Search

The reduction of energy production due to heliostat soiling causes generation losses and therefore reduces the contribution and competitiveness of CSP. The optimal cleaning plan for CSP plants must strike the proper balance between the cost for cleaning activities and revenue losses due to a reduced power production. However, uncertainty in future soiling rates, weather conditions, and electricity sale prices (e.g. due to electricity markets, uncertainty in future purchasing agreements) significantly complicate the cleaning optimization. To address these challenges, a bi-objective simulation- optimization approach is developed to optimize the cleaning policy to simultaneously minimize cleaning costs and expected generation losses in the presence of stochastic weather, dust. The outcome of the proposed approach is the optimal number of cleaning trucks and periodic cleaning schedule for non-dominating solutions (Pareto front), which may be used to study the sensitivity of the cleaning schedule to variations in electricity prices. The approach is applied to a case study for a CSP plant located in Woomera, South Australia. The results show that for electricity sale prices between 40-100 AUD/MWh, the optimal number of trucks is 3 and the total cleaning costs are sensitive to the number of trucks purchased, but not to the cleaning frequency. However, as electricity prices increase above 300 AUD/MWh, the total costs become quite sensitive to the cleaning frequency. Finally, a comparison of the simulation-optimization approach and a (simpler) deterministic optimization indicates that the latter approach is suitable for most practical electricity prices.

This paper describes methods for predicting track geometry degradation to provide support for maintenance planning. The track geometry data recorded by the Track Recording Car (TRC) and the maintenance work orders were used for degradation modelling. The degradation indicator considered in this study was the deviation of the longitudinal level from the design value sampled every 100 m-long track segment. A Wiener process model was built to model the time-evolution of the geometry condition indicator for a track segment and the parameters were estimated via standard maximum likelihood estimation techniques. It was found that the proposed degradation model provided good fit to the data, except in some extreme cases where the TRC data exhibits large changes in the degradation indicator between two runs. Nevertheless, even in these cases the Wiener process model “responds” to this lack of confidence about the slope in an intuitively appealing manner: the variance on the slope is increased to reflect a lack of confidence in the degradation rate. Finally, the utility of the model in maintenance decision making is demonstrated by predicting the expected number of different maintenance interventions and policies.

Boiler heat exchanger tubes lose thickness over time, resulting in costly ruptures and losses in capacity as thin tubes are taken out of operation. To mitigate the costs associated with thickness loss, boiler tubes are inspected and can be preventively taken out of service (i.e. “plugged”) to avoid in-service ruptures. Moreover, there is an economic dependence among thousands of tubes of heat exchangers in inspection and maintenance activities where a large setup cost may be induced when inspection is conducted for any tube. This paper extends the authors’ previous study on boiler tube inspections to include a dynamic inspection and preventive maintenance grouping policy to pursue additional savings on setup costs. The inspection policy is said to be condition-based since the time for next inspection is based on the current inspected state. A heuristic thickness loss threshold for plugging is computed for each tube by balancing the risk of an in-service rupture with the lost revenue due to capacity loss and an optimal inspection grouping strategy is developed using the Markov Decision Process paradigm. The policy is applied to a case study of a boiler operating in an Australian sugar factory and bench-marked with policies that represent current practices and do not consider grouping. The results show that the proposed condition-based grouping inspection policy yields significant savings compared to tube individual inspection policies.

Recent studies have demonstrated the economic benefit of exploiting partial opportunities for maintenance, which arise from the reduction of costs of doing partial maintenance utilizing the opportunity given. A typical example is an externally induced production stop with a duration that may not be sufficient to a complete maintenance activity. This paper combines partial opportunities and condition-based maintenance (CBM) strategies and proposes an innovative maintenance optimization method considering time-varying economic conditions. This scenario naturally fits a broad range of assets with a finite design life (including long-life machinery and infrastructure), operating under variable economic conditions and usage-intensities. The maintenance optimization problem is formulated in this study as a finite-horizon Markov decision process, where the randomly occurring opportunities are accounted for by augmenting the time-varying, decision-dependent transition probabilities. A dynamic programming approach is subsequently used to obtain the optimal CBM policy, consisting of time-varying thresholds on equipment condition and the cost of conducting maintenance during the arrived opportunity. A case study of such a maintenance policy for an induced draft fan of a sugar production system located in Queensland, Australia, is undertaken to evaluate the benefits of the proposed methodology against more traditional approaches that neglect opportunities or consider only replete opportunity durations.

In sugar production, the vacuum pans used for crystallisation constitute one of most important systems and replacement costs are a significant capital expenditure. The maintenance strategy for pans is currently conducted on a time-based approach. This paper proposes a condition-based maintenance (CBM) model for pans, which is applied to a real multi-component pan operating in an Australian sugar mill. The maintenance policy is formulated as a Markov Decision Process (MDP) and solved via approximate dynamic programming. The results show a 10% saving on total maintenance costs compared to the time-based maintenance strategy.

Introduction - Staffing levels strongly influence economics of CSP plants•NREL SAM spreadsheets: ~10 staff each for solar field and power block (supervisor, foreman, techs) - Salaries ∼100 kUSDper year (burdened labour cost, based on CA labour costs) - Similar in in Australia? - Basic trade-off: More staff →higher availability of CSP systems and more revenue, but larger direct labour costs - Purpose of this study: propose methodologies to calculate appropriate staffing - Examples on Solar Field and Power Block –how much can be saved?

This paper aims to develop an opportunistic maintenance (OM) policy for the generator of a hypothetical wind turbine using methods developed recently by the authors. The OM policy considers external opportunities caused by low wind speeds which produce little-to-no electric power. The results show that some cost savings are achievable by taking maximal advantage of these low-speed wind events, particularly when electricity prices are at their peak cycle.

This thesis proposed a novel Opportunistic Maintenance policy where uncertain and uncontrollable maintenance opportunities and the benefits may arrive from sources external to a production system. Optimisation models were developed and analysed for both time- and condition-based approaches with the impact of using short-term forecasts of external opportunities. The new models could result in significant savings in maintenance costs that are demonstrated using numerical examples and real case studies in energy systems for different industries. This thesis discussed special "cost-free" maintenance opportunities, which are prevalent in solar energy systems where rain events can provide free cleaning of solar collectors.

Cleaning costs of mirrors contribute a significant portion in operation and maintenance costs (O&M) of concentrating solar power (CSP) systems. The optimal cleaning policy is obtained from a tradeoff between revenue received from generating electricity and the costs of conducting cleaning operations (e.g. water, labor, etc.). However, this balance depends strongly on the local electricity market and weather conditions and currently available cleaning policies have not considered variation in electricity prices nor the potential for “natural” cleaning events (e.g. rain). In this study, a Condition-Based Cleaning (CBC) policy is developed for mirrors whose degradation is stochastic and subject to seasonal variations. The optimal policy is determined by formulating and solving a finite-horizon Markov decision process whose time-varying transition matrices describe stochastic soiling, rain events, and imperfect cleanings. The optimal cleaning policy is therefore a time-varying reflectivity threshold, below which cleaning is triggered. The methodology has been applied to a case study on a hypothetical plant in Brisbane, Australia. Using publicly available electricity price and weather data, the optimized CBC policy was found to save 5–30% of total cleaning costs compared with a fixed-time strategy. Importantly, higher CBC savings are achieved when the direct cleaning costs are high, indicating that the policy could be particularly significant for countries with high labor or resource (water, etc.) costs (e.g. Australia). Though applied to CSP in this study, the methodology is also applicable to optimal cleaning of other solar collectors (e.g. photovoltaic collectors), albeit with different efficiency models.

Many systems and manufacturing processes undergo intermittent operation due to external factors (e.g. weather, low market prices), offering opportunities to conduct maintenance with reduced production losses. Making use of appropriate opportunities can thus lead to significant reduction in the total cost of maintenance and improvement in productivity. In this paper, an opportunistic maintenance (OM) model is developed considering two critical properties of real world opportunities: (i) non-homogeneous opportunity arrivals and (ii) stochastic opportunity duration. The model enables exploiting downtime cost savings from "partial" opportunities (stops shorter than the required maintenance time) thus extending the potential benefit of OM. The criteria for accepting maintenance opportunities are found by minimizing the single-cycle total cost. A closed form expression of the single-cycle total cost is derived for a given PM/OM policy and then a Genetic Algorithm is used to solve the optimization problem. Numerical results are presented to assess the benefit of opportunistic maintenance and the marginal benefit of considering partial opportunities. Results indicate that significant savings can be achieved by considering OM. Moreover, it is shown that the novel consideration of partial opportunities significantly increase the benefit of OM.

In production activities, external events (e.g. weather, low market prices) may provide some opportunities to do the preventive maintenance with reduced cost. Implementing an appropriate maintenance policy considering these opportunities can result a significant savings in total maintenance cost. This study focuses on developing a joint preventive and opportunistic maintenance strategy for a production system. The cost savings of opportunities is considered random with a time-varying distribution. A new approach for accepting (or rejecting) an arrived opportunity is also proposed based on minimizing the single-cycle cost rate. A Simulation - Optimization procedure is proposed to determine the optimal preventive maintenance time and acceptance of opportunities. The simulation results presented in this work show that the joint optimization of preventive and opportunistic maintenance can yield a significant cost savings.