Recommending Mitigation Measures
Chapter 2 estimated the risk of a new connection and decided with a guaranteed rate of error whether it could be approved without a study, and Chapter 3 ordered the transformers of a network by the risk they are expected to carry as adoption grows. Ranked separately for the thermal and the voltage constraint, the transformers form two programmes of work, and these results identify where a network will need attention but not what attention it needs. A transformer that is expected to breach its limits requires a mitigation measure, and a distribution network operator can choose among several: settings of the customers’ inverters, a static limit on their export, the management of their charging, a change of the transformer’s tap and the replacement of the transformer by a larger one. The measures differ in cost, in the violations they relieve and in the customers on whom they place a burden. The choice of a measure for each transformer is the recommendation step of the chain of Chapter 1, whose potential outcomes, costs and regret this chapter applies to a real network, and the simulator provides what recommender systems seldom have, the outcome of every measure on every transformer.
This chapter recommends mitigation measures for the transformers of HPK11. It first establishes what each measure does on the network and what it costs, and the ratio of costs at which reinforcement becomes the measure of least cost. It then compares ways of recommending a measure: a rule, a policy learned within the violation patterns of Part 5, the nearest simulated transformers, learners fitted on a field record in which each transformer received a single measure chosen by the utility, and a recommender that verifies its proposals with a few network studies. It measures how the curtailment imposed by a measure is distributed among customers, by their position and by the demand clusters of Part 5. Chapter 3 ordered the transformers and the violations; this chapter chooses what is done about them, which completes the progression of this part from the planning of a connection to the operation of the network.
Mitigation Measures in Distribution Planning
The mitigation measures of a distribution network act on different causes of violation. The voltage rise caused by rooftop Photovoltaic (PV) can be limited at its source by curtailing the active power of the inverters, through a volt-watt response or a static export limit [1], or by absorbing reactive power through a volt-VAr response [2]; AS/NZS 4777.2 requires both responses of every inverter in Australia. The voltage can also be lowered at the transformer by its off-circuit tap changer, which a field crew sets a few times a year rather than adjusting under load, so that a change of tap shifts the voltage of every customer at every hour. In terms of the voltage rise of Equation 7.1, the export limit and the volt-watt response reduce the injected power \(P_q\), the volt-VAr response sets \(Q_q\) against it, and the tap change lowers the sending voltage to which the rise is added. The evening violations caused by Electric Vehicle (EV) charging are relieved by moving the charging to the night [3]. Reinforcement, the replacement of a transformer by a larger one, relieves thermal violations at every hour. Figure 9.1 shows where each measure acts and which violations it relieves. A measure that relieves one violation can worsen another, since the tap change that lowers the midday voltage also lowers the evening voltage.
The choice among these measures has been formulated as an optimisation over a network model and as the reuse of past solutions. Xu and colleagues applied case-based reasoning to coordinated voltage control, retrieving the control actions that resolved similar past states of a distribution network with distributed generation and verifying them online before applying them [4]. The potential-outcome framework of Chapter 1 adds a consideration that case-based reasoning leaves implicit. A record of past measures is a log of the utility’s own decisions, in which a measure was applied where the problem appeared most severe, so that the outcomes that were not observed are missing not at random. Schnabel and colleagues treated recommendations as treatments and corrected both the evaluation and the training of recommenders on such logs by inverse-propensity weighting [5]. A mitigation measure also distributes a burden among customers, and the volt-watt response curtails the customers at the end of a feeder more than those near the transformer, because their voltage rises first; Liu and colleagues demonstrated this effect on the networks of AusNet and compared curtailment schemes by household measures of fairness [6], and Gerdroodbari and colleagues proposed a decentralised volt-watt control that distributes curtailment more equally among inverters [7]. During operation, finally, violations begin while others are in progress, and process industries treat the corresponding problem as the rationalisation of alarm floods, in which alarms are ranked by priority;1 Parvez and colleagues ranked alarm floods by their similarity to historical ones [8]. The sections of this chapter take up these four subjects in turn, beginning with what each measure does on HPK11.
Outcomes and Costs of the Mitigation Measures
The outcome of mitigation measure \(a\) on transformer \(j\) is the potential outcome \(Y_j(a)\), the hours of violation that remain under the measure, and its uplift is the number of hours it removes, Equation 6.20. Every measure was simulated on every transformer of HPK11 with popo, over four seasonal weeks beginning on 10 January, 11 April, 10 July and 10 October, with PV and EVs at 60% and at 80% of the residential customers and the AS/NZS 4777.2 inverter settings, and with the adopters, arrays and charging sessions drawn from the same seeded generator, so that the outcomes differ only by the measure. The measures are no action; a static export limit of 5 kW for every array; managed charging, which delays the evening charging of every vehicle to a start drawn between 22:00 and 02:00 while preserving its power profile; a change of the tap of every LV transformer by one step, which lowers its secondary voltage by approximately 2.5%; and the replacement of every LV transformer by the next standard rating. A further simulation without volt-watt response or export limit provides the generation that the measures curtail. The results are stated at 80% adoption, summed over the four weeks, and the section on adoption below repeats the comparison at 60%.
The measures acted on the violations that Figure 9.1 attributes to them (Table 9.1). The export limit reduced over-voltage from 3,526.5 to 3,011 transformer-hours and overload from 3,003.5 to 2,563, and curtailed 244.9 MWh of generation, so that it relieved 15% of the over-voltage at the price of a curtailment 32 times that of the volt-watt response. The tap change reduced over-voltage to 1,470.5 hours, a reduction of 58%, and raised under-voltage from 82.5 to 447 hours: lowering the sending voltage of Equation 7.1 lowers the voltage of every customer at every hour, so that the margin gained at midday is lost in the evening. Managed charging reduced under-voltage to 76.5 hours and overload to 2,944, small reductions because evening under-voltage was rare on HPK11 in every season. Reinforcement reduced overload to 1,528 hours, by 49%, and raised over-voltage by 7%, to 3,776 hours. The increase appeared in each of the four weeks and at both levels of adoption. The additional hours began between 09:00 and 15:00, in the hours of export, and the volt-watt response curtailed 13.1 MWh under reinforcement against 7.6 MWh with no action, so that the customers’ voltage was higher at midday. The increase follows from the volt-VAr response of the inverters. When the voltage is high, the inverters absorb reactive power, which the transformer imports while it exports active power, so that in Equation 7.1 the active term \(R P\) raises the voltage and the reactive term \(X Q\) lowers it. The LV transformers of HPK11 have reactances between 3.78% and 4.88% and resistances between 1.1% and 1.47% on their own rating, and with a reactance between 2.6 and 4.4 times the resistance the reactive term prevails, so that the voltage falls across the transformer even in the hours of export. The upgraded transformer keeps its percentage impedance, its impedance in ohms falls with the larger rating, and the fall across it is reduced, which raises the voltage of every customer it supplies. A comparison of one July day with and without reinforcement confirmed the mechanism: the transformers absorbed reactive power at midday, and the rise of the busbar voltage under reinforcement followed the reduction of \(R P + X Q\) transformer by transformer. The higher voltage in turn engages the volt-watt response, which explains the additional curtailment. Reinforcement thus interacts with the reactive-power settings of the inverter standard, and on a network whose transformers carry the reactive power absorbed by volt-VAr inverters, a larger transformer raises the voltage at midday rather than lowering it. Reinforcement therefore relieves the thermal constraint but not the voltage constraint, and where both bind a voltage measure is required in addition. The volt-watt response alone curtailed 7.6 MWh, and over-voltage was almost unchanged in its absence, at 3,530 hours, so that the volt-watt response at the settings of the standard curtails little and leaves the over-voltage of the network largely unresolved.
Pitfall: a measure judged by the violation it targets
A mitigation measure is usually proposed for one kind of violation, and an evaluation that counts only that kind credits it with its full effect. On HPK11 the tap change removed 58% of the hours of over-voltage and multiplied the hours of under-voltage by five, and reinforcement removed half of the overload and added 7% to the over-voltage. An evaluation restricted to the targeted violation would have overstated the benefit of both measures and missed the interaction of reinforcement with the volt-VAr response. The outcome \(Y_j(a)\) of Equation 6.20 therefore counts every kind of violation, and a study of a measure reports its effect on each kind separately before they are combined in the cost of Equation 6.27.
A measure is chosen by its total cost, Equation 6.27: the annualised capital cost of the measure, the energy it curtails at a value \(c_{\mathrm{e}}\) per unit, and the violation that remains at a penalty \(\mu\) per hour. Absolute costs differ between utilities by an order of magnitude with the rating of the transformer,2 and the costs are therefore expressed relative to the annualised cost \(c_{\mathrm{r}}\) of reinforcing the transformer,
\[ \tilde C_j(a) = \frac{C_j(a)}{c_{\mathrm{r}}}, \tag{9.1}\]
with the results reported as functions of the ratios \(c_{\mathrm{e}} / c_{\mathrm{r}}\) and \(\mu / c_{\mathrm{r}}\). Here the capital costs relative to reinforcement are 0.05 for a tap change and 0.10 for managed charging, and curtailed energy is valued at 0.0002 per kWh. The penalty at which reinforcement becomes the measure of least cost for transformer \(j\) is its switching ratio,
\[ \mu_j^{\star} = \min \bigl\{ \mu : C_j(\mathrm{reinforcement}) \le \min_{a} C_j(a) \bigr\}, \tag{9.2}\]
the weight that a utility must give to a violation hour before it prefers investment to an operational measure.
The measure of least cost depended on the penalty assigned to a violation hour (Figure 9.2). At the lowest penalty examined, \(10^{-4}\) reinforcement annuities per violation hour, no action was the measure of least cost for 65% of the transformers and the tap change for the remaining 35%, on which the curtailment and the violations that the tap change avoids outweighed its capital cost. At a penalty of 0.01 the tap change was the measure of least cost for 98% of the transformers. Reinforcement became the measure of least cost for 7 of the 43 transformers within the range examined, with switching ratios, Equation 9.2, between 0.028 and 0.18, and for 16% at a penalty of one annuity per hour, at which the tap change remained the measure of least cost for 79% and the export limit for 5%. Managed charging was not the measure of least cost for any transformer at any penalty, because evening under-voltage amounted to 82.5 transformer-hours against 3,526.5 hours of over-voltage. At the central penalty, the median of the seven switching ratios, 0.079, the tap change was the measure of least cost for 39 of the 43 transformers and reinforcement for 4. On HPK11 the tap change is the measure of least cost for most transformers over a wide range of penalties, and the decision that requires a recommendation is which few transformers need reinforcement instead.
Recommending a Measure for Each Transformer
A recommendation must be made before the outcomes of every measure are known for a transformer, and the recommenders differ in what they draw on. Not every measure applies to every violation: with \(\mathcal{A}_{\kappa}\) the set of measures that relieve violations of kind \(\kappa\), given by Figure 9.1, the recommendation for a transformer whose violations are of kind \(\kappa\) is restricted to
\[ \hat a_j \in \mathcal{A}_{\kappa} \cup \{0\}. \tag{9.3}\]
The violation-pattern rule of Part 5 (Violation Patterns) applies this restriction directly. It assigns each transformer the measure that relieves its dominant violation with no action: the export limit for over-voltage, managed charging for under-voltage and reinforcement for overload. The violation pattern-cluster policy uses the same grouping as clusters in the sense of Chapter 2, and assigns each transformer the measure of least mean cost among the other transformers of its group, a recommendation learned within each cluster rather than fixed by a rule, as cluster-based collaborative filtering recommends to a user the items preferred by its cluster [9]. The case-based policy of Chapter 1 estimates the cost of every measure for a transformer as the mean cost on its five nearest transformers, described by their rating, customers, impedance, demand and hours of each violation with no action, and selects the least.
Three recommenders learn instead from a field record, generated as in Chapter 1 by a utility that applied reinforcement with a probability rising from 0.15 to 0.85 with the overload of the transformer and an operational measure otherwise, and redrawn 200 times. The T-learner of Chapter 1 fits, for each measure, a regression of the cost on the features of the transformers that received it. Matrix factorisation, the principal method of collaborative filtering, estimates the missing entries of a matrix of users and items from its observed entries [10]. With transformers as users and measures as items, it estimates the cost of transformer \(j\) under measure \(a\) as
\[ \hat C_{ja} = \bar C + \theta_j + \theta_a + \mathbf{u}_j^{\top} \mathbf{v}_a, \tag{9.4}\]
with the mean cost \(\bar C\) of the record, a baseline \(\theta_j\) for the transformer and \(\theta_a\) for the measure, and latent factors \(\mathbf{u}_j\) and \(\mathbf{v}_a\). It is fitted on the observed entries \(\mathcal{O}\) of the record, without weights and, in a second variant, with each entry weighted by the inverse of the propensity with which the utility selected that measure for that transformer [5],
\[ \min \sum_{(j, a) \in \mathcal{O}} \frac{\bigl( C_{ja} - \hat C_{ja} \bigr)^2}{\rho_a(\mathbf{x}_j)} + \gamma \bigl( \theta_j^2 + \theta_a^2 + \lVert \mathbf{u}_j \rVert^2 + \lVert \mathbf{v}_a \rVert^2 \bigr). \tag{9.5}\]
The last recommender verifies its proposals before it decides. Reasoning models in machine learning improve their answers by generating several candidates and retaining the one that a verifier scores highest, with a gain that grows with the number of candidates examined [11]. A utility can apply the same principle with its network model: the \(N\) measures of least estimated cost are simulated, and the measure of least simulated cost among them is selected,
\[ \hat a_j^{(N)} = \arg\min_{a \in \mathcal{T}_N(j)} C_j(a), \tag{9.6}\]
where \(\mathcal{T}_N(j)\) is the set of the \(N\) measures of least case-based estimate for transformer \(j\). Its regret, Equation 6.28, decreases to zero as \(N\) reaches the number of measures (Figure 9.3). Every recommender is evaluated at the central penalty by its mean regret, in units of the reinforcement annuity, and by the share of transformers to which it assigns the measure of least cost.
The recommenders differed by more than an order of magnitude in regret (Table 9.2). The violation-pattern rule assigned the measure of least cost to 9.3% of the transformers, with a mean regret of 2.93 reinforcement annuities, because it assigns the export limit to the 24 transformers dominated by over-voltage and reinforcement to the 19 dominated by overload, whereas the tap change relieves both at lower cost on this network; the restriction of Equation 9.3 is correct in the measures it admits but does not rank them. The violation pattern-cluster policy, which uses the same grouping but learns the measure within each group, and the case-based policy each assigned the tap change to every transformer, and each gave the measure of least cost to 90.7% of the transformers, with a mean regret of 0.065. Both therefore recover the dominant measure of the network, and neither identifies the four transformers that require reinforcement, since a measure of least mean cost among similar transformers is the measure of the majority. The T-learner, fitted on field records in which reinforcement was applied where the overload was most severe, reached 57.7%, with a mean regret of 1.35: in these records each operational measure, the tap change among them, is observed on about one transformer in eight, so that its cost is estimated from about five transformers, a weak form of the overlap condition of Equation 6.24.
Matrix factorisation did not exploit the field record. With one observed measure for each transformer, the baseline of the transformer absorbs its single entry, the latent factors of Equation 9.4 are not identified, and the fitted model assigns every transformer the measure of least mean cost in the record, which was the measure of least cost for 69.4% of the transformers. Weighting by inverse propensity, Equation 9.5, corrects the mean for the utility’s selection but not for the absence of entries, and reached 63.9%, below the unweighted model, since the weights increase the variance of a mean estimated from few entries. Collaborative filtering therefore requires several measures recorded for each transformer, a condition that field records seldom satisfy, and a utility with a record of one measure per asset gains more from a network model than from a factorisation of its record.
Verification removed the regret that remained after the case-based estimate. Simulating the two measures of least case-based estimate and selecting the one of least simulated cost assigned the measure of least cost to every transformer, with zero regret, because the second measure proposed for the four transformers that require reinforcement was reinforcement. Two studies for each transformer thus recover the full value of studying every measure, in agreement with the decrease of Figure 9.3, and the case-based estimate serves to select which measures to study rather than to decide. For a utility, the implication is that a model of the network is most useful when it verifies a short list of measures for each asset, and the cost of the programme of studies falls from five to two simulations per transformer.
The conclusions held at 60% adoption. Over-voltage with no action fell to 2,693 transformer-hours and overload to 1,053, a third of its value at 80%, so that overload grows faster with adoption than over-voltage. The tap change was the measure of least cost for 42 of the 43 transformers, the violation-pattern rule assigned the measure of least cost to 2.3% of them, the case-based policy to 97.7%, and two verified studies to 97.7%. At the lower adoption, fewer transformers require reinforcement, the dominant measure is more nearly the measure of every transformer, and a policy that recommends the measure of the majority loses less. The value of verification therefore rises with adoption, as the transformers that need reinforcement become more numerous. The measure of least cost, however, need not be the measure that treats customers equally, which the next section examines.
Equity of Curtailment
The burden of a measure on customer \(i\) is the share of the customer’s generation that it curtails,
where \(E_i^{\mathrm{f}}\) is the energy generated by the customer’s array without volt-watt response or export limit and \(E_i(a)\) the energy generated under measure \(a\). The inequality of the burden across the \(n\) customers of a network is measured by the Gini coefficient of Equation 6.29 and by Jain’s index,
\[ J(\mathbf{b}) = \frac{\bigl( \sum_{i} b_i \bigr)^2}{n \sum_{i} b_i^2}, \tag{9.8}\]
which equals one when the burden is shared equally and \(1/n\) when a single customer bears all of it [6]. Figure 9.4 contrasts the two forms of curtailment on a stylised feeder: the volt-watt response curtails the customers farthest from the transformer, and a static export limit curtails every array alike.
On HPK11 the curtailment was distributed unequally among customers (Table 9.3). Without any measure, the volt-watt response required by the inverter standard curtailed 0.13% of the generation of the average customer, a small burden but a concentrated one, with a Gini coefficient of 0.92 and a Jain’s index, Equation 9.8, of 0.06, close to its value when a few customers bear all of the burden. It fell on the customers at the highest impedance, from less than 0.01% in the two nearest quintiles to 0.41% in the farthest (Figure 9.5), in agreement with the findings of Liu and colleagues on other AusNet networks [6] and with the voltage rise of Equation 7.1, which grows with the impedance between the customer and the transformer. The export limit curtailed 2.9% of the generation of the average customer and distributed it evenly across positions, between 2.8% and 3.0% in every quintile of impedance. Its burden was nonetheless unequal, with a Gini coefficient of 0.76, because a static limit of 5 kW curtails only the arrays whose output exceeds the limit, and it therefore transfers the burden from the customers far from the transformer to the customers with large arrays.
Across the demand clusters of Part 5 the curtailment was nearly even, between 2.8% and 2.9% of generation under the export limit and between 0.11% and 0.14% with no action, consistent with the finding of Chapter 2 that the outcomes set by position are independent of the clusters of household demand. The export limit was modelled as a limit on the output of each array rather than on its net export, so the curtailment of households with high daytime demand is overstated. The burden of a measure is thus set by where a customer is connected and by the size of its array, and an equitable programme must be designed against these two quantities rather than against the type of household.
Further Applications of Recommendation on the Network
The chain of this chapter, the simulated outcome of every action on every unit, a cost that combines them, a recommender evaluated by its regret and verified by a few studies, and an equity measure of the burden it imposes, applies to decisions other than the choice of a measure for a transformer. Three decisions of a distribution network operator follow it directly.
The first is the selection of customers for managed charging. Part 7 detects the households that charge an electric vehicle behind an import meter, and an operator with a limited programme must choose which of them to enrol. The unit is then the household, the action its enrolment, and the uplift of Equation 6.20 the evening hours of under-voltage and overload that its enrolment removes from its transformer, which a power-flow simulation provides for each candidate as it provided the outcomes of this chapter. A ranking by uplift concentrates the programme on the weakest transformers, and the equity bound of Equation 6.30, applied to the burden of delayed charging measured by Jain’s index of Equation 9.8, limits how far the programme may concentrate on the customers of a few feeders. The design is stated here and not evaluated, because the evening violations of HPK11 were too few in every season to distinguish the candidates.
The second is the recommendation of storage for a household with PV. The actions are a battery of each standard size and no battery, the outcome the over-voltage hours and the curtailment that remain, and the cost the annualised capital of the battery against the curtailed energy it recovers. The verification of Equation 9.6 applies unchanged: a case-based estimate proposes two sizes, and two simulations select between them.
The third is the restoration of supply after a fault, in which the feeders to be restored form a list and the order of restoration is a ranking. The ranking of live violations in Chapter 3 is its model: a score learned from past events orders the feeders by the customer-hours that remain to be restored, and the NDCG of Equation 6.15 measures how much of that loss the first crews to be dispatched recover.
Conclusion
This chapter recommended mitigation measures for the transformers of HPK11 from the simulated outcome of every measure on every transformer over four seasonal weeks. The measures acted on the violations that their physics predicts: the export limit and the tap change on over-voltage, managed charging on evening under-voltage and reinforcement on overload, with the tap change worsening under-voltage and reinforcement raising over-voltage by 7%. With costs expressed relative to reinforcement, the tap change was the measure of least cost for most transformers over a wide range of penalties, and reinforcement for 7 of the 43 within the range examined.
The recommenders differed by more than an order of magnitude in regret. The violation-pattern rule of Part 5 assigned the measure of least cost to 9.3% of the transformers, because it does not consider the tap change; the violation pattern-cluster and case-based policies assigned it to 90.7%, by recommending the measure of the majority to every transformer; and the simulation of the two measures of least case-based estimate assigned it to every transformer. Learners fitted on biased field records reached 57.7%, and matrix factorisation, with one measure recorded for each transformer, reduced to the assignment of the measure of least mean cost. At 60% adoption the same ordering held. The curtailment of the volt-watt response was small but fell on the customers of highest impedance, whereas a static export limit distributed it across positions and onto the customers with large arrays; across the demand clusters of Part 5 it was nearly even.
The results rest on four seasonal weeks and on costs relative to reinforcement, and a utility applying them substitutes its own ratios \(c_{\mathrm{e}} / c_{\mathrm{r}}\) and \(\mu / c_{\mathrm{r}}\). Evening under-voltage was rare on HPK11 in every season, so managed charging was never the measure of least cost; on a network with longer feeders or a higher adoption of EVs its place would be examined again. The field records are generated by a stated policy, and the recommenders have not been evaluated on the records of an operating network. The increase of over-voltage under reinforcement arises because the volt-VAr response makes the transformer carry reactive power, across a reactance about three times its resistance, so that a utility that reinforces a transformer subject to both constraints pairs the reinforcement with a change of tap or a review of the volt-VAr settings. Part 7 turns from violations that the simulator predicts to the detection of anomalous behaviour in measured data.
Exercises
- Raise the capital cost of managed charging in the
costscell from 0.10 to 0.30 reinforcement annuities. For how many transformers does the measure of least cost change at the central penalty?
The standards ANSI/ISA-18.2 and EEMUA 191 recommend that approximately 5% of alarms be of high priority, 15% of medium priority and 80% of low priority.↩︎
The Australian Energy Regulator publishes the value of curtailed exports as the customer export curtailment value, estimated half-hourly by region, and the unit rates of distribution transformers in the regulatory proposals of each network, in which the cost of a pole-mounted transformer of 25 kVA is approximately an order of magnitude below that of a kiosk of 500 kVA.↩︎