Smart PV Workgroup of EITCI SESG

Artificial Intelligence assisted Smart Photovoltaics Workgroup of the EITCI SESG

The objective of the Smart PV Workgroup is in initiating ICT reference standards development combining recent progress in Artificial Intelligence based on Neural Networks and Deep Machine Learning with automated management of renewable energy generated in grid-connected photovoltaic (PV) systems along with their operation-and-maintenance (O&M) and their smart on-grid integration control.

A grid-connected PV system is generating electricity from the solar irradiation while being interconnected to the utility electric power grid. It generally consists of solar panels (PV modules), inverters, power conditioning units and grid connection equipment. Such PV systems range from small residential and commercial rooftop installations to large industrial-scale solar power-plants. Unlike stand-alone (off-grid) PV power systems, a grid-connected system usually does not include integrated batteries. Thus whenever the solar irradiation conditions admit it, the grid-connected PV system automatically supplies the excess power beyond consumption by the connected load, to the utility electric grid, turning a consumer into a prosumer and transforming the energy market to a highly distributed model.

Increasing automation of the PV solar power generated in-grid feeding control, operative optimization and maintenance has been recently dubbed smart PV.

The activity of Smart PV SESG WG aims at supporting international standardization efforts at a higher level of abstraction for the state of the art framework standard for Artificial Intelligence assisted smart control over PV systems in solar power plants, PV integrated industrial buildings and the prosumer residential homes PV installations.

The standardization efforts in smart PV assisted by AI in Neural Network models trained on a feedback loop of operational parameters is expected to add value to already developed digital and smart energy standards and support uptake of the smart energy technologies of crucial importance for the EU climate and energy policy framework, especially in view of the recent emphasis on joining Digital and Green agendas as two major pillars for the EU COVID-19 economic recovery.

H2020 StandICT action for AI assisted Smart PV

The StandICT will support initiation and coordination of an ICT reference standards development combining recent progress in AI automated management of renewable energy generated in grid-connected photovoltaic (PV) systems along with their operation-and-maintenance (O&M) and their smart on-grid integration control. The activity aims at supporting international standardization efforts at a higher level of abstraction for the state of the art framework standard for AI assisted smart control over PV systems in solar power plants, PV integrated industrial buildings and the prosumer residential homes PV installations. The standardization efforts in smart PV assisted by AI in Neural Network models trained on a feedback loop of operational parameters is expected to add value to already developed digital and smart energy standards and support uptake of the smart energy technologies of crucial importance for the EU climate and energy policy framework, especially in view of the recent emphasis on joining Digital and Green agendas as two major pillars for the EU COVID-19 economic recovery. The standardization activity will be hosted under EITCI SESG WG in cooperation with European Solar Network.

Increasing automation of the PV solar power generated in-grid feeding control, operative optimization and maintenance has been recently dubbed smart PV.

Magnitude of various PV modules and inverters equipment producers develop their own systems of automated O&M and control processes. Many solar modules producers embed electronics into PV modules. Such systems (smart modules) enable maximum power point tracking along with monitoring of performance data for fault detection at module level. Some of these systems make use of power optimizers to maximize generated power outputs. With recent PV advancements related electronics can compensate e.g. for shadows falling partially on a section of a solar module causing drop of electrical output of one or more strings of cells, but not zeroing the output of the entire module. A smart PV system should automatically control all its sophisticated operation parameters, including central or module-level MPPT, discover, diagnose and neutralize faults, hence improving its total efficiency, lowering O&M costs and increasing revenues. Main features of smart PV systems are automation, digitization and intelligence, optimally based on latest developments in AI applications (neural-networks big data learning comprising constant feedback input of all operational parameters of PV systems and their on-grid interconnection to AI enabled management system).

The Smart PV WG of EITCI SESG aims at supporting international standardization efforts at a higher level of abstraction for the state of the art framework standard for deep-learning NN AI assisted smart control over PV systems in solar power plants, PV integrated industrial buildings and the prosumer residential homes PV installations. The proposed Smart-PV standards development effort in a newly organized SDO Smart PV Workgroup hosted under European IT Certification Institute EITCI jointly with the European Solar Network ESN will conceive two Request for Comments documents that will be iterated among engaged experts and disseminated to other international SDOs active in the area of Smart Grids and Smart Metering standards with focus on solar power. The newly proposed smart PV standard will aim at systemizing conceptual architecture and implementation specification in software domain to define compatibility requirements between interfaces of PV modules and their associated electronic equipment control systems with inclusion of AI and cloud technologies. It aims in filling gaps in general smart-grid uniform communication standards mainly pursued by international SDOs in this field. The standardisation efforts in smart PV assisted by AI in neural network model trained on feedback loop of operational parameters is expected to contribute to growing digital energy standards inventory and support uptake of the smart energy technologies of crucial importance for the EU climate and energy policy framework, especially in view of recent emphasis on joing digital agenda and green agenda as two major pillars for the EU COVID-19 economic recovery.

The relevance of the action in a context of current European and international challenges is in a direct answer to the EU Rolling-Plan 2020 for ICT standardisation overviewing the needs for digital standards in support of EU policy for Smart Grids and Smart Metering in focus on smart PV solar systems. Accordingly with the EU Rolling-Plan 2020 ICT standards in energy are needed to cover smart grid management, grid-balancing and interfacing with millions of new renewable sources in particular optimizing efficiency in complex processes of renewable energy systems control. These standards mainly focus on uniform communication and cybersecurity protocols (providing plug-and-play compatibility for new devices entering the grid, from renewable sources to electric cars or other smart devices and IoT enhancing the smart homes, buildings and cities of the future).

The current dynamic EU energy transformation is driven by two main factors:

  1. The energy systems becoming clean (i.e. environmentally neutral accordingly with goals of the EU climate and energy framework and the European Green Deal policy) based on renewable and consumer-centric sources, primarily in form of solar power.
  2. The ongoing digital/smart transformation of the energy and electrical grid sectors.

The first factor is due to the EU energy policy encouraging stakeholders to adapt to an increasing number of means of generating electricity from a variety of renewable energy sources with minimizing environmental impact (clean energy transformation). The key policy milestones for this transformation are the EU's energy and climate targets for 2030 which emphasize Europe's leading role in the global fight against climate change. These 2030 EU climate and energy framework targets include at least 40% EU domestic reduction in greenhouse gas emissions compared to 1990 (with an increased ambition to 55% reduction as a part of the European Green Deal of September 2020), at least 32% share of renewable energy consumed in the EU, at least 32,5% improvement of energy efficiency and an electricity interconnection targeted at 15%. In this context both the PV systems and the electricity networks are of key importance. In 2012 electricity represented 22% of the EU's energy consumption with renewables accounting for a share of 24% of gross production (with ca. 3% increase on 2011, while reaching as high as 30.2% in 2016 and expected to grow up to 55% in 2030, correspondingly with the 2030 energy and climate goals and the Paris Agreement). Furthermore the consumer position in the energy value chain has considerably changed. The energy consumer can now easily become a prosumer, deploying grid-connected renewable energy source (e.g. a PV system), feeding the surplus of the generated energy into the utility grid. For this end with smart optimization of energy efficiency the digital and energy technologies need to overlap taking advantage of most recent developments in big data enabled AI control methods, smart homes and cities applications, energy intelligent products, the IoT, 5G networks, etc. It is for a reason that the EU COVID-19 strategic response is summarized in prioritizing two pillars: the single energy market and the digital single market combined as strongly interdependent and being both critical to the policy of the EU. This is where the second factor of EU energy transformation through smart (AI assisted) digitization is pronounced.

The Smart PV WG targets a specific sector of this outlined in-demand technical standards of smart PV systems assisted by feedback loop trained neural networks based AI. An important concept for the proposed standards is defining a common cloud-based platform specification for distributed Smart PV operational data aggregation that will enable NN deep-learning not only on individual operative systems but also on the whole ecosystem of AI enabled Smart PV devices (with properly addressed security and privacy issues).

Smart PV Workgroup coordination

The coordinator of this WG is Agnieszka E. Rządkowska, a co-founder and chair of the European Solar Network seated in Bruges (membered by over 300 experts and several industrial/research institutions in solar energy), Member of the Board of the International Solar Energy Society seated in Freiburg (founded in 1954 joining 2000+ solar energy experts from 110+ countries), Member of the Editorial Board of the joint Elsevier & ISES Solar Energy Advances Journal and International Policy Director for Smart Energy Systems at EITCI Institute (joining 3000+ ICT experts since 2008 working on digital standards and certifications). Academically she is affiliated with her Alma Mater, University of Wrocław, Poland.

Agnieszka Rządkowska has 15 years of professional experience in both digital and energy EU policy and standardisation, with a research experience in AI assisted analytical modeling of energy systems and with a main focus on the solar power. She is an author/co-author of circa 20 publications and 3 books (her newest book is the 2020 "Europe - Energy - Climate: The quest for the clean energy transition in the EU"). In terms of expertise towards the current proposal, the applicant has been contributing to AI standards for smart energy development in EITCI since 2008 (with dedication to multi-agent AI systems applied to international energy relations simulations and decision-making support on state-actors-level energy-mix shaping, upon coordinating dedicated National Research Center scientific grant budgeted at €150k between 2012-2018). Subsequently the applicant participated in establishing a cooperation between ESN and EITCI resulting in a joint Smart Energy Standards Group SESG organized in 2019 (one of statutory goals pursued by ESN is in technical standards drafting and development towards increasing rate of smart solar energy deployment and its integration with buildings, transport infrastructure and industry). For this end the applicant upon her AI digital standardization engagement at EITCI coordinated participation in Google’s Data for Change NGOs support program and secured a grant of €40k for use of Google’s Cloud AI resources based on the TensorFlow platform to research neural-networks enabled smart energy systems (which was further extended to an annually-perpetual €100k dissemination grant, planned to be used in dissemination of the currently proposed standard of the AI enabled Smart-PV after its acceptance and publishing). In her most recent engagement the applicant has been invited to the organizing committee of the EuroSun 2020 Conference in Athens as a chair of the theme Renewable Energy Strategies, Policies, SfF where she has addressed development of international smart-energy standards as one of main energy transformation drivers. The proposal is a direct extensions of applicant’s activities aimed at supporting smart energy and smart grids international standard setting efforts focused on smart-PV systems enabled by applicant’s expertise joining digital and energy fields for its successful coordination.

Impact of Smart PV WG initiated standardization action

The PV solar systems are deployed in many different configurations with regard to their relationship to electronic power inverter systems, external electric grids, battery packs, or other electrical loads. Their output efficiencies are conditioned not only by solar cell technology, but ultimately also by a real multitude of external and internal factors ranging from environmental (atmospheric transmittance of irradiation, sun position, temperature, etc.) to operational (different load profiles of electric loads, on-grid integration parameters and demand situation, etc.). Regardless of the destination of the PV generated power, efficiency of power transfer from the solar panel (module) depends on the amount of sunlight falling on the solar panel, the temperature of the solar panel and the electrical characteristics of the load. Variation of these conditions affect the load characteristic giving the highest power transfer efficiency. Hence the load characteristic should be optimized to keep the efficiency of the system at its highest power transfer profile (so-called maximum power point tracking, MPPT). MPPT electrical circuits are designed to self-adaptively-generate arbitrary loads to the PV cells and convert the voltage/current/frequency to suit other devices or systems. Other factors that benefit of AI assisted optimization of operative parameters are complex relationship between PV systems operating temperatures and optimizing efficiencies with proper resistance loads in the I-V curve (power output) sampling.

The complex interplay of many external and internal factors affecting the overall efficiency of the grid-connected PV system calls for modern optimization methods proven most effective when based on deep-learning neural-networks AI model.

Automation aims at eliminating manual operations and in PV systems is mainly achieved with technological advancement of PV cells and inverters to be prone to technical failures (which drive costly manual O&M). Digitization of PV systems refers to centralized operations parameters control in PV management system, as well as computable detections on strings along with monitoring of the components operational data (at lowest granularity possible). Advanced power line communication (PLC) and wireless communication technologies are used for intra-array transmission and intra-power-station communication. For instance high-precision sensors are used for string-fault detection (with high frequency differential compensation algorithms smart PV systems accurately monitoring two-dimensional information of the string voltage and current in real time, automatically detecting failures). Intelligence of smart PV systems in proposed new technical standards is achieved with using big data analysis to enable AI assisted efficiency optimization exceeding fault-detection by operative parameters control. The proposed standards are be based on active mining of PV systems components low-granulated-data operational parameters together with prevision sensor reading external conditions (including transmittance, irradiation angle, temperature, devices energy load profiles, on-grid load-demand situation) combined with neural network deep-learning methodologies. The feedback loop will train NNs with time of operation to optimize overall efficiency. The standards refer to a higher layer of devices operation interface above the hardware components which partially automate operational modes of the PV systems, but only if managed jointly in view of the sum of all factors can lead to the peak efficiency. The AI assisted control algorithms would enable the smart PV power installation to better adapt to changes of grids and the varying grid load profiles.

The answer to this purpose is in the current action to initiate international work on technical standardization of Smart-PV assisted by deep-learning in neural-networks (also in the cloud model), simultaneously supporting Europe’s position on international SDOs/SSOs forum and its interests in future uptake of this technology. Europe's Agenda 2020 asseses that standardisation is important instrument for innovation adoption (COM/2010/2020) and emphasises relationship between R&D projects and standardisation activities that boost impact of the results and their market uptake (COM(2018) 26).

In 10.2014 the CEN/CENELEC/ETSI's Smart Grid Coordination Group (SG-CG) successfully completed requirements of the EC M/490 mandate, with industry representatives confirming their will to take over and implement the results of the Expert-Group-1 work on first iteration of the Smart Grid standards. Consequently, EG1 of the Smart Grids Task Force assessed in 2016 the interoperability, standards and functionalities applied in the large scale roll out of smart-energy metering in Member States and in particular the status of implementation of the required standardised interfaces, along with EC recommended functionalities related to the provision of information to consumers (summarising report was published in 10.2015). Further coordination of standardisation efforts related to Smart Meters was due to the Smart Meters Coordination Group (SM-CG) established under the M/441 mandate. The SM-CG has returned the reference architecture (TR-50572) and an overview of technical requirements, continuing to liaise with its successor CG-SEG (since end of 2016, the CEN-CENELEC-ETSI Smart Energy Grid Coordination Group took over and cooperates with the EC-SGTF). In 09.2017 EC issued a proposal for a regulation on ENISA on Cybersecurity certification (Cybersecurity Act) as a voluntary mechanism framework enabling creation of individual EU-wide certification schemes (with a scheme indicating a specific product/service, an assurance level and a standard for evaluation). Such schemes are now developed to verify security properties of digital energy systems. The EC fostered conceiving a common interoperability language SAREF - a standard of ETSI and OneM2M. CEN-CENELEC-ETSI is endowed to further align SAREF with the data models developed at ISO and IEC. These are initial steps to enable smart-energy grid and its adaptive demand-response operation mode.

The standards of the current action will mainly provide an added value as extensions of the CENELEC / IEC-TC CLC/TC-82 (Solar photovoltaic energy systems) and the CLC/TC-57 (Power systems management and associated information exchange) for power systems control equipment and systems including EMS (Energy Management Systems) and SCADA (Supervisory Control And Data Acquisition). Furthermore they will also build on CLC/TC-57 in providing amendments to the ENs on (Communication networks and systems for power utility automation – EN-61850), along with Application integration at electric utilities (prEN-61968), energy management system application program interface (EMS-API) (prEN-61970) and on Power systems management and associated information exchange (EN-62351). The added value will also address the CEN-CENELEC-ETSI Coordination Group on Smart Energy Grids, CG-SEG (incl. the M/490 and its iteration) and EN-IEC-61850 (Distributed Energy Resources).

SESG SmartPV Workgroup's StandICT Workplan

The initial activity will last 6 months, starting 1st March 2021 and will be divided into 2 phases.

Phase I – standards drafting as RFC documents — COMPLETED (MAY 2021)

During the first 3 months (starting 1st March 2021 and finalized on 31st May 2021) AI Assisted Smart PV Standards drafting will take place in a form of 2 Request for Comments documents on:

  1. AI Assisted Smart PV Conceptual Framework (Definitions, Architectures, Use Cases)
  2. AI Assisted Smart PV Technical Specification of Processes and Devices

The draft standards for the Smart AI Assisted Photovoltaic Systems – Conceptual Framework (Definitions, Architectures, Use Cases) aim to combine recent progress in Artificial Intelligence with improving performance of PV installations on many planes. Some of these account for automated management of solar energy generated in grid-connected photovoltaic (PV) systems along with their operation-and-maintenance (O&M) and their smart on-grid integration control, other related to AI assisted methods in solar cells design and manufacturing (for optimized efficiency), while other relate to AI aided mapping of solar irradiation in low-data available regions. The standardization effort in AI assisted smart PV aligns with the strategy of the European Union joining Digital and Green agendas as two major pillars for the COVID-19 economic recovery in the EU and is a part of the EU funded standardization action under the H2020 StandICT programme coordinated by the author and hosted by the Smart Energy Standards Group of the European Information Technologies Certification Institute (EITCI SESG) in cooperation with the European Solar Network. The reference standard aim to contribute to one of the four primary objectives of the European Green Deal, i.e. to achieve a fully integrated, interconnected and digitalized EU energy market by increasing research oriented towards technical reference standardization aimed at consolidation of the expert community and the technology uptake.

Phase II – RFCs reiteration, establishing of a dedicated Smart-PV-SESG WG, acceptance, publication and dissemination of the AI Assisted Smart PV Reference Standards

Upon the second part of the project, starting 1st June 2021 and lasting for the subsequent 3 months until 31st August 2021, the RFC documents will be disseminated with relevant SDOs workgroups. The dedicated Smart-PV workgroup will be established under coordination of the applicant and hosted by EITCI SDO (it is planned that Smart-PV WG will be a cross-SDO WG organized as a part of the Smart Energy Standards Group SESG, established at EITCI in 2019 with joint participation of the European Solar Network).

Following establishment of a dedicated Smart-PV WG the RFCs will be reiterated among its members upon a process of correcting and extending contributions to be finally vote-accepted as published Reference Standards before the end of the project. The accepted RSs will be further circulated with the relevant SDOs workgroups to work of which the proposal aims to contribute in initiating cross-SDOs cooperation towards adopting of standards in Smart-PV assisted by deep-learning AI in system operational-parameters feedback loop.

In particular the Request for Comments documents will be distributed beyond the established cross-SDOs WG to other relevant WGs of international SDOs/SSOs, including at least the CENELEC / IEC workgroups (developing international standards in collaboration with the IEC): CLC/TC 82 (Solar photovoltaic energy systems) and the CLC/TC 57 (Power systems management and associated information exchange), for power systems control equipment and systems including EMS (Energy Management Systems) and SCADA (Supervisory Control And Data Acquisition), also providing amendments to the ENs on Communication networks and systems for power utility automation - EN 61850 series) and covering the Application integration at electric utilities (prEN 61968 series), energy management system application program interface (EMS-API) (prEN 61970 series) along with Power systems management and associated information exchange (EN 62351 series), as well as the EN IEC 61850 series (Distributed Energy Resources), EN IEC 62746, EN IEC 61689-5, EN IEC 62325 with IEEE Std 2030.5-2013 and OASIS OpenADR – with a goal of inviting contribution to standards development aimed at setting internationally trusted industry specifications for Smart-PV AI assisted systems.

EITCI Smart-PV SESG WG invitations will be directed to relevant SDOs WGs’ members and also relevant industry and academic experts. Upon invitations to Smart-PV WG there is planned reaching a number of at least 35 experts actively participating on iteration of the Smart-PV RFCs towards RS acceptance, publication and dissemination (the remaining 3 months of the project implementation will be spent on reiteration of the RFCs within the Smart-PV SESG WG ending with vote-acceptance of the final Reference Standards specifications, along with publishing and dissemination taking place before 31st August 2021).

The proposed workplan will result in stimulating further development of Smart-PV AI enabled standards in joint efforts of the international SDOs/SSOs. Upon the accepted Reference Standards distribution to relevant SDOs/SSOs any members of related technical standards committees or workgroups interested in Smart-PV standards development towards reaching final industrial specifications will be invited to join the established dedicated Smart-PV WG to participate in further post-proposal activities.

EITCI Smart-PV-SESG WG invitations will be directed to related smart energy technology international SDOs/SSOs WGs members as well as to leading smart energy researchers and relevant engineering experts. Upon invitations to Smart-PV-SESG WG there is planned expanding general SESG membership by at least 20 new experts participating in reiteration of the AI Assisted Smar PVt RFCs towards RS acceptance, publication and dissemination (the remaining 3 months of the project implementation will be spent on reiteration of the RFCs within the WG ending with vote-acceptance of the final Reference Standards specifications, along with publishing and dissemination taking place before 31st August 2021).

The proposed workplan will result in stimulating further development of AI Assisted Smart PV standards in joint efforts of the international SDOs/SSOs. In particular the proposed activity will be coordinated in already initiated cooperation with CEN-CENELEC, supporting EU-stemming international standards in smart energy technologies. The purpose of this cooperation will be to ensure interaction between relevant EU stakeholders upon joint efforts in international standardisation in area overlapping between smart grids and smart PV to support prospects of global industrial uptake of AI assisted smart PV systems. The Smart-PV-SESG Reference Standards will be implemented with CEN-CENELEC as CWAs as part of the post-proposal results dissemination activities. Upon the accepted Reference Standards distribution to relevant SDOs/SSOs any members of related technical standards committees or WGs interested in smart photovoltaics standards development towards final industrial specifications will be invited to join dedicated Smart-PV-SESG to participate in further post-proposal activities.

The activity of the Smart-PV-SESG hence involves the two-step procedure of the Request for Comments on technical specifications reiterations and acceptance of the Reference Standards. The initial AI assisted Smart-PV technical specifications drafting concluded on 31st May 2021 is a basis for the RFC (Request for Comments) publication and corrective iterations. The final phase begins with June 2021 when these reiterated RFC drafts upon possible extensions and corrections become candidates for Smart-PV-SESG Reference Standards upon the SESG Members vote lasting until 31st August 2021. The vote will conclude in either acceptance or forwarding for further correction of the Reference Standards considered to be accepted by the Smart-PV-SESG Members.

The acceptance vote for Smart-PV-SESG Reference Standards will be based on the iterated RFC initial technical specifications drafts with ongoing corrections, that will include:

  1. AI Assisted Smart PV Conceptual Framework (Definitions, Architectures, Use Cases)
  2. AI Assisted Smart PV Technical Specification of Processes and Devices

Accordingly to the Smart-PV-SESG voting procedure only Smart-PV-SESG Members are entitled to participate in vote. If a Member would like to vote against acceptance of the RFC drafts as RS, such objection should be emailed along with indication of vote cast against at sesg@eitci.org. Upon the EITCI SESG proceeding rules, a lack of response is interpreted as a vote in favour of acceptance. All further corrections are still accepted during vote being sent also at sesg@eitci.org and are added to the improvements list for further reiterations of the Reference Standards, if admitted by the vote. If a Member would like to vote against accepting of the current drafts as RS, it musy be clearly indicated in the message, because proposing just the corrections is not counted as a vote against (rather the corrections are added to further improvements of the accepted RS).

During the 3 months before the vote the RFC drafts are to be circulated with the Smart-PV-SESG Members and distributed to the relevant Working Groups of SDOs, with an invitation to submit the comments on the drafts and join the Smart-PV-SESG activities, including further reiteration of the RFCs upon community collaboration ending with Reference Standards acceptance and publishing by the Smart Energy Standards Group under EITCI Institute as the hosting SDO. The accepted Reference Standards are then subject to continous development and improvements.

The EITCI Smart-PV-SESG activities are aimed at stimulation of international SDOs' own WGs with further iterations towards increasing collaboration and reaching of the consensus for adopted international standards in smart PV domain. The AI assisted smart PV standardization drafting process will be open and RFC drafts and final standards specification documents will be of open access (while some of the technically documented concepts may remain under IP protection, different competing solutions will be a part of consensus reaching process in the research and industry community towards smart PV technology standardization).

Return to EITCI Smart Energy Standards Group.