Using Risk Models and Automated Defect Characterization Algorithms to Convert PACP Data into Capital Upgrading Programs for ALCOSAN
Bibliographic record
Abstract
ALCOSAN provides regional wastewater conveyance and treatment for 83 municipalities in Alleghany County, PA. They own and operate approximately 90 mi of interceptor sewers and the 250 million gallon per day Woods Run Wastewater Treatment Plant. There are over 4,000 mi. of combined and separated collection sewers located within the ALCOSAN service area owned by the municipalities. Utilizing EPA’s Integrated Watershed Approach, large portions of the collection system are being regionalized to facilitate the development of a more efficient and equitable upgrading and long term operating model. Integrating data from such a diverse array of sources is a very challenging task. Maintaining a high quality data capture with such an immense condition assessment data set is both a challenge and essential component of fully understanding the technical and financial ramifications of the condition assessment process accurately and expeditiously. The condition assessment program in support of regionalization is a traditional CCTV data capture with defect characterization of the conveyance infrastructure and its appurtenances being carried out with NASSCO’s PACP and MACP data logging techniques. The data capture with PACP/MACP involves the recording of literally millions of observations and requires considerable innovation to process the data efficiently and in a consistent manner such that clarity of risk and upgrading priorities is maintained. A traditional WRc/ASCE MOP 62 Risk Model was initially developed to lay the foundation for establishing condition assessment priorities, intervention timing, and policy. The model used readily available surrogate data to develop overall cost factors (an indication of the direct and indirect cost of collapse without rehabilitation intervention) for each reach of sewer pipe with the entire conveyance system. To facilitate a consistent and common approach to condition assessment and rehabilitation rationalization amongst a diverse pool of reviewers and end stakeholders; the writers mapped out the desired roadmap from defect patterns to rehabilitation assignment in some detail with the aid of master algorithms that analyze the specific pattern of defects in each MH-to-MH reach to suggest an optimum rehabilitation work stream and work limits in the reach. While trained practitioners can modify the rehabilitation assignment in quality assurance (QA) review, the master algorithms have been very successful in attaining a consistent approach in rehabilitation assignment between reviewers and promote a transparency in policy to all stakeholders that is well aligned with best practice in rehabilitation selection. By attaching a cost estimating model to the defect database and integrating it with the prioritization model and intervention policy; the Owner is able to have an initial understanding of the financial ramifications of the defects, the nature of upgrading programs that are required and the optimum timing associated with the program; instantaneously upon QA of the condition data which is typically long before final QA reviews of the rehabilitation assignments. The paper provides an overview of the development of the risk model and of the innovative approaches utilized in rehabilitation rationalization from the Pilot Regionalization program that was carried out in 2015 and 2016.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.009 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".