Workplace Environment Measurement | Thai Factories 2026

Your Thai plant files its workplace environment measurement report on time every year, and yet the complaints from the shop floor never stop. It is too loud near the press line. The forming area is unbearable in the afternoon. Operators in the inspection booth say they cannot see what they are doing. This is not negligence by your safety officer. It is a direct consequence of how the rule is designed. Thailand’s statutory measurement of heat, light and noise is a spot check performed at least once a year, which means that for the remaining eleven months and change, nothing is recorded at all. This article lays out what Thai law actually requires, how the newer hazard assessment obligation interacts with it, and where continuous IoT monitoring belongs in that picture, in the order a factory manager would need to work through it.

What workplace environment measurement means under Thai law

Workplace environment measurement is the act of measuring the physical and chemical conditions of the space where people work, checking those values against legally defined limits, and reporting the results to the authorities. Manufacturers arriving from Japan often assume Thailand mirrors the Japanese system with its national working environment measurement expert licence. There is no equivalent standalone national qualification in Thailand.

The legal basis here is the Occupational Safety, Health and Environment Act B.E. 2554, which came into force in 2011, together with the ministerial regulations and Department of Labour Protection and Welfare (DLPW) notifications issued under it. The instrument you will reference most often in practice is the Ministerial Regulation on the Standard of Management and Operation on Safety, Occupational Health and Workplace Environment regarding Heat, Light and Noise B.E. 2559, issued in 2016.

Three points matter most for a foreign-owned plant.

• Three parameters are in scope, namely heat, light and noise. An establishment with a heat-generating process measures heat, an establishment with a noise-generating process measures noise, and every workplace measures illuminance.
• The frequency is at least once a year. That is a floor, not a ceiling. You are free to measure more often.
• Results must be reported to the DLPW within 30 working days of completing the measurement.

So the entire Thai scheme resolves to a simple loop. Once a year a competent measurer visits the site, takes readings, and a report goes to the authority within 30 working days. Note that 30 working days is roughly a month and a half on the calendar once weekends are excluded, and longer still if Thai public holidays fall inside the window. Do not confuse it with 30 calendar days when you build your compliance schedule. There is nothing unusual about this framework. Periodic spot measurement is the global norm. The real question is whether your factory understands precisely what that framework guarantees and, more importantly, what it does not.

Thailand applies the same structural pattern of periodic external verification to other domains as well. Electrical installations follow an equivalent cycle with their own transitional provisions, which we cover in statutory inspection of electrical equipment in Thai factories. It is worth reading alongside this one when you assemble the annual compliance calendar.

Is once a year actually enough

In practice, “at least once a year” gets read on the shop floor as “once a year is fine,” and budgets and manpower are planned accordingly. Before accepting that, it is worth looking at what the phrase means as a quantity of time.

A year contains 8,760 hours. Depending on the size of the site, the measurer is physically present and running instruments for somewhere between a few hours and a single working day. Even if you generously assume a full eight-hour day of active measurement, that is 8 divided by 8,760, or roughly 0.09 percent of the year. For the other 99.9 percent, the statutory scheme observes nothing whatsoever.

That is not a defect. It is the design intent. Statutory measurement exists to confirm that steady-state conditions fall within limits under representative conditions. It does not exist to prove that conditions were always within limits. That is exactly why measurement is scheduled on a representative working day during a representative time window. The flip side is that non-representative conditions leave no trace at all unless they happen to coincide with the measurement date. Consider the following.

• A peak production week where an older machine you normally keep idle was run in parallel with the main line.
• A dry-season afternoon with maximum radiant heat through the roof, on a day when several ventilation fans happened to be down.
• A stretch of night shifts during which the adjacent press line was running a heavier die for a different part number.
• The slow six-month decline in task lighting as LED fixtures failed one at a time and nobody replaced them.

An annual snapshot misses all of these by construction. The value of continuous IoT monitoring lies in closing that gap, not in replacing the statutory measurement. An argument that conflates the two breaks down immediately, whether it is a sales conversation or an internal capital request.

Thai limits for heat, light and noise

Thailand did not simply import Japanese or Western limits. The values are locally set. Start with heat and noise.

The heat limits are where multinational safety teams most often get caught out. Thailand first set thermal limits in 1976 and revised them in 2006, arriving at the current 34, 32 and 30 degrees C thresholds. Because the derivation and the underlying reasoning differ from the global standards a head office typically adopts, such as the ACGIH TLV framework, it is entirely normal to end up in a state where a work area is legally compliant in Thailand but non-conforming against corporate policy. Decide in advance which document governs before that discrepancy surfaces in an audit.

Illuminance works differently. A DLPW notification dated 27 November 2017 sets required average illuminance by location and by type of work. Rather than a single limit, it is a classification table.

Separately, where work takes place in dark environments such as pits, tunnels or underground mines, employers must issue safety helmets fitted with a lamp capable of illuminating at least 20 lux over three metres ahead. That is a performance requirement for personal protective equipment, not a workplace illuminance limit, and the two should not be filed together.

Noise needs a similar caution. The 85 dB(A) figure applies to an eight-hour time-weighted average, not to an instantaneous reading. A press impact can momentarily exceed 100 dB and still produce a TWA below 85 if exposure duration is short. Conversely, keeping someone in a 90 dB environment for more than 2 hours 31 minutes puts you over the limit. Both failure modes appear in real plants. One team buys a sound level meter, reads 90, and concludes the site is illegal. Another team hears a loud bang, decides it is only momentary, and never checks the exposure duration at all.

Protective equipment carries explicit performance requirements too. Earplugs must provide at least 15 dB(A) of attenuation and earmuffs at least 25 dB(A).

One further development deserves a mention. A new DLPW notification was published in the Royal Gazette on 1 May 2026 and took effect on 1 July 2026. It explicitly prohibits passing the cost of occupational safety and health compliance, including protective equipment and mandatory training, on to workers. It repeals the previous notification dated 11 November 2011 and requires that the rights and duties of both employer and employee be posted, and that hazard warning symbols and safety signs conform to recognised standards such as TIS, ISO, EN, ANSI or JIS.

The 2021 amendment tightened who may measure and with what

If you procure workplace environment measurement by comparing quotation totals alone, you will eventually get burned. A DLPW notification (No. 2) published in the Royal Gazette on 11 January 2021, prescribing the methods of measurement and analysis for heat, light and noise in establishments and the business categories in scope, raised the bar on both the measurer’s credentials and the instruments used.

The requirements added or clarified by that amendment include the following.

• The person performing the measurement must hold expert knowledge in the relevant field. In practice this means a professional-level safety officer, or a holder of a bachelor’s degree or higher in occupational health, certifying the results.
• The measurer must own the instruments used. Measuring with equipment borrowed from another company is not contemplated by the rule.
• The serial number of each instrument used must be stated in the report, so that readings can be traced back to a specific physical device.
• Calibration must be performed before every use in accordance with the manufacturer’s instructions.
• For standard calibration, service providers offering measurement to third parties must calibrate annually, while private establishments that own their own instruments must calibrate once every two years.

The practical consequence is significant. If the low-cost provider you selected turns up with borrowed instruments, omits serial numbers and cannot produce calibration certificates, the report itself may fail to meet the requirements. When an auditor or an inspector flags that, the party who has to redo the work is you, not the vendor. Confirm instrument ownership, serial number disclosure and calibration certificate attachment at the quotation stage, in writing.

How to scope and budget the measurement

“What does workplace environment measurement cost in Thailand” is the single most common question, and the honest answer is that no official tariff and no industry rate card exists. There is no single market price to quote. What you can do instead is understand which variables move a quotation, so that competing bids become comparable. Six variables do most of the work.

• Number of measurement points. Pricing accumulates per point, not per building. Line count, operator placement and the number of separate rooms determine the total.
• Number of parameters. A site that only needs heat is a different job from one that needs heat, light and noise across the board.
• Whether personal exposure monitoring is included. Fitting dosimeters to workers and following them for a full shift, on top of fixed-point measurement, adds significantly to the man-days required.
• Days on site and shift coverage. Covering two or three shifts including nights multiplies the days required.
• Report language and format. A Thai submission pack for the authority is mandatory. An English or Japanese internal version is an extra deliverable.
• Whether corrective recommendations are included. Root cause analysis and countermeasure proposals for any exceedance are separate work.

The discipline is to fix all six as your own specification before you approach anyone. If you simply ask for “a quote for workplace environment measurement” with no specification attached, each vendor invents its own assumptions, the totals diverge by a factor of two or three, and the comparison becomes meaningless. The same structural problem shows up in factory system procurement, and the reasoning we set out in digitizing records for audit response transfers directly.

The hazard assessment obligation with its November 2025 deadline

Alongside workplace environment measurement, a newer obligation landed on Thai manufacturers between 2024 and 2025. It is the Ministry of Labour notification on hazard assessment, study of workplace environment impact, and preparation of a management plan for employees and the establishment.

The timeline and requirements are as follows.

The 48 sectors in Schedule 2 cover the principal manufacturing fields, namely food, textiles, plastics, metals, electronic components, machinery and automotive, and extend further to recycling, hotels, hospitals and amusement parks. Given the 20-employee threshold, almost every foreign-owned manufacturer operating in Thailand should assume it is in scope.

The two rows that matter operationally are the last two. A review every three years and an update within 30 days of any operational change. These run on a completely different clock from the annual measurement cycle. Three years sounds like a long interval, but no factory stands still for three years. Equipment gets added. Line layouts change. New products are introduced. Shift patterns are restructured. When those things happen, is your management plan being updated within 30 days?

And the trigger for that update is noticing the change in the first place. If nobody notices, nobody pulls the trigger. That is the second role for continuous monitoring data.

Why annual measurement alone does not prevent occupational disease

Step away from the regulation for a moment and look at the health outcomes. What does the exposure an annual measurement misses actually do to a worker over time?

Noise-induced hearing loss is the classic case. Hearing loss does not begin on the measurement date. It develops over years of exposure above 85 dB(A), degrading gradually. An analysis based on the Global Burden of Disease study (GBD 2021), published on 22 September 2025, reports that disability-adjusted life years (DALYs) from occupational noise-induced hearing loss reached approximately 7.847 million globally in 2021, an increase of 104.46 percent over 1990. The age-standardised rate stands at 91.12 per 100,000. Southeast Asia alone accounts for roughly 893,000 DALYs, and the projection points to about 9.78 million DALYs worldwide by 2040.

Thai epidemiology adds concrete local numbers. A study of 700 sawmill workers in southern Thailand reported a noise-induced hearing loss prevalence of 22.8 percent, corresponding to 42 affected individuals. A separate study of welding operations in an automotive parts factory confirmed that exposure levels of 86 to 90 dB(A), and levels above that band, significantly raised the risk of hearing impairment in at least one ear.

Heat tells a similar story. Research validating occupational heat standards in Thai workplaces found that among 168 workers studied, only 55 percent worked in environments that met the applicable standard. Foundry floors averaged a WBGT of 32.4 degrees C with a maximum of 36.9. Construction sites averaged 31.8 degrees C with a maximum of 34.2. Depending on work intensity, both exceed the limit. On the incident side, 2,473 heat-related illness cases were recorded in outdoor work in 2016, a rate of 4.12 per 100,000 workers, and 2,150 cases along with 22 deaths were recorded across industrial sectors between 2009 and 2015.

What these figures demonstrate is that a meaningful gap can exist between the record stating “within limits at the annual measurement” and the exposure workers are actually receiving. A peak of 36.9 degrees C never appears in any report unless the annual visit happens to land on the right day.

Drawing the line between statutory measurement and IoT monitoring

With all of that established, it is worth being explicit about which job belongs to which tool.

The row that matters most is legal standing. Continuous monitoring with IoT sensors does not replace statutory workplace environment measurement. The obligation to measure and to report continues to sit with a measurer who meets the qualification requirements. Factory system vendors, ourselves included, are not authorized statutory measurement bodies, and continuous monitoring data cannot be submitted as a statutory report to the authority.

So what is continuous monitoring for? Three roles.

• Filling the space between measurements. For the 99.9 percent of the year the annual visit does not observe, you record whether exceedances occurred, how long they lasted and under what conditions. You can align the timestamp of an exceedance against production records and see what was running.
• Feeding the hazard assessment update obligation. The three-year review and the 30-day update after an operational change both need a trigger. Continuous data lets you evidence a change with numbers rather than impressions. If daytime average noise rose by 2 dB the week after a machine was added, that is a fact you can take into an internal review.
• De-risking the annual measurement in advance. An exceedance found during the statutory visit costs you remediation plus a re-measurement. If you already know from continuous data which points are marginal, you can add ventilation or acoustic treatment before the measurer arrives and avoid the rework.

This pattern of supplementing statutory spot measurement with continuous monitoring is not unique to the work environment. In Thai industrial wastewater, pairing statutory sampling and analysis with automatic BOD and COD instrumentation has become a standard design, which we cover in designing a factory wastewater monitoring system. The same structure applies to stack emissions, discussed in CEMS for factory exhaust gas monitoring and DIW reporting. The relationship between regulation and monitoring is essentially the same shape in all three.

A practical sequence for monitoring noise and heat with IoT

Here is how to actually introduce continuous monitoring, in order.

Step 1, select the measurement points

Do not attempt to blanket the whole plant. Open last year’s statutory measurement report and pick the three to five points with the smallest margin against the limit. A point recording a noise TWA of 82 dB(A) will cross 85 dB(A) as soon as production conditions shift slightly. That is a point worth watching. You do not need to instrument every point, and the ones with comfortable margins can wait.

Step 2, choose sensors and mounting

For noise, use an industrial sound sensor conforming to IEC 61672 Class 2 or Class 1. For heat, use a unit that captures the globe, wet bulb and dry bulb temperatures needed to compute WBGT. For light, use an illuminance sensor. Match mounting height and position to the conditions used at the statutory measurement point as closely as you can. If you do not, you lose the ability to explain any later discrepancy where the statutory result is compliant but the continuous data shows exceedance.

Step 3, set sampling intervals and retention

Match the interval to how fast the parameter moves. One-minute equivalent continuous sound level for noise, five-minute WBGT for heat and fifteen-minute readings for illuminance are a reasonable starting point. You do not need to keep everything forever. Tier it, for example raw data for three months and daily aggregates for three years or more, and both bandwidth and storage stay within practical limits. Since the hazard assessment retention obligation is at least three years, aligning aggregate retention to that figure is the natural choice.

Step 4, design thresholds and notification routing

This step takes the most thought. Set thresholds below the legal limit, not at it. For noise, a two-stage design works well, such as a caution when the one-hour equivalent level exceeds 83 dB(A) against an eight-hour TWA limit of 85 dB(A), and an alarm above 85 dB(A). Decide routing by asking who has to receive the alert for behaviour to change. Ringing a safety officer’s phone in the middle of the night accomplishes nothing if that person cannot reach the floor.

Step 5, correlate with production data

This is where continuous monitoring earns its keep. At the timestamp of a noise peak, which part number was running, on which machine, and at what rate? Overlay the two and causal patterns emerge, such as an exceedance that occurs only on days when a particular die is installed. For how to start collecting equipment-side data in the first place, moving factory data loggers to IoT is a useful reference.

On starting small and defining exit criteria before you begin, we go into detail in how to run an IoT PoC and set the decision criteria first. Work environment monitoring is harder than productivity IoT to justify in pure financial return, which makes designing those criteria up front even more important.

Frequently asked questions

How often is workplace environment measurement required in Thailand

At least once a year. Under the regulations issued below the Occupational Safety, Health and Environment Act B.E. 2554, an establishment with a heat-generating process measures heat, an establishment with a noise-generating process measures noise, illuminance is measured in every workplace, and results are reported to the DLPW within 30 working days of completion. If your parent company standard or your ISO 45001 operating rules specify a higher frequency, you comply with both, not just the legal minimum.

What does workplace environment measurement cost in Thailand

There is no official tariff and no industry rate card, so no single market figure can be quoted. Cost is driven by six variables, namely the number of measurement points, the number of parameters, whether personal exposure monitoring is included, days on site and shift coverage, report language, and whether corrective recommendations are included. Fix all six as your own specification before requesting quotations so that bids are comparable. Additionally, confirm that the provider owns its instruments, states serial numbers in the report and can attach calibration certificates. The 2021 amendment made these mandatory, and a report that falls short can become a problem later.

Can IoT sensor data be used for statutory measurement reporting

No. Statutory workplace environment measurement must be performed by a measurer meeting the qualification requirements, using calibrated instruments that the measurer owns, with serial numbers disclosed in the report. Continuous IoT monitoring does not satisfy those requirements. Its role is not to substitute for statutory measurement but to detect exceedances occurring between measurements and feed the hazard assessment review and the update triggered by operational change. Statutory reporting still has to be commissioned from a qualified measurement body.

How far can continuous noise monitoring realistically go

Continuously logging the equivalent continuous sound level at one-minute intervals is technically well established. However, the statutory evaluation metric, the eight-hour time-weighted average, cannot be computed rigorously without knowing how many hours a given worker spent at that point. A fixed sensor tells you how the noise level at that location behaved. It does not tell you an individual worker’s exposure dose. Once that distinction is understood, combining fixed-point data with staffing and placement information produces something genuinely usable.

How does hazard assessment differ from workplace environment measurement

Workplace environment measurement is the act of measuring specific physical quantities, namely heat, light and noise, with instruments, at an annual frequency, reported within 30 working days. Hazard assessment is the systematic identification of hazards present at the establishment, the study of their impact on the work environment, and the preparation of a management plan. Under the Ministry of Labour notification published on 22 November 2024 and effective 21 May 2025, establishments in scope had to implement it by 17 November 2025, report within 60 days of completion, retain the records for at least three years and review every three years. Measurement is one input. Assessment is the wider framework that consumes it.

Does continuous monitoring let us reduce the statutory measurement frequency

No. “At least once a year” is a floor, and installing continuous monitoring is not a basis for relaxing statutory measurement to every other year. What continuous monitoring gives you is information that sits outside the statutory scheme. In fact the effect usually runs the other way. If continuous data reveals signs of exceedance, the sensible response is to measure more often, or to commission an additional measurement after remediation.

Summary

Thailand’s workplace environment measurement regime, grounded in the Occupational Safety, Health and Environment Act B.E. 2554, requires heat, light and noise to be measured at least annually and reported within 30 working days. The 2021 amendment tightened the qualification requirements for measurers and the ownership, serial number and calibration requirements for instruments. On top of that, the hazard assessment and management plan obligation took effect in May 2025 with a 17 November 2025 implementation deadline, bringing with it a three-year review cycle and a 30-day update requirement whenever operations change. The regulation is steadily becoming more granular and more continuous.

At the same time, statutory measurement observes only a sliver of the 8,760 hours in a year. With global DALYs from occupational noise-induced hearing loss having roughly doubled since 1990, and Thai field research finding only a little over half of workplaces meeting the heat standard, it is worth holding two ideas apart. An annual report showing compliance is not the same thing as workers actually spending the year in a safe environment. Continuous IoT monitoring does not perform your statutory measurement. What it does is fill the space between measurements, surface the signs of change that should trigger a hazard assessment update, and reduce rework before the annual visit. Those three roles are valuable precisely because they sit outside the statutory scheme.

TOMAS TECH is based in Bangkok and designs and builds systems that handle shop floor data for manufacturers operating in Thailand. On work environment monitoring, we can help with the groundwork long before any hardware is chosen, such as reading your existing statutory measurement report to identify which points are marginal and deciding where to start. Early-stage discussions are welcome even if you have not decided to proceed. Get in touch through our contact page and we will review your current measurement report and come back with a concrete view on which points to monitor and in what order.

References

• Enviliance ASIA, Thailand amends requirements on heat, noise and lighting measurement — https://enviliance.com/regions/southeast-asia/th/report_5312
• Enviliance ASIA, Thailand mandates hazard assessment of work environment by Nov 2025 — https://enviliance.com/regions/southeast-asia/th/report_12808
• Enviliance ASIA, Thailand, Management of heat, noise and lighting at Workplace — https://enviliance.com/regions/southeast-asia/th/th-osh/th-osh-heat
• Thailand Law Office, Occupational Health and Safety in Relation to Heat, Light and Noise. Covers WBGT limits, protective equipment requirements and lamp-fitted safety helmets for underground work. Parts of the text describe the earlier ministerial regulation B.E. 2549, so read the reporting deadline as the current 30 working days — https://www.thailandlawoffice.com/legal-articles/277
• LeKise, Light intensity standards in factories and warehouses. Illuminance classification table from the DLPW notification dated 27 November 2017 — https://www.lekise.com/en/blog/2022/06/light-intensity-standards-in-factories-and-warehouses
• Geonoise, Noise legislation and noise limits in Thailand — https://www.geonoise.com/noise-legislation-and-noise-limits-in-thailand/
• Mahanakorn Partners Group, Thailand Issues New Notification on Occupational Safety, Rights, Duties, Warning Symbols and Safety Signs — https://mahanakornpartners.com/thailand-issues-new-notification-on-occupational-safety-rights-duties-warning-symbols-and-safety-signs/
• Frontiers in Public Health, Global burden and trends of occupational noise-induced hearing loss 1990-2021 and projection to 2040, published 22 September 2025 — https://www.frontiersin.org/journals/public-health/articles/10.3389/fpubh.2025.1682413/full
• Safety and Health at Work, Climate Warming and Occupational Heat and Hot Environment Standards in Thailand, March 2021 issue — https://pmc.ncbi.nlm.nih.gov/articles/PMC7940126
• PubMed, Occupational noise exposure and hearing defects among sawmill workers in the south of Thailand — https://pubmed.ncbi.nlm.nih.gov/29058570/
• PubMed, Occupational noise-induced hearing loss in auto part factory workers in welding units in Thailand — https://pubmed.ncbi.nlm.nih.gov/27885242/