TDX-ASSIST Portuguese Demonstrator (CIRED 2019)

Simão, T., Gama, P., Louro, M., Carvalho, L., Glória, G., Pestana, R., Reis, F., Soares, T., Silva, J. (2019). “TDX-ASSIST: Beyond State of Art in TSO-DSO Interoperability – The Portuguese Demonstrator.” CIRED 2019, Madrid, 3–6 June 2019, Paper 872. Authors from EDP Distribuição (Portuguese DSO), REN (Portuguese TSO), INESC TEC, and R&D Nester. Open-access PDF via the CIRED conference repository.

TDX-ASSIST (“Coordination of Transmission and Distribution data eXchanges for renewables integration…”) was an EU Horizon 2020 project building ICT tools for scalable, secure TSO-DSO data exchange based on international smart-grid standards (CIM, IEC 61970/61968/62325). This paper covers only the Portuguese demonstrator. Sourcing note: the project’s literature-review deliverable, D1.1 “State of the Art – TSO-DSO Interoperability” (Dec 2017), is a separate, earlier document and is not the source of the technical results below — but D1.1 is the source for the grant number (774500) and coordinator acronym (BUL) in this paragraph, since the CIRED2019 paper itself doesn’t state them. The project’s full duration, exact EU funding amount, “Brunel University London” full name, and the specific claim that demonstrators ran in exactly Portugal/Slovenia/France (D1.1 actually lists five participant countries — Slovenia, Germany, Portugal, UK, France — for its country profiles, without confirming which had field demonstrators specifically) are not verifiable against either raw document on hand and should be treated as unconfirmed outside knowledge.

What the paper covers

Business use-case taxonomy. TDX-ASSIST classified TSO-DSO coordination needs into 11 business use cases (BUCs) spanning three timeframes — real-time system operation, operational planning (intraday/day-ahead/week-ahead), and long-term planning (year-ahead+). Examples: activation of DSO-connected resources by the TSO for balancing (BUC1), coordination of distributed flexibility in a shared-resource marketplace (BUC2), active power management for congestion by either operator (BUC3 DSO, BUC4 TSO), reactive power management for voltage control by either operator (BUC5 DSO, BUC6 TSO), and coordination of long-term network planning (BUC9). This is a clean, timeframe-based way to classify any TSO-DSO coordination need — a useful organizing structure distinct from Network Code on Demand Response‘s role-based (procuring/connecting/impacted SO) framework.

The “observability area” concept. The Portuguese demo is explicitly built around each operator defining an observability area — the part of the adjacent operator’s network it needs visibility into to do its own job (e.g., REN needing DSO-side load/generation forecasts to plan transmission operation). This is the same term the EU’s Network Code on Demand Response later formalizes as a mandatory DSO obligation (Title VII) — TDX-ASSIST’s 2018–2019 demonstrator work is an early operational precedent for a concept NC DR subsequently made regulatory. See TSO-DSO Coordination — The Central Design Problem‘s “Observability areas” section.

Real-time data exchange architecture. REN and EDPD (TSO/DSO) built a redundant, dual-server real-time link using ICCP (Inter-Control Centre Protocol — the standard SCADA-to-SCADA real-time data exchange protocol between grid control centres) connecting both companies’ SCADA systems. Non-real-time exchange (forecasts, oscillography, maintenance plans, short-circuit levels) ran over the same physical interface but a separate, security-segregated VLAN, using SFTP and web-services rather than ICCP.

Concrete exchange parameters demonstrated:

• Short-circuit power at the TSO-DSO physical interface: TSO computes and sends 30-minute-granularity values for the next 24h based on market/topology data; DSO adds its own three-phase short-circuit contribution at each interface bus and returns the aggregated total via web-services.
• Generation forecast at bus level, disaggregated by generation type (wind/hydro/PV/CHP) and market player type, across ~80 interconnection points on the Portuguese network — 15-minute sampling, 72-hour horizon, refreshed every 24 hours, DSO → TSO.
• Fault location: TSO’s existing tools (SRAO, gathering oscillography automatically; SAFIR, an incident-analysis application) relay fault-location and oscillography data to the DSO via ICCP (real-time) and web-services (detailed report), targeting a 5-minute delay for detailed fault-type/location information — improving the DSO’s ability to dispatch field teams to lines it owns but that are directly connected to TSO transmission bays.

Reactive-power coordination tools (New ICPF / New SOPF). Two tools — evolved from the earlier FP7 EvolvDSO project — compute how much reactive-power flexibility the DSO can contract from distributed generation to meet a TSO-required profile at the interface substation:

• New ICPF (Interval Constrained Power Flow): builds a 24h-ahead reactive-power flexibility map at interface substations, accounting for discrete control actions (switched capacitor banks, OLTC tap positions) and network limits. Computed by the DSO using an equivalent transmission-network model (so the DSO never sees TSO-private data), then sent to the TSO, which selects its desired reactive-power profile from the map.
• New SOPF (Sequential Optimal Power Flow): once the TSO has picked a profile, the DSO runs a two-stage stochastic optimization to contract the reactive-power flexibility (capacitor banks, OLTC, DER reactive production) needed to hold it, under renewable-generation uncertainty. Active-power DER dispatch is fixed in advance; only reactive-power flexibility is optimized, with active-power curtailment allowed but heavily penalized.

Validation status — important caveat. At publication, only the New SOPF had been tested, and only on a simulated academic network (an 11kV, 37-bus system with 1,908 consumers, adapted from an IEEE reliability test system, using Portugal’s actual reactive-power regulatory limits: tan φ = 0 off-peak, 0.3 peak, ±5% tolerance). Results: the tool kept tan φ within the required band across all wind-generation scenarios while limiting capacitor-bank switching to ≤4/day; the OLTC tap stayed at the same position throughout the full 24 hours in the SOPF case — a tighter result than the capacitor-bank figure, not identical to it. This is a methodology validation, not a field-measured demonstrator result — the paper explicitly frames both tools as “currently under development,” and no MW-scale field figures from the live Portugal/Slovenia/France demonstrators are reported here.

Relevance to this wiki

• TSO-DSO Coordination — The Central Design Problem — adds a concrete, standards-based (ICCP, IEC 62559/62913) international precedent for a real-time + non-real-time TSO-DSO data architecture, and a documented origin point for the “observability area” concept the NC DR later mandates. Portugal-specific and 2019-era — cited as architectural precedent, not Swedish practice.
• Relevant background for a possible future TSO/DSO coordination capability guide: the BUC taxonomy and the real-time/non-real-time channel split are both reusable framing devices.
• The New ICPF/SOPF reactive-power tools are a narrower technical thread (DSO-to-TSO reactive-power flexibility contracting) not closely tied to Sweden’s active-power-focused local flexibility markets; not pursued further here unless a reactive-power-specific need arises.