Automotive lidars for rotating wind turbine blade monitoring

Jin, Liqin; Mann, Jakob

doi:10.5194/wes-2026-104

Preprints

https://doi.org/10.5194/wes-2026-104

Preprints

03 Jul 2026

| 03 Jul 2026

Status: this preprint is currently under review for the journal WES.

Automotive lidars for rotating wind turbine blade monitoring

Liqin Jin and Jakob Mann

Abstract. Permanently integrated sensor systems, such as strain gauges and fiber optic sensors, are the predominant means of measuring deflection in full-scale wind turbine blades. However, these approaches suffer from several key limitations, including complex calibration procedures, labor-intensive installation, and the inability to repair sensors once the blade structure is sealed. Furthermore, they are severely limited in measuring torsional deformation, a parameter of increasing importance for large wind turbine blades. To address these limitations, this study presents a novel non-contact monitoring framework based on a synchronized array of three automotive-grade lidars, enabling full-scale measurement of blade deflection and torsional deformation under diverse operating conditions. Lidar-derived flapwise deflection measurements (sampled at 33.3 Hz) are validated against co-located strain gauge data acquired at 1.4 m from the rotor plane center (sampled at 50 Hz), while lidar-based pitch angle estimates are validated against SCADA measurements after both signals are resampled to 2 Hz. The measured blade torsional deformation reaches approximately 0.8° under above-rated wind speed conditions, consistent with expected aerodynamic behavior. The dependence of median flapwise deflection on mean hub-height wind speed, rotor azimuth angle, and wind shear is also systematically analyzed. The results demonstrate that the proposed lidar-based system can accurately capture both flapwise deflection and pitch deformation along the blade span. These findings highlight the potential of cost-effective automotive lidar sensors for reliable, high-resolution monitoring of wind turbine structural dynamics under challenging field conditions.

Received: 12 Jun 2026 – Discussion started: 03 Jul 2026

Competing interests: At least one of the (co-)authors is a member of the editorial board of Wind Energy Science.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Liqin Jin and Jakob Mann

Status: final response (author comments only)

RC1:
'Comment on wes-2026-104', Anonymous Referee #1, 18 Jul 2026
The manuscript presents a new wind turbine rotor blade monitoring concept based on the use of automotive-grade lidars. The flapwise deflection measurement data recorded with the lidar were validated against the strain measurement data of one of the three rotor blades. This is an important and relevant research topic and this reviewer found the manuscript well-written and very interesting to read.
Still, this reviewer has some comments and questions regarding the procedure and the results, which is why some clarifications should be made before the paper is accepted for publication.
Lines 58ff: I see and agree with the limitations of embedded sensors, but to my knowledge, monitoring a structure using sensors belongs in the category of NDE/NDT, regardless whether they are embedded or mounted?

Lines 69ff: maybe also mention that camera-based systems usually require a solid ground to stand on and are, thus, often restricted to onshore use. However, your approach can also be applied offshore (you did not mention this advantage with regard to camera-based systems anywhere I believe)

General comment: You often write “structural monitoring” – is this on purpose or do you mean “structural health monitoring”? If so, please remain consistent in the wording.

You often mention that the considered blade is relatively stiff and larger blades exhibit larger deflection – wouldn’t the lidar miss some deflections, especially at the blade tip, because the blade is simply out of sight at some point? Did this at no time happen in the current measurements?

I read the short paragraph in the conclusion section about this, maybe consider discussing this earlier on and in even more detail (I thought about this often while reading)

Lines 161ff: Please mention that this is a local optimization algorithm. As such, it is dependent on the starting point. What do you “give” as the starting point in your case(s)?

Is it possible to place the figures more near the actual mentioning in the text? They are often two pages down.

Figure 7:

Can you visualize the centroid for each segment – because you refer to it in the text quite often

Can you turn the figure (similar to Figure 6) and have all three subfigures fit into one so that z is continuous and you don’t need the subfigures?

Lines 165ff: This simply means that it is not applicable in real time at the moment, but with a slight time gap, correct?

Table 2: Please add numbers for the different cases and refer to those in the following text (and figures) – this would simply make it easier for readers to follow which case you refer to rather than search for the correct date.

Regarding the motion of the boom:

If I understood correctly, you correct the boom motion by removing the displacement you measured for the root segment – but using what assumed direction? Or for each “measured line” (cf. Figure 6) individually? I suppose you assume one direction of boom motion for this, but what if the boom motion is also crosswise to this direction? Would/does this affect the correction?

Would, e.g., additional accelerometer(s) at the boom tip (where the lidars are installed) be of interest to find out the exact motion even more accurate?

Lines 195ff: I find the order a little confusing here – the following paragraph is the explanation of the sentence in lines 197-199, this should belong together?

Line 210f: How exactly did you normalize? Maybe even give the formula?

Line 213: Where exactly inside the blade Is the strain gauge embedded?

Figure 8: Maybe an even shorter time interval would be interesting to look at exemplary for even better comparison? It looks a little like the amplitude of the strain gauge is consistently a little lower?

Figure 9: A trend is visible that the higher frequencies identified from the strain measurements are a slightly lower than frequencies identified from the lidar measurements. Is there a possible explanation?

I believe a short summary (maybe in bullet points) of the findings and observations would be nice at the end - just a suggestion.

Also the last sentences could be a nice benefit of this approach – consider mentioning this also earlier on, maybe even in the introduction as I believe this is an advantage

How can I get to the mentioned video supplements? I don’t see it anywhere in the files.

Some additional (minor) comments
Line 14: You probably mean monitoring of the structural dynamics of wind turbine rotor blades (and not the whole wind turbine)?

Line 166: However, …

Line 186: Maybe give a (numbered) equation here?

Line 194: You write denoted as “0” – but in Table 2, it is denoted as “-“.

Line 242f: Number of equation missing

Line 261: time-stamp

Line 262: I guess the word “that” is too much in this sentence.
Citation: https://doi.org/10.5194/wes-2026-104-RC1
RC2: 'Comment on wes-2026-104', Anonymous Referee #2, 27 Jul 2026

The paper under review presents a very interesting and original approach for monitoring wind turbine blade bending deflections and torsional rotations.
The proposed methodology is still at an early stage of development, with room for improvement (as fairly acknowledged by the authors). Nevertheless, it has already demonstrated its capability to provide valuable results for characterizing both the static and dynamic structural behaviour of wind turbine blades under real operating conditions.
The reviewer believes that the paper contains sufficient innovative contributions to merit publication in this journal. However, there is still scope for improvement regarding clarity, completeness of the presented information, and the accuracy of some statements and derivations.
Therefore, the reviewer suggests that the authors address the following points:
- Title: considering that the main advantage of the proposed monitoring methodology is its low cost, I suggest highlighting this aspect in the paper title, for example: Cost-Effective Automotive-Grade Lidar for Rotating Wind Turbine Blade Monitoring.
- Measurement campaign: A description of the strain gauge monitoring system used for validation would be welcome in this section, as the details of this installation are important for understanding the results presented later. The strain gauge configuration (quarter-bridge, half-bridge, or full-bridge) and the calibration procedure should be described.
- Measurement campaign (line 138): It should be explained why the sensor angles are slightly different for each blade, how these angles were determined, and whether their adjustment could further improve the results through better blade coverage. It would be also nice to have a better idea of the work associated with the system installation in the field, needed resources and time.
- Figure 6: The differences observed between blades should be explained. Presumably, they are related to slight variations in the sensor installation angles.
- Lines 168–171: The concepts of flatwise and edgewise may be misleading. It should be explicitly stated that, in this paper, these directions are defined relative to the rotor plane rather than the blade chord, which is the more commonly adopted definition.
- Results: The analysis is focused exclusively on Blade C, which is acceptable for the objectives presented. However, it would also be important to demonstrate the consistency of the results across all three blades in order to assess the reliability and repeatability of the monitoring system.
- Results: If the system operates continuously, why are only a few events of 10–15 minutes analysed? A larger dataset would enable a more comprehensive statistical analysis. Furthermore, for two of the selected events, wind mast data are unavailable. The rationale behind the selection of these specific events should be clarified.
- Results (lines 207–210): This paragraph lacks accuracy and important supporting information. The relationship between blade tip deflection and blade root bending moment should be established more rigorously. Tip displacement depends on the double spatial integration of the bending moment distribution along the blade span. Therefore, in theory, the same root bending moment may lead to substantially different tip deflections. Under specific assumptions regarding load distribution, these quantities can be correlated. Moreover, the relationship between strain gauge measurements and bending moments is not straightforward, as it depends on the sensor installation and calibration, which should be characterized in Section 2.
- Results (Section 4.2): Please justify the use of the median instead of the mean.
- Results (Section 4.2): The relationship between wind speed and flapwise deflection should be discussed in greater detail. Wind loading generates a distributed aerodynamic load that depends on the local aerodynamic characteristics of each airfoil section. This load distribution produces a bending moment distribution, which in turn determines the deformation profile. Consequently, assuming a simple quadratic relationship between tip deflection and wind speed may not always be valid.
- Results (line 247): The maximum thrust force is typically observed at the onset of pitch control, which for this turbine is presumably around 10 m/s, slightly below rated wind speed. Since SCADA data are available, this behaviour could be analysed and incorporated into the case study description in Section 2.1.
- Results (Figure 10): These results are very interesting; therefore, presenting equivalent results for all three blades would significantly strengthen the paper.
- Results (Figure 13a): Could the vibration observed in this figure be associated with the torsional natural frequency of the boom? It appears to be concentrated within a relatively narrow frequency band.
- Future work (lines 379–380): The last proposed alternative architecture is not clearly described and should be explained in greater detail.
- Future work: It would be interesting to investigate the use of these measurements for performing modal identification of the blades.
Minor Corrections
Abstract (line 3): “…eventual inability to repair a sensor…” (sensors installed in some blade locations may be inaccessible and therefore difficult to repair).
Figure 1 caption: “…turbine blade collapse…”
Line 41: Remove the final closing parenthesis “)”.
Lines 51–52: The statement regarding the need for 200 strain gauges and three weeks of installation appears exaggerated. Naturally, the required number of sensors depends on the intended application. For example, accurate estimation of blade root bending moments can often be achieved with only a few strategically placed strain gauges and just one day for installation.
Line 170: z-axis (positive outward).
Line 268: “blade gravity” ? gravitational load acting on the blade?

Citation: https://doi.org/10.5194/wes-2026-104-RC2

Liqin Jin and Jakob Mann

Viewed

Total article views: 197 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
143	38	16	197	10	12

HTML: 143
PDF: 38
XML: 16
Total: 197
BibTeX: 10
EndNote: 12

Views and downloads (calculated since 03 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	112	29	12	153
Aug 2026	31	9	4	44

Cumulative views and downloads (calculated since 03 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	112	29	12	153
Aug 2026	31	9	4	44

Viewed (geographical distribution)

Total article views: 147 (including HTML, PDF, and XML) Thereof 147 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 15 Aug 2026

Download

Preprint (19559 KB)
Metadata XML

Short summary

Wind turbine blades bend and twist during operation, making reliable monitoring important for performance and safety. This study presents a novel non-contact measurement system using three low-cost automotive lidar sensors to track blade motion on a full-scale wind turbine. The results agree well with existing monitoring systems and reveal how blade movement changes with operating conditions. The findings demonstrate a practical and affordable method for monitoring large wind turbines.


Total:	0
HTML:	0
PDF:	0
XML:	0