Every day the city ecological environment monitoring center keeps an eye on a large number of discharge outlets, with online monitoring data for COD, ammonia nitrogen, SO2, and NOx reported in real time. The data all appears to be there, but what front-line operations staff find most painful is not “failing to see an exceedance” but “seeing the exceedance and still being unable to say where it came from, or whether anything else is involved.”
At two in the morning, COD at a chemical park outlet suddenly spikes past 400. Is the instrument drifting, or is someone really discharging illegally? Around the same time, ammonia nitrogen at a downstream outlet also goes over its limit. Is it the same water, or coincidence? At an electronics park stack, SO2 tops 300. What went wrong with the desulfurization tower? The SO2 at the downwind ambient air station starts climbing too. Does it have anything to do with the stack?
There is enough data, but it is scattered across different systems: wastewater on one dashboard, exhaust gas on another, weather on a third. An exceedance triggers an alarm, but the alarms are not linked to one another. Faced with a pile of pop-ups, on-duty staff can hardly piece together a complete causal chain in a short time.
A single alarm triggers, but the pollution chain remains unclear
Exceedances are easy to see, illegal discharge is hard to identify
Wastewater online monitors raise threshold alarms for COD, ammonia nitrogen, and total phosphorus, and a pop-up appears whenever a value exceeds 100 mg/L. The problem is that nighttime illegal discharge does not show up as simply “one metric running high.” It shows up as COD, ammonia nitrogen, and total phosphorus rising together, pH falling, and instantaneous flow rising abnormally. If you only watch COD, the operator on duty is likely to write it off as instrument fluctuation. By the time the downstream outlet also exceeds its limit, the pollution plume has already flowed past.
The outlet exceeds its limit, but the environmental impact is unclear
An alarm triggers when SO2 at an exhaust outlet exceeds 200 mg/m³, but where does the SO2 go once it is released? Is the downwind ambient air station affected? How much? In a traditional system, outlet data and air station data are separated, and staff have to manually match timestamps and dig through trend charts, which is very inefficient.
Tracing relies on experience, and the evidence chain is incomplete
Once an exceedance is found, enforcement authorities require a complete evidence chain: when it started, what the peak was, how long it lasted, where the source is, and which downstream sites were affected. This information is scattered across different systems and logs, and patching it together by hand is both slow and prone to omissions.
From “over-limit alarm” to “chain judgment”: reorganizing the data
Object-based modeling: wastewater, exhaust gas, and weather return to a unified operational view
To answer “where does the exceedance come from and where does its impact go,” how the data is organized is the key. TDengine IDMP uses Industrial Ontology modeling to map wastewater outlets, exhaust outlets, weather stations, and ambient air stations into a clear tree-structured asset hierarchy. Each outlet is not an isolated collection of measurement points but an operational object bound to its park, monitoring category, and geographic location.
Operators no longer face a jumble of signal streams but device objects with physical hierarchy and topology: the chemical park outlet sits upstream, the general park outlet downstream, with upwind and downwind weather stations on either side. The spatial relationship between outlets and environmental stations is clear at a glance, and this provides the logical skeleton for downstream pollution tracing.
Real-time analysis and event correlation: alarms return to the full process context
On top of the asset model, the IDMP real-time analytics engine continuously aggregates the monitoring data stream. When the system detects that a measurement point has exceeded its limit, it dynamically generates an “event context” and pushes the abnormal duration, the slope of change, and the affected neighboring parameters to the operator console as a single package.
More importantly, IDMP supports event detection with multiple combined conditions. A “suspected nighttime illegal discharge event,” for instance, is not triggered by a single COD threshold but by four conditions holding at once: COD over its limit, ammonia nitrogen over its limit, abnormal rise in instantaneous flow, and the current time being nighttime. Combined conditions like these cut the false alarm rate sharply, so operators already know, the moment they see the alarm, that this is not ordinary instrument fluctuation.
Process analysis and AI-assisted insights: judgment shifts from experience-driven to evidence-driven
Once multi-source time-series data is correlated as objects, pollution tracing shifts from “senior staff guessing from experience” to “data investigation aligned on a shared timeline.”
When a complex case comes up, operators can start an AI-assisted interpretation with one click. The AI overlays the current operating curve on a normal baseline, quantifies the deviation, traces back the precursor features, and produces an analysis report with the root-cause conclusion and time-series evidence, giving environmental enforcement complete data support.
Closing the tracing loop in two typical scenarios
Scenario 1: a company in a chemical park discharges outside permitted conditions during the early-morning hours, and COD spikes to 450 mg/L
At two in the morning, COD at wastewater outlet 1 of the chemical park (WW-CH-01) climbs suddenly from a normal 85 mg/L, breaks the 100 mg/L red line within 20 minutes, and reaches a peak of 450 mg/L two hours later. Ammonia nitrogen rises from 12 to 78 mg/L, total phosphorus from 2.5 to 12.5 mg/L, and pH drops from 7.2 to 5.5. At the same time, instantaneous flow jumps from 120 to 210 m³/h. Normal nighttime flow should be falling, so this abnormal rise in flow is a key signal of illegal discharge.
Thirty minutes later, COD at chemical park outlet 2 (WW-CH-02) starts rising too, from a normal 80 to 239 mg/L. Ninety minutes later, COD at the general park outlet far downstream (WW-ZH-01) also shows a slight rise of 10-15 mg/L.
The on-duty team, after the COD alarm triggers, runs a four-step tracing in IDMP:
Step 1: Confirm the nature of the exceedance. Look at the COD, ammonia nitrogen, and total phosphorus trend charts for WW-CH-01. The three metrics rise together and pH falls, ruling out instrument failure, since instrument drift usually affects only a single metric. The steep slope of the COD rise also rules out a gradual process deterioration and points to a sudden discharge.
Step 2: Use flow to corroborate the illegal discharge. Overlay the instantaneous flow curve. WW-CH-01 flow rises from 120→210 m³/h. At night (02:00-06:00) the chemical park’s discharge volume should be lower than during the day, so the abnormal rise in flow confirms the illegal discharge.
Step 3: Confirm the source through downstream evidence. Look at the COD curves for WW-CH-02 and WW-ZH-01 and confirm the delayed response. WW-CH-02 lags by about 30 minutes and WW-ZH-01 by about 90 minutes. The lag times match the hydraulic transport distance, pinning the source to a company inside the chemical park.
Step 4: AI-assisted diagnosis. The AI interprets the multi-dimensional data and produces a complete report on the illegal discharge event: start and end times of the exceedance, peaks of each metric, scope of impact (two downstream outlets), and evidence of abnormal flow during the discharge window. This directly supports evidence collection for enforcement.
Full transmission chain: a company discharges high-concentration organic wastewater at night → COD, ammonia nitrogen, and total phosphorus exceed their limits together at WW-CH-01, pH falls, and flow rises abnormally → COD rises at WW-CH-02 after a 30-minute lag (hydraulic transport) → COD rises slightly at WW-ZH-01 after a 90-minute lag (attenuation downstream).
Scenario 2: a desulfurization pump fails at a company in an electronics park, and SO2 reaches 520 mg/m³
At ten in the morning, SO2 at exhaust outlet 1 of the electronics industrial park (WG-EL-01) starts climbing slowly from a normal 80 mg/m³. The initial rise is small, and the on-duty staff do not take notice. But 30 minutes later SO2 breaks the 200 mg/m³ red line and keeps climbing to 520 mg/m³. At the same time, flue gas velocity falls from 15 to 8 m/s, exhaust temperature rises from 120°C to 138°C, and O2 content rises from 8.5% to 9.2%. This set of parameter changes points to one clear physical cause: the desulfurization facility’s efficiency has dropped.
Even more noteworthy, wind speed at the upwind weather station (WS-01) drops from 3.2 to 1.5 m/s. Stagnant weather means poor conditions for pollutant dispersion, so the released SO2 does not disperse quickly. After 60 minutes, SO2 at the general park exhaust outlet (WG-ZH-01) starts to rise with a delay; 30 minutes later, SO2 at the city-center ambient air station (AQ-01) climbs from a normal 15 μg/m³ to nearly 100 μg/m³, a value already close to the limit set by the ambient air quality standard.
After the SO2 alarm triggers, the on-duty team runs a four-step tracing in IDMP:
Step 1: Confirm the desulfurization facility anomaly. Look at the SO2, flue gas velocity, and exhaust temperature trend charts for WG-EL-01. The three data series change in sync: SO2 rises sharply, velocity falls, and temperature rises, a typical combined signature of a drop in desulfurization efficiency. The velocity falling from 15→8 m/s indicates that the desulfurization absorbent circulation pump is running below its required output, not that the boiler load has changed.
Step 2: Assess the meteorological dispersion conditions. Look at the wind speed curve for WS-01. The current weather is stagnant (wind speed below 2 m/s), which is bad for pollutant dispersion. The released SO2 will accumulate near the ground, and the impact on ambient air quality will be larger than under normal weather.
Step 3: Trace the atmospheric dispersion impact. Look at the SO2 curves for WG-ZH-01 and AQ-01 and confirm the delayed response. The general park outlet lags by about 60 minutes and the ambient air station by about 30 minutes. SO2 at AQ-01 is close to 100 μg/m³, already a real impact on the city’s air quality.
Step 4: AI-assisted diagnosis. The AI interprets the emission source, weather, and environment data and produces a complete report on the desulfurization facility fault: the root cause is bearing wear in the desulfurization absorbent circulation pump, stagnant weather has intensified ground-level accumulation, the environmental impact already covers the city-center area, and emergency response should be started immediately.
Full transmission chain: bearing wear in the desulfurization circulation pump → the absorbent’s pH falls and desulfurization efficiency collapses → SO2 at WG-EL-01 climbs from 80→520 mg/m³, flue gas velocity falls, and exhaust temperature rises → stagnant weather is bad for dispersion → SO2 rises at WG-ZH-01 with a delay (atmospheric transport) → SO2 at AQ-01 reaches nearly 100 μg/m³ (environmental impact).
From post-event enforcement to keeping the process under control
The ceiling on the value of urban industrial pollution discharge monitoring is not set by how much data is connected or how many big screens are built, but by whether the data can turn into decisions that can actually be acted on. Concretely:
- Exceedance detection: upgraded from single-variable threshold alarms to multi-condition combined event detection. The four-way combination of COD, ammonia nitrogen, flow, and time window flags suspected illegal discharge directly and cuts false alarms sharply
- Pollution tracing: upgraded from digging through three systems after the fact to three-way tracing across upstream, downstream, and the environment. The lag signatures of hydraulic transport and atmospheric dispersion are mapped onto the asset tree topology, pinning down the source in minutes
- Impact assessment: upgraded from watching only the outlet to looking at the emission source, weather conditions, and environmental quality together, seeing how much was discharged, whether the weather allows it to disperse, and whether the environment has taken a hit
TDengine IDMP turns scattered wastewater, exhaust gas, and weather data into operational objects, turns alarms into explainable process judgments, and turns data capability into a regulatory and enforcement capability that supports exceedance tracing, illegal discharge identification, cross-park pollution transmission tracking, and environmental impact assessment.
TDengine comes with a high-performance, distributed time-series database, Industrial Ontology modeling, and an Industrial Agent Runtime, providing a full-stack solution for industrial data streams from collection and storage to real-time analytics, visualization, event management, and root-cause analysis. To learn more about TDengine, visit www.tdengine.com and try it for free.
Try it yourself
Install and deploy TDengine Visit the TDengine Download Center, select TDengine All-in-One, choose the deployment platform and architecture that matches your environment, and follow the guided steps to complete the installation.
Load the sample data
On first activation, choose Urban Industrial Wastewater and Exhaust Emission Monitoring on the sample data loading screen and wait for it to finish loading.
If you have already activated the product, click your avatar in the top-right corner, select Management Console, choose Sample Data on the left, then select Urban Industrial Wastewater and Exhaust Emission Monitoring. Wait a few minutes for the data to load.
Urban Industrial Wastewater and Exhaust Emission Monitoring
