How Do Manual and Automatic Monitoring Differ?

Manual and automatic monitoring differ in who takes the readings and how often. Both mostly use the same instruments and answer the same engineering question: how far has the ground or structure moved so far? In manual monitoring, staff carry a readout unit or a level instrument to each measuring point at the frequency specified in the project specification, record the values and bring them back to process into graphs. In automatic monitoring, the sensors stay wired to a datalogger on site, which takes readings on its own at preset times 24 hours a day and sends the data to a server over a cellular network or WiFi.

The most common misconception is that an automatic system gives more accurate readings. In fact, accuracy comes from the sensor itself and how it is installed, not from the system reading itself. The same sensor read with a handheld readout or a datalogger gives readings of similar resolution. Where an automatic system is clearly superior is resolution along the time axis: hourly or minute-by-minute data reveal short-term trends, the effect of each construction activity, and changes at night or on holidays that periodic manual readings cannot capture.

Conversely, manual reading offers something an automatic system cannot: an engineer's eyes on site. Every time staff go in to take readings, they see new cracks, seepage, soil loss behind the retaining wall, loose struts and the condition of the sensors themselves. This observational information is often worth more than the numbers. For an overview of every instrument type and what each one does, see the geotechnical instrumentation guide

The 3D view below places the two kinds of site side by side. On the left, staff walk from point to point taking readings; on the right, sensors are wired into a datalogger that sends data to the cloud. Pay particular attention to the graph in the lower panel: the value jumps between rounds 3 and 4. The automatic system sees it and issues an alert at that time, whereas manual reading only finds out at round 4.

3D model comparing a manual monitoring site with an automatic one and the data frequency each provides Conceptual 3D view of a site divided into two soil blocks. The left block is manual: staff carry a readout unit from point to point, with a probe-type inclinometer casing showing the cable and probe inside, an observation well with a water level meter tape lowered into it, and a settlement plate whose rod projects above the ground for a level instrument on a tripod to read. The right block is automatic: a vibrating wire piezometer embedded in a borehole, an in-place inclinometer with several sensors left in the casing, and a tiltmeter mounted on a structural wall. Blue signal cables from every sensor run to a datalogger cabinet on a pole fitted with a solar panel and an antenna. Data are sent to the cloud, displayed as graphs on a computer screen and alerts are sent by SMS, Line and email. The lower panel is a conceptual graph of measured values over time: the continuous orange line is data from the automatic system and the five blue dots are periodic manual readings. Between rounds three and four the value jumps above the threshold line; the automatic system sees it and issues an alert at that time, while manual reading only finds out at round four. 3D model: Manual vs Automatic site Not to scale 1 2 3 Data reach engineer after each round but site seen first-hand every round e.g. cracks, seepage 4 5 6 7 Cloud 8 Automatic alerts SMS · Line · Email Manual: staff read at set intervals ① Probe-type inclinometer ② Observation well / standpipe + dip meter ③ Settlement plate + level Automatic: datalogger reads 24/7 ④ VW Piezometer ⑤ In-place Inclinometer ⑥ Tiltmeter (or strain gauge) ⑦ Datalogger + solar ⑧ Cloud Data frequency over time (conceptual) Time Engineer-set threshold ! Round 1 Round 2 Round 3 Round 4 Measured value Automatic: continuous line Manual: periodic points ! Alert when over threshold Change between manual rounds is found at the next round
3D model: on the manual side, ① probe-type inclinometer, ② observation well or standpipe read with a water level meter, ③ settlement plate read with a level instrument; on the automatic side, ④ vibrating wire piezometer, ⑤ in-place inclinometer, ⑥ tiltmeter on the structure, ⑦ datalogger with solar panel and antenna, ⑧ cloud that displays graphs and sends alerts. The graph below is conceptual, not actual measured data (Image: SPN Soil Engineering)

Which Instruments Can Be Read Automatically and Which Need People?

Instruments that output an electrical signal can be read automatically, while instruments that require surveying or a manually lowered probe still need people to read them. If a sensor outputs an electrical signal that can be connected to a datalogger, it can be automated straight away. This table summarizes the instruments in both groups.

← Scroll to see the full table →

EquipmentMeasuresReading methodAutomatable
Vibrating Wire PiezometerPore water pressureElectrical signal to dataloggerYes
Strain GaugeStrain in reinforcement or strutsElectrical signal to dataloggerYes
TiltmeterTilt of buildings and retaining wallsElectrical signal to dataloggerYes
In-place InclinometerLateral movementA string of sensors left in the casing, wired to a dataloggerYes
Traversing inclinometer (Probe-type)Lateral movementStaff lower the probe depth interval by depth intervalManual
Standpipe Piezometer / Observation WellWater level in pipeWater level meter lowered into pipeManual
Settlement Plate / Surface Settlement Marker / Multi-level Settlement PointSettlementLevel surveyManual

The automatable group comprises 4 types: the vibrating wire piezometer, which measures pore water pressure directly and is the instrument most often automated; the strain gauge, which measures strain in reinforcement or struts; the tiltmeter, which measures the tilt of buildings and retaining walls; and the in-place inclinometer, which uses a string of sensors left in the inclinometer casing for the whole project instead of lowering a probe each time.

The group that still needs people includes settlement plates, surface settlement markers and multi-level settlement points, which are read by level surveying; standpipe piezometers and observation wells, read with a water level meter lowered into the pipe; and probe-type inclinometers following ASTM D6230, which require staff to lower the probe interval by interval on both the down and up passes. For more on these two groups, see Piezometers: measuring pore water pressure and Inclinometers: measuring lateral movement

💡 A common wrong decision

Many projects decide to go automatic only after the instruments have been installed, which is often too late, because the choice of sensor type at the outset determines whether it can be connected to an automatic system. For example, if standpipe piezometers have already been installed but continuous data are needed later, new vibrating wire piezometers have to be drilled and installed. This decision should be made together with the layout of measuring points.

What Does an Automatic Monitoring System Consist Of?

An automatic monitoring system is a chain of 4 linked data components: the sensors on site, the datalogger, data transmission, and the cloud platform with alerts. If any one of them fails, the data stop immediately. There must also be a fifth component: site visits by staff to maintain the system and check actual conditions. System design must therefore consider the robustness of every component, not just selecting good sensors.

Components of an automatic (real-time) monitoring system: electrical sensors, datalogger, data transmission, cloud and alerts, and site visits by staff The figure shows a deep excavation with a retaining wall and struts. Permanently wired sensors that output electrical signals include a tiltmeter on the adjacent building wall, a vibrating wire piezometer in the soil behind the wall, an in-place inclinometer left in its casing, and strain gauges on the struts. All cables run to a datalogger in a weatherproof cabinet on site, which takes readings at preset intervals, stores them in memory, runs on a battery with a solar panel, and sends the data over 4G or WiFi to a cloud platform that displays graphs, compares readings with thresholds and sends alerts by SMS, Line or email for the engineer to review before giving instructions. The fifth component is staff who still visit the site periodically to check the cabinet, cables and solar panel, take manual readings to compare with the system's values, and look for cracks and seepage that the sensors cannot see. The lower-right panel summarizes the often-overlooked items: battery and solar panel sizing for the rainy season, lightning and surge protection along long cables, temperature compensation for vibrating wire sensors and strain gauges, and on-device data storage when the signal drops, with later back-filling. Conceptual illustration; the actual components are set by each project's monitoring specification. Automatic monitoring: 4-part data chain + site staff Not to scale Tiltmeter 1 Wired electric sensors ↑ Strain gauge on strut ← In-place Inclinometer ← VW Piezometer 5 2 3 Cloud 4 ! SMS / Line / Email Often overlooked • Battery + solar sized for rainy season • Lightning / surge on long cables • Log temp. to correct VW / SG • Signal lost → store, send later • Over limit → engineer checks first 1 Sensors Electrical output 2 Datalogger Reads on schedule 3 Transmit data 4G / WiFi 4 Cloud Graphs + alerts 5 Site staff Check real site
An automatic system is a chain of 4 linked parts (① sensors ② datalogger ③ data transmission ④ cloud and alerts) and must always be overseen by staff visiting the site (⑤) — conceptual illustration

Often overlooked are power and protection. Most field systems run on batteries with solar panels. If the panel and battery are undersized for the rainy season, when sunshine is limited, the system shuts down intermittently and data are lost exactly when they are needed most. The other issue is lightning and surge protection along signal cables tens of meters long, which is the most frequent cause of datalogger and sensor damage on Thai sites.

Another issue is temperature compensation. Vibrating wire sensors and strain gauges also respond to temperature, so the system must record temperature alongside the measured values and apply the correction given by the manufacturer on each sensor's calibration certificate. Otherwise, readings will rise and fall in a daily cycle with air temperature and be mistaken for structural movement. For vibrating wire pore pressure sensors there is also guidance on acceptance testing before installation in ASTM D7764, which helps weed out faulty sensors before they go into the borehole.

How Do Manual and Automatic Monitoring Compare on Each Aspect?

Manual monitoring has a lower upfront cost and works with every instrument type, while automatic monitoring provides continuous data and issues alerts on its own but only works with sensors that output an electrical signal. The table below summarizes the differences that drive real project decisions, assuming the same sensor type is used in both cases.

← Scroll to see the full table →

TopicManual MonitoringAutomatic (Real-time) Monitoring
Reading methodStaff visit the site and read with a handheld readout or level instrumentDatalogger reads automatically at preset times
Data frequencyDiscrete points at each site visitContinuous 24 hours a day; adjustable from every 1 minute to daily
Speed of detecting abnormal readingsKnown at the next reading roundAutomatic alert at the preset transmission interval
Accuracy of readingsSimilar, because it depends on the sensor type and installation, not on who reads it
Applicable instrumentsAll types, including level surveyingOnly sensors that output an electrical signal
Upfront costLowerHigher, due to the datalogger, power system and cabling
Cost of more frequent readingsIncreases with the number of reading visitsAlmost no increase
Seeing actual site conditionsYes cracks, seepage, abnormal strutsNo sees only sensor readings
Risk of data gapsGaps when the site cannot be accessed, e.g. rain or restricted areasGaps when power runs out, the signal fails or cables are damaged
Best forLong-term earth fills, slowly changing behavior, limited budgetsUrban deep excavations, work near adjacent buildings, work needing immediate warning

The takeaway from the table is that the two approaches do not compete on data quality but on acceptable risk. If the damage from a problem would be high and behavior changes quickly, finding out a week late can cost more than the entire automatic system. The clearest example of work that meets this condition is deep excavation in urban areas, explained in Deep excavation monitoring in Bangkok soft clay

How Does an Automatic Alert System Work?

An automatic alert system compares each reading with preset thresholds and sends a message to the responsible people when a reading exceeds a threshold. This, more than the graphs on screen, is the heart of a real-time system. International practice sets thresholds in stages, typically 3 levels.

  • Alert level: readings start trending the wrong way; increase monitoring frequency
  • Action level: review work methods and prepare remedial measures
  • Alarm level (stop work): stop the activity causing it and carry out remedial work immediately
Three alert levels (Alert, Action, Alarm) of an automatic monitoring system and false-alarm prevention The left graph shows measured values over time against three threshold lines: Alert at the bottom, Action above it and Alarm at the top. At first the value fluctuates slightly around the baseline, with a single one-off spike caused by signal noise; a system configured to require several consecutive exceedances does not send an alert for this value. Later the value climbs steadily and exceeds the Alert level for several consecutive readings, so the system sends an alert by SMS or Line. The engineer checks the readings and site conditions and orders measures, and the value stabilizes below the Action level. The right panel summarizes what to do at each level: Alert — stay vigilant and increase monitoring frequency; Action — review work methods and prepare remedial measures; Alarm — stop the activity causing it and fix it immediately. The bottom strip notes that each threshold value is set by the design engineer for the project's conditions and is not copied from other projects, that false alarms are prevented by requiring several consecutive exceedances, and that while the system can send alerts on its own, decisions are the responsibility of the engineer in charge. Conceptual illustration. 3 alert levels: system warns, engineer decides Not to scale Measured value vs thresholds (conceptual) Alarm Action Alert Time → Measured value → One-off spike (noise) → no alert ! Exceeds several rounds → Send SMS / Line Measures → stable Baseline What to do at each level Alarm Stop the causing activity Fix immediately Action Review work method Prepare remedies Alert Stay vigilant Monitor more often 1 Thresholds Set by design engineer 2 No false alarms Must exceed repeatedly 3 Decision Engineer, not system
The system can send alerts on its own, but deciding what to do next remains the responsibility of the engineer in charge.

The values for each level have no universal standard. They must be set by the design engineer from the analysis of that particular project, e.g. the design movement values, the distance to adjacent buildings and the sensitivity of surrounding structures. Copying threshold values from another project is a direct risk.

The most common problem with alert systems is false alarms. A one-off spike from signal noise, a rapid temperature change or a battery voltage drop makes the system send messages even though the structure is fine. When this happens often, everyone stops paying attention to alerts. The remedy is to require several consecutive exceedances before an alert is sent, route alerts to different recipients by severity, and assign someone to check every alert, with the final conclusion always coming from an engineer, not from the system.

Which Should You Choose? 5 Decision Criteria

Choose by asking one question: "How much damage would finding out a week late cause?", not "Which one is better?" The following 5 criteria can support the decision.

1. Consequences if a problem occurs. Deep excavations next to old buildings, work near underground utilities, or work with people working beneath the at-risk zone should use an automatic system, because the time to detection directly affects safety.

2. Rate of the behavior being monitored. Settlement of earth fill on soft clay changes over weeks to months, so periodic manual reading is sufficient. This is unlike the movement of a retaining wall during excavation, which can change within a single day.

3. Instrument types specified in the design. If the monitoring plan relies mainly on settlement plates and level surveying, an automatic system adds little. But if vibrating wire piezometers or strain gauges are already specified, connecting them to a datalogger is a worthwhile add-on.

4. Site access and project duration. Measuring points in the middle of roads with traffic, in restricted areas or far from the site supervision office greatly raise the cost and risk of manual readings. The longer the project and the more frequent the readings, the sooner an automatic system breaks even.

5. Requirements of the owner and relevant authorities. Many projects already specify the monitoring method and reporting frequency in the contract or in the construction permit conditions. Check these requirements before planning the budget.

In practice, most projects settle on a hybrid approach: automatic monitoring only at critical sections and for instruments that can be connected, periodic manual reading for the remaining points, and staff taking manual check readings against the system's values from time to time. This is the best way to catch sensor drift, and it is why an automatic system has never fully replaced site visits.

What Services Does SPN Offer, and How Much Do They Cost?

SPN provides both forms of monitoring, from planning the measuring-point layout with the design engineer, supplying and installing instruments, taking baseline and periodic readings, and processing them into graphs and reports, to installing dataloggers with cloud data transmission and setting alert thresholds by SMS, Line and email based on the values specified by the design engineer. Every report is certified by a licensed civil engineer (COE license) with signature and registration number, and can be used for submission to the relevant authorities. See the full scope of work on the instrumentation and monitoring services page

Instruments within our service scope that can be connected to an automatic system include vibrating wire piezometers, strain gauges, tiltmeters and in-place inclinometers. Instruments read by staff include settlement plates and settlement markers, observation wells and standpipe piezometers, as well as probe-type inclinometers.

Fees depend on the number of measuring points, sensor types, monitoring duration, number of reading rounds and the chosen data transmission method, so they are estimated per project. Contact us for a free quotation within 24 hours. Prices exclude travel and equipment mobilization costs, which are charged according to the distance to the site.

⚠️ This article provides general information

This content explains the principles of comparing manual and automatic monitoring for general understanding. The choice of monitoring method, the type and number of sensors, the reading frequency, the alert threshold for each level and the measures to take when readings exceed thresholds must be specified by the design engineer and the engineer in charge according to the specific conditions of each project.