Learnings

Outcomes

Deliverable 1.3: Report of E-field sensor performance 
 
The work within this deliverable highlighted the electric field sensor performance, such as usability, sensitivity and reproducibility, against standard electric field profiles including plane-plane, sphere-plane, sphere-sphere geometries.  
The three considered test geometries provided examples of uniform and non-uniform electric field profiles, with repeatable results. The simple geometries were also simulated within Comsol, to provide a baseline electric-field profile, which was used to verify the measurements obtained using the electric field probes. Measurements were performed within the HV labs, with voltages up to 15 kVrms across the geometries, and generating electric fields up to 12 kV/cm. The voltages were increased in 1 kV steps, to understand the sensitivity of the sensors to change in electric fields. The sensor probes were placed within the field in the tests and were stationary throughout the voltage application.  
The results highlighted a difference between 0.2 and 20% between the measured and simulated results, with the largest margin observed for the case of non-uniform geometry. This difference can be explained by the nature of the simulations, and also the difficulty in accurately mapping the locations of the probes in a 3D space for comparison with idealised scenario simulations. Despite the observed difference in the measured/simulated values, the sensor itself responds to small changes in fields, highlighting the capability to be able to develop electric field profiles of more complex geometries. The current exercise also allowed the team to understand the use of the probes, with significant learnings in data interpretation, repetition and positioning of the sensor.  
The research team also identified problems with one of the three sensors and recognised that erroneous measurements were being recorded by one of the probes. This was identified during the above testing exercises and later sent for replacement from the parent company. However, the two working probes provided evidence to enable progression to the next part of the project and start measuring electric field profiles of insulator samples. 

Deliverable 2.1: Bespoke rig building with design documentation and CAD drawings (Month 12) 
The key tasks of this deliverable were to build a complex rig, that could hold onto a small number of cap and pin insulators (currently set at 5 sheds).  
 
The idea behind such a system was to employ an actuator stand capable of holding an insulating arm "rod"  on which the electric field sensors’ probes would be installed. To measure the electric field on the surface of healthy and unhealthy "faulty" insulators from various angles, the actuator stand can be moved vertically (up/down) to position the probes at each cap and pin unit. The original idea also involved a motorized head, which would hold onto the small-scale samples. This motorized head would rotate the insulator along the central axis.   
 
However, without suitable tension within the system, the motorised head would only rotate the first cap and pin section, leaving the subsequent pins stationary. To avoid this, a circular track was designed, along which the vertical actuator system would move. This would provide the rotational aspect of the sensor,. enabling the electric field probes to measure the electric field at each unit with distinct coordinates. However, further analysis of the initial design highlighted some challenges that made the system difficult to implement and construct. Some of the challenges are highlighted below,  

1) The framework and stand included significant metal components. The large metal framework could potentially cause electric field interference, which could possibly skew the results of the insulator samples.  

2) Whilst the vertical movement of the actuator was not an issue, the horizontal movement provided challenges. A simultaneous control of both these axes required significant precision, highlighting the need for a complex automated system and adding further risk with component failure (and delays in the project). 

3) The sheer complexity of the rig required several suppliers to provide different components to be brought and assembled (and automated) at the University. This added further delays and costs to the rig.  

To overcome the above challenges, it was decided to move to an automated robot arm, which was capable of performing all the complex measurements and actuations required for probe manipulation along the small-scale insulators.  
 
The insulators itself are now held using an engine block, as the robotic arm provides the necessary actuation around the circumference of the sheds. 
Initial results have highlighted positive feedback between the arm and the controller, allowing for more accurate readings to be performed within a 3D space. 
 
 
 
Deliverable 2.2: small-scale test cell demonstration (Dec-24) 
This deliverable was done on a NG visit to The University of Manchester on 18/11/2024. The small-scale string comprised of five glass cap and pin insulator units. The probes moved with a pre-programme robot arm holding an insulator rod to do 1) sweeping line scans along the string, and 2) point measurements including points between insulator sheds. Artificial defects were also made present in the string. The commercial e-filed measurement systems were further developed to be automatic, giving real-time curve plotting of voltage outputs of EO probes. The plots showed noticeable differences between a healthy string and strings with defects. It showed the methodology works of using e-field as an index of insulator working conditions. Videos showing robot arm movement along the small-scale string in lab have been documented.   
 
Deliverable 3.1: report on small test cell results against FEA models (Dec-24) 
The report has a detailed description of the small test cell design with automatic data acquisition units developed by The University of Manchester. The probe movement modes were the sweeping line scan and point measurements as demonstrated in D2.2. The report summarises e-field profiles of test results in the small-scale insulator string, comparing a healthy string and strings with defects. It also included simulation parameters of 3D FEA models in COMSOL. Both the test results and FEA models confirmed the methodology of using e-field distribution changes to locate defects in the small-scale string. Measurements on two components of the e-filed vector were analysed, leading to conclusions 1) sweeping line scans by one probe measuring the e-filed component parallel to the central line of the string could recognise the defects, 2) when probes go closer between the insulator sheds in point measurement mode, the other component vertical to the string central line could accurately locate the faulty units. It also discussed how to decide the distance between the probes and the insulator shed edges.  
 
Deliverable 4.1: initial report on UAV requirements (Dec-24)  
This report layouts the concept of UAV operations in general (e.g., general system requirements, policy and regulatory requirements, security requirements and training requirements) and initial system architecture including the EO probes on a drone. It also includes as an attachment a bespoke EO probe system design provided by the probe manufacturer to fit UAV installation. The report also lists next steps, including using a smaller UAV to do initial tests for flight trials and to obtain the hover accuracy of the UAV and its performance with payload mounted.  
 
Deliverable 2.3: full scale test demonstration (Aug-25) 
 
This deliverable was done on a NG visit to The University of Manchester on 19-20/02/2025. A 400 kV full-scale glass cap and pin insulator string comprising of 24 units was set up in the HV lab. The probes were attached to an insulator rod which moved with a MEWP (Mobile Elevating Work Platform). Sweeping line scans along the string were conducted.   The voltage output plotting of EO probes showed noticeable differences between a healthy string and strings with defects. It confirms the methodology of using e-field as an index of insulator working conditions in a full-scale string. Videos showing the MEWP movement along the full-scale string in HV lab have been documented. It must be highlighted that a MEWP was used in place of a Drone, due to the complexities of flying a drone within a shielded high voltage lab. The lack of a GPS signal meant the sensor was mounted onto a remote-controlled MEWP to test the scanning protocols of the device. Future work will need to take into account the inherent uncertainty of the drone vibration. 
 
Deliverable 3.2 (partially): report on full scale test results & FEA models + UAV influence (Aug-25) 
 
D3.2 requires: Report summarising the results of the electric field measurements on the full-scale test rig against the FEA simulations. This deliverable will also report the UAV influence on the electric field. 
Below is a summary of the completed two parts, FEA models and the UAV influence assessment. 
FEA models of 24 insulator units forming a string have been built in COMSOL, with and without defects. UAV vibration, ±2.5 cm in the vertical direction, ± (2.5-5) cm on the horizontal plane, is simulated by multiple lines parallel to the string and a rectangular box of a certain dimension filled with points. FEA calculations lead to the conclusion - While an ideal straight-line measurement path parallel to the string over a distance of 6.5 cm between the line and glass shed edge could identify the defects, the variation in e-field profiles caused by drone vibration could misinterpret defect locations. Possible solutions included to hover a drone at target positions and then take the average of measurements as the e-field reading, and/or develop onboard accessories for drones to ensure a straight-line scanning of the string. 
 
 
Recommendations for further work 
As per original proposal, no further work required at this stage. 

Lessons Learnt

The electric field probe performance was tested on three different electrode geometries that provided distinct electric field profiles on which sensing performance can be evaluated. The fibre optic probes were used in experiments to measure electric fields within these electrode geometries, and hence test the sensitivity, usability and repeatability of e-field measurements. In parallel with experiments COMSOL FEA models of tested geometries were developed to obtain predicted localised electric field data for the considered geometries. The obtained COMSOL predicted e-field magnitudes were compared to the corresponding probe measured data for practical assessment of sensing performance. In all three geometries, the two probes yielded repeatable results and magnitudes. With increase in the voltage magnitude and associated electric field increase, the measurements formed a linear response for all electrode geometries. Differences between the COMSOL magnitudes and those recorded using the probes were observed, however these were associated with the placement of the probe within the electrode geometries, and variations within the actual COMSOL models. The results contained in this report provide sufficient evidence to deploy the sensors for the next stage of this project, which is the small-scale insulator string measurements. Detailed results of the sensor tests against electrode geometries are provided within the report/Document “D1.3 Report on sensor performance against standard electric field profiles, sensitivity and usability.  
 
Dissemination   
1) The project has its first conference paper accepted at IEEE ICD in July. 2024. The paper titled 'Electric Field Mapping by Electro-optical Probes in known Geometries under High Voltage'.  

2) The project has its first journal paper submitted to IEEE Trans. on Power Delivery in Jan. 2025, under review currently, titled ‘Condition Monitoring of Cap and Pin Insulator String by Electric Field Mapping using Electro-optical Probes with FEA Model Validation’ 

3)The project has two conference papers which will be shown at IEEE CEIDP in Sep. 2025.    ‘Evaluation of the Impact of UAV Vibration on Defect Recognition of a Cap and Pin Insulator String’ and ‘Condition Assessment of Composite Insulators-Based Electric Field Measurement Using Electro-Optic Probe’ .