, panel space, very small currents | Partly filled pipe, low velocity, special pipe materials |
Read that table and it becomes clear why so many factories have visibility on electricity and nothing on water. It was not that engineers dismissed water. The retrofit difficulty was simply on another level.
Crucially, that difference in difficulty does not lead to the conclusion “so give up on water”. It leads to a different conclusion: because you cannot casually add points later, the quality of the initial measurement point design matters even more than it does for electricity. That is the core message of this article.
Where water disappears in a factory, and why you cannot see it
Before getting into measurement point design, it is worth mapping where factory water actually goes. Order flow meters while this remains vague and you end up having measured something that tells you nothing.
The only visible point is the receiving tank inlet
In most factories, exactly one water measurement point exists in practice. The billing meter on the incoming line from the estate or the waterworks authority, or the meter at the receiving tank inlet. At month end somebody reads that number, the maintenance team records the difference against last month, and that is the whole system.
What that single point tells you is one thing only, which is how many cubic metres the whole factory consumed this month. If the figure is up 10 percent on the previous month, you can never read from that number alone whether the increase came from higher production volume, from somebody changing the cooling tower blowdown setting, or from a leak somewhere.
Against the per-panel breakdown you have for electricity, the asymmetry is stark
In a factory that has already deployed electrical monitoring, the picture is completely different. On top of the total at the incoming point, you have breakdowns by power and lighting circuits, by major production line, by compressor, by air conditioning. That is what allows statements like “half of last month’s electricity increase was the compressor”.
For water, that entire breakdown layer is missing. You have the total and nothing beneath it. The decomposition that is routine for electricity is simply never performed for water. That is the reality in a great many Thai factories.
Where factory water actually goes
To build the breakdown, start by listing the candidates. The weighting varies by industry, but a general manufacturing plant splits roughly as follows.
| Use | Typical equipment | How the water is consumed |
|---|---|---|
| Cooling tower make-up | Cooling towers, chillers | Leaves the system through evaporation, drift and blowdown |
| Washing and rinsing | Parts washers, CIP, mould cleaning | Leaves the system as effluent |
| Boiler feedwater | Steam boilers, hot water plant | Dissipates as steam, discharged via blowdown |
| Incorporation into product | Mixing, plating, pre-treatment for painting | Absorbed into product or work in progress |
| Domestic use | Canteen, toilets, showers | Fixed consumption unrelated to production volume |
| Leakage and unknown | Buried pipe, aged fittings, taps left open | Consumption with nobody using it |
The row to focus on in that table is the last one. Leakage and unidentified losses are never found without measurement points. In a factory that only watches the total, leakage dissolves into a vague sense that “consumption seems high this month”.
Design so that “total minus measured equals unmeasured” always remains visible
If there is one principle to state about measurement point design, it is this. Always keep the remainder after subtracting individual measured values from the total as an explicit “unmeasured” line in your ledger.
Fit a handful of flow meters by process and their sum will never match the total exactly. That mismatch is normal in itself. The question is whether you leave the difference unattended or manage it as a traceable number. Plot unmeasured volume on a monthly chart and, when the unmeasured portion alone jumps in a given month, you have a trigger to suspect a leak or a changed setting. Without that column, the increase hides inside one of the process figures and disappears.
The thinking here is identical to the fundamentals of energy management. The energy review in ISO 50001 asks you to evaluate current energy use and identify significant energy uses, and organising how much of that use is actually captured by measurement makes the work considerably easier. See ISO50001 Energy Management 2026 – Payback Starts With Power Visibility for a fuller treatment, which is useful reading if you are considering certification.
Measurement point design for water usage monitoring, three tiers expanded gradually
Now to the main subject. Since you cannot casually add points the way you can with electricity, the design quality of where you install first determines your return on investment outright. We recommend thinking about water measurement points in three distinct tiers.
Tier 1, intake, the total and the entry point of each source
The first tier is the water entering the factory itself. Even where a billing meter already exists, there is value in digitising this point. Simply moving a figure you can only read monthly to something you can read hourly or daily changes what you see.
For example, if consumption on a Sunday with production stopped does not fall to zero, that residue comes from one of several causes, namely leakage, domestic use, equipment somebody forgot to shut off, cooling towers and chillers running on standby, or make-up through the receiving tank ball tap. Viewing the total meter on a time axis alone gets you to the starting point for separating them. In factories that also draw on well water or recycled water, measure each source separately.
Tier 1 amounts to roughly one to three measurement points, and in most cases adding a pulse output to the existing meter, or installing one flow meter just after the incoming connection, is sufficient. The construction burden is relatively small, which makes this the tier to start with.
Tier 2, process level, the branches feeding heavy water users
The next tier is where the real breakdown is created, and it is also the hardest tier to design. The governing principle is “install at the root of each branch, in descending order of consumption”.
In practice, the priority order in most factories is as follows.
- Cooling tower make-up line. It flows continuously and is definitively consumed through evaporation and blowdown, so the cost impact is large
- Boiler feedwater line. In plants that use steam, this ranks alongside the cooling tower as a major consumer
- Supply line to washing processes. Because these run in batches, there is substantial scope to revise the operating method
- Domestic water system. It does not track production volume, so separating it makes the specific consumption calculation far more accurate
The key here is to think in terms of branch roots, not processes. Chase a perfect process-by-process breakdown and your measurement points swell to 20 or 30, at which point the installation cost alone freezes the project. Check drawings and walk the site to establish where the existing pipework branches, and look for locations where a small number of points can separate large blocks of consumption. For systems where the branching has become too finely divided, measure them together at a single upstream point and defer the breakdown to a later phase.
Keeping Tier 2 to somewhere between three and eight points is realistic. If, once this tier is designed, you can account for most of the factory’s water, that is enough. As a rule of thumb from our own project experience, once the remainder that the breakdown cannot explain has fallen to around 20 percent of the total, you are ready to move to the next stage.
Tier 3, discharge, effluent outlets and recoverable volume
The third tier is the discharge side. Many factories measure only intake, but without measuring discharge you cannot evaluate reuse or recycle out of the 3R set.
There are three reasons to measure discharge. First, the difference between intake and discharge volumes lets you estimate how much water never comes back, through evaporation or incorporation into product. Second, when you evaluate reuse, you cannot size recovery equipment unless you know which outlet discharges how much. Third, there are situations where records of discharge volume are required for reporting to the industrial estate or for environmental permitting.
The discharge side, however, has plenty of open channels and gravity-flow pipes that do not run full, and ultrasonic clamp-on frequently cannot be used. For an open channel you install a primary device such as a weir or a Parshall flume, measure the water level above it with an ultrasonic level meter, and convert that level to flow through the level-discharge relationship. For a pressurised full-bore line, an electromagnetic flow meter. The method has to be selected point by point. Keep in mind that what an ultrasonic level meter measures is level, not flow.
Bringing the three tiers together
Setting the three tiers out in a table makes the relationship between priority and construction burden explicit.
| Tier | What is measured | Indicative point count | Main purpose | Construction burden |
|---|---|---|---|---|
| Tier 1 intake | Incoming total, entry point per source | One to three points | Hourly view of the total, holiday leak detection | Light |
| Tier 2 process | Cooling tower, boiler, washing, domestic | Three to eight points | Building the breakdown, calculating specific consumption | Heavy |
| Tier 3 discharge | Main effluent outlets, reuse candidate lines | One to four points | Evaluating reuse volume, meeting reporting duties | Moderate |
As the table suggests, there is no need to start construction on all three tiers at once. Installing Tier 1, running it for a few months, and then deciding Tier 2 positions based on the imbalances that emerge is the approach that wastes the fewest measurement points.
Select the flow meter technology point by point
Once the measurement points are fixed, the next decision is the measuring technology. Committing to “everything will use this method” guarantees a bad fit somewhere.
| Method | Pipe cutting | Where it works well | Cautions |
|---|---|---|---|
| Electromagnetic | Required | Conductive water, main lines that need accuracy | Unusable on pure water or oil, needs straight run and proper grounding |
| Ultrasonic clamp-on | Not required | Existing lines that cannot be shut down, interim measurement | Pipe material, internal condition, straight run and full pipe are conditions |
| Paddle wheel (impeller) | Required | Small bore branches, points where cost must be contained | Vulnerable to debris, moving parts need maintenance, straight run and minimum velocity limits apply |
| Mechanical (positive displacement and similar) | Required | Replacing an existing billing meter | Vulnerable to debris so a strainer is needed, causes pressure loss, confirm whether a pulse output is available |
| Weir or flume plus ultrasonic level meter | Not required | Discharge volume in open channels | Requires a primary device, never for full-bore pressurised pipe |
When choosing a method, prioritise “can this point be read reliably for years” over headline accuracy. Put a method that requires a water shutdown for annual calibration onto a line that is critical to production, and within a few years it becomes a measurement point nobody touches.
Feed flow meter monitoring data into your existing utility monitoring
Flow meter outputs are almost always pulse, 4-20 mA or Modbus. In a factory that already has an electrical monitoring system, putting water onto the same collection platform is both the cheapest route and the one most likely to survive in daily operation.
How to choose between wireless and wired when adding new sensors to an existing network, and how to select sensors starting from the sampling period rather than from headline accuracy, are covered in Factory IoT Sensor Types 2026 – Sampling Rate Decides, Not Accuracy. Water measurement does not require as short a sampling period as electricity does, which changes the range of options open to you.
Being able to view electricity and water on the same screen produces unexpected findings. Overlay the hours when cooling tower make-up rises with the hours when chiller power is high, and you can separate a genuinely high thermal load from water escaping inside the system. Seeing causation that neither dataset revealed on its own is the practical dividend of integration.
Assume that industrial water in Thailand is not as cheap as you think
Before moving to the investment case, it is worth calibrating the cost of water. Carrying a home-country assumption that “water is cheap” into Thailand leads to poor decisions.
As one reported example of industrial water pricing inside an industrial estate, a level of 24.75 baht per cubic metre has been cited. The household base tariff of the Provincial Waterworks Authority (PWA), meanwhile, starts from 10.2 baht per cubic metre.
This needs careful reading. Prices inside an industrial estate and household water tariffs are different systems with different supplying entities, so it is not appropriate to conclude simply that “industrial water costs more than double the household rate”. Prices also vary by contract terms and by estate. The figures above are reported price examples in the public domain, and should not be treated as confirmed current prices applying to every estate today.
Even so, there is something these two numbers do support. Multiplied by the volume a factory actually consumes, water inside an industrial estate adds up to an annual sum that cannot be ignored. The premise that “water is cheap, so it is not worth measuring” no longer holds up well, at least in Thai industrial estates.
Installation cost and payback, a model calculation
The figures from here on are a model based on assumptions we have set in order to demonstrate the shape of the calculation. They are neither an actual quotation nor the recorded performance of a real factory. Please read them separately from the researched facts in the preceding sections.
Model assumptions
| Item | Model value | Note |
|---|---|---|
| Factory consumption | 1,000 cubic metres per day | Assumes a mid-sized plant with meaningful water use |
| Operating days per year | 300 days | 300,000 cubic metres per year |
| Unit price of water | 24.75 baht per cubic metre | The reported price example above used as a model input |
| Annual water cost | About 7.43 million baht | 300,000 cubic metres multiplied by 24.75 baht |
On these assumptions, suppose the design comprises three points in Tier 1, six in Tier 2 and two in Tier 3, for a total of 11 measurement points. Estimating flow meter hardware plus installation at a range of 80,000 to 200,000 baht per point puts the initial investment broadly in the range of 880,000 to 2.2 million baht. Data collection hardware and software costs sit on top of that.
How to estimate the saving
Effects expressed as a single saving percentage will almost always miss. Break the effect down and build it up as follows.
| Type of effect | How it arises | Model estimate |
|---|---|---|
| Early leak detection | Detected through a rising holiday baseline | 1 to 3 percent of annual consumption |
| Cooling tower operation improvement | Correcting blowdown settings and cycles of concentration | 2 to 5 percent of annual consumption |
| Revision of washing processes | Optimising rinse time, flow rate and frequency | 1 to 4 percent of annual consumption |
| Fewer outlets left running | Making consumption visible during non-operating hours | Under 1 percent of annual consumption |
Adding the four items together straightforwardly gives 4 to 13 percent, but it is rare for every effect to reach its maximum at the same time. Setting the total saving in this model at 5 to 12 percent gives an annual saving of about 370,000 to 890,000 baht. Against an initial investment of 880,000 to 2.2 million baht, simple payback falls in a range of roughly 1 to 6 years. Note that this payback period puts only the flow meter hardware and installation work in the denominator. Include the data collection hardware and software and the payback stretches out further.
The width of that range may feel unsatisfying, but it is close to reality. Payback on water investment depends entirely on how much waste currently exists. In a factory where cooling tower management is already disciplined and there are no leaks, payback stretches out. Conversely, in a factory where holiday consumption runs at 30 percent of a weekday, Tier 1 alone can pay for the whole project.
Effects excluded from the model
The calculation above leaves out three effects that the model does not put a monetary value on. As decision criteria, they are sometimes the larger consideration.
- Lower effluent treatment cost. Less intake means less discharge, which reduces treatment chemical and sludge disposal costs
- Capacity to respond to industrial estate requests. Being able to present consumption and reuse rates as numbers becomes decisive in situations such as expansion applications
- Alignment with production planning. If water use restrictions actually arrive, you can decide in advance which processes to throttle first
IEAT’s 3R push and the supply infrastructure moves inside the estates
So far the discussion has been about the factory. There are also moves on the supply side. Knowing this context means you will not be caught off guard when a request arrives from your estate.
IEAT is driving 3R inside the estates
IEAT is advancing measures based on the 3R approach of reduce, reuse and recycle within industrial estates. In other words, the emphasis is not only on increasing the volume supplied but also on circulating water within the estate.
The implication for factories is clear. Promoting 3R will eventually show up as specific requests to tenant companies. What will be asked for at that moment is not a statement that “we are making efforts to conserve water” but numbers showing where, how much and by what means you reduced consumption. This is another reason to measure discharge as well as intake.
Storage and source connection show the supply side is acting too
A concrete example on the supply side is Amata U, which handles the water business in the Amata group. The company holds over 30 million cubic metres across 17 reservoirs, and Amata U CEO Chawalit Tippawanich has stated that this would allow supply to factories to be maintained for two years even if a drought hit the east of the country.
Amata U has also signed an MOU with Eastern Water Resources Development and Management to connect water sources by pipeline and to study more efficient supply. The study is expected to complete in about one year.
How should this be read? Not as evidence that the supply side has it covered and factories need do nothing. Investment in reservoirs and pipelines ultimately feeds through into the cost of supplying water. And the very premise that supply can be sustained for two years during a drought is conditional on consumption in normal times staying within the assumed range. The estate’s incentive to ask tenants to restrain consumption gets stronger, not weaker, as supply side investment advances.
How to roll it out, four steps
Once the design philosophy for measurement points is settled, the remaining question is execution order. Because water involves heavy construction, getting the order wrong leaves you with rework you cannot undo cheaply.
Step 1, trace the pipework on drawings and on site
The first task is not sensor selection but understanding the pipework. Even where as-built drawings exist, factories that have gone through repeated extensions almost never match them. Walk the route from the receiving tank to each item of equipment, confirming where the branches are, which sections are buried, and what the bore and pipe material are, and update the drawings as you go.
Skip this step and you get rework: the flow meter you ordered does not match the site bore, the straight run cannot be secured, or a section turns out to be buried and needs excavation. The single most cost-effective activity in a water monitoring project is, in fact, this survey.
Step 2, install Tier 1 only and run it for a few months
Do not order every point at once. First get the total into a state where it can be read hourly, and run it for two or three months. That period alone yields three pieces of critical information, namely holiday consumption, the overnight baseline, and correlation against production volume.
With that data in hand, you can choose Tier 2 measurement points on the basis of “there is a system here that keeps flowing on holidays” rather than “this is probably where the big consumption is”.
Step 3, install Tier 2 in phases by importance
For Tier 2, we recommend splitting installation into two or three rounds rather than doing everything at once. Plan the shutdown work to coincide with annual maintenance periods or long holidays and the impact on production stays contained.
Review the data from the points installed in the first round, then fine-tune the positions for the second. Being able to repeat that loop is the advantage of phased deployment.
Step 4, define operating rules and thresholds
Adding measurement points reduces nothing on its own. Decide who looks at what, when, and what triggers action. At minimum, settling these three items keeps the practice alive.
- The baseline value for non-operating days, and the verification procedure when it is exceeded
- A daily ceiling for cooling tower make-up volume, and the inspection items when it is exceeded
- Monthly specific consumption, meaning water per unit of product, and who receives the month-on-month report
Common failure patterns
Here are the patterns we have seen where water visibility projects fail to deliver.
Digitising the total and stopping there
This is the most common shape of failure. The moment a pulse output is fitted to the existing meter and the value appears on a monitoring screen, the project is treated as complete. Being able to see the total on an hourly basis is genuine progress, but with no breakdown it does not translate into improvement actions. Share the framing at the outset that Tier 1 exists primarily as the design input for Tier 2.
Splitting by process so finely that installation cost explodes
The opposite pattern. Chasing an ideal breakdown pushes the count to 25 points, and the plan is frozen the moment the quotation lands. Unlike electricity, with water the point count feeds straight into construction cost. Separate the large blocks first, and subdivide only after the effect has been confirmed.
Planning around ultrasonic clamp-on that turns out to be unusable on site
Looking only at the bore on the drawing, somebody plans “clamp-on will be fine here”, and then on site the pipe turns out to be mortar-lined cast iron, or heavily scaled internally, or the only available position is immediately downstream of an elbow. This rework happens frequently. For any point where clamp-on is assumed, always carry out a trial measurement on site before finalising the design.
Evaluating reuse without ever measuring the discharge side
A 3R request arrives, evaluation of reuse equipment begins, and then nobody knows which outlet discharges how much, so the recovery equipment cannot be sized. Adding discharge measurement in a hurry at that stage stops the equipment study itself for several months. In factories where reuse is a realistic prospect, plan Tier 3 at the same time as Tier 2.
Nobody is assigned to look at the numbers
The measurement points multiply but nobody reviews them monthly, and six months later there is no one left who opens the charts. This happens across energy monitoring generally, not only with water, but water tends to be neglected more easily because it changes more slowly than electricity. Do not treat the operating rules in Step 4 as an afterthought.
Frequently asked questions
How do you monitor water usage in a factory?
Start by making a total measurement point, such as the receiving tank inlet, readable on an hourly basis, then add flow meters at the branches serving heavy consumers such as cooling tower make-up, boiler feedwater and washing processes. Rather than building a fine process-level breakdown first, watching the total for a few months and then deciding where the breakdown points go wastes fewer measurement points.
How much does it cost to install a flow meter?
Cost varies greatly with pipe bore, material, whether the line is buried, and whether a shutdown is feasible, so quoting a single market rate would be misleading. The model calculation in this article assumes a range of 80,000 to 200,000 baht per point, but that is a working assumption used to show the shape of the calculation. In reality you need a site survey and a point-by-point quotation.
Is there a way to measure without cutting the pipe?
Ultrasonic clamp-on meters measure from outside the pipe. They do, however, require that several conditions be satisfied, namely suitable pipe material and wall thickness, limited internal scale, adequate straight run upstream and downstream, a full pipe, and flow velocity above the minimum threshold. On existing pipework it is not unusual for a point to fail these conditions, so an advance site check is essential.
Can water be integrated into an existing energy monitoring system?
In most cases, yes. Flow meter outputs are almost always pulse, 4-20 mA or Modbus, and if the data collection hardware used for electrical monitoring supports these, water can ride on the same platform. Once electricity and water can be viewed on a shared time axis, things that neither dataset showed alone become visible, such as separating a load variation from a leak.
What should you do first if a Thai industrial estate asks you to save water?
Start by organising current monthly consumption and, if possible, the difference between holidays and working days. Then check the operating conditions of the two large consumers, the cooling tower and the washing processes. Those two items alone are enough to draft an initial response to the request. After that, begin designing measurement points so that you can keep demonstrating progress with numbers.
How long does it take before water visibility delivers results?
In factories with obvious waste such as leaks or outlets left running, it is not unusual to make a discovery as soon as Tier 1 is installed and the first holiday data comes in. Improvements to cooling tower operation and revisions to washing processes, on the other hand, involve changing operating rules and therefore run on a timescale of months. Set expectations separately for short-term discoveries and medium-term improvement.
Summary
There were technical reasons why Thai factories advanced electricity visibility while leaving water behind. Electricity can be retrofitted live with a CT clamp, whereas water usually requires cutting the pipe and inserting a flow meter, so the burden of adding one point is completely different.
In 2026, however, the situation is that 41 industrial estates and around 6,000 factories in the EEC carry demand of 1.5 million cubic metres per day while IEAT is preparing 1.18 million cubic metres. Data centre investment is pushing demand higher still, and IEAT is driving 3R within the estates. Occasions on which an industrial estate asks you about consumption and reuse rates are going to become more frequent.
Precisely because you cannot casually add points afterwards, the initial measurement point design is decisive for water. Start with the total at intake, expand to the large process-level consumers, and finally cover the discharge side. Progress through these three tiers in phases rather than all at once. Always retain the unmeasured balance obtained by subtracting measured values from the total. Hold to this design philosophy and you can reach a state where a limited number of points still explains the majority of your factory’s water.
We supply energy monitoring and production management systems to Japanese-affiliated manufacturers in Thailand, and we support the whole path from measurement point design for utility monitoring, water included, through to coordinating the on-site installation work. Even at the exploration stage, such as having received a conservation request from your industrial estate, or simply not knowing how many points to start with, we are happy to talk. After looking at the condition of your pipework on site, we can begin with a proposal on priorities, and you can reach us through the contact page.
